Datasets:

florentgbelidji

/

ncbi_extracted_running

like 0

PMC6021049

1 Table A. Linear mixed-effect model results for transformed minimum reaction times for all sprints in 2016.

On the apparent decrease in Olympic sprinter reaction times.

06-27-2018

Mirshams Shahshahani, Payam,Lipps, David B,Galecki, Andrzej T,Ashton-Miller, James A

eng

PMC10688325

Supplementary File 3: P-values of the Wilcoxon-Mann-Whitney tests assessing the null hypothesis that it is equally likely that a value chosen at random from one year is greater or less than a value chosen at random from another year’s population. Top 100 Table 1 Men’s 100m Table 2 Men‘s 110m hurdles Table 3 Men‘s 200m Table 4 Men‘s 400m Table 5 Men‘s 400m hurdles 2016 2017 2018 2019 2021 2017 1 2018 1 0.638283 2019 1 1 1 2021 1 0.309905 1 1 2022 1 0.00015 0.00196 0.003115 0.062573 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 0,714507 2021 0,097206 0,025 0,000356 0,003474 2022 0,003464 0,000552 4,56E-06 3,78E-05 0,726513 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 1 1 1 1 2022 0,052459 0,001119 0,046565 0,014442 0,047502 2016 2017 2018 2019 2021 2017 1 2018 0,572175 1 2019 1 1 1 2021 1 1 1 1 2022 0,052919 0,627806 1 0,078112 0,272402 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 1 1 1 1 2022 0,972897 1 1 0,388271 0,973264  Table 6 Women‘s 100m Table 7 Women‘s 100m hurdles Table 8 Women‘s 200m Table 9 Women‘s 400m Table 10 Women‘s 400m hurdles 2016 2017 2018 2019 2021 2017 1 2018 1 0,466331 2019 1 1 1 2021 0,03227 0,02574 0,139897 0,011156 2022 4,53E-07 4,06E-06 2,3E-06 4,28E-08 0,003582 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 1 0,004164 0,065499 0,129516 2022 0,746016 0,001377 0,023704 0,042453 1 2016 2017 2018 2019 2021 2017 1 2018 1 0,492091 2019 1 1 1 2021 1 0,265376 1 0,002601 2022 0,085241 0,000304 0,043264 2,27E-06 0,265376 2016 2017 2018 2019 2021 2017 1 2018 0,804407 0,371019 2019 1 0,702147 1 2021 1,35E-05 2,2E-07 0,001172 6,98E-05 2022 4,93E-05 5,75E-07 0,002364 0,000161 1 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 1 1 0,346209 0,375669 2022 1 0,608207 0,098102 0,080571 1  Top 20 Table 11 Men‘s 100m Table 112 Men‘s 110m hurdles Table 13 Men‘s 200m Table 14 Men‘s 400m Table 15 Men‘s 400m hurdles 2016 2017 2018 2019 2021 2017 1 2018 1 0,800315 2019 1 1 1 2021 0,972321 0,017573 0,297023 0,059996 2022 1 0,021544 0,33573 0,078011 1 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 0,303898 0,57473 0,003671 0,082601 2022 0,109341 0,290332 0,003215 0,062877 1 2016 2017 2018 2019 2021 2017 1 2018 1 0,062617 2019 1 0,175687 1 2021 1 0,685787 1 1 2022 0,175687 0,000813 0,232567 0,269417 0,154942 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 1 1 1 1 2022 1 1 1 1 1 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 0,407642 0,524398 0,53255 0,160764 2022 0,197812 0,338821 0,407642 0,160764 1  Table 16 Women‘s 100m Table 17 Women‘s 100m hurdles Table 18 Women‘s 200m Table 19 Women‘s 400m Table 20 Women‘s 400m hurdles 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 1 0,517847 0,006641 0,012547 2022 0,317904 0,018162 4,46E-05 0,004229 0,494987 2016 2017 2018 2019 2021 2017 0,711735 2018 1 1 2019 0,711735 1 0,699526 2021 0,045484 0,231794 0,114568 0,614191 2022 0,001485 0,001485 0,005646 0,014589 0,076741 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 0,012536 0,016489 0,003203 0,007309 2022 0,007309 0,007309 0,001137 0,005665 1 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 0,026829 0,011861 0,054195 0,010109 2022 0,467379 0,615339 0,757428 0,054195 1 2016 2017 2018 2019 2021 2017 1 2018 1 1 2019 1 1 1 2021 1 1 0,074017 0,574967 2022 0,103321 0,574967 0,005963 0,074017 1

The potential impact of advanced footwear technology on the recent evolution of elite sprint performances.

11-27-2023

Mason, Joel,Niedziela, Dominik,Morin, Jean-Benoit,Groll, Andreas,Zech, Astrid

eng

PMC5325470

RESEARCH ARTICLE Comparison of wrist-worn Fitbit Flex and waist-worn ActiGraph for measuring steps in free-living adults Anne H. Y. Chu1*, Sheryl H. X. Ng1, Mahsa Paknezhad2, Alvaro Gauterin2, David Koh1,3, Michael S. Brown4, Falk Mu¨ller-Riemenschneider1,5 1 Saw Swee Hock School of Public Health, National University of Singapore, Singapore, Singapore, 2 Department of Computer Science, School of Computing, National University of Singapore, Singapore, Singapore, 3 PAPRSB Institute of Health Sciences, Universiti Brunei Darussalam, Jalan Tungku Link, Gadong, Brunei Darussalam, 4 Department of Electrical Engineering and Computer Science, Lassonde School of Engineering, York University, Toronto, Ontario, Canada, 5 Institute of Social Medicine, Epidemiology and Health Economics, Charite´ University Medical Centre Berlin, Berlin, Germany * [email protected], [email protected] Abstract Introduction Accelerometers are commonly used to assess physical activity. Consumer activity trackers have become increasingly popular today, such as the Fitbit. This study aimed to compare the average number of steps per day using the wrist-worn Fitbit Flex and waist-worn Acti- Graph (wGT3X-BT) in free-living conditions. Methods 104 adult participants (n = 35 males; n = 69 females) were asked to wear a Fitbit Flex and an ActiGraph concurrently for 7 days. Daily step counts were used to classify inactive (<10,000 steps) and active (10,000 steps) days, which is one of the commonly used physical activity guidelines to maintain health. Proportion of agreement between physical activity categoriza- tions from ActiGraph and Fitbit Flex was assessed. Statistical analyses included Spear- man’s rho, intraclass correlation (ICC), median absolute percentage error (MAPE), Kappa statistics, and Bland-Altman plots. Analyses were performed among all participants, by each step-defined daily physical activity category and gender. Results The median average steps/day recorded by Fitbit Flex and ActiGraph were 10193 and 8812, respectively. Strong positive correlations and agreement were found for all partici- pants, both genders, as well as daily physical activity categories (Spearman’s rho: 0.76– 0.91; ICC: 0.73–0.87). The MAPE was: 15.5% (95% confidence interval [CI]: 5.8–28.1%) for overall steps, 16.9% (6.8–30.3%) vs. 15.1% (4.5–27.3%) in males and females, and 20.4% (8.7–35.9%) vs. 9.6% (1.0–18.4%) during inactive days and active days. Bland-Altman plot indicated a median overestimation of 1300 steps/day by the Fitbit Flex in all participants. PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 1 / 13 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Chu AHY, Ng SHX, Paknezhad M, Gauterin A, Koh D, Brown MS, et al. (2017) Comparison of wrist-worn Fitbit Flex and waist- worn ActiGraph for measuring steps in free-living adults. PLoS ONE 12(2): e0172535. doi:10.1371/ journal.pone.0172535 Editor: Maciej Buchowski, Vanderbilt University, UNITED STATES Received: August 18, 2016 Accepted: February 6, 2017 Published: February 24, 2017 Copyright: © 2017 Chu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: Due to ethical restrictions set by the National University of Singapore Institutional Review Board, study data cannot be made publicly available. Requests for data may be sent to Anne Chu (email: anne.chu@u. nus.edu), Falk Mu¨ller-Riemenschneider (email: [email protected]) or the National University of Singapore Institutional Review Board (email: [email protected]). Fitbit Flex and ActiGraph respectively classified 51.5% and 37.5% of the days as active (Kappa: 0.66). Conclusions There were high correlations and agreement in steps between Fitbit Flex and ActiGraph. However, findings suggested discrepancies in steps between devices. This imposed a chal- lenge that needs to be considered when using Fibit Flex in research and health promotion programs. Introduction New wearable technologies have helped raise individual self-awareness about physical activity behavior. Among all the functionalities that a range of wearable devices have, step counting is the most fundamental and consistently found feature. Step counts have been proposed as a health indicator for population studies [1], and even community-based health-promotion pro- grams [2]. The 10,000 steps/day guideline is one of the commonly used physical activity indices [3]. Various government/professional organizations around the world have used the 10,000 daily steps recommendation as an index of high physical activity level. This daily step-based rec- ommendation has been endorsed by the World Health Organization (WHO), National Heart Association of Australia, US Centers for Disease Control and Prevention, and American Heart Association to improve overall health. For healthy adults, it appears that this guideline is a real- istic estimate of an appropriate daily physical activity level [4, 5]. It was suggested that those achieving the goal of 10,000 steps per day were more likely to meet physical activity guidelines as compared to those with lower step counts [2]. Furthermore, health promotion programs that included a daily step goal were reportedly more successful in increasing physical activity than those without this component [6]. The use of step data (usually as steps/day) is a simple means of reflecting habitual physical activity pattern, and this approach has become acceptable to many researchers and practitioners [1, 6]. Moreover, walking activity has been reported as a prevalent form of leisure-time physical activity and a functional task in the daily lives [7]. Among all the accelerometers commonly used in research, the ActiGraph (Pensacola, FL, USA) is well-validated and has been extensively used for assessing physical activity under free- living conditions [8–11]; The ActiGraph accelerometers use algorithms to quantify and con- textualize the resultant acceleration signals of human motion. They have shown high accuracy for moderate-to-high walking speed stepping in the laboratory (compared to direct observa- tions, ICC: 0.72–0.99) and under free-living conditions (compared to the Yamax Digiwalker, ICC: 0.90) [12]. The ActiGraph has been used in large-scale epidemiological studies such as the US National Health and Nutrition Examination Survey (NHANES) [13], and the Women’s Health Study (WHS) [14]. Recently, consumer-based activity trackers (e.g. Fitbit, Jawbone UP, LUMOback, Nike + Fuelband, Omron Walking Style Pro pedometer, etc.) and in-built accelerometers in smart- phones have become increasingly popular [15, 16]. It was forecasted that the smart wearables market could reach 170 million units by 2017 [17]. Fitbit (San Francisco, CA, USA) is one of the most commonly used brands amongst the consumer-based activity trackers. As of 2015, Fitbit had reached 9.5 million active users [18]. Among their products, the wrist-worn Fitbit Flex has become popular in recent years either for aesthetic reasons or wearing comfort. The Fitbit Flex is sleek and displays only LED with a tap screen. Users are able to monitor and Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 2 / 13 Funding: This research was supported by a grant WBS: R-608-000-117-646 from the National University of Singapore. Competing interests: The authors have declared that no competing interests exist. access data on the number of steps, sleep quality, and other personal metrics through the Fitbit dashboard. This could be useful for targeted physical activity interventions designed to achieve healthy behaviors. It was suggested that wrist-worn accelerometers allowed for monitoring of low-intensity activities, and were associated with considerable increases in wearing compliance and data quality [19]. A number of studies have validated wireless consumer-based monitors of different brands in measuring step counts and energy expenditure [16, 20–23]. A recent systematic review con- cluded high validity for the Fitbit Classic, One and Zip compared to accelerometry-based step counts (particularly in laboratory settings) [24]. It was further highlighted that more field- based studies are needed. Evaluation of the trackers in assessing free-living physical activity (non-controlled environment outside a lab setting) is particularly important, as the results are more likely to reflect usual day-to-day activities. To date, sample sizes of studies on the Fitbit Flex validity under free-living conditions have been relatively small (ranging from 14 to 25 par- ticipants) and based on young adults [16, 25–27]. Of note, one similar study was limited by a small sample size of one adult only [28]. However, despite the high correlation between activity trackers, these studies generally showed that Fitbit Flex has measurement limitations regarding the overestimation and underestimation of activity levels compared with the reference device, depending on different study settings and types of activity [26, 27]. Given these considerations and highlighted gaps, this study aimed to make standardized comparisons based on step counts from the consumer-oriented Fitbit Flex and the research- grade ActiGraph wGT3X-BT. Differences in levels and types of physical activity between males and females have been reported [29, 30]. It was reported that more males than females tended to practise sports (e.g. soccer, basketball, etc.), whereas females were more likely to engage in yoga, dancing, aerobics, etc. [31]. Because these differences may influence their accu- racy in measurement, we further performed gender specific analysis. Hence, the objectives of this study were: 1. To compare free-living steps/day recorded by the Fitbit Flex and the ActiGraph wGT3X-BT accelerometers in all participants, by each step-defined daily physical activity category and gender. 2. To compare the agreement between devices in classifying participants’ step-defined daily physical activity categories. Materials and methods Study design and participants This was a cross-sectional study. The present study was a part of a previously published study [32], whereby a convenience sample of 107 employees who completed both ActiGraph and Fit- bit Flex measures were included. Participants from a large public University and a hospital in Singapore were recruited between February 2014 and June 2014. Individuals were residing in Singapore and were of various ethnicities (Chinese, Malay, Indian and others). Participants were invited to take part in this study through mass e-mailing. Individuals who indicated inter- est were approached and interviewed by the researcher. The inclusion criteria were: 1. Males and females aged 21 to 65 years 2. Either students or working adults 3. Absence of physical disabilities or illness that would create abnormal gait patterns. Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 3 / 13 The study was approved by the National University of Singapore Institutional Review Board (NUS-IRB Ref No.: B-14-021). Participants provided their written informed consent to participate in this study. Procedure The goals and procedures of the study were explained to each participant by the researcher via face-to-face interview. Participants’ information on gender, age, education level, height and weight were self-reported. Instructions were given to the participants by trained personnel on how to put on a wrist-worn Fitbit Flex and a waist-worn ActiGraph concurrently for 7 days. Instruction manuals on the proper use of the ActiGraph and Fitbit Flex were also given to par- ticipants for additional guidance. Participants were instructed that the devices had to be worn for at least 10 hours/day, and could be removed at night depending on their comfort level. They were asked to complete a daily time sheet to record each wearing day when both devices were worn while maintaining their normal activities. Information required on the time sheet comprised of the dates they started and stopped wearing the devices. ActiGraph wGT3X-BT The ActiGraph™ wGT3X-BT monitor (ActiGraph, LLC, Pensacola, Florida, USA) is a triaxial accelerometer (Dimensions: 4.6cm x 3.3cm x 1.5cm; weight: 19 grams) worn on the waist using an elastic belt to secure above the right hip bone for quantifying the amount and fre- quency of human movements. The monitor was initialized at a sample rate of 30Hz to record activities for free-living conditions. Participants were instructed to wear the ActiGraph for 7-day. They were allowed to remove the ActiGraph only while bathing or immersing the body in water. ActiGraph data were downloaded using ActiLife 6 software (ActiGraph, LLC, Pensa- cola, FL, USA) by the researchers upon collection of the devices. Downloaded data were inte- grated into 60-sec epochs. Fitbit Flex Fitbit FlexTM (Dimensions: 22.2cm x 6.0cm x 6.0cm; weight: 100 grams) is a wrist-worn wear- able wireless sensor with a triaxial accelerometer that records physical activity throughout the day. It can sync with a smartphone application/computer. Participants were instructed to wear the Fitbit Flex on their non-dominant wrist, for the same duration as the ActiGraph (up to 7-day) concurrently. In general, Fitbit Flex requires the creation of individual user accounts to download stored data using a Web-based software application. However, for the purpose of our study, anonymous user accounts were created by the study team which could only be accessed by the researchers. Steps data were therefore stored on the devices, and the minute- by-minute Fitbit Flex data were downloaded at the end of each participant’s wearing period by the study team. Data reduction For wear time validation, because the ActiGraph accelerometer is an established device to mea- sure physical activity with many validation studies determining their accuracy [33, 34], valid wear time determined by the ActiGraph was regarded as the reference. A detailed description of the procedures on ActiGraph wear time validation and removal of sleep time can be found elsewhere [32]. Then, a valid day was defined as having an accumulation of 1500 steps/day with 10 hours/day restricted only to common wear time based on both ActiGraph and Fitbit Flex. The 1500 steps/day criterion was based on a previous research conducted by Tudor- Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 4 / 13 Locke et al. comparing accelerometers positioned at different locations under free-living con- ditions [35]. All participants with 4 valid days of data were included in the analysis. Addi- tionally, wear time was also verified based on the daily time sheets. Statistical analysis All statistical procedures were performed using SPSS software (version 20.0). The significance level was set at P<0.05. Descriptive characteristics were presented as mean (standard devia- tion; SD) or median (interquartile range; IQR). Shapiro-Wilk test was used to determine whether the data was normally distributed. Differences in the characteristics between genders were detected by non-parametric tests. Mann-Whitney U test (for continuous variables), chi- squared test (for categorical variables) and Fisher’s exact test (for categorical variables with cells having an expected frequency of five or less) were used. Analyses of the relationship between ActiGraph and Fitbit Flex were performed across: all participants, by each category of step-defined daily physical activity, and gender. Because there could be potential within-subject variations, comparison of step counts for the magnitude of relationship between the two devices was done on a day-to-day basis. Spearman’s correlation coefficient (rho) and intraclass correlation coefficient (ICC) were used to assess correlation and agreement, respectively in steps between ActiGraph and Fitbit Flex. An ICC value of 0.75 implied excellent, 0.60–0.74 good, 0.40–0.59 fair and <0.40 poor agreement [36]. Median of absolute percentage error (MAPE) between devices was calculated: (absolute error/ observed steps) × 100%. The difference in MAPE by each category of step-defined daily physi- cal activity and gender was compared using Mann-Whitney U test. ActiGraph derived steps/ day was used to classify two step-defined activity categories for the assessments of Spearman’s rho and ICC. The classification of days into two step-defined activity categories was adapted based on previous studies: valid days with a cumulative of 10,000 steps/day were considered as active days, and <10,000 steps/day were inactive days [5, 37, 38]. As for the Bland-Altman analysis, a non-parametric approach was adopted since the differences between the two devices were non-normally distributed. Bland-Altman plots were presented as median, 10th and 90th percentiles to display variance around differences between two devices. Proportion of agree- ment in achievement of 10,000 steps per day produced by ActiGraph and Fitbit Flex was assessed using Kappa. Results Out of 107 recruited participants, 104 were included because they met the wear time criteria and provided 682 days of data. Table 1 shows participants’ sociodemographic characteristics of the study. Participants had a median age of 31.0 years (IQR: 26.0–42.8), predominantly female (66.3%), and had a university degree (74.0%). On average, 6.6 valid wear days were recorded per participant and there was no significant difference between males and females. The Acti- Graph and Fitbit Flex steps were significantly higher in males than females (P = 0.03 and 0.01 for ActiGraph and Fitbit Flex, respectively). Fitbit Flex recorded a significantly higher (P < 0.001) number of daily step counts than that from the ActiGraph across all participants, by gender and each category of step-defined daily physical activity (Table 2). Males reflect significantly higher daily step counts from Fitbit Flex (P = 0.01) and ActiGraph (P = 0.028) compared to females. The magnitude of the correlation and agreement in step counts between ActiGraph and Fit- bit Flex were assessed (Table 2). Good to excellent significant positive correlations and agree- ment were shown in all participants, by gender and category of step-defined daily physical activity. Table 3 shows the number of days that were misclassified as active or inactive Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 5 / 13 according to the Fitbit Flex. The proportion of overall agreement of devices in classifying days as active or inactive was estimated, reporting a kappa of 0.66, indicating a moderate agreement (Table 3). Fig 1 shows the MAPE in number of steps between the two devices. Significant differences in the MAPE of step counts were found between devices across step-defined physical activity categories (P<0.001), but not for gender (P = 0.17). Figs 2 and 3A–3D present Bland-Altman plots on the median of differences, and the 10th and 90th percentiles between steps/day obtained from Fitbit Flex and ActiGraph. The bias (median difference) is 1300 steps/day for all participants. In general, the Fitbit Flex overesti- mated steps/day relative to ActiGraph (median differences range: 1166–1509 steps/day by gen- der and 1280–1312 by step-defined physical activity categories). Discussion This study focused on the direct comparison of steps obtained from the Fitbit Flex and Acti- Graph. The results show positive correlations and agreement in step counts of free-living adults as measured by the waist-worn ActiGraph and wrist-worn Fitbit Flex activity monitors. At the same time, overestimation of step counts and classification as active days by Fitbit Flex were found. This may have important public health implications if consumers or participants of health promotion programs are identified as being active when in fact they are not. Table 1. Characteristics of study population. All (n = 104) Males (n = 35) Females (n = 69) P-valuea Age (Med; IQR) 31.0; 26.0–42.8 33.0; 27.0–50.0 30.0; 25.5–40.5 0.05 Height, cm (Med; IQR) 163.0; 157.0–169.8 170.0; 168.0–175.0 160.0; 155.0–163.0 <0.001 Weight, kg (Med; IQR) 60.0; 53.0–69.9 65.0; 60.0–80.0 56.6; 50.0–66.0 <0.001 BMI (Med; IQR) 22.6; 20.3–25.5 23.1; 20.8–25.8 22.1; 20.2–25.1 0.3 Education, n (%) 0.01 Secondary 7 (6.8) 0 (0) 7 (10.2) Technical school/diploma 20 (19.2) 3 (8.6) 17 (24.6) University 77 (74.0) 32 (91.4) 45 (65.2) Organization, n (%) 0.51 Public university 70 (67.3) 24 (68.6) 46 (66.7) University hospital 34 (32.7) 11 (31.4) 23 (33.3) 0.92 Valid wearing day/week (M±SD) 6.6 ± 0.9 6.6 ± 1.0 6.5 ± 0.9 BMI, body mass index; IQR, interquartile range; M, mean; Med, median; SD standard deviation. a Test of significant difference between males and females. doi:10.1371/journal.pone.0172535.t001 Table 2. Comparison, relative agreement and median of absolute error in step counts between ActiGraph and Fitbit Flex: all participants, by gen- der and category of step-defined daily physical activity. Step count/day All (682 days) Males (229 days) Females (453 days) Inactive (426 days) Active (256 days) Fitbit Flex (Med; IQR) 10193; 7490–12898a 11030; 7604–14838a 9992; 7397–12509a 8235; 6267–10003a 14075; 11948–16864a ActiGraph (Med; IQR) 8812; 6152–11471a 9409; 6268–12897a 8599; 6053–11118a 6856; 4982–8465a 12716; 11112–14505a Spearman’s rho 0.89* 0.91* 0.87* 0.76* 0.76* ICC (95% CI) 0.85 (0.58–0.93) 0.87 (0.56–0.94) 0.83 (0.56–0.92) 0.73 (0.68–0.77) 0.82 (0.77–0.85) CI, confidence interval; IQR, interquartile range. a ActiGraph and Fitbit Flex estimates are significantly different (P < 0.05). * P < 0.01 doi:10.1371/journal.pone.0172535.t002 Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 6 / 13 Recently, a number of studies have investigated the accuracy of various consumer-based physical activity trackers, recognizing the role they may play in physical activity promotion. For instance, Case et al. [16], Storm et al. [20], and Diaz et al. [21] have validated consumer wearables for measuring steps. However, to date very few studies have investigated the accu- racy of these monitors under free-living conditions [24]. This is highly important because the accuracy of devices may differ considerably in day-to-day life as compared to under highly controlled and short protocols of activities. Recently, several studies have been conducted with regard to this important research question [25–27]. Dierker et al. [25] assessed the validity of Fitbit Flex among 17 college-aged adults and found that although the steps measured by Fitbit Flex (9596 ± 2361 steps) were higher than the ActiGraph GT3X+ (7766 ± 2388 steps), the dif- ference was not statistically significant (P = 0.052). However, the authors instructed the partici- pants to remove the devices while they were exercising over the 7-day monitoring period; hence it is possible that not all free-living movements have been captured as in the present Table 3. Agreement between ActiGraph and Fitbit Flex for categorizing step-defined daily physical activity. No. of days (%)a ActiGraph Fitbit Flex Inactive Active Inactive 320 (46.9) 11 (1.6) Active 106 (15.5) 245 (35.9) Total 426 (62.5) 256 (37.5) Kappa (95% CI) 0.66 (0.61–0.71) a Physical activity categories are based on ActiGraph daily step counts: inactive <10,000 steps/day and active 10,000 steps/day [5]. doi:10.1371/journal.pone.0172535.t003 Fig 1. MAPE (%) between ActiGraph and Fitbit Flex. Error bars indicate IQR of MAPE. MAPE, median absolute percentage error. doi:10.1371/journal.pone.0172535.g001 Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 7 / 13 study. In another study by Dominick et al. [26], the Fitbit Flex registered a total of 10286 ± 3760 free-living steps/day as compared to the ActiGraph of 9639 ± 3456 steps/day (albeit no significant difference was found between devices) among 19 participants. In contrast, Sus- hames et al. [27] reported a larger absolute difference of over 3000 steps (47.0%) in free-living steps between Fitbit Flex and ActiGraph among 25 adults, of which the Fitbit Flex has underes- timated step counts. The reason for this underestimation from Fitbit Flex is unclear, but it could be related to the variability in participants’ movements or undercounting of steps by the Fitbit Flex. Different study settings and reference methods could contribute to the discrepancies in out- comes. Kooiman et al. [39] assessed the validity of Fitbit Flex over 1 day in a smaller sample of free-living adults and found high agreements in steps with the activPAL. They found a notice- ably smaller mean absolute percentage difference of 3.7% against the activPAL [39]. In accor- dance with our findings, another recent study comparing Fitbit Flex and ActiGraph on 48 cardiac patients (mean age: 65.5 years), in which high correlations and a difference in step counts of 1038 steps/day in the total population over 4 days of monitoring period were reported. Thus, comparing findings among different populations can provide an implication Fig 2. Bland-Altman plot of differences between waist-worn ActiGraph and wrist-worn Fitbit Flex against the mean according to all participants. The solid line represents median of the differences between devices, dotted lines are 10th and 90th percentiles of the differences. doi:10.1371/journal.pone.0172535.g002 Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 8 / 13 of how reproducible and valid this device is. It was also noted that the overestimation in step counts by the Fitbit Flex in this study resulted in a considerable misclassification of days as being active, which may have important public health implications. As shown in our analysis, the differences in steps between Fitbit Flex and ActiGraph were larger on inactive days as com- pared to active days. Hypothetically, as most lifestyle activities include movements at the wrist, people might have performed movements such as hand waving that could be identified as potential false pos- itive events/steps by Fitbit Flex. It was apparent that wrist-movements could reflect arm/ Fig 3. Bland-Altman plots of differences between waist-worn ActiGraph and wrist-worn Fitbit Flex against the mean according to: (A) Males, (B) Females, (C) Inactive days, and (D) Active days. The solid lines represent median of the differences between devices, dotted lines are 10th and 90th percentiles of the differences. doi:10.1371/journal.pone.0172535.g003 Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 9 / 13 forearm motions with a relatively small mass (while sitting), or they could be classified as step counts (while walking or running) [40]. Tudor-Locke et al. [35] found a large difference even using the same ActiGraph device placed between different attachment sites. They further reported that the difference between mean steps from the wrist and waist was 2558 steps under free-living conditions, with a higher average step counts on the wrist [35]. In line with this, Hilderbrand et al. [41] found a 200% higher step activity from the wrist-worn GENEActiv than the waist-worn ActiGraph in some adults. These observations suggest room for further progress, since recent studies reported using wrist-worn monitors resulted in improved wear- ing compliance due to comfort issues and without having the need to remove them intermit- tently [42]. Ultimately, prolonged wear time would improve data quality as the issue of missing data due to non-compliance could be minimized. Strengths Despite the growing body of evidence, this study expands substantially on previous studies. Most importantly, as highlighted earlier, the comparison of the devices was done under free-living conditions for estimation of unstructured lifestyle activities. Secondly, the relationship between these devices were assessed for 7-day of wearing protocol. Thirdly, this study was conducted among a relatively large sample of adults. Fourthly, the performance of the devices was compared across different subgroups (males vs. females and step-defined physical activity categories). Limitations This study may have limited generalizability as participants were predominantly females, rela- tively young and healthy. Furthermore, the use of ActiGraph as the reference instrument has its drawbacks. It is possible that the difference in steps between devices could be attributable to not only the Fitbit Flex, but also the ActiGraph, which is not the gold standard for measuring step counts [43]. However, the ActiGraph has been shown to be a valid tool to assess step count (as compared with the Omron pedometer and Yamax Digiwalker [11, 12]), and it is practical for use in epidemiological studies [44]. Careful consideration should also be given to the effects of movement artefact and signal noise due to the use of devices that are not attached directly to the skin (i.e. Fitbit Flex worn on a wrist-band and ActiGraph on a waist-belt), which might have affected the devices’ functionality to accurately measure step count. Being limited to only step count data, there was no indication as to whether the activities performed were of light-, moderate- or vigorous-intensity level. In general, step counts from accelerome- ters of different attachment sites (i.e. wrist- and waist-worn) might not be ideal for a direct comparison; nonetheless, results of this study were more likely to reflect the performances of these devices in real-world practice. Conclusions Positive correlation and agreement in step counts were found between wrist-worn Fitbit Flex and waist-worn ActiGraph in free living adults, which is consistent with the existing evidence mainly from laboratory studies. However, a considerable overestimation of Fitbit Flex was noted, which resulted in substantial misclassification by Fitbit Flex when applying common step count recommendations. This can have important practical implications for the use of these devices by researchers, practitioners and health promoters, which often use the achieve- ment of certain step count goals or increases in step counts as desired outcomes. Evidence pre- sented in this paper adds to the existing literature on the validity of consumer devices for physical activity monitoring and these cautionary limitations should be considered in the design of study data collection and health promotion strategies. Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 10 / 13 Acknowledgments We thank our colleagues and participants for their involvement in this study. Author Contributions Conceptualization: AC FMR. Data curation: AC FMR AG. Formal analysis: AC FMR SN AG. Funding acquisition: FMR. Investigation: AC FMR. Methodology: AC FMR SN. Project administration: AC FMR. Resources: AC FMR MSB MP AG. Software: AC FMR SN AG MSB. Supervision: FMR DK. Validation: AC FMR. Visualization: AC. Writing – original draft: AC FMR SN DK. Writing – review & editing: AC FMR SN DK. References 1. Tudor-Locke C, Johnson WD, Katzmarzyk PT. Accelerometer-determined steps per day in US adults. Med Sci Sports Exerc. 2009; 41(7):1384–91. doi: 10.1249/MSS.0b013e318199885c PMID: 19516163 2. Shephard RJ, Tudor-Locke C. The objective monitoring of physical activity: Contributions of accelero- metry to epidemiology, exercise science and rehabilitation. Cham: Springer; 2016. 3. Tudor-Locke C, Craig CL, Brown WJ, Clemes SA, De Cocker K, Giles-Corti B, et al. How many steps/ day are enough? For adults. Int J Behav Nutr Phys Act. 2011; 8:79. doi: 10.1186/1479-5868-8-79 PMID: 21798015 4. Hardman AE, Stensel DJ. Physical activity and health: The evidence explained. 2nd ed. London: Rout- ledge Taylor and Francis Group; 2009. 5. Tudor-Locke C, Bassett DR Jr. How many steps/day are enough? Preliminary pedometer indices for public health. Sports Med. 2004; 34(1):1–8. PMID: 14715035 6. Bravata DM, Smith-Spangler C, Sundaram V, Gienger AL, Lin N, Lewis R, et al. Using pedometers to increase physical activity and improve health: a systematic review. JAMA. 2007; 298(19):2296–304. doi: 10.1001/jama.298.19.2296 PMID: 18029834 7. Tudor-Locke C, Ham SA. Walking behaviors reported in the American Time Use Survey 2003–2005. JAMA. 2008; 5(5):633–47. 8. Calabr MA, Lee JM, Saint-Maurice PF, Yoo H, Welk GJ. Validity of physical activity monitors for assess- ing lower intensity activity in adults. Int J Behav Nutr Phys Act. 2014; 11(1):1–9. 9. Aadland E, Ylvisåker E. Reliability of the Actigraph GT3X+ accelerometer in adults under free-living conditions. PloS One. 2015; 10(8):e0134606. doi: 10.1371/journal.pone.0134606 PMID: 26274586 10. Freedson P, Bowles HR, Troiano R, Haskell W. Assessment of physical activity using wearable moni- tors: recommendations for monitor calibration and use in the field. Med Sci Sports Exerc. 2012; 44(1 Suppl 1):S1–4. Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 11 / 13 11. Rosenberger ME, Buman MP, Haskell WL, McConnell MV, Carstensen LL. 24 Hours of sleep, seden- tary behavior, and physical activity with nine wearable devices. Med Sci Sports Exerc. 2016; 48(3):457– 65. doi: 10.1249/MSS.0000000000000778 PMID: 26484953 12. Lee JA, Williams SM, Brown DD, Laurson KR. Concurrent validation of the Actigraph GT3X+, Polar Active accelerometer, Omron HJ-720 and Yamax Digiwalker SW-701 pedometer step counts in lab- based and free-living settings. J Sports Sci. 2015; 33(10):991–1000. doi: 10.1080/02640414.2014. 981848 PMID: 25517396 13. Troiano RP, Berrigan D, Dodd KW, Masse LC, Tilert T, McDowell M. Physical activity in the United States measured by accelerometer. Med Sci Sports Exerc. 2008; 40(1):181–8. doi: 10.1249/mss. 0b013e31815a51b3 PMID: 18091006 14. Lee IM, Shiroma EJ. Using accelerometers to measure physical activity in large-scale epidemiologic studies: issues and challenges. Br J Sports Med. 2014; 48(3):197–201. doi: 10.1136/bjsports-2013- 093154 PMID: 24297837 15. Fanning J, Mullen SP, McAuley E. Increasing physical activity with mobile devices: a meta-analysis. J Med Internet Res. 2012; 14(6):e161. doi: 10.2196/jmir.2171 PMID: 23171838 16. Case MA, Burwick HA, Volpp KG, Patel MS. Accuracy of smartphone applications and wearable devices for tracking physical activity data. JAMA. 2015; 313(6):625–6. doi: 10.1001/jama.2014.17841 PMID: 25668268 17. Ranck J. The wearable computing market: a global analysis. Gigaom Pro. 2012. 18. Smart wearables—Statista Dossier. Fitbit—number of active users 2012–2015. 2015. Available: http:// www.statista.com/statistics/472600/fitbit-active-users/. 19. Troiano RP, McClain JJ, Brychta RJ, Chen KY. Evolution of accelerometer methods for physical activity research. Br J Sports Med. 2014; 48(13):1019–23. doi: 10.1136/bjsports-2014-093546 PMID: 24782483 20. Storm FA, Heller BW, Mazza C. Step detection and activity recognition accuracy of seven physical activity monitors. PloS One. 2015; 10(3):e0118723. doi: 10.1371/journal.pone.0118723 PMID: 25789630 21. Diaz KM, Krupka DJ, Chang MJ, Peacock J, Ma Y, Goldsmith J, et al. Fitbit(R): An accurate and reliable device for wireless physical activity tracking. Int J Cardiol. 2015; 185:138–40. doi: 10.1016/j.ijcard.2015. 03.038 PMID: 25795203 22. Nelson MB, Kaminsky LA, Dickin DC, Montoye AH. Validity of consumer-based physical activity moni- tors for specific activity types. Med Sci Sports Exerc. 2016. 23. Schneider M, Chau L. Validation of the Fitbit Zip for monitoring physical activity among free-living ado- lescents. BMC Res Notes. 2016; 9(1):448. doi: 10.1186/s13104-016-2253-6 PMID: 27655477 24. Evenson KR, Goto MM, Furberg RD. Systematic review of the validity and reliability of consumer-wear- able activity trackers. Int J Behav Nutr Phys Act. 2015; 12(1):159. 25. Dierker K, Smith B. Comparison between four personal activity monitors and the Actigraph GT3X+ to measure daily steps. Med Sci Sports Exerc. 2014; 46(5):792. 26. Dominick G, Winfree K, Pohlig R, Papas M. Physical activity assessment between consumer-and research-grade accelerometers: a comparative study in free-living conditions. JMIR MHealth UHealth. 2016; 4(3):e110. doi: 10.2196/mhealth.6281 PMID: 27644334 27. Sushames A, Edwards A, Thompson F, McDermott R, Gebel K. Validity and reliability of fitbit flex for step count, moderate to vigorous physical activity and activity energy expenditure. PloS One. 2016; 11 (9):e0161224. doi: 10.1371/journal.pone.0161224 PMID: 27589592 28. Dontje ML, de Groot M, Lengton RR, van der Schans CP, Krijnen WP. Measuring steps with the Fitbit activity tracker: an inter-device reliability study. J Med Eng Technol. 2015; 39(5):286–90. doi: 10.3109/ 03091902.2015.1050125 PMID: 26017748 29. Abel T, Graf N, Niemann S. Gender bias in the assessment of physical activity in population studies. Soz Praventivmed. 2001; 46(4):268–72. PMID: 11582854 30. Marquez DX, McAuley E. Gender and acculturation influences on physical activity in Latino adults. Ann Behav Med. 2006; 31(2):138–44. doi: 10.1207/s15324796abm3102_5 PMID: 16542128 31. Australian Bureau of Statistics. Women’s participation in sport and physical activity. Canberra, ACT; 2006. 32. Chu AHY, Ng SH, Koh D, Mu¨ller-Riemenschneider F. Reliability and validity of the self-and interviewer- administered versions of the Global Physical Activity Questionnaire (GPAQ). PloS One. 2015; 10(9): e0136944. doi: 10.1371/journal.pone.0136944 PMID: 26327457 Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 12 / 13 33. Choi L, Liu Z, Matthews CE, Buchowski MS. Validation of accelerometer wear and nonwear time classi- fication algorithm. Med Sci Sports Exerc. 2011; 43(2):357–64. doi: 10.1249/MSS.0b013e3181ed61a3 PMID: 20581716 34. Sasaki JE, John D, Freedson PS. Validation and comparison of ActiGraph activity monitors. J Sci Med Sport. 2011; 14(5):411–6. doi: 10.1016/j.jsams.2011.04.003 PMID: 21616714 35. Tudor-Locke C, Barreira TV, Schuna JMJ. Comparison of step outputs for waist and wrist accelerome- ter attachment sites. Med Sci Sports Exerc. 2015; 47(4):839–42. doi: 10.1249/MSS. 0000000000000476 PMID: 25121517 36. Fleiss JL, Levin B, Paik MC. The measurement of interrater agreement—Statistical methods for rates and proportions. 1981; 2:212–236. 37. Tudor-Locke C. Steps to better cardiovascular health: How many steps does it take to achieve good health and how confident are we in this number? Curr Cardiovasc Risk Rep. 2010; 4(4):271–6. doi: 10. 1007/s12170-010-0109-5 PMID: 20672110 38. Barriera TV, Tudor-Locke C, Champagne CM, Broyles ST, Johnson WD, Katzmarzyk PT. Comparison of GT3X accelerometer and YAMAX pedometer steps/day in a free-living sample of overweight and obese adults. J Phys Act Health. 2013; 10(2):263–70. PMID: 22821951 39. Kooiman TJ, Dontje ML, Sprenger SR, Krijnen WP, van der Schans CP, de Groot M. Reliability and validity of ten consumer activity trackers. BMC Sports Sci Med Rehabil. 2015; 7:24. doi: 10.1186/ s13102-015-0018-5 PMID: 26464801 40. Rosenberger ME, Haskell WL, Albinali F, Mota S, Nawyn J, Intille S. Estimating activity and sedentary behavior from an accelerometer on the hip or wrist. Med Sci Sports Exerc. 2013; 45(5):964. doi: 10. 1249/MSS.0b013e31827f0d9c PMID: 23247702 41. Hildebrand M, Van Hees VT, Hansen BH, Ekelund U. Age-group comparability of raw accelerometer output from wrist-and hip-worn monitors. Med Sci Sports Exerc. 2014; 46(9):1816–24. doi: 10.1249/ MSS.0000000000000289 PMID: 24887173 42. Tudor-Locke C, Barreira T, Schuna J, Mire E, Chaput J-P, Fogelholm M, et al. Improving wear time compliance with a 24-hour waist-worn accelerometer protocol in the International Study of Childhood Obesity, Lifestyle and the Environment (ISCOLE). Int J Behav Nutr Phys Act. 2015; 12(1):11. 43. Welk G. Physical activity assessments for health-related research. Champaign, IL: Human Kinetics; 2002. 44. John D, Freedson P. Actigraph and actical physical activity monitors: a peek under the hood. Med Sci Sports Exerc. 2012; 44(1 Suppl 1):S86–S9. Comparison of Fitbit Flex and ActiGraph for steps in free-living adults PLOS ONE | DOI:10.1371/journal.pone.0172535 February 24, 2017 13 / 13

Comparison of wrist-worn Fitbit Flex and waist-worn ActiGraph for measuring steps in free-living adults.

02-24-2017

Chu, Anne H Y,Ng, Sheryl H X,Paknezhad, Mahsa,Gauterin, Alvaro,Koh, David,Brown, Michael S,Müller-Riemenschneider, Falk

eng

PMC10721660

Vol.:(0123456789) Sports Medicine (2023) 53 (Suppl 1):S7–S14 https://doi.org/10.1007/s40279-023-01876-3 REVIEW ARTICLE Carbohydrate Nutrition and Skill Performance in Soccer Ian Rollo1,2  · Clyde Williams2 Accepted: 8 June 2023 / Published online: 8 July 2023 © The Author(s) 2023 Abstract In soccer, players must perform a variety of sport-specific skills usually during or immediately after running, often at sprint speed. The quality of the skill performed is likely influenced by the volume of work done in attacking and defending over the duration of the match. Even the most highly skilful players succumb to the impact of fatigue both physical and mental, which may result in underperforming skills at key moments in a match. Fitness is the platform on which skill is performed during team sport. With the onset of fatigue, tired players find it ever more difficult to successfully perform basic skills. Therefore, it is not surprising that teams spend a large proportion of their training time on fitness. While acknowledging the central role of fitness in team sport, the importance of team tactics, underpinned by spatial awareness, must not be neglected. It is well established that a high-carbohydrate diet before a match and, as a supplement during match play, helps delay the onset of fatigue. There is some evidence that players ingesting carbohydrate can maintain sport-relevant skills for the duration of exercise more successfully compared with when ingesting placebo or water. However, most of the assessments of sport- specific skills have been performed in a controlled, non-contested environment. Although these methods may be judged as not ecologically valid, they do rule out the confounding influences of competition on skill performance. The aim of this brief review is to explore whether carbohydrate ingestion, while delaying fatigue during match play, may also help retain sport soccer-specific skill performance. Key Points The successful execution of repeated skilled actions is a fundamental requirement for soccer performance. Soccer players experience, to different degrees, physical and mental fatigue that have a negative impact on the performance of specific skills. Increasing muscle and liver glycogen stores before and ingesting carbohydrate during competition delays the onset of fatigue and is conducive to maintaining the execution of soccer-specific skills. Ingesting carbohydrate, at key times during competition, could counter negative feelings and improve concentra- tion, helping players maintain skill execution over the duration of exercise. 1 Introduction In soccer, players must perform a variety of sport-specific skills usually during or immediately after running at vari- ous speeds. There is an obvious link between sport-spe- cific fitness and the players’ ability to execute the relevant skill as and when it is appropriate, when defending and attacking. In all sport, skill is used as an umbrella term that includes not only physical performance of a particu- lar skill but also the complex interaction of cognitive and technical abilities to respond to the multitude of scenarios that occur in every match. While technical skills can be taught to the point of being instinctive, the cognitive skill of being able to ‘read the game’ is one that is developed over the sporting lifespan of successful players. Both the skill proficiency of the player and the number of specific technical actions reduce as a match progresses [1, 2]. In addition, the higher the tempo of a match, the sooner players begin to experience both physical (run, sprint, jump) and mental (concentration, decision-making) * Ian Rollo [email protected] 1 Gatorade Sports Science Institute, PepsiCo Life Sciences, Global R&D, Leicestershire, UK 2 School of Sports Exercise and Health Sciences, Loughborough University, Loughborough, UK S8 I. Rollo, C. Williams effects of fatigue, which often results in a decrease in skill performance [3, 4]. This is often to the frustration of coaches as well as spectators, who, for example, observe a misplaced shot, an ill-timed pass or a poor decision just when the team need it least. Therefore, teams dedicate a large proportion of their training time to fitness [5, 6]. Fatigue during prolonged exercise is closely associated with the depletion of the carbohydrate store (glycogen) in skeletal muscles (for full review see Ref. [7]). In a recent study of fatigue in a football match, Mohr et al. reported critically low glycogen levels in the skeletal muscles after 90 min of play and a further significant reduction following 30 min of extra time. Players ran less and per- formed standard skills with less accuracy than earlier in the game [8]. An early reduction in muscle and liver gly- cogen stores, during prolonged exercise, can be prevented by carbohydrate ingestion before and during exercise. Using this nutritional strategy, fatigue is delayed and per- formance sustained for longer than in the absence of this intervention [9]. In addition, several previous reviews have concluded carbohydrate ingestion also facilitates the pres- ervation of skill performance when players are fatigued [10–12]. The aim of this paper is to discuss the most recent studies investigating the effects of carbohydrate inges- tion on soccer-specific skills, and the possible role that carbohydrate ingestion plays in negating the impact that more recently reported mental fatigue has on skill perfor- mance. To inform this review article an electronic litera- ture search was undertaken using three online databases (PubMed, Web of Science, SPORTDiscus). Searches were performed using keywords from existing relevant papers. Search terms were ‘Soccer’, ‘Football’, ‘Carbohydrate’, ‘Skill’ and ‘Performance’ phrased as appropriate. Refer- ence lists of all studies and relevant systematic reviews were examined manually to identify relevant studies for this review. 2 Skill Assessment Skilled movements are physically complex but even more so when performed during match play because they involve an interaction between the physical and cognitive qualities nec- essary to achieve successful outcomes [13]. The acquisition of skills and their retention is a process that begins early in the career of soccer players. By the time they become pro- fessional players they will have achieved superior levels of soccer-specific skills, both technical and cognitive. Further- more, hours of team training and competitions help players consolidate and extend the tactical execution of their skills. Therefore, it is not surprising that the defining characteristics of professional players are their levels of sport-specific skills in addition to their superior physical attributes [14–16]. Traditionally, a team’s and players’ level of soccer-spe- cific skills have been assessed by the ‘experienced eye’ of coaches who know what is expected of professional soccer players. The technical components of skill fall into two large categories: closed (free kick, corners, penalties, throw-in) and open (passing, tackling, heading, goal shooting) skills [17]. In the modern game, skill performance is typically cap- tured via team metrics from competitive matches, for exam- ple, pass completion, interceptions, shots on target, chal- lenges won and number of interceptions [18]. An important metric is ball possession during match play. Individual play- ers must work cohesively to create space, pass and control the ball repeatedly whilst being challenged by the opposi- tion. Although percentage ball possession does not guaran- tee success, those teams with greater percentage ball pos- session perform more passes, touches per possession, shots, dribbles and final-third entries in comparison with teams with low percentage ball possession [19]. On-field analyses allow comparisons of how the speed and skill of the game changes, from match to match and beyond. For example, an analysis of the Men’s World Cup finals between 1966 and 2010 reported a 35% increase in the number of passes per minute of play, which was accompanied by a 15% increase in the speed of the match [20]. Nonetheless, while the team metrics obtained by ever more sophisticated match analysis technology are hugely informative, the impact of training, rehabilitation and nutritional intervention on individual players may be better understood by assessing their skills by objective assessments. Desirable as this is, it is difficult to design objective skill tests that reproduce all that goes into the successful execution of skills in competition. As a result, some studies have used isolated tests of soccer skill, for example, ball juggling [21], wall volley [22], heading [23], shooting [13, 24], passing [24–27] and dribbling [28]. Some laboratory-based studies provide controlled envi- ronments to investigate isolated skills and also attempt to simulate the physical demands of the sport. For example, the Soccer Match Simulation (SMS) protocol embeds soccer- specific skills to enhance the ecological validity of a previ- ously validated simulated assessment of the energy demands of a soccer match [29, 30]. However, while objective tests of skill have many advantages, they are not without several limitations. Rodriguez et al. discuss the importance of play- ing surface on the ecological validity of soccer skills tests [27, 28]. For example, dribbling a ball at speed on a smooth floor is likely a greater challenge than executing this skill on grass. Correspondingly, the footwear worn for differ- ent surfaces may not be optimal for the skill under assess- ment, such as boots versus trainers when testing shooting skill. Furthermore, the use of sport-specific materials that S9 Carbohydrate Nutrition and Skill Performance in Soccer are familiar to players, such as soccer mannequins instead of target boxes, should also be utilised [31]. Ali [17] has described the strengths and limitations of tests of soccer skill performance. 3 Carbohydrate Ingestion and Skill Fitness and skill go ‘hand-in-glove’; as players tire, they are less able to perform the relevant skills when needed [1, 2]. As mentioned earlier, there is a close association between the development of fatigue during a match and the depletion of players’ muscle glycogen stores, which becomes criti- cal should the match go into extra time, extending play to 120 min [8]. Nutritional strategies to increase the body’s glycogen stores by providing carbohydrate before and during exercise improves endurance by delaying the depletion of this essential fuel. The effectiveness of carbohydrate inges- tion applies not only to constant pace running and cycling but also to intermittent high-speed running [9], which is the common activity pattern in team sport, especially in soccer. How much carbohydrate should be consumed, and when, are questions that have led to tried and tested recommendations [5, 28, 32–37] (Table 1). While adopting nutritional strategies to delay a rapid loss of the body’s glycogen stores helps players maintain their work rate during matches, the question is whether it also helps prevent a loss of skill? A simple answer would be that if players tire less readily, after implementing a carbo- hydrate feeding strategy, then they would be better able to execute the necessary skills in match play. Unfortunately, there are too few studies to provide a definitive answer to this question. However, one study reported that when male professional soccer players ingested either a 7% carbo- hydrate–electrolyte or placebo beverage before (5 ml per kilogram body mass) and every 15 min (2 ml per kilogram body mass) during a 90 min on-field soccer match and then completed the assessment of four skills, dribbling speed, coordination, precision and power, there was a significantly improved retention of dribbling speed and precision follow- ing carbohydrate ingestion [38]. In an innovative study on the impact of carbohydrate ingestion on skill, tests were undertaken on players’ domi- nant and non-dominant limbs. Using a soccer-specific pro- tocol, higher passing scores were achieved by both dominant and non-dominant feet following the ingestion of carbohy- drate (30 g, before and at half time, compared with placebo whilst drinking water ad libitum) [27]. This effect was evi- dent from 60 min onwards. Importantly, improved perfor- mance was attained without loss of passing speed, which was better maintained in the non-dominant foot with carbo- hydrate ingestion. This observation is of interest because it is consistent with other studies in sports such as tennis, where Table 1 Carbohydrate intake recommendations for team sport Team sport exercise scenario Objectives Desired adaptation/outcome Suggested daily carbohydrate inges- tion range Considerations In-season training (1 game per week) To delay physical and mental fatigue To maintain physical qualities (and improve where possible/appropriate) To keep players injury and illness free To maintain aerobic and anaerobic fitness To at least maintain strength, power, speed To maintain lean body mass To support physical and technical perfor- mance 4–8 g/kg body mass Range accommodates variations in loads across the micro-cycle (e.g. low load days and match day − 1 carbohydrate loading protocols) as well as individual training goals (e.g. manipulation of body composi- tion to accommodate weight loss and fat loss or weight gain and lean mass gain). Practice competition carbohydrate ingestion regime Match day − 1, match day and match day + 1 6–8 g/kg body mass to elevate muscle glycogen stores Ingest 1–3 g of carbohydrate per kilogram body mass 3–4 h before a match to replenish liver glycogen stores Ingest 30 g of carbohydrate following the warm-up and during the half-time interval Ingest 1 g carbohydrate per kilogram body mass per hour with fluids after a match to start restoration of glycogen and rehydration S10 I. Rollo, C. Williams non-dominant or weaker side (backhand) shots respond posi- tively to carbohydrate ingestion, especially when fatigued [39]. The assessment of complex skilled actions on the non- dominant side may require greater activation of the central nervous system (CNS) and therefore be more susceptible to fatigue [27]. Furthermore non-dominant skilled actions may be more likely influenced by the arousal level of the player [40]. Thus, the performance of players’ non-dominant sides appears to have a greater sensitivity to carbohydrate inges- tion [27], even though the ‘non-dominant’ side is likely to be inferior in performing skills. 4 Carbohydrate Ingestion and Mental Fatigue The physiology of fatigue has been extensively studied [41]. A recent model of motor or cognitive task induced fatigue proposes that no single factor is responsible for declines in skill performance. Instead, fatigue is considered a psycho- physiological condition. Motor fatigue and perceived fatigue are interdependent but hinge on various determinants and depend on modulating factors such as age, sex and specific skill characteristics [42]. Mental fatigue is defined as a psy- chobiological state that arises during prolonged demanding cognitive activity and results in an acute feeling of tired- ness and/or a decreased cognitive ability as well as mood changes [43, 44]. Mental fatigue can reduce physical capac- ity, assessed through reduced time to exhaustion and ele- vated rating of perceived exertion (RPE) [45], and has been shown to fluctuate throughout a competitive season [46]. To highlight this point, mental fatigue has been found, in one review, to have a negative influence on 37% of soccer- specific skills (n = 92) [43]. Mental fatigue has been recognised as a key considera- tion in team sport, due to the associated negative impact on physical, technical, decision-making and tactical perfor- mance [47]. Contributing factors to mental fatigue in team sport environments include but are not limited to prolonged cognitive demands, team meetings, travel and the inability to ‘switch off’ [48, 49]. Of note is the approach taken in laboratory studies which use the repeated execution of inherent sport-specific skills to induce mental fatigue [50]. Thus, tracking skill execution may also be important because it might reflect the presence of both mental and physical fatigue. Correspondingly, moni- toring mental fatigue has been recommended in team sport to provide an overall picture of how players are coping with the demands of training and competition [51]. Therefore, strategies are used to help avoid mental fatigue, for example, displacement activities, such as changes in training routines, environment and, of course, adequate rest and recovery. Increasing dietary carbohydrate while improving exercise capacity both in training and in competition may also be a mood-changing countermeasure to mental fatigue [52, 53]. If players are feeling good rather than bad (pleasure–dis- pleasure) and energized (i.e. in an activated state) before and during matches, then it is more likely that they will per- form better [40, 54]. For example, Backhouse et al. have shown that the ingestion of carbohydrate elevated perceived activation during the final 30 min of 120-min of intermit- tent running exercise [55] and also attenuated the decline in pleasure–displeasure during a 120-min bout of cycling [56]. Administering both a Feeling Scale (FS) and an RPE scale allows a measure of not only ‘what’ (RPE) but also ‘how’ (FS) a person feels [57] but is rarely administered during skill intervention studies or applied settings. A recent review identified mouth rinsing and expectorat- ing a carbohydrate beverage as a potential acute counter- measure to mental fatigue [58]. The recognition of carbo- hydrate in the mouth, when administered immediately after a mentally fatiguing task, was linked to increased excitabil- ity of corticomotor pathways [59, 60]. Furthermore, there appears to be a direct link between improvements in task- specific activity and activation within the primary senso- rimotor cortex in response to oral carbohydrate signalling [61]. These results contribute to a possible explanation for improved high-intensity intermittent running performance in response to mouth rinsing with a 10% carbohydrate bev- erage [62, 63]. Although not all studies report this effect [64], central activation mediated by the ingestion of carbo- hydrate may contribute to the better retention of sprint and technical performance observed early in exercise or in the absence of hypoglycaemia [27, 28, 65]. While mouth rins- ing with a carbohydrate beverage has been shown to benefit complex whole-body skilled actions in fencers, compared with taste-matched placebos [66], the impact on soccer skill performance is yet to be investigated. Furthermore, it is also important to note that mouth rinsing with non-sweet car- bohydrate activates the reward centres of the brain and so may contribute to the ‘feel good’ sensation that may counter mental fatigue [67]. Nevertheless, these findings should be considered as an additional benefit to carbohydrate inges- tion, during or after exercise, when substrate delivery and replenishment of glycogen stores are the respective priorities [68–70]. These responses to carbohydrate ingestion may not be sur- prising bearing in mind that glucose is the main fuel for the brain and CNS [71]. For optimum functioning of the brain and CNS, glucose homeostasis must be maintained even dur- ing a wide range of conditions. Should blood glucose fall to hypoglycaemic levels, then the neural drive to skeletal mus- cles will be compromised; however, it is restored following the ingestion of carbohydrate [72]. During exercise, the rate of glucose release from the liver into the blood increases to match the glucose uptake by contracting muscle [73]. In most S11 Carbohydrate Nutrition and Skill Performance in Soccer team sport, blood glucose concentrations are well maintained over the duration of competition (80–90 min) and extra time (120 min in soccer) in well-fed individuals [74]. Nevertheless, carbohydrate ingestion at the onset of exercise is an effec- tive strategy not only to top up muscle glycogen stores but also because it temporarily inhibits hepatic glucose release in a dose-dependent manner, and so conserves liver glycogen stores [75, 76]. Carbohydrate ingestion, as a means of pre- serving the finite store of liver glycogen, will maintain blood glucose concentrations and performance late in exercise. This strategy is particularly beneficial when matches extend to extra time [8, 77]. Of interest is the observation that elevated blood glucose concentrations are associated with improved skill performance in comparison with euglycaemia [27, 28, 65, 78]. An immediate explanation for this observation is not apparent other than that glucose is a fuel for the brain [79, 80]. However, the brain is sensitive to changes in blood glucose, and the rate of change may act to monitor the availability of whole-body carbohydrate stores. 5 Conclusion Participants in team sport experience, to different degrees, physical and mental fatigue that have a negative impact on the performance of sport-specific skills. The complex series of events between brain and skeletal muscle that interact to minimise the impact of physical and mental fatigue on the performance of skills during competition, following carbo- hydrate feeding, is summarised in Fig. 1. Nutritional strate- gies that increase muscle and liver glycogen stores prior to competition and provide carbohydrate during competition maintain work rate by delaying the onset of fatigue. This effect of carbohydrate ingestion is, in itself, conducive to maintaining the execution of sport-specific skill. Further- more, ingesting carbohydrate, at key times during competi- tion, could counter negative feelings and improve concentra- tion, thereby helping players maintain skill execution over the duration of exercise. Acknowledgements This supplement is supported by the Gatorade Sports Science Institute (GSSI). The supplement was guest edited by Lawrence L. Spriet, who convened a virtual meeting of the GSSI Expert Panel in October 2022 and received honoraria from the GSSI, a division of PepsiCo, Inc., for his participation in the meeting. Dr Spriet received no honoraria for guest editing this supplement. Dr Spriet Fig. 1 Translating thoughts into skilled actions. The electro-chemical chain of events between the brain and skeletal muscles, and how car- bohydrate ingestion may impact skill performance. BM body mass, SR sarcoplasmic reticulum, Ca2+ calcium, Na+/K+ sodium–potassium pump, ATP adenosine triphosphate. ‘+’ = positive influence upon, ‘−’ = negative influence upon. Mood, motivation, RPE [52, 55, 58], facilitation of corticomotor outputs [60, 61], blood glucose availabil- ity, hepatic glycogen preservation [75, 76, 81, 82], muscle innerva- tion: SR calcium handling [83], ATP generation [83–85] S12 I. Rollo, C. Williams suggested peer reviewers for each paper, which were sent to the Sports Medicine Editor-in-Chief for approval, prior to any reviewers being approached. Dr Spriet provided comments on each paper and made an editorial decision based on comments from the peer reviewers and the Editor-in-Chief. Where decisions were uncertain, Dr Spriet consulted with the Editor-in-Chief. The views expressed in this manuscript are those of the authors and do not necessarily reflect the position or policy of PepsiCo, Inc. The authors would like to acknowledge and thank all previous and existing colleagues and collaborators. Declarations Funding This article is based on a presentation by Ian Rollo to the GSSI Expert Panel in October 2022. No honorarium for participation in or preparation of the article for that meeting was provided by the GSSI. No other sources of funding were utilized by the authors in the preparation of the article for this supplement. Conflict of interest Ian Rollo is an employee of the Gatorade Sports Science Institute. However, the views expressed in this manuscript are those of the authors and do not necessarily reflect the position or policy of PepsiCo, Inc. Clyde Williams declares no conflicts of inter- est relevant to the content of this review. While this author previously presented to the GSSI Expert Panel in 2015, and funding for participa- tion in that meeting together with an honorarium were provided by the GSSI, the honorarium was donated to charity. Author contributions IR conceived the idea for this review. IR and CW conducted the literature search and selected the articles for inclusion in the review. IR and CW co-wrote the first draft and revised the original manuscript. Both authors read and approved the final version. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. References 1. Harper LD, West DJ, Stevenson E, Russell M. Technical perfor- mance reduces during the extra-time period of professional soccer match-play. PLoS ONE. 2014;9(10): e110995. 2. Rampinini E, Impellizzeri FM, Castagna C, Coutts AJ, Wisloff U. Technical performance during soccer matches of the Italian Serie A league: effect of fatigue and competitive level. J Sci Med Sport. 2009;12(1):227–33. 3. Mohr M, Krustrup P, Bangsbo J. Fatigue in soccer: a brief review. J Sports Sci. 2005;23(6):593–9. 4. Smith MR, Zeuwts L, Lenoir M, Hens N, De Jong LM, Coutts AJ. Mental fatigue impairs soccer-specific decision-making skill. J Sports Sci. 2016;34(14):1297–304. 5. Anderson L, Orme P, Di Michele R, Close GL, Morgans R, Drust B, Morton JP. Quantification of training load during one-, two- and three-game week schedules in professional soccer players from the English Premier League: implications for car- bohydrate periodisation. J Sports Sci. 2016;34(13):1250–9. 6. Papadakis L, Tymvios C, Patras K. The relationship between training load and fitness indices over a pre-season in professional soccer players. J Sports Med Phys Fit. 2020;60(3):329–37. 7. Mohr M, Vigh-Larsen JF, Krustrup P. Muscle glycogen in Elite Soccer—a perspective on the implication for performance, fatigue, and recovery. Front Sports Active Living. 2022;4: 876534. 8. Mohr M, Ermidis G, Jamustas AZ, Vigh-Larsen J, Poulios A, Draganidis D, Papanikolaou K, Tsimeas P, Batsilas D, Loules G, Batrakoulis A, Sovatzidis A, Nielsen JL, Tzatzakis T, Deli CK, Nybo L, Krustrup P, Fatouros IG. Extended match time exacerbates fatigue and impacts physiological responses in male soccer players. Med Sci Sports Exerc. 2022;55(1):80–92. 9. Williams C, Rollo I. Carbohydrate nutrition and team sport per- formance. Sports Med. 2015;45(Suppl 1):S13-22. 10. Hills SP, Russell M. Carbohydrates for soccer: a focus on skilled actions and half-time practices. Nutrients. 2017;10(1):22–32. 11. Baker LB, Rollo I, Stein KW, Jeukendrup AE. Acute effects of carbohydrate supplementation on intermittent sports perfor- mance. Nutrients. 2015;7(7):5733–63. 12. Russell M, Kingsley M. The efficacy of acute nutritional interventions on soccer skill performance. Sports Med. 2014;44(7):957–70. 13. Haaland E, Hoff J. Non-dominant leg training improves the bilateral motor performance of soccer players. Scand J Med Sci Sports. 2003;13(3):179–84. 14. Slimani M, Nikolaidis PT. Anthropometric and physiological characteristics of male soccer players according to their competi- tive level, playing position and age group: a systematic review. J Sports Med Phys Fit. 2019;59(1):141–63. 15. Cometti G, Maffiuletti NA, Pousson M, Chatard JC, Maffulli N. Isokinetic strength and anaerobic power of elite, subelite and ama- teur French soccer players. Int J Sports Med. 2001;22(1):45–51. 16. Gouveia JN, França C, Martins F, Henriques R, Nascimento MM, Ihle A, Sarmento H, Przednowek K, Martinho D, Gouveia ÉR. Characterization of static strength, vertical jumping, and isokinetic strength in soccer players according to age, competi- tive level, and field position. Int J Environ Res Public Health. 2023;20(3):1799. https:// doi. org/ 10. 3390/ ijerp h2003 1799 17. Ali A. Measuring soccer skill performance: a review. Scand J Med Sci Sports. 2011;21(2):170–83. 18. Andrzejewski M, Oliva-Lozano JM, Chmura P, Chmura J, Czarniecki S, Kowalczuk E, Rokita A, Muyor JM, Konefal M. Analysis of team success based on match technical and running performance in a professional soccer league. BMC Sports Sci Med Rehabil. 2022;14(1):82. 19. Bradley PS, Lago-Penas C, Rey E, Gomez Diaz A. The effect of high and low percentage ball possession on physical and technical profiles in English FA Premier League soccer matches. J Sports Sci. 2013;31(12):1261–70. 20. Wallace JL, Norton KI. Evolution of World Cup soccer final games 1966–2010: game structure, speed and play patterns. J Sci Med Sport. 2014;17(2):223–8. 21. Hoare DG, Warr CR. Talent identification and women’s soccer: an Australian experience. J Sports Sci. 2000;18(9):751–8. 22. Vanderford ML, Meyers MC, Skelly WA, Stewart CC, Hamilton KL. Physiological and sport-specific skill response of Olympic youth soccer athletes. J Strength Cond Res. 2004;18(2):334–42. 23. Rosch D, Hodgson R, Peterson TL, Graf-Baumann T, Junge A, Chomiak J, Dvorak J. Assessment and evaluation of football per- formance. Am J Sports Med. 2000;28(5 Suppl):S29-39. 24. Ali A, Williams C. Carbohydrate ingestion and soccer skill per- formance during prolonged intermittent exercise. J Sports Sci. 2009;27(14):1499–508. S13 Carbohydrate Nutrition and Skill Performance in Soccer 25. Rostgaard T, Iaia FM, Simonsen DS, Bangsbo J. A test to evaluate the physical impact on technical performance in soccer. J Strength Cond Res. 2008;22(1):283–92. 26. Bendiksen M, Bischoff R, Randers MB, Mohr M, Rollo I, Suetta C, Bangsbo J, Krustrup P. The Copenhagen Soccer Test: physi- ological response and fatigue development. Med Sci Sports Exerc. 2012;44(8):1595–603. 27. Rodriguez-Giustiniani P, Rollo I, Witard OC, Galloway SDR. Ingesting a 12% carbohydrate-electrolyte beverage before each half of a soccer match simulation facilitates retention of passing performance and improves high-intensity running capacity in academy players. Int J Sport Nutr Exerc Metab. 2019;29(4):397–405. 28. Harper LD, Stevenson EJ, Rollo I, Russell M. The influence of a 12% carbohydrate-electrolyte beverage on self-paced soccer- specific exercise performance. J Sci Med Sport. 2017;12:1123–9. 29. Russell M, Rees G, Benton D, Kingsley M. An exercise pro- tocol that replicates soccer match-play. Int J Sports Med. 2011;32(7):511–8. 30. Nicholas CW, Nuttall FE, Williams C. The Loughborough Inter- mittent Shuttle Test: a field test that simulates the activity pattern of soccer. J Sports Sci. 2000;18(2):97–104. 31. Rodriguez-Giustiniani P, Rollo I, Galloway SDR. A preliminary study of the reliability of soccer skill tests within a modified soc- cer match simulation protocol. Sci Med Footb. 2021;6(3):363–71. 32. Collins J, Maughan RJ, Gleeson M, Bilsborough J, Jeukendrup A, Morton JP, Phillips SM, Armstrong L, Burke LM, Close GL, Duffield R, Larson-Meyer E, Louis J, Medina D, Meyer F, Rollo I, Sundgot-Borgen J, Wall BT, Boullosa B, Dupont G, Lizarraga A, Res P, Bizzini M, Castagna C, Cowie CM, D’Hooghe M, Geyer H, Meyer T, Papadimitriou N, Vouillamoz M, McCall A. UEFA expert group statement on nutrition in elite football current evidence to inform practical recommendations and guide future research. Br J Sports Med. 2021;55(8):416. 33. Thomas DT, Erdman KA, Burke LM. Position of the Academy of Nutrition and Dietetics, Dietitians of Canada, and the American College of Sports Medicine: nutrition and athletic performance. J Acad Nutr Diet. 2016;116(3):501–28. 34. Rollo I, Randell RK, Baker L, Leyes JY, Medina Leal D, Lizarraga A, Mesalles J, Jeukendrup AE, James LJ, Carter JM. Fluid bal- ance, sweat Na+ losses, and carbohydrate intake of elite male soccer players in response to low and high training intensities in cool and hot environments. Nutrients. 2021;13(2):401. https:// doi. org/ 10. 3390/ nu130 20401 35. Moss SL, Randell RK, Burgess D, Ridley S, ÓCairealláin C, Alli- son R, Rollo I. Assessment of energy availability and associated risk factors in professional female soccer players. Eur J Sport Sci. 2020;6:861–70. 36. Funnell MP, Dykes NR, Owen EJ, Mears SA, Rollo I, James LJ. Ecologically valid carbohydrate intake during soccer-specific exercise does not affect running performance in a fed state. Nutri- ents. 2017;9(1):39. https:// doi. org/ 10. 3390/ nu901 0039 37. Burke LM, Hawley JA, Wong SH, Jeukendrup AE. Carbohy- drates for training and competition. J Sports Sci. 2011;29(Suppl 1):S17-27. 38. Ostojic SM, Mazic S. Effects of a carbohydrate-electrolyte drink on specific soccer tests and performance. J Sports Sci Med. 2002;1(2):47–53. 39. McRae KA, Galloway SD. Carbohydrate-electrolyte drink inges- tion and skill performance during and after 2 hr of indoor tennis match play. Int J Sport Nutr Exerc Metab. 2012;22(1):38–46. 40. McMorris T, Graydon J. The effect of exercise on cogni- tive performance in soccer-specific tests. J Sports Sci. 1997;15(5):459–68. 41. Enoka RM, Baudry S, Rudroff T, Farina D, Klass M, Duchateau J. Unraveling the neurophysiology of muscle fatigue. J Electromyogr Kinesiol. 2011;21(2):208–19. 42. Behrens M, Gube M, Chaabene H, Prieske O, Zenon A, Broscheid KC, Schega L, Husmann F, Weippert M. Fatigue and human per- formance: an updated framework. Sports Med. 2023;53(1):7–31. https:// doi. org/ 10. 1007/ s40279- 022- 01748-2 43. Habay J, Van Cutsem J, Verschueren J, De Bock S, Proost M, De Wachter J, Tassignon B, Meeusen R, Roelands B. Mental fatigue and sport-specific psychomotor performance: a systematic review. Sports Med. 2021;51(7):1527–48. 44. Roelands B, Kelly V, Russell S, Habay J. The physiological nature of mental fatigue: current knowledge and future avenues for sport science. Int J Sports Physiol Perform. 2022;17(2):149–50. 45. Marcora SM, Staiano W, Manning V. Mental fatigue impairs physical performance in humans. J Appl Physiol (1985). 2009;106(3):857–64. 46. Russell S, Jenkins DG, Halson SL, Juliff LE, Kelly VG. How do elite female team sport athletes experience mental fatigue? Com- parison between international competition, training and prepara- tion camps. Eur J Sport Sci. 2022;22(6):877–87. 47. Smith MR, Thompson C, Marcora SM, Skorski S, Meyer T, Coutts AJ. Mental fatigue and soccer: current knowledge and future directions. Sports Med. 2018;48(7):1525–32. 48. Thompson CJ, Noon M, Towlson C, Perry J, Coutts AJ, Harper LD, Skorski S, Smith MR, Barrett S, Meyer T. Understanding the presence of mental fatigue in English academy soccer players. J Sports Sci. 2020;38(13):1524–30. 49. Thompson CJ, Smith A, Coutts AJ, Skorski S, Datson N, Smith MR, Meyer T. Understanding the presence of mental fatigue in elite female football. Res Q Exerc Sport. 2022;93(3):504–15. 50. Bian C, Ali A, Nassis GP, Li Y. Repeated interval Loughborough soccer passing tests: an ecologically valid motor task to induce mental fatigue in soccer. Front Physiol. 2021;12: 803528. 51. Thompson CJ, Fransen J, Skorski S, Smith MR, Meyer T, Bar- rett S, Coutts AJ. Mental fatigue in football: is it time to shift the goalposts? An evaluation of the current methodology. Sports Med. 2019;49(2):177–83. 52. Achten J, Halson SL, Moseley L, Rayson MP, Casey A, Jeuken- drup AE. Higher dietary carbohydrate content during intensified running training results in better maintenance of performance and mood state. J Appl Physiol. 2004;96(4):1331–40. 53. Killer SC, Svendsen IS, Jeukendrup AE, Gleeson M. Evidence of disturbed sleep and mood state in well-trained athletes during short-term intensified training with and without a high carbohy- drate nutritional intervention. J Sports Sci. 2017;35(14):1402–10. 54. Acevedo E, Gill D, Goldfarb A, Boyer B. Affect and per- ceived exertion during a two-hour run. Int J Sport Psychol. 1996;27:286–92. 55. Backhouse SH, Ali A, Biddle SJ, Williams C. Carbohydrate ingestion during prolonged high-intensity intermittent exercise: impact on affect and perceived exertion. Scand J Med Sci Sports. 2007;17(5):605–10. 56. Backhouse SH, Bishop NC, Biddle SJ, Williams C. Effect of car- bohydrate and prolonged exercise on affect and perceived exer- tion. Med Sci Sports Exerc. 2005;37(10):1768–73. 57. Hardy CJ, Rejeski W. Not what, but how ones feels: the meas- urement of affect during exercise. J Sport Exerc Psychol. 1989;11:304–17. 58. Proost M, Habay J, De Wachter J, De Pauw K, Rattray B, Meeusen R, Roelands B, Van Cutsem J. How to tackle mental fatigue: a sys- tematic review of potential countermeasures and their underlying mechanisms. Sports Med. 2022;52(9):2129–58. 59. Bailey SP, Harris GK, Lewis K, Llewellyn TA, Watkins R, Weaver MA, Roelands B, Van Cutsem J, Folger SF. Impact of a carbohy- drate mouth rinse on corticomotor excitability after mental fatigue S14 I. Rollo, C. Williams in healthy college-aged subjects. Brain Sci. 2021;11(8):972. https:// doi. org/ 10. 3390/ brain sci11 080972 60. Gant N, Stinear CM, Byblow WD. Carbohydrate in the mouth immediately facilitates motor output. Brain Res. 2010;1350:151–8. 61. Turner CE, Byblow WD, Stinear CM, Gant N. Carbohydrate in the mouth enhances activation of brain circuitry involved in motor performance and sensory perception. Appetite. 2014;80:212–9. 62. Rollo I, Homewood G, Williams C, Carter J, Goosey-Tolfrey VL. The influence of carbohydrate mouth rinse on self-selected intermittent running performance. Int J Sport Nutr Exerc Metab. 2015;25(6):550–8. 63. Kasper AM, Cocking S, Cockayne M, Barnard M, Tench J, Parker L, McAndrew J, Langan-Evans C, Close GL, Morton JP. Carbo- hydrate mouth rinse and caffeine improves high-intensity interval running capacity when carbohydrate restricted. Eur J Sport Sci. 2016;16(5):560–8. 64. Gough LA, Faghy M, Clarke N, Kelly AL, Cole M, Lun Foo W. No independent or synergistic effects of carbohydrate-caffeine mouth rinse on repeated sprint performance during simulated soccer match play in male recreational soccer players. Sci Med Footb. 2022;6(4):519–27. 65. Ali A, Williams C, Nicholas CW, Foskett A. The influence of carbohydrate-electrolyte ingestion on soccer skill performance. Med Sci Sports Exerc. 2007;39(11):1969–76. 66. Rowlatt G, Bottoms L, Edmonds CJ, Buscombe R. The effect of carbohydrate mouth rinsing on fencing performance and cogni- tive function following fatigue-inducing fencing. Eur J Sport Sci. 2017;17(4):433–40. 67. Chambers ES, Bridge MW, Jones DA. Carbohydrate sensing in the human mouth: effects on exercise performance and brain activity. J Physiol. 2009;587(Pt 8):1779–94. 68. Rollo I, Gonzalez JT, Fuchs CJ, van Loon LJC, Williams C. Pri- mary, secondary, and tertiary effects of carbohydrate ingestion during exercise. Sports Med. 2020;50(11):1863–71. 69. Erith S, Williams C, Stevenson E, Chamberlain S, Crews P, Rushbury I. The effect of high carbohydrate meals with different glycemic indices on recovery of performance during prolonged intermittent high-intensity shuttle running. Int J Sport Nutr Exerc Metab. 2006;16(4):393–404. 70. Rollo I, Williams C, Nevill M. Influence of ingesting versus mouth rinsing a carbohydrate solution during a 1-h run. Med Sci Sports Exerc. 2011;43(3):468–75. 71. Mergenthaler P, Lindauer U, Dienel GA, Meisel A. Sugar for the brain: the role of glucose in physiological and pathological brain function. Trends Neurosci. 2013;36(10):587–97. 72. Nybo L. CNS fatigue and prolonged exercise: effect of glucose supplementation. Med Sci Sports Exerc. 2003;35(4):589–94. 73. Wasserman DH. Four grams of glucose. Am J Physiol-Endocrinol Metab. 2009;296(1):E11-21. 74. Harper LD, Briggs MA, McNamee G, West DJ, Kilduff LP, Stevenson E, Russell M. Physiological and performance effects of carbohydrate gels consumed prior to the extra-time period of prolonged simulated soccer match-play. J Sci Med Sport. 2016;19(6):509–14. 75. Jeukendrup AE, Wagenmakers AJ, Stegen JH, Gijsen AP, Brouns F, Saris WH. Carbohydrate ingestion can completely suppress endogenous glucose production during exercise. Am J Physiol. 1999;276(4 Pt 1):E672–83. 76. Newell ML, Wallis GA, Hunter AM, Tipton KD, Galloway SDR. Metabolic responses to carbohydrate ingestion during exercise: associations between carbohydrate dose and endurance perfor- mance. Nutrients. 2018;10(1):37. https:// doi. org/ 10. 3390/ nu100 10037 77. Field A, Naughton RJ, Haines M, Lui S, Corr LD, Russell M, Page RM, Harper LD. The demands of the extra-time period of soccer: a systematic review. J Sport Health Sci. 2022;11(3):403–14. 78. Ali A, Williams C. Carbohydrate ingestion and soccer skill per- formance during prolonged intermittent exercise. J Sports Sci. 2009;27(14):1499–508. https:// doi. org/ 10. 1080/ 02640 41090 33347 72 79. Lopez-Gambero AJ, Martinez F, Salazar K, Cifuentes M, Nualart F. Brain glucose-sensing mechanism and energy homeostasis. Mol Neurobiol. 2019;56(2):769–96. 80. van Praag H, Fleshner M, Schwartz MW, Mattson MP. Exercise, energy intake, glucose homeostasis, and the brain. J Neurosci. 2014;34(46):15139–49. 81. Gonzalez JT, Fuchs CJ, Betts JA, van Loon LJ. Liver glycogen metabolism during and after prolonged endurance-type exercise. Am J Physiol-Endocrinol Metab. 2016;311(3):E543–53. 82. Fuchs CJ, Gonzalez JT, Beelen M, Cermak NM, Smith FE, Thelwall PE, Taylor R, Trenell MI, Stevenson EJ, van Loon LJ. Sucrose ingestion after exhaustive exercise accelerates liver, but not muscle glycogen repletion compared with glucose ingestion in trained athletes. J Appl Physiol (1985). 2016;120(11):1328–34. 83. Ortenblad N, Nielsen J, Saltin B, Holmberg HC. Role of glyco- gen availability in sarcoplasmic reticulum Ca2+ kinetics in human skeletal muscle. J Physiol. 2011;589(Pt 3):711–25. 84. Duhamel TA, Stewart RD, Tupling AR, Ouyang J, Green HJ. Mus- cle sarcoplasmic reticulum calcium regulation in humans during consecutive days of exercise and recovery. J Appl Physiol (1985). 2007;103(4):1212–20. 85. Nielsen J, Holmberg HC, Schroder HD, Saltin B, Ortenblad N. Human skeletal muscle glycogen utilization in exhaustive exer- cise: role of subcellular localization and fibre type. J Physiol. 2011;589(Pt 11):2871–85.

Carbohydrate Nutrition and Skill Performance in Soccer.

07-08-2023

Rollo, Ian,Williams, Clyde

eng

PMC8171865

RESEARCH ARTICLE Effects of a period without mandatory physical training on maximum oxygen uptake and anthropometric parameters in naval cadets A´ lvaro Huerta OjedaID*☯, Guillermo Barahona-FuentesID☯, Sergio Galdames Maliqueo☯ Grupo de Investigacio´n en Salud, Actividad Fı´sica y Deporte ISAFYD, Escuela de Educacio´n Fı´sica, Universidad de Las Ame´ricas, sede Viña del Mar, Chile ☯ These authors contributed equally to this work. * [email protected] Abstract The effects of a period without physical training on the civilian population are well estab- lished. However, no studies show the effects of a period without mandatory physical training on maximum oxygen uptake (VO2 max) and anthropometric parameters in naval cadets. This study aimed to investigate changes in VO2 max and anthropometric parameters after 12 weeks without mandatory physical training in naval cadets. The sample was 38 healthy and physically active naval cadets. The measured variables, including VO2 max and anthro- pometric parameters, were evaluated through the 12-minute race test (12MRT) and the somatotype. Both variables had a separation of 12 weeks without mandatory physical train- ing. A t-test for related samples was used to evidence changes between the test and post- test; effect size was calculated through Cohen’s d-test. Distance in 12MRT and VO2 max showed significant decreases at the end of 12 weeks without mandatory physical training (p < 0.001). Likewise, the tricipital skinfold thickness and the endomorphic component showed significant increases (p < 0.05). 12 weeks without mandatory physical training significantly reduces the VO2 max in naval cadets. Simultaneously, the same period without physical training increases both the tricipital skinfold thickness and the endomorphic component in this population. Introduction Increased physical capabilities through strength training [1, 2] and aerobic capacity [3] have been associated with health, quality of life, and sports performance benefits [1–3]. In this sense, people included in strength training have shown neuronal and morphological adapta- tions [4]; these two adaptations, generated by strength training, allow for the improvement of both the metabolic health [5] and the quality of life of people [6]. At the same time, aerobic training has reported significant decreases in cardiovascular risk factors [7], as well as an increase in maximum oxygen uptake (VO2 max) [3]. Specifically, the VO2 max has a direct PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 1 / 15 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Huerta Ojeda A´, Barahona-Fuentes G, Galdames Maliqueo S (2021) Effects of a period without mandatory physical training on maximum oxygen uptake and anthropometric parameters in naval cadets. PLoS ONE 16(6): e0251516. https:// doi.org/10.1371/journal.pone.0251516 Editor: Randy Wayne Bryner, West Virginia University, UNITED STATES Received: October 10, 2020 Accepted: April 27, 2021 Published: June 2, 2021 Copyright: © 2021 Huerta Ojeda et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: The data underlying this study are publicly available at: https://doi.org/ 10.6084/m9.figshare.14049590. Funding: The author(s) received no specific funding for this work. Competing interests: The authors have declared that no competing interests exist. association with the quality of life of people [8]. These morphological and metabolic changes, triggered by force training or aerobic training, are experienced by both the civilian population [9] and the military and naval population [10–12]; in the latter, they provide specific physical characteristics that allow missions to be carried out efficiently and with a low risk of injury [13]. Scientific evidence shows that physical training acts as a physiological stressor, increasing energy expenditure [14], anabolic hormone concentrations [15], arterial diameter, and blood flow [16]. These responses to physical training contribute to a physiological adaptation of the body [17], specifical adaptations of muscles [18], and bone tissues [19]. In this sense, a recently published meta-analysis showed the benefits of eccentric strength training through isoinertial devices; the study results showed increases in strength, power, and muscle size with this train- ing [20]. Concerning aerobic training, these stimuli have been considered as the primary method to improve markers of cardiorespiratory fitness, mainly VO2 max [21]. Additionally, physical training carried out regularly, and with the principles of intensity, volume, and fre- quency, will minimize muscular fatigue [22] and favor the physiological adaptations of the body [17]. Despite the above, there is also a transition phase in sports periodization [23]; this stage corresponds to the interruption of physical training [24], which can be short term (less than four weeks) or long term (more than four weeks) [25]. However, if professionals do not control the transition phase, there is a high probability of provoking a detraining [25]. In this way, a period without physical training can generate a partial or total loss of morphological adaptations, physiological adaptations, and physical performance [26], as well as cause alter- ations in the psychological well-being of the population [27]. The sports transition phase is an opportunity for the physical recovery of athletes [23]. However, there are unplanned situations that generate periods of non-physical training in the population [28–30], for example, the period of vacation experienced by students each year [28] or the current period of confinement generated by COVID-19 [30]. Regardless of the reasons, an extended time-period without physical training has been shown to negatively influence ath- letes’ body composition [23], increasing fat mass and decreasing lean mass [31–33]. It has also been shown that a period without physical training of fewer than eight weeks leads to a decrease in muscle cross-section [34], decreases in maximum strength [35], and a reduction in VO2 max in both the civilian [36] and naval [37] populations. Currently, naval personnel has been the subject of several research studies [38, 39]. One of the reasons for the growing number of investigations in this sample is that the Chilean Navy comprises more than 25,000 personnel. Of this number, 9.6% (equivalent to 2,400 personnel) corresponds to naval officers, all trained at the Arturo Prat Naval Academy [40]. These figures show several aspects, such as the high number of officers [40] and, therefore, the need for this population to be studied from a psychological [11, 13], health [12] and physical [10, 38] perfor- mance perspective. This last dimension includes the transition phase considering that we hypothesize that naval cadets decrease their physical condition, associated with VO2 max and anthropometric parameters, after a period without mandatory physical training; thus, with correctly applied training loads, physical fitness loss in this phase could be avoided [23–25]. Despite the existence of studies showing a decrease in the physical condition and anthropo- metric parameters after a period without physical training in some segments of the population [23, 31–33], the available evidence in the naval population is scarce and limited [37]. Likewise, and as far as knowledge goes, no studies evidence the effects of periods without physical train- ing on VO2 max and anthropometric parameters in naval cadets from 18 to 25 years old. Con- sequently, this study aimed to evidence the changes in VO2 max and anthropometric parameters after 12 weeks without mandatory physical training in naval cadets from 18 to 25 years old. PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 2 / 15 Materials and methods Research design This study was empirical research with a manipulative, quasi-experimental strategy with a lon- gitudinal design with repeated means [41]. To highlight the changes in VO2 max and anthro- pometric parameters, the 12-minute race test (12MRT) and the somatotype were evaluated 12 weeks apart, a period without mandatory physical training (Fig 1). Procedures As a first action, all participants who voluntarily accepted to be part of the study (non-probabi- listic sample) were recruited. The purpose and procedures of the study were indicated in an informative talk. The inclusion criteria were that all participants should be healthy, physically active [21] and between 18 and 25 years of age, while the exclusion criteria were: prevalence of musculoskeletal injuries, pre-existing cardiac pathologies, abnormal respiratory and cardiac responses during the familiarization period and inability to perform the 12MRT. All partici- pants were asked not to engage in physical training that would generate nervous or musculo- skeletal fatigue 48 hours before the measurements and refrain from ingesting caffeine or any substance that could increase their metabolism during the assessment. Finally, only those par- ticipants who signed informed consent were subjected to 12MRT and somatotype evaluations. Participants Thirty-eight healthy and physically active naval cadets volunteered to participate in this study (Table 1). The type of sampling was non-probabilistic for convenience. All participants were informed of the study objective and possible risks of the experiment. Indeed, all participants signed the informed consent form before the implementation of the protocols. The informed consent and the study were approved by the Human Research Committee of the University of Las Americas (registry number CEC-FP-2020011). The informed consent and the study were conducted under the Declaration of Helsinki (WMA 2000, Bosˇnjak 2001, Tyebkhan 2003), which sets out the fundamental ethical principles for research with human subjects. Fig 1. Research design. 12MRT: 12-minute race test. https://doi.org/10.1371/journal.pone.0251516.g001 Table 1. Characterization of the participants. Women (n = 8) Men (n = 30) All (n = 38) mean ± SD (min–max) Mean ± SD (min–max) mean ± SD (min–max) Age (years) 21.0 ± 1.51 (19–23) 20.5 ± 1.22 (18–24) 20.6 ± 1.28 (18–24) BMI (kg/m2) 21.9 ± 1.79 (20.2–25.5) 22.7 ± 1.69 (20.4–26.7) 22.5 ± 1.72 (20.2–26.7) % Fat 23.3 ± 4.7 (18.5–33.1) 12.6 ± 2.2 (9.3–18.1) 14.9 ± 5.2 (9.3–33.1) VO2 max (mLO2
kg–1
min–1) 46.7 ± 3.9 (42.6–51.5) 59.3 ± 4.7 (50.9–65.5) 56.6 ± 6.9 (42.6–65.5) SD: standard deviation; kg/m2: kilograms per square meters; min: minimum; max: maximum; %: percentage; VO2 max: maximum oxygen uptake; mLO2
kg–1
min–1: milliliters of oxygen per kilogram of body mass per minute. https://doi.org/10.1371/journal.pone.0251516.t001 PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 3 / 15 Somatotype evaluation The somatotype corresponds to the shape of the human body. It is obtained by analyzing the arm and leg’s circumferences, the humerus and femur’s diameters, four skinfolds (tricipital, subscapular, supra-iliac, and mid-calf), and the weight and height of a person. Body shape can be represented two-dimensionally through the somatochart or three-dimensionally through the compogram; the latter representation corresponds to three numerical values representing the endomorphic, mesomorphic, and ectomorphic components of a participant (always in that order) [42]. To represent a participant’s morphology, Berral [42] recommends using both the somatochart and the compogram since using only the somatochart can generate an error in interpreting the results; for example, values 3–5–3 and 4–6–4 would be represented with the same point on the somatochart [42]. Body mass and height. The method used to determine the participants’ somatotype was pro- posed by Carter & Heath [43]. The body mass (kg) was evaluated through a Tanita Inner Scan BC-5541 digital scale, with the participants barefoot, in shorts, and wearing a light shirt. The height was measured through a Seca1 stadiometer from the feet to the vertex (Frankfort plane) [44]. Circumferences. Arm and leg circumferences, humeral and femoral diameters, and skin folds were evaluated with the FAGA SLR1 anthropometric kit. The circumference of the right leg was evaluated in this segment’s bulkiest area, in a standing position and with the gas- trocnemius relaxed; in contrast, the circumference of the right arm was evaluated in the bulki- est area of the contracted biceps; this evaluation was performed standing with the elbow in front and bent at 90 [43]. Diameters. The humeral epicondyle distance was considered the humerus’s diameter, which is the distance between the epicondyle and the right arm’s epitrochlea. For this evalua- tion, participants were standing with the elbow bent at 90˚. The distance between the femoral condyles (medial and distal) was considered the femur’s diameter, which evaluation was per- formed in a sitting position with the right knee bent at 90˚ [43, 44]. Skinfold thickness. Four skinfolds were considered to determine the participants’ somatotype: tricipital, subscapular, supra-iliac, and mid-calf [43–45]. Body Mass Index (BMI). The BMI’s interpretation was made according to anthropomet- ric standards to evaluate nutritional status [46]. Percentage of fat (%). The fat percentage was evaluated through impedance measurement with the Tanita Inner Scan BC-5541 digital scale. Waist-Hip Index (WHI). The WHI was obtained by dividing the waist perimeter, mea- sured at a point equidistant from the lower edge of the last rib and the iliac crest, by the perim- eter of the hips, measured at the greatest prominence of the buttocks [44, 47]. 12 weeks without mandatory physical training In regular class periods, the naval cadets had an average of two hours of daily mandatory phys- ical training (Monday through Saturday). This physical training was mandatory and consid- ered loads with the orientation to all physical capacities (strength, power, flexibility, speed, aerobic capacity and aerobic power). However, upon leaving school, whether for vacation or unplanned situations such as the current COVID 19 pandemic [30], the physical training regi- men was not mandatory. During the 12 weeks without mandatory physical training, the naval cadets voluntarily took part in walking, cycling, and ball games, among other activities. Standardized warm-up For both the first and the second evaluation of the 12MRT, the warm-up consisted of 10 min- utes of jogging, then 5 minutes of ballistic movements of the lower limb (adduction, abduction, PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 4 / 15 flexion, and extension of hips, and flexion and extension of knees and ankles). To finish, par- ticipants performed three 80-meter accelerations. After this warm-up and before running the 12MRT, there was a 5-minute break. 12-minute race test The evaluation of the 12MRT was carried out on a 400-meter athletic track. Before the evalua- tion, participants were instructed to perform as much distance as possible within the test’s 12 minutes. During the application of the test, the participants received verbal incentives from the researchers. The distance achieved in meters was converted into kilometers, and then the VO2 max was obtained through the following equation [48]: VO2 max ðmLO2
kgthe tricipital skinfold participants, a very large, negative correlation was observed between both variables (r = -0.76, p = 0.01). At the end of the 12 weeks without mandatory physical training (post-test), a very large, negative correlation was observed between VO2 max and the participants’ tricipital skinfold (r = -0.81, p = 0.01). The graphic representation of these analy- ses is presented in Fig 4. Discussion Concerning this study’s primary objective, the variables of VO2 max and anthropometric parameters showed changes after the 12 weeks without mandatory physical training in naval cadets from 18 to 25 years old. The findings revealed that the analysis initial point relates phys- ical training to quality of life [6, 8] and sports performance [1–3]. In this way, detrimental physiological changes and a decline in performance observed after a period without physical training can be reversed by applying correct training loads and professional supervision [17]. Specifically, the present study’s findings showed a significant decrease in the VO2 max of naval cadets, both men and women, after 12 weeks without mandatory physical training (p < 0.001, ES = 0.34). Similarly, Liguori et al. [37] determined changes in VO2 max after a vacation period without mandatory training; at the end of the vacation period, the researchers reported signifi- cant decreases in relative (p = 0.009) and absolute (p = 0.001) VO2 max in both men and women. Likewise, Sotiropoulos et al. [33] evaluated changes in VO2 max after a four-week transition period in soccer players. The experimental group (EG) conducted a directed Table 2. Mean values and SD before and after 12 weeks without mandatory physical training (n = 38). Test mean ± SD Post test mean ± SD Related differences Mean SD SEM 95% confidence interval t p d Lower Upper Weight (kg) 67.1 ± 8.0 67.5 ± 8.3 -0.32 1.78 0.28 -0.91 0.25 -1.13 ns 0.01 BMI (kg/m2) 22.5 ± 1.7 22.7 ± 1.8 -0.16 0.58 0.09 -0.35 0.02 -1.78 ns 0.10 % Fat 14.9 ± 5.2 14.9 ± 5.4 0.05 1.24 0.2 -0.35 0.46 0.26 ns 0.01 WHI 0.84 ± 0.05 0.83 ± 0.04 0.00 0.03 0.00 0.00 0.01 0.71 ns 0.08 WHeI 0.46 ± 0.03 0.46 ± 0.02 0.00 0.01 0.00 0.00 0.00 0.84 ns 0.08 Tricipital skinfold (mm) 11.1 ± 3.9 11.8 ± 4.0 -0.69 1.83 0.29 -1.29 -0.09 -2.34 ns 0.18 Subscapular skinfold (mm) 10.7 ± 3.1 10.9 ± 3.0 -0.26 1.32 0.21 -0.7 0.17 -1.22 ns 0.09 Suprailiac skinfold (mm) 9.4 ± 3.4 10.4 ± 3.8 -0.97 3.02 0.49 -1.97 0.01 -1.99 ns 0.27 Mid-calf skinfold (mm) 10.2 ± 4.6 9.9 ± 3.6 0.30 2.34 0.37 -0.46 1.07 0.79 ns 0.07 Arm circumference (cm) 31.6 ± 2.9 31.8 ± 3.0 -0.16 1.62 0.26 -0.7 0.36 -0.62 ns 0.06 Leg circumference (cm) 36.7 ± 2.0 36.8 ± 2.1 -0.11 0.77 0.12 -0.37 0.13 -0.91 ns 0.06 Humerus diameter 6.77 ± 0.42 6.76 ± 0.40 0.00 0.15 0.02 -0.04 0.05 0.21 ns 0.01 Femur diameter 9.76 ± 0.53 9.69 ± 0.52 0.06 0.17 0.02 0.01 0.12 2.4 ns 0.13 Endomorphic component 3.12 ± 0.96 3.32 ± 1.00 -0.20 0.55 0.08 -0.38 -0.02 -2.32 ns 0.21 Mesomorphic component 5.07 ± 0.96 5.10 ± 0.93 -0.02 0.39 0.06 -0.15 0.10 -0.41 ns 0.03 Ectomorphic component 2.51 ± 0.76 2.44 ± 0.77 0.06 0.27 0.04 -0.02 0.15 1.39 ns 0.08 12MRT (m) 3100.8 ± 348.6 2978.1 ± 364.7 122 115 18.6 84.9 160.5 6.57  0.34 VO2 max (mLO2
kg–1
min–1) 56.6 ± 6.9 54.2 ± 7.2 2.45 2.3 0.37 1.69 3.21 6.57  0.34 SD: standard deviation; SEM: standard error of the mean; WHI: waist-hip index; WHeI: waist-height index; BMI: muscle mass index; kg/m2: kilograms per square meters; 12MRT: 12-minute race test; mm: millimeters; cm: centimeters; m: meters; VO2 max: maximum oxygen consumption; mLO2
kg–1
min–1: milliliters of oxygen per kilogram of body mass per minute  p < 0.002; ns: not significant; d: Cohen’s d. https://doi.org/10.1371/journal.pone.0251516.t002 PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 6 / 15 training program, while the control group (CG) executed a free training program. At the end of the research, the EG decreased from 57.66 ± 2.56 to 56.85 ± 2.52 mLO2
kg-1
min-1. In con- trast, the CG decreased from 58.08 ± 2.60 to 54.52 ± 2.80 mLO2
kg-1
min-1. Additionally, the researchers reported significant decreases in VO2 max when comparing the EG to the CG in the post-test (t = 16.06; p < 0.0001). Likewise, the endomorphic somatotype has a greater fat mass than the mesomorphic and ectomorphic somatotype [43], and subjects with endomor- phic predominance have shown a lower VO2 max than subjects with a mesomorphic or ectomorphic predominance (endomorphic: 37.3 ± 0.77; mesomorphic: 40.2 ± 0.46; and ecto- morphic: 43.5 ± 0.52) [51]. For this reason, the increase in the endomorphic component observed in naval cadets after 12 weeks without mandatory physical training could condition the decrease of VO2 max at the end of this period (p < 0.001, TE = 0.34). However, it is impor- tant to analyze the ES for each variable studied, which allows us to observe each phenomenon’s degree of presence, independent of the alpha level calculated [52]. In this study, like in research by Parpa & Michaelides [24], all ES in the tests with significant differences in VO2 max, includ- ing men and all data analysis, oscillated between 0.2–0.6. This was considered a small effect. Fig 2. Changes in VO2 max and anthropometric parameters before and after 12 weeks without mandatory physical training. 12MRT: 12-minute race test; mLO2
Kg–1
min–1: milliliters of oxygen per kilogram of body mass per minute; mm: millimeters; cm: centimeters; kg: kilograms; : p < 0.002. https://doi.org/10.1371/journal.pone.0251516.g002 PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 7 / 15 On the other hand, the significant differences in women had an ES between 0.6–1.2 (which was considered as a moderate effect). Furthermore, the large and negative correlation between VO2 max and the fat percentage observed in the test (r = -0.69, p = 0.01) increased after the period without mandatory physical training (r = -0.75, p = 0.01). Up to this point, the decrease in VO2 max has been attributed to two leading causes; on the one hand, a transition period without mandatory and controlled physical training, while on the other hand, an increase in fat mass, reflected in the endomorphic component of naval cadets [51]. Periods without physical training have also been associated with a decrease in muscle cross- section [34]. This unfavorable consequence could be related to lower levels of muscle strength [35]. In this case, Koundourakis et al. [31] examined the effects of six weeks without physical training on performance parameters in soccer players; at the end of the study, the researchers reported significant decreases in both squat jump (Team A: 39.70 ± 3.32 vs 37.30 ± 3.08 kg; p < 0.001; Team B: 41.04 ± 3.34 vs 38.18 ± 3.03 kg; p < 0.001) and countermovement jump (Team A: 41.04 ± 3.99 vs 39.13 ± 3.26%; p < 0.001); Team B: 42.82 ± 3.60 vs 40.09 ± 2.79 kg; p < 0.001) in both experimental groups. The researchers also concluded that the observed reductions in jumping ability (considered to be a negative effect) could be related to mis- matches of rapidly contracting muscle fibers [25, 53]. In parallel, the endomorphic somatotype has lesser muscle mass than the mesomorphic and ectomorphic somatotype [43]. In turn, Mir- oshnichenko et al. [51] showed a high correlation between the predominance of the mesomor- phic component and VO2 max. Likewise, an increase in the endomorphic component and lower muscle mass could be associated with a lower VO2 max of the participants. Therefore, Fig 3. Somatotype before and after 12 weeks without mandatory physical training. https://doi.org/10.1371/journal.pone.0251516.g003 PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 8 / 15 an increment of the endomorphic component in naval cadets may decrease the lower extremi- ties’ strength, generating biomechanical and neuronal changes [54]. These last changes could affect the economy of the race [55] and, consequently, decrease the performance in 12MRT (p < 0.001, ES = 0.34). Although the evidence shows the negative influence of periods without training on strength and muscular power [31, 35], mainly due to loss of muscle mass [34, 51], the present study did not consider assessing naval cadets’ anaerobic capacity. Therefore, the possible effects of 12 weeks without mandatory physical training on strength or power in both the lower and upper extremities should be considered in future studies. On the other hand, this study also showed increases in some anthropometric parameters after 12 weeks without mandatory physical training, specifically in the tricipital skinfold thick- ness in men (p = 0.02, ES = 0.18), arm circumference in women (p = 0.04, ES = 0.19) and the endomorphic component in both men and women (p = 0.02, ES = 0.25). In this sense, evi- dence shows that a period without physical training leads to increased fat mass and a decreased lean mass [31–33]. Also, the tricipital fold, together with the subscapular and suprailiac folds, are anthropometric indicators with a high explanatory power of VO2 max in both sexes [56]. We evidenced that those naval cadets with a higher tricipital fold had a reduced VO2 max Fig 4. Correlation between VO2 max and anthropometric parameters before and after 12 weeks without mandatory physical training. mLO2
Kg–1
min–1: milliliters of oxygen per kilogram of body mass per minute; % fat: percentage of fat; mm: millimeter. https://doi.org/10.1371/journal.pone.0251516.g004 PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 9 / 15 (Test: r = 0.76, p = 0.01; post test: r = 0.81, p = 0.01). Likewise, an elevated tricipital fold condi- tions an elevated endomorphic component [42]. Consequently, anthropometric parameters influence cardiorespiratory fitness, independent of sex, age, and obesity level [57]. Related to this, Sotiropoulos et al. [33] evaluated changes in body weight and body fat percentage after a four-week transition period in soccer players (The EG conducted a directed training program and the CG a free training program). At the end of the study, the EG increased from 78.14 ± 4.77 to 78.74 ± 5.00 kg, while the CG increased from 76.48 ± 2.65 to 77.90 ± 2.82 kg (t = -4.91; p < 0.005); and, also reported increased percentage of body fat (EG from 7.92 ± 1.68 to 8.17 ± 1.81%; CG from 7.77 ± 1.79 to 8.59 ± 1.80%; t = -8.42; p < 0.005). On the other hand, Ormsbee et al. [58] examined the effect of five weeks without physical training on body com- position in swimmers. At the end of the study, significant differences were observed in body weight (68.96 ± 9.7 vs. 69.8 ± 9.8 kg; p = 0.03), fat mass (14.7 ± 7.6 vs. 16.5 ± 7.4 kg; p = 0.001), and waist circumference (72.7 ± 3.1 vs. 73.8 ± 3.6 cm; p = 0.03). Also, Koundourakis et al. [31] examined the effects of six weeks without physical training on the body composition of soccer players; at the end of the study, the researchers reported significant increases in both body weight (Team A: 77.60 ± 5.88 vs. 79.13 ± 6.16 kg; p < 0.001; Team B: 77.89 ± 8.75 vs. 79.49 ± 8.95 kg; p < 0.001) and in the fat percentage (Team A: 9.2 ± 3.33 vs. 11.01 ± 4.11%; p < 0.001; Team B: 9.43 ± 3.55 vs. 10.40 ± 4.08 kg; p < 0.001) in both experimental groups. Although some studies have established the body composition of armed forces personnel in some countries [59] and anthropometric changes have been documented concerning soldiers’ physical training [60], the effects of 12 weeks without mandatory physical training on anthro- pometric parameters have not been reported for naval cadets. Consequently, in connection with the studies referred to above, our study’s findings show the importance of verifying and controlling body composition after a period without mandatory physical training in naval cadets [61], especially somatotype indicators [43]. However, it is essential to mention that the present study did not control the participants’ caloric intake [62]. For this reason, we are not sure that the changes in anthropometric parameters were only due to a decrease in physical training [63–65]; there is a possibility that higher caloric intake, above the daily energy needs, has also influenced these physical changes [62, 66]. Finally, the data show that VO2 max is an essential parameter of the physical condition [38], and a higher VO2 max allows the efficient performance of physical tasks associated with military personnel [13, 60]. It has also been demonstrated that subjects with a higher percent- age of body fat have lower VO2 max, lower strength levels, and lower fatigue tolerance [67]. As demonstrated in this study, a vacation period without mandatory physical training generates decreases in the VO2 max [37] and negatively affects anthropometric parameters [51]. There- fore, the vacation periods must be adapted into a transition phase [24, 25]. In this way, with controlled and directed physical training, both athletes and naval cadets will have optimal physical recovery and maintenance; this condition will allow them to face better the next cycle of physical training [23]. One of the limitations of this study was the sample used. As mentioned above, the sample was by convenience, which would not allow us to generalize the data. However, armed forces personnel are more homogeneous in body structure [68] and eating behavior [69]. For this reason, in this specific case, the results could be generalized to this population. Conclusions Twelve weeks without mandatory physical training significantly decreases the VO2 max in naval cadets from 18 to 25 years old. Simultaneously, the same period without mandatory training increases skinfold thickness and the endomorphic component in this population. PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 10 / 15 Practical applications After evidence of decreases in VO2 max and negative increases in some anthropometric parameters after 12 weeks without mandatory physical training, it is suggested that training loads in the transition phase [25], whether due to vacations [28] or to unforeseen events [30]. Acknowledgments We thank the 38 naval cadets for their voluntary and disinterested participation in the Arturo Prat Naval Academy. Author Contributions Conceptualization: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Data curation: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Formal analysis: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Funding acquisition: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Investigation: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Methodology: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Resources: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Supervision: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Validation: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Visualization: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Galdames Maliqueo. Writing – original draft: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Gal- dames Maliqueo. Writing – review & editing: A´lvaro Huerta Ojeda, Guillermo Barahona-Fuentes, Sergio Gal- dames Maliqueo. References 1. Ojeda A´ H, Rı´os LC, Barrilao RG, Rios IC, Serrano PC. Effect of variable resistance on post-activation potentiation: A systematic review. Arch Med Deporte. 2016; 33(5):338–45. 2. Barahona-Fuentes G, Huerta Ojeda A´ , Galdames Maliqueo S. The influence of High-Intensity Interval Training Based Plyometric Exercise on Jump Height and Peak Power of Under-17 Male Soccer Players. Educacio´n Fı´sica y Ciencia. 2019; 21(2):e080. https://doi.org/10.24215/23142561e080. 3. Huerta A´ C, Galdames S, Cataldo M, Barahona G, Rozas T, Ca´ceres P. Effects of a high intensity inter- val training on the aerobic capacity of adolescents. Rev Med Chil. 2017; 145(8):972–9. https://doi.org/ 10.4067/s0034-98872017000800972 PMID: 29189854 4. Gacesa JZ, Jakovljevic DG, Kozic DB, Dragnic NR, Brodie DA, Grujic NG. Morpho-functional response of the elbow extensor muscles to twelve-week self-perceived maximal resistance training. Clin Physiol Funct Imaging. 2010; 30(6):413–9. https://doi.org/10.1111/j.1475-097X.2010.00957.x PMID: 20670339 PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 11 / 15 5. Ingle L, Stephenson A, Sandercock GR. Physical activity profiles and selected muscular fitness vari- ables in English schoolchildren: A north–south divide? Eur J Sport Sci. 2016; 16(8):1187–96. https:// doi.org/10.1080/17461391.2016.1183714 PMID: 27220086 6. Sabido R, Peñaranda M, Herna´ndez-Davo´ JL. Comparison of acute responses to four different hyper- trophy-oriented resistance training methodologies. Eur J Hum Mov. 2016;37109–21. 7. Cordova A, Villa G, Sureda A, Rodriguez-Marroyo JA, Sa´nchez-Collado MP. Physical Activity and Car- diovascular Risk Factors in Spanish Children Aged 11–13 Years. Rev Española Cardiol. 2012; 65 (7):620–6. https://doi.org/10.1016/j.recesp.2012.01.026 PMID: 22633280 8. Ga´lvez Casas A, Rodrı´guez Garcı´a PL, Garcı´a-Canto´ E, Rosa Guillamo´n A, Pe´rez-Soto JJ, Tarraga Marcos L, et al. Aerobic capacity and quality of life in school children from 8 to 12. Clin Investig Arterios- cler. 2015; 27(5):239–45. https://doi.org/10.1016/j.arteri.2015.01.001 PMID: 25814171 9. Patel A V., Hodge JM, Rees-Punia E, Teras LR, Campbell PT, Gapstur SM. Relationship between mus- cle-strengthening activity and cause-specific mortality in a large US cohort. Prev Chronic Dis. 2020;171–9. https://doi.org/10.5888/pcd17.190408. 10. Ojeda A´ H, Rı´os LC, Barrilao RG, Serrano PC. Acute effect of a complex training protocol of back squats on 30-m sprint times of elite male military athletes. J Phys Ther Sci. 2016; 28(3):752–6. https://doi.org/ 10.1589/jpts.28.752 PMID: 27134353 11. Webber BJ, Flower AM, Pathak SR, Burganowski RP, Pawlak MT, Gottfredson RC, et al. Physical and Mental Health of US Air Force Military Training Instructors. Mil Med. 2019; 184(5–6):e248–54. https:// doi.org/10.1093/milmed/usy418 PMID: 30690457 12. Vrijkotte S, Roelands B, Pattyn N, Meeusen R. The Overtraining Syndrome in Soldiers: Insights from the Sports Domain. Mil Med. 2019; 184(5–6):e192–200. https://doi.org/10.1093/milmed/usy274 PMID: 30535270 13. Taylor MK, Markham AE, Reis JP, Padilla GA, Potterat EG, Drummond SPA, et al. Physical fitness influences stress reactions to extreme military training. Mil Med. 2008; 173(8):738–42. https://doi.org/ 10.7205/milmed.173.8.738 PMID: 18751589 14. Schneider KL, Spring B, Pagoto SL. Exercise and energy intake in overweight, sedentary individuals. Eat Behav. 2009; 10(1):29–35. https://doi.org/10.1016/j.eatbeh.2008.10.009 PMID: 19171314 15. Lovell DI, Cuneo R, Wallace J, McLellan C. The hormonal response of older men to sub-maximum aero- bic exercise: The effect of training and detraining. Steroids. 2012; 77(5):413–8. https://doi.org/10.1016/ j.steroids.2011.12.022 PMID: 22248672 16. Stebbings GK, Morse CI, McMahon GE, Onambele GL. Resting arterial diameter and blood flow changes with resistance training and detraining in healthy young individuals. J Athl Train. 2013; 48 (2):209–19. https://doi.org/10.4085/1062-6050-48.1.17 PMID: 23672385 17. Mujika I. The influence of training characteristics and tapering on the adaptation in highly trained individ- uals: A review. Int J Sports Med. 1998; 19(7):439–46. https://doi.org/10.1055/s-2007-971942 PMID: 9839839 18. Flann KL, Lastayo PC, McClain DA, Hazel M, Lindstedt SL. Muscle damage and muscle remodeling: No pain, no gain? J Exp Biol. 2011; 214(4):674–9. https://doi.org/10.1242/jeb.050112. 19. Gunter K, Baxter-Jones ADG, Mirwald RL, Almstedt H, Fuchs RK, Durski S, et al. Impact exercise increases BMC during growth: an 8-year longitudinal study. J Bone Miner Res. 2008; 23(7):986–93. https://doi.org/10.1359/jbmr.071201 PMID: 18072874 20. Maroto-Izquierdo S, Garcı´a-Lo´pez D, Fernandez-Gonzalo R, Moreira OC, Gonza´lez-Gallego J, de Paz JA. Skeletal muscle functional and structural adaptations after eccentric overload flywheel resistance training: a systematic review and meta-analysis. J Sci Med Sport. 2017; 20(10):943–51. https://doi.org/ 10.1016/j.jsams.2017.03.004 PMID: 28385560 21. Garber CE, Blissmer B, Deschenes MR, Franklin BA, Lamonte MJ, Lee IM, et al. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults: Guidance for prescribing exercise. Med Sci Sports Exerc. 2011; 43(7):1334– 59. https://doi.org/10.1249/MSS.0b013e318213fefb PMID: 21694556 22. Barahona-Fuentes G, Ojeda A´ H, Jerez-Mayorga D. Effects of different methods of strength training on indicators of muscle fatigue during and after strength training: a systematic review. Motriz J Phys Educ. 2020; 26(3):e10200063. https://doi.org/10.1590/S1980-6574202000030063. 23. Silva JR, Brito J, Akenhead R, Nassis GP. The transition period in soccer: a window of opportunity. Sport Med. 2016; 46(3):305–13. https://doi.org/10.1007/s40279-015-0419-3 PMID: 26530720 24. Parpa K, Michaelides MA. The effect of transition period on performance parameters in elite female soc- cer players. Int J Sports Med. 2020; 41(8):528–32. https://doi.org/10.1055/a-1103-2038 PMID: 32059247 PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 12 / 15 25. Mujika I, Padilla S. Detraining: loss of training-induced physiological and performance adaptations. Part I. Sport Med. 2000; 30(2):79–87. https://doi.org/10.2165/00007256-200030020-00002. 26. Musumeci G. Sarcopenia and exercise “the state of the art.” J Funct Morphol Kinesiol. 2017; 2(4):. https://doi.org/10.3390/jfmk2040040. 27. Maugeri G, Castrogiovanni P, Battaglia G, Pippi R, D’Agata V, Palma A, et al. The impact of physical activity on psychological health during Covid-19 pandemic in Italy. Heliyon. 2020; 6(6):e04315. https:// doi.org/10.1016/j.heliyon.2020.e04315 PMID: 32613133 28. Katzmarzyk PT, Barreira T V., Broyles ST, Champagne CM, Chaput JP, Fogelholm M, et al. Physical activity, sedentary time, and obesity in an international sample of children. Med Sci Sports Exerc. 2015; 47(10):2062–9. https://doi.org/10.1249/MSS.0000000000000649 PMID: 25751770 29. Prentice-Dunn H, Prentice-Dunn S. Physical activity, sedentary behavior, and childhood obesity: a review of cross-sectional studies. Psychol Health Med. 2012; 17(3):255–73. https://doi.org/10.1080/ 13548506.2011.608806 PMID: 21995842 30. Paoli A, Musumeci G. Elite athletes and COVID-19 lockdown: future health concerns for an entire sec- tor. J Funct Morphol Kinesiol. 2020; 5(2):10–2. https://doi.org/10.3390/jfmk5020030 PMID: 33467246 31. Koundourakis NE, Androulakis NE, Malliaraki N, Tsatsanis C, Venihaki M, Margioris AN. Discrepancy between exercise performance, body composition, and sex steroid response after a six-week detraining period in professional soccer players. PLoS One. 2014; 9(2):. https://doi.org/10.1371/journal.pone. 0087803. 32. Reinke S, Karhausen T, Doehner W, Taylor W, Hottenrott K, Duda GN, et al. The influence of recovery and training phases on body composition, peripheral vascular function and immune system of profes- sional soccer players. PLoS One. 2009; 4(3):1–7. https://doi.org/10.1371/journal.pone.0004910. 33. Sotiropoulos AS, Travlos AK, Gissis I, Souglis AG, Grezios A. The effect of a 4-week training regimen on body fat and aerobic capacity of professional soccer players during the transition period. J Strength Cond Res. 2009; 23(6):1697–703. https://doi.org/10.1519/JSC.0b013e3181b3df69 PMID: 19675494 34. Hakkinen K. Effect of combined concentric and eccentric strength training and detraining on force-time, muscle fiber and metabolic characteristics of leg extensor muscles. Scand J Sport Sci. 1981;350–8. 35. Izquierdo M, Ibanez J, Gonzalez-Badillo JJ, Ratamess N, Kraemer WJ, Hakkinen K, et al. Detraining and tapering effects on hormonal responses and strength performance. J Strenght Cond Res. 2007; 21 (3):768–75. https://doi.org/10.1519/R-21136.1 PMID: 17685721 36. De Paiva PRV, Casalechi HL, Tomazoni SS, MacHado CDSM, Ribeiro NF, Pereira AL, et al. Does the combination of photobiomodulation therapy (PBMT) and static magnetic fields (sMF) potentiate the effects of aerobic endurance training and decrease the loss of performance during detraining? A rando- mised, triple-blinded, placebo-controlled trial. BMC Sports Sci Med Rehabil. 2020; 12(1):1–11. https:// doi.org/10.1186/s13102-020-00171-2. 37. Liguori G, Krebsbach K, Schuna J. Decreases in maximal oxygen uptake among army reserve officers’ training corps cadets following three months without mandatory physical training. Int J Exerc Sci. 2012; 5(4):354–9. PMID: 27182392 38. Ojeda A´ H, Maliqueo SG, Serrano PC. Validacio´n del test de 6 minutos de carrera como predictor del consumo ma´ximo de oxı´geno en el personal naval. Rev Cuba Med Mil. 2017; 46(4):1–11. 39. Galdames S, Huerta A´ , Pastene A. Effect of acute sodium bicarbonate supplementation on perfor- mance on the obstacle run in professional military pentathlete. Arch Med Deporte. 2020; 37(4):220–6. 40. Armada-de-Chile. ¿Quie´nes la componen? Armada de Chile. 2014; 41. Ato M, Lo´pez JJ, Benavente A. A classification system for research designs in psychology. An Psicol. 2013; 29(3):1038–59. https://doi.org/10.6018/analesps.29.3.178511. 42. Berral F. Protocolo de medidas antropome´tricas. Jornadas Me´dico Sanit. sobre Atlet., 2004;, p. 115– 22. 43. Carter JEL, Heath BH. Somatotyping: development and applications. vol. 5. New York, USA: Cam- bridge university press; 1990; 44. National Health and Nutrition Examination Survey. Anthropometry procedures manual. 2005; 45. Durnin J, Womersley J. Body fat assessed from total body density and its estimation from skinfold thick- ness: measurements on 481 men and women aged from 16 to 72 years. Br J Nutr. 1974;3277–97. https://doi.org/10.1079/bjn19740060 PMID: 4843734 46. Barrera A, Gladys M. Esta´ndares antropome´tricos para evaluacio´n del estado nutritivo. INTA. Santi- ago, Chile: 2004; 47. Mederico M, Paoli M, Zerpa Y, Briceño Y, Go´mez-Pe´rez R, Martı´nez JL, et al. Reference values of waist circumference and waist/hip ratio in children and adolescents of Me´rida, Venezuela: comparison PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 13 / 15 with international references. Endocrinol y Nutr. 2013; 60(5):235–42. https://doi.org/10.1016/j.endoen. 2012.12.006. 48. Cooper KH. A means of assessing maximal oxygen intake: correlation between field and treadmill test- ing. Jama. 1968; 203(3):201–4. PMID: 5694044 49. Hopkins WG, Marshall SW, Batterham AM, Hanin J. Progressive statistics for studies in sports medicine and exercise science. Med Sci Sports Exerc. 2009; 41(1):3–12. https://doi.org/10.1249/MSS. 0b013e31818cb278 PMID: 19092709 50. Martı´nez Camblor P. P-value adjustment for multiple comparisons. Rev Chil Salud Pu´blica. 2012; 16 (3):225–32. https://doi.org/10.5354/0717-3652.2012.23127. 51. Miroshnichenko VM, Furman YM, Brezdeniuk OY, Onyshchuk VE, Gavrylova N V, Salnykova S-V. Cor- relation of maximum oxygen consumption with component composition of the body, body mass of men with different somatotypes aged 25–35. Pedagog Phys Cult Sport. 2020; 24(2):290–6. https://doi.org/ 10.15561/26649837.2020.0603. 52. Casado A, Prieto L, Alonso J. El tamaño del efecto de la diferencia entre dos medias: ¿estadı´stica- mente significativo o clı´nicamente relevante? Med Clin (Barc). 1999; 112(15):584–8. PMID: 10365387 53. Perez-Gomez J, Rodriguez GV, Ara I, Olmedillas H, Chavarren J, Gonza´lez-Henriquez JJ, et al. Role of muscle mass on sprint performance: gender differences? Eur J Appl Physiol. 2008; 102(6):685–94. https://doi.org/10.1007/s00421-007-0648-8 PMID: 18084774 54. Trowell D, Vicenzino B, Saunders N, Fox A, Bonacci J. Effect of strength training on biomechanical and neuromuscular variables in distance runners: a systematic review and meta-analysis. Sport Med. 2020; 50(1):133–50. https://doi.org/10.1007/s40279-019-01184-9 PMID: 31541409 55. Skovgaard C, Christensen PM, Larsen S, Andersen TR, Thomassen M, Bangsbo J. Concurrent speed endurance and resistance training improves performance, running economy, and muscle NHE1 in mod- erately trained runners. J Appl Physiol. 2014; 117(10):1097–109. https://doi.org/10.1152/japplphysiol. 01226.2013 PMID: 25190744 56. de Andrade E, Gimenes H, Santos D. Which body fat anthropometric indicators are most strongly asso- ciated with maximum oxygen uptake in adolescents? Asian J Sports Med. 2017; 8(3):e13812. https:// doi.org/10.5812/asjsm.13812. 57. Yanek LR, Vaidya D, Kral BG, Dobrosielski DA, Moy TF, Stewart KJ, et al. Lean mass and fat mass as contributors to physical fitness in an overweight and obese african american population. Ethn Dis. 2015; 25(2):214–9. PMID: 26118151 58. Ormsbee MJ, Arciero P. Detraining increases body fat and weight and decreases VO2peak and meta- bolic rate. J Strength Cond Res. 2012; 26(8):2087–95. https://doi.org/10.1519/JSC. 0b013e31823b874c PMID: 22027854 59. Nelson R, Cheatham J, Gallagher D, Bigelman K, Thomas DM. Revisiting the United States Army body composition standards: a receiver operating characteristic analysis. Int J Obes. 2019; 43(8):1508–15. https://doi.org/10.1038/s41366-018-0195-x PMID: 30181655 60. Dyrstad SM, Soltvedt R, Halle´n J. Physical fitness and physical training during norwegian military ser- vice. Mil Med. 2006; 171(8):736–41. https://doi.org/10.7205/milmed.171.8.736 PMID: 16933814 61. Pl¸avin¸a L, Umbrasˇko S. Analysis of physical fitness tests and the body composition of the military per- sonnel. Pap Anthropol. 2016; 25(1):27. https://doi.org/10.12697/poa.2016.25.1.03. 62. Thomas DT, Erdman KA, Burke LM. Position of the academy of nutrition and dietetics, dietitians of can- ada, and the american college of sports medicine: nutrition and athletic performance. J Acad Nutr Diet. 2016; 116(3):501–28. https://doi.org/10.1016/j.jand.2015.12.006 PMID: 26920240 63. Franckle R, Adler R, Davison K. Accelerated weight gain among children during summer versus school year and related racial/ethnic disparities: a systematic review. Prev Chronic Dis. 2014; 11(12):1–10. https://doi.org/10.5888/pcd11.130355 PMID: 24921899 64. Moreno JP, Johnston CA, Woehler D. Changes in weight over the school year and summer vacation: results of a 5-year longitudinal study. J Sch Health. 2013; 83(7):473–7. https://doi.org/10.1111/josh. 12054 PMID: 23782089 65. von Hippel PT, Workman J. From kindergarten through second grade, U.S. children’s obesity preva- lence grows only during summer vacations. Obesity. 2016; 24(11):2296–300. https://doi.org/10.1002/ oby.21613 PMID: 27804271 66. Cooper JA, Tokar T. A prospective study on vacation weight gain in adults. Physiol Behav. 2016;15643–7. https://doi.org/10.1016/j.physbeh.2015.12.028 PMID: 26768234 67. Crawford K, Fleishman K, Abt JP, Sell TC, Lovalekar M, Nagai T, et al. Less body fat improves physical and physiological performance in army soldiers. Mil Med. 2011; 176(1):35–43. https://doi.org/10.7205/ milmed-d-10-00003 PMID: 21305957 PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 14 / 15 68. Maldonado I, Calero S. Anthropometric profile and body composition in aspirants of the Army Soldiers Training School. Rev Cuba Investig Biome´dicas. 2017; 36(2):1–13. 69. Kildal CL, Syse KL. Meat and masculinity in the Norwegian Armed Forces. Appetite. 2017; 112(Novem- ber 2013):69–77. https://doi.org/10.1016/j.appet.2016.12.032 PMID: 28040506 PLOS ONE Period without physical training on maximum oxygen uptake PLOS ONE | https://doi.org/10.1371/journal.pone.0251516 June 2, 2021 15 / 15

Effects of a period without mandatory physical training on maximum oxygen uptake and anthropometric parameters in naval cadets.

06-02-2021

Huerta Ojeda, Álvaro,Barahona-Fuentes, Guillermo,Galdames Maliqueo, Sergio

eng

PMC10703220

Reference number: PONE-D-23-18858 (previous submission PONE-D-23-16206) Exploring running styles in the field through cadence and duty factor modulation Dear dr. L. A. Peyré-Tartaruga and editorial office, Thank you for evaluating our manuscript and for giving us the opportunity to resubmit. We have made the following changes to the manuscript in response to the concerns of the editorial office: - We have included the reference numbers of both the original ethical application (VCWE-2019– 006R1) and the amendment (VCWE-2021-043) in the method section of the manuscript. - We have included the original ethical application (VCWE-2019–006R1) in Dutch, with the English translations in comments in the pdf file. - We have included the amendment (VCWE-2021-043) in Dutch, with the English translations in comments in the pdf file. - We have included the approval emails from our IRB for both the original ethical application (VCWE-2019–006R1) and the amendment (VCWE-2021-043). - We have included the informed consent form for this study in English. - We have included the participant information form for this study in English. We believe that by making those changes and including the additional files we have addressed the concerns of the editorial office. Please find below the message from the editorial office. Thank you for your time and consideration. Sincerely, Anouk Nijs, Msc. [email protected] Dr. Melvyn Roerdink [email protected] Prof. Dr. Peter J. Beek [email protected] Department of Human Movement Sciences, Faculty of Behavioural and Movement Sciences, Vrije Universiteit Amsterdam, The Netherlands PONE-D-23-16206 Exploring running styles in the field through cadence and duty factor modulation PLOS ONE Dear Dr. Nijs, I am writing to you about your appeal on the editorial decision for your submission to PLOS ONE above. After careful consideration of the manuscript, the reasons for the previous rejection, and your reasons for appealing, we are upholding the decision to reject the manuscript. As you are aware, manuscripts submitted to PLOS ONE are assessed based on the journal’s publication criteria. We have concerns on the contents of the manuscript including that the approval document provided did not match the study presented in the manuscript. Furthermore, the approval number on the manuscript (VCWE-2019–006R1) did not match the one provided in the email from your IRB (VCWE- 2021-043) Considering those concerns, the manuscript does not currently meet our criteria for publication requiring that the research meets all applicable standards for the ethics of experimentation and research integrity. However, if you are able to provide a copy of the original approval document issued by your IRB (i.e. VCWE-2019–006R1) and an English translation, as well as the documents that you have mentioned in your appeal email, we do feel that a revised manuscript may be suitable for consideration. This would however need to be considered as a new submission. If you are able to revise the manuscript as indicated above and submit a new manuscript to PLOS ONE, please refer to the original submission in the cover letter. Thank you for your interest in PLOS ONE. Best wishes, Anushmathi PM Editorial Office PLOS ONE

Exploring running styles in the field through cadence and duty factor modulation.

12-07-2023

Nijs, Anouk,Roerdink, Melvyn,Beek, Peter Jan

eng

PMC7309010

sensors Article Effects of Novel Inverted Rocker Orthoses for First Metatarsophalangeal Joint on Gastrocnemius Muscle Electromyographic Activity during Running: A Cross-Sectional Pilot Study Rubén Sánchez-Gómez 1 , Carlos Romero-Morales 2,* , Álvaro Gómez-Carrión 1, Blanca De-la-Cruz-Torres 3 , Ignacio Zaragoza-García 1 , Pekka Anttila 4, Matti Kantola 4 and Ismael Ortuño-Soriano 1 1 Nursing Department, Faculty of Nursing, Physiotherapy and Podiatry, Universidad Complutense de Madrid, 28040 Madrid, Spain; [email protected] (R.S.-G.); [email protected] (Á.G.-C.); [email protected] (I.Z.-G.); [email protected] (I.O.-S.) 2 Faculty of Sport Sciences, Universidad Europea de Madrid, Villaviciosa de Odón, 28670 Madrid, Spain 3 Department of Physiotherapy, University of Seville, c/Avicena, s/n, 41009 Seville, Spain; [email protected] 4 Applied Science of Metropolia Univesity, Podiatry Department, 01600 Helsinki, Finland; pekka.anttila@metropolia.fi (P.A.); Matti.Kantola@metropolia.fi (M.K.) * Correspondence: [email protected] Received: 15 April 2020; Accepted: 3 June 2020; Published: 5 June 2020



 Abstract: Background: The mobility of the first metatarsophalangeal joint (I MPTJ) has been related to the proper windlass mechanism and the triceps surae during the heel-off phase of running gait; the orthopedic treatment of the I MPTJ restriction has been made with typical Morton extension orthoses (TMEO). Nowadays it is unclear what effects TMEO or the novel inverted rocker orthoses (NIRO) have on the EMG activity of triceps surae during running. Objective: To compare the TMEO effects versus NIRO on EMG triceps surae on medialis and lateralis gastrocnemius activity during running. Study design: A cross-sectional pilot study. Methods: 21 healthy, recreational runners were enrolled in the present research (mean age 31.41 ± 4.33) to run on a treadmill at 9 km/h using aleatory NIRO of 6 mm, NIRO of 8 mm, TMEO of 6 mm, TMEO of 8 mm, and sports shoes only (SO), while the muscular EMG of medial and lateral gastrocnemius activity during 30 s was recorded. Statistical intraclass correlation coefficient (ICC) to test reliability was calculated and the Wilcoxon test of all five different situations were tested. Results: The reliability of values was almost perfect. Data showed that the gastrocnemius lateralis increased its EMG activity between SO vs. NIRO-8 mm (22.27 ± 2.51 vs. 25.96 ± 4.68 mV, p < 0.05) and SO vs. TMEO-6mm (22.27 ± 2.51 vs. 24.72 ± 5.08 mV, p < 0.05). Regarding gastrocnemius medialis, values showed an EMG notable increase in activity between SO vs. NIRO-6mm (22.93 ± 2.1 vs. 26.44 ± 3.63, p < 0.001), vs. NIRO-8mm (28.89 ± 3.6, p < 0.001), and vs. TMEO-6mm (25.12 ± 3.51, p < 0.05). Conclusions: Both TMEO and NIRO have shown an increased EMG of the lateralis and medialis gastrocnemius muscles activity during a full running cycle gait. Clinicians should take into account the present evidence when they want to treat I MTPJ restriction with orthoses, and consider the inherent triceps surae muscular cost relative to running economy. Keywords: triceps surae; first metatarsophalangeal joint; surface electromyography 1. Introduction Coterill [1] was the first author who described painful osteoarthritis (OA) of the first metatarsophalangeal joint (IMTPJ), which is known as hallux rigidus (HR). HR is the last stage Sensors 2020, 20, 3205; doi:10.3390/s20113205 www.mdpi.com/journal/sensors Sensors 2020, 20, 3205 2 of 12 of the IMTPJ degeneration, with functional hallux limitus [2] (FHL) at the beginning of the pathological progress [3]. Joint disease is thought to be caused by repetitive impacts on the dorsal aspect of the base of the proximal phalanx of the hallux by the first metatarsal head during the propulsion phase of gait and running in feet with multifactorial biomechanical and/or structural deficits [4]. The limitation of IMTPJ has been linked to gait problems [5] and its consequences on ankle, knee, hip, or low back during running [6]. The treatment of this injury has been addressed in several conservative non-surgical and surgical ways. Non-surgical management is valid to treat HR in the earliest stages [7,8] and includes ultrasound therapy, infiltrative drugs, shoe modifications, hallux bandages, manual mobilization, flexor strengthening, and orthoses to improve the joint problems. There are a few references on treatment of OA using plantar insoles in HR and FHL. Traditional Morton’s extensions are orthoses with a flat light modification under the first ray that has been used to treat HR [9–11] to avoid the impact between the proximal phalanx and first metatarsal bones. This opens the IMTPJ dorsally but restricts its dorsiflexion movement, while rocker-sole footwear modifications have shown a reduction in the peak pressure under the IMTPJ. This decreases the average gait cycle that is spent in the stance phase [12] and increases muscle activity of the lower limb [13]. However, there is no reference to either the inverted rocker-sole orthoses effects or the effect of footwear modifications on muscle activity during running. On the other hand, running economy (RE) has been described as the oxygen cost of running at a given speed in every case [14] and factors such as biomechanics and muscular fatigue can influence the RE [15]. Additionally, barefoot running has shown differences in biomechanical behaviour [16] and muscular responses [17,18] when it is compared with classical running shoes. Compared to fatigue, strength training added to a normal training program for distance running can improve RE between 2% and 8%. An increase in muscle mass training programs around the proximal region of the lower limb, such as quadriceps or hamstring [19], or around the distal regions, such as the triceps surae [20] with plantarflexion and dorsiflexion ankle exercises, has shown some benefits on RE. Accordingly, triceps surae and its relationship with the windlass mechanism [21] in the propulsion phase of gait and running has been reported to provide between 8% and 17% of the elastic energy that is needed for the heel-off phase [22,23] toward a suitable IMTPJ dorsiflexion [24,25]. However, the electromyography (EMG) effects in the triceps surae with limited dorsiflexion of the IMTPJ that is induced by any orthotic dorsiflexion restriction has never been studied. Understanding the EMG activity of this muscle will allow us to understand if the subjects could be increasing their energy cost during running, which is very important for an efficient RE [19]. However, no previous research has studied the effect of a novel inverted rocker orthoses (NIRO) on the EMG activity of the triceps surae compared to traditional Morton’s extension orthoses (TMEO) during running in the healthy population. Because of the restricting IMTPJ effect of TMEO and its influence on the windlass mechanism that is linked with the triceps surae [24,25], we hypothesized that TMEO (6 mm and 8 mm) may increase the EMG activity of the gastrocnemius medialis and lateralis muscles compared to the shoe only (SO) condition during running activity; in addition, regarding previous muscular activity changes that are reported with classical rocker soles [13], we hypothesized that NIRO (6 mm and 8 mm) may reduce the EMG of gastrocnemius medialis and lateralis compared to TMEO (6 mm and 8 mm), and this may increase EMG compared to SO in healthy people during running activity. 2. Materials and Methods The public institutional review board at Virgen Macarena-Virgen del Rocío hospitals, reviewed and approved the present study (certificate number f7f4a6567676d7ba7163bce0d15e7f98c9f33354). Ethical and human criteria were followed according to the Declaration of Helsinki, and signed informed consent was obtained from all subjects. Sensors 2020, 20, 3205 3 of 12 2.1. Design and Sample Size The statistics unit at the Spanish public university used software to assess the suitable sample size to perform this cross-sectional observational study and to study the difference in the EMG changes in the gastrocnemius medialis and lateralis muscles between SO, NIRO 6 mm, NIRO 8 mm, TMEO 6 mm, and TMEO 8 mm groups during running. Previous data on the triceps surae showed 7.0 ± 0.6 millivolts (mV) wearing 9-mm heel lifts compared to 4.9 ± 0.6 mV wearing typical shoes [26]. Taking into account a statistical power of 80%, β = 20%, a 95% confidence interval (CI), and α = 0.05, 30 subjects were needed to complete the study. Considering the typical loss of 20% subjects, 24 participants were recruited. However, three individuals were excluded from the study because they felt pain and discomfort during the EMG assessment. Reporting of Observational Studies in Epidemiology (STROBE) [27] criteria and a randomly consecutive sampling technique were followed to develop the present research. 2.2. Subjects The following inclusion criteria were used to select the participants: (1) healthy participants, between 18 and 30 years old; (2) recreational runners with 3–4 h of training per week with more than 1 year of experience; (3) neutral foot posture index (FPI) with values between 0 and +5 points according to a validity tool [28]; and (4) no injuries or pain at the time of the test. The exclusion criteria were as follows: (1) any lower limb injury during the last 6 months; (2) less movement in either foot joint than what is required to perform the optimal biomechanics according to normal values [29,30]; or (3) under the influence of any drugs effects at the time of the measurements. Body mass index (BMI) was taken into account to select a homogeneous sample, using Quetelet’s equation as follows: BMI = weight (kg)/height (m2) [31]. 2.3. Instrumentation and Assessments Neurotrac® Simplex Plus (Verity Medical Ltd., Braishfield, UK) EMG electronic device with a USB-Bluetooth [32] was used to study the triceps surae activity during the running test. The recording range on the device was 0.2 mV to 2000 mV, with a sensitivity of 0.1 mV RMS, 10 m of free wireless (Bluetooth) connection range and an accuracy of 4% of the reading from mV +/− 0.3 mV to 200 Hz, with a bandpass filter of 18 Hz +/− 4 Hz to 370 Hz +/− 10% for readings below 235 mV. The signal was assessed using self-adhesive circular surface electrodes that were 30 mm in diameter and made of high-quality hydrogel and conductive carbon film to detect the electrical action of the muscle fibers. The signal from each electrode was captured by the receiver module and filtered automatically by the Neurotrac® software (Verity Medical Ltd., Braishfield, UK). It was sent by a unidirectional radioelectric secure connection to the computer and it was digitally transformed by the software to generate activity patterns data for each electrode. 2.4. Materials NIRO was made using a flat sheet of ethylene-vinyl acetate (EVA) with a semi-rigid density that was 3 mm thick, without any orthotic element that could interface with normal biomechanical behaviour of the foot. NIRO had an inverted rocker composed of EVA medium that was 5 cm long, 2 cm wide, and 6 mm thick. Its proximal and distal edges were smoothly polished, and it was placed on the IMTPJ. The whole orthotic was covered with an EVA soft layer that was 1 mm thick (Figure 1). The TMEO was made with the same flat sheet of semi-rigid EVA that was 3 mm thick without any orthotic element and with a rectangular flat piece of EVA medium (6 mm thick) that was placed under the IMTPJ area and it was covered with an EVA soft layer that was 1 mm thick (Figure 2). The neutral SOs were “New Feel PW 100M medium grey” (ref. number: 2018022). NIRO and TMEO were made in an external orthopedic laboratory that was blinded to the study protocol. Sensors 2020, 20, 3205 4 of 12 Sensors 2020, 20, x FOR PEER REVIEW 4 of 12 Figure 1. Novel inverted rocker orthotic (NIRO). A flat sheet of ethylene-vinyl acetate (EVA) with an inverted rocker piece of EVA medium 6 mm thick under IMTPJ (bulked raised shape) covered with a yellow EVA soft layer that was 1 mm thick. Figure 2. Typical Morton’s extension orthotic (TMEO). A flat sheet of ethylene-vinyl acetate (EVA) with a rectangular flat piece of EVA medium 6 mm thick under IMTPJ covered with a black EVA soft layer that was 1 mm thick. 2.5. Procedure The podiatric clinician researcher (RSG) performed a physical assessment of the subjects and applied the eligibility criteria. To visualize the muscle belly, each subject was asked to perform plantarflexion of the ankle joint for a few seconds. The surface electrodes were then placed longitudinally onto the most prominent bulge of the gastrocnemius medialis and lateralis, based on the “European recommendations for surface EMG” [33]. The subjects were then asked to stand on one leg in the tip-toe position using their dominant foot for 5 s to set the maximal voluntary contractions that were needed in the strongest limb to calibrate the software and to normalize EMG data amplitudes for each test [34]. This was followed by acclimatization of subjects to a motorized treadmill at 5.17 km/h for 3 min [17]. The participants were divided randomly in gastrocnemius lateralis or medialis group by choosing a sealed envelope that assigned them to one group or another to begin the test; after that, they selected one of the five sealed envelopes with each of the five different conditions of the study (SO, NIRO 6 mm, NIRO 8 mm, TMEO 6 mm, TMEO 8 mm) to set randomly the order of the test. The 11 subjects who began with medialis gastrocnemius assessments, did the lateralis test following the same randomized protocol for each of the five different conditions and vice versa for the 12 participants who began with the lateralis test (Figure 3). Three running trials at 9 km/h [35] under five different conditions (SO, NIRO 6 mm, NIRO 8 mm, TMEO 6 mm, and TMEO 8 mm) on the same day were randomly performed. The duration of each trial was 1 min. For each subject, the mean EMG muscle activity pattern [36] of the gastrocnemius medialis of the dominant leg was recorded during the last 30 s of each 1-min trial, which was performed three times, leaving 5 min of rest between each test [37]. To avoid a potential imbalance, the same condition was added to contralateral foot. The same protocol was performed to Figure 1. Novel inverted rocker orthotic (NIRO). Sensors 2020, 20, x FOR PEER REVIEW 4 of 12 Figure 1. Novel inverted rocker orthotic (NIRO). A flat sheet of ethylene-vinyl acetate (EVA) with an inverted rocker piece of EVA medium 6 mm thick under IMTPJ (bulked raised shape) covered with a yellow EVA soft layer that was 1 mm thick. Figure 2. Typical Morton’s extension orthotic (TMEO). A flat sheet of ethylene-vinyl acetate (EVA) with a rectangular flat piece of EVA medium 6 mm thick under IMTPJ covered with a black EVA soft layer that was 1 mm thick. 2.5. Procedure The podiatric clinician researcher (RSG) performed a physical assessment of the subjects and applied the eligibility criteria. To visualize the muscle belly, each subject was asked to perform plantarflexion of the ankle joint for a few seconds. The surface electrodes were then placed longitudinally onto the most prominent bulge of the gastrocnemius medialis and lateralis, based on the “European recommendations for surface EMG” [33]. The subjects were then asked to stand on one leg in the tip-toe position using their dominant foot for 5 s to set the maximal voluntary contractions that were needed in the strongest limb to calibrate the software and to normalize EMG data amplitudes for each test [34]. This was followed by acclimatization of subjects to a motorized treadmill at 5.17 km/h for 3 min [17]. The participants were divided randomly in gastrocnemius lateralis or medialis group by choosing a sealed envelope that assigned them to one group or another to begin the test; after that, they selected one of the five sealed envelopes with each of the five different conditions of the study (SO, NIRO 6 mm, NIRO 8 mm, TMEO 6 mm, TMEO 8 mm) to set randomly the order of the test. The 11 subjects who began with medialis gastrocnemius assessments, did the lateralis test following the same randomized protocol for each of the five different conditions and vice versa for the 12 participants who began with the lateralis test (Figure 3). Three running trials at 9 km/h [35] under five different conditions (SO, NIRO 6 mm, NIRO 8 mm, TMEO 6 mm, and TMEO 8 mm) on the same day were randomly performed. The duration of each trial was 1 min. For each subject, the mean EMG muscle activity pattern [36] of the gastrocnemius medialis of the dominant leg was recorded during the last 30 s of each 1-min trial, which was performed three times, leaving 5 min of rest between each test [37]. To avoid a potential imbalance, the same condition was added to contralateral foot. The same protocol was performed to Figure 2. Typical Morton’s extension orthotic (TMEO). A flat sheet of ethylene-vinyl acetate (EVA) with an inverted rocker piece of EVA medium 6 mm thick under IMTPJ (bulked raised shape) covered with a yellow EVA soft layer that was 1 mm thick. A flat sheet of ethylene-vinyl acetate (EVA) with a rectangular flat piece of EVA medium 6 mm thick under IMTPJ covered with a black EVA soft layer that was 1 mm thick. 2.5. Procedure The podiatric clinician researcher (RSG) performed a physical assessment of the subjects and applied the eligibility criteria. To visualize the muscle belly, each subject was asked to perform plantarflexion of the ankle joint for a few seconds. The surface electrodes were then placed longitudinally onto the most prominent bulge of the gastrocnemius medialis and lateralis, based on the “European recommendations for surface EMG” [33]. The subjects were then asked to stand on one leg in the tip-toe position using their dominant foot for 5 s to set the maximal voluntary contractions that were needed in the strongest limb to calibrate the software and to normalize EMG data amplitudes for each test [34]. This was followed by acclimatization of subjects to a motorized treadmill at 5.17 km/h for 3 min [17]. The participants were divided randomly in gastrocnemius lateralis or medialis group by choosing a sealed envelope that assigned them to one group or another to begin the test; after that, they selected one of the five sealed envelopes with each of the five different conditions of the study (SO, NIRO 6 mm, NIRO 8 mm, TMEO 6 mm, TMEO 8 mm) to set randomly the order of the test. The 11 subjects who began with medialis gastrocnemius assessments, did the lateralis test following the same randomized protocol for each of the five different conditions and vice versa for the 12 participants who began with the lateralis test (Figure 3). Three running trials at 9 km/h [35] under five different conditions (SO, NIRO 6 mm, NIRO 8 mm, TMEO 6 mm, and TMEO 8 mm) on the same day were randomly performed. The duration of each trial was 1 min. For each subject, the mean EMG muscle activity pattern [36] of the gastrocnemius medialis of the dominant leg was recorded during the last 30 s of each 1-min trial, which was performed three times, leaving 5 min of rest between each test [37]. To avoid a potential imbalance, the same condition was added to contralateral foot. The same protocol was performed to assess another gastrocnemius EMG activity pattern. Subjects were blinded to which of the five random conditions that they were wearing, and the results were used to test the hypothesis. Sensors 2020, 20, 3205 5 of 12 Sensors 2020, 20, x FOR PEER REVIEW 5 of 12 assess another gastrocnemius EMG activity pattern. Subjects were blinded to which of the five random conditions that they were wearing, and the results were used to test the hypothesis. Figure 3. Randomized flow chart. Abbreviations: SO = shoe only; NIRO = novel inverted rocker orthoses; and TMEO = traditional Morton extension´s orthoses. 2.6. Statistical Analysis To test for reliability in the present research, within-day trial-to-trial intraclass correlation coefficient (ICC) and the standard error of measurement (SEM) were calculated for the subjects under the five conditions for each muscle during the running test [14]. According to Landis and Koch [38], coefficients of ICC that were lower than 0.20 indicated a slight agreement, 0.20–0.40 indicated fair reliability, 0.41–0.60 indicated moderate reliability, 0.61–0.80 indicated substantial reliability, and 0.81–1.00 indicated almost perfect reliability. The authors considered coefficients of ≥0.81 to be appropriate to consider the results of the study as valid. SEM assessed the minimal detectable change (MDC) for all measurements. This is known as reliable change index (RCI), and it was used to determine the clinical significance of the data [39]. The Shapiro–Wilks test was used to assess the normality of the sample, and normal a distribution was present if p >0.05. Demographic values were presented as the mean and standard deviation (±SD). The p-values for multiple comparisons were corrected with a non-parametric paired Friedman test to prove that all SOs, NIROs, and TMEOs conditions were different between them. The Wilcoxon test with Bonferroni’s correction was performed to analyze differences between the five different conditions, indicating statistically significant differences when p < 0.05 with a 95% CI. All the values that were generated using NeuroTrac® software were loaded into Excel® template (Windows® 97–2003), and they were analyzed using SPSS version 19.0 (SPSS Science, Chicago, IL, USA). 3. Results The Shapiro–Wilks test showed a non-normal distribution of the sample (p < 0.05), while the Friedman test showed that values were different between the five conditions (p < 0.05). All subjects were recruited from a biomechanical clinic in Madrid (Spain) over a two-month period (October to Figure 3. Randomized flow chart. Abbreviations: SO = shoe only; NIRO = novel inverted rocker orthoses; and TMEO = traditional Morton extension’s orthoses. 2.6. Statistical Analysis To test for reliability in the present research, within-day trial-to-trial intraclass correlation coefficient (ICC) and the standard error of measurement (SEM) were calculated for the subjects under the five conditions for each muscle during the running test [14]. According to Landis and Koch [38], coefficients of ICC that were lower than 0.20 indicated a slight agreement, 0.20–0.40 indicated fair reliability, 0.41–0.60 indicated moderate reliability, 0.61–0.80 indicated substantial reliability, and 0.81–1.00 indicated almost perfect reliability. The authors considered coefficients of ≥0.81 to be appropriate to consider the results of the study as valid. SEM assessed the minimal detectable change (MDC) for all measurements. This is known as reliable change index (RCI), and it was used to determine the clinical significance of the data [39]. The Shapiro–Wilks test was used to assess the normality of the sample, and normal a distribution was present if p >0.05. Demographic values were presented as the mean and standard deviation (±SD). The p-values for multiple comparisons were corrected with a non-parametric paired Friedman test to prove that all SOs, NIROs, and TMEOs conditions were different between them. The Wilcoxon test with Bonferroni’s correction was performed to analyze differences between the five different conditions, indicating statistically significant differences when p < 0.05 with a 95% CI. All the values that were generated using NeuroTrac® software were loaded into Excel® template (Windows® 97–2003), and they were analyzed using SPSS version 19.0 (SPSS Science, Chicago, IL, USA). Sensors 2020, 20, 3205 6 of 12 3. Results The Shapiro–Wilks test showed a non-normal distribution of the sample (p < 0.05), while the Friedman test showed that values were different between the five conditions (p < 0.05). All subjects were recruited from a biomechanical clinic in Madrid (Spain) over a two-month period (October to November 2019). Forty-five subjects were asked to participate in the experiment and assessed for eligibility; 24 did not meet the study entry requirements and three withdrew from the study because of pain and discomfort. Ultimately, 21 participants (10 males and 11 females) were enrolled into the study. The participants’ flow chart following the STROBE guidelines, is shown in Figure 4. Sociodemographic data are shown in Table 1. Sensors 2020, 20, x FOR PEER REVIEW 6 of 12 November 2019). Forty-five subjects were asked to participate in the experiment and assessed for eligibility; 24 did not meet the study entry requirements and three withdrew from the study because of pain and discomfort. Ultimately, 21 participants (10 males and 11 females) were enrolled into the study. The participants’ flow chart following the STROBE guidelines, is shown in Figure 4. Sociodemographic data are shown in Table 1. Figure 4. Participant flow chart. Table 1. Participant demographics. Variable n = 21 Mean ± SD (95% CI) Age 31.41 ± 4.33 (32.26–35.09) FPI (scores) 3.12 ± 0.17 (2.07–3.41) Weight (kg) 67.50 ± 8.06 (62.36–70.06) Height (cm) 170.08 ± 6.91 (166.9–172.43) BMI (kg/m2) 23.15 ± 3.05 (21.7–24.7) Abbreviations: SD = standard deviation; CI = confidence interval; FPI = foot posture index; and BMI = body mass index. The reliability of the data obtained from the EMG activity of muscles during running under five different conditions is presented as the ICC and SEM, which are shown in Table 2. Most of the values reached cut-off values over of 0.81 in the ICC data, which suggests “almost perfect reliability” [38], with 0.971 for NIRO-8 mm as the highest value and 0.458 for TMEO-8 mm as the lowest for the gastrocnemius lateralis, and 0.894 for TMEO-8 mm as the highest and 0.767 for NIRO-8 mm as the lowest for the gastrocnemius medialis. Considering the reference that was chosen by the authors, we dismissed TMEO-8 mm values for gastrocnemius lateralis. For SEM, 0.817 mV was the lowest value set for NIRO-8 mm, and 3.766 mV was the lowest value for TMEO-6 mm for the gastrocnemius lateralis, and 2.083 mV was the highest value for NIRO-8 mm and 0.326 Figure 4. Participant flow chart. Table 1. Participant demographics. Variable n = 21 Mean ± SD (95% CI) Age 31.41 ± 4.33 (32.26–35.09) FPI (scores) 3.12 ± 0.17 (2.07–3.41) Weight (kg) 67.50 ± 8.06 (62.36–70.06) Height (cm) 170.08 ± 6.91 (166.9–172.43) BMI (kg/m2) 23.15 ± 3.05 (21.7–24.7) Abbreviations: SD = standard deviation; CI = confidence interval; FPI = foot posture index; and BMI = body mass index. Sensors 2020, 20, 3205 7 of 12 The reliability of the data obtained from the EMG activity of muscles during running under five different conditions is presented as the ICC and SEM, which are shown in Table 2. Most of the values reached cut-off values over of 0.81 in the ICC data, which suggests “almost perfect reliability” [38], with 0.971 for NIRO-8 mm as the highest value and 0.458 for TMEO-8 mm as the lowest for the gastrocnemius lateralis, and 0.894 for TMEO-8 mm as the highest and 0.767 for NIRO-8 mm as the lowest for the gastrocnemius medialis. Considering the reference that was chosen by the authors, we dismissed TMEO-8 mm values for gastrocnemius lateralis. For SEM, 0.817 mV was the lowest value set for NIRO-8 mm, and 3.766 mV was the lowest value for TMEO-6 mm for the gastrocnemius lateralis, and 2.083 mV was the highest value for NIRO-8 mm and 0.326 mV was the lowest value for TMEO-8 mm for the gastrocnemius medialis. The highest MDC value for TMEO-8 mm was 5.798 mV and 2.264 mV were the lowest value for the gastrocnemius lateralis. Additionally, 5.775 mV was the highest value in the NIRO-8 mm group and 0.904 mV was the lowest value in the TMEO-8 mm group for gastrocnemius medialis. EMG mean muscle activity in the gastrocnemius medialis and lateralis in SO compared to NIRO-6 mm and 8 mm and TMEO-6 mm and 8 mm are shown in Table 3. In the gastrocnemius lateralis, the EMG activity significantly increased between the SO and NIRO-8 mm (22.27 ± 2.51 vs. 25.96 ± 4.68 mV; p < 0.05). There was another statistically significant increase between SO and TMEO-6 mm (22.27 ± 2.51 vs. 24.72 ± 5.08 mV, p < 0.05) and vs. TMEO-8 mm (25.49 ± 1.97, p < 0.001), but the low ICC of the last value invalidated the reliability of this value. For the gastrocnemius medialis, a statistically significant increase in the EMG activity was noted for SO vs. NIRO-6 mm (22.93 ± 2.1 vs. 26.44 ± 3.63, p < 0.001), vs. NIRO-8 mm (28.89 ± 3.6, p < 0.001), vs. TMEO-6 mm (25.12 ± 3.51, p < 0.05), and vs. TMEO-8 mm (26.38 ± 3.02, p < 0.05). The latter was not considered because of its low ICC value. In addition, the relationship between NIROs and TMEOs showed that there was a statistically significant increase in NIRO-6 mm and NIRO-8 mm (26.44 ± 3.63 vs. 28.89 ± 3.6, p < 0.05), and a statistically significant decrease in NIRO-8 mm vs. TMEO-6 mm (28.89 ± 3.6 vs. 25.12 ± 3.51, p < 0.001) and in NIRO-8 mm vs. TMEO-8 mm (28.89 ± 3.6 vs. 26.38 ± 3.02, p < 0.05), although the latter could not be considered because of its low ICC values. Sensors 2020, 20, 3205 8 of 12 Table 2. Reliability ICC of variables with “shoe only” versus 6- and 8-mm of novel inverted rocker orthoses (NIRO) and traditional Morton extension orthoses (TMEO). Variables SO NIRO-6 mm NIRO-8 mm TMEO-6 mm TMEO-8 mm ICC (95% CI) MDC ICC (95% CI) MDC ICC (95% CI) MDC ICC (95% CI) MDC ICC (95% CI) MDC SEM 0.950 SEM 0.950 SEM 0.950 SEM 0.950 SEM 0.950 Gastrocnemius lateralis (mV) 0.839 0.932 0.971 0.937 0.458 (0.651–0.935) 1.010 3.560 (0.852–0.973) 1.254 3.477 (0.938–0.988) 0.817 2.264 (0.861–0.975) 1.359 3.766 (0.148–0.777) 2.092 5.798 Gastrocnemius medialis (mV) 0.848 0.832 0.767 0.872 0.894 (0.649–0.94) 0.913 2.530 (0.637–0.931) 1.707 4.731 (0.501–0.905) 2.083 5.775 (0.723–0.948) 1.408 3.904 (0.77–0.957) 0.326 0.904 Abbreviations: ICC = intraclass correlation coefficient; CI = confidence interval; SEM = standard error of measurement; MDC = minimal detectable change; (mV) = millivolts; SO = shoe only; and mm = millimeters. Table 3. Signal amplitudes and comparison values of the mean gastrocnemius lateralis and medialis muscle activities. SO NIRO 6 mm NIRO 8 mm TMEO 6 mm TMEO 8 mm p-Value SO p-Value SO p-Value SO p-Value SO p-Value NIRO 6 mm p-Value NIRO 6 mm p-Value NIRO 6 mm p-Value NIRO 8 mm p-Value NIRO 8 mm p-Value TMEO 6 mm Variable mean (mV) mean(mV) mean (mV) mean (mV) mean (mV) vs. vs. vs. vs. vs. vs. vs. vs. vs. vs. gastrocnemius lateralis ±SD (95% CI) ±SD (95% CI) ±SD (95% CI) ±SD (95% CI) ±SD (95% CI) NIRO 6 mm NIRO 8 mm TMEO 6 mm TMEO 8 mm NIRO 8 mm TMEO 6 mm TMEO 8 mm TMEO 6 mm TMEO 8 mm TMEO 8 mm 22.27 ± 2.51 24.65 ± 4.51 25.96 ± 4.68 24.72 ± 5.08 25.49 ± 1.97 (20.77–23.279) (22.41–26.897) (23.634–28.29) (23.675–27.35) (22.19–27.253) 0.085 <0.05 * <0.05 * <0.001 ** 0.39 0.88 0.356 0.372 0.67 0.913 22.93 ± 2.1 26.44 ± 3.63 28.89 ± 3.6 25.12 ± 3.51 26.38 ± 3.02 gastrocnemius medialis (21.88–23.97) (24.63–28.24) (27–30.68) (23.37–26.87) (24.88–27.89) <0.001 ** <0.001 ** <0.05 * <0.05 * <0.05 * 0.06 0.67 <0.001 ** <0.05 * 0.22 Abbreviations: mV = millivolts; SO = shoe only; NIRO = novel inverted rocker orthoses; TMEO = traditional Morton extension orthoses; mm = millimeters; ±SD = standard deviation; p < 0.05 * (95% CI) was considered statistically significant; and p < 0.001 ** (95% CI) was considered statistically significant. Sensors 2020, 20, 3205 9 of 12 4. Discussion 4.1. TMEO and NIRO Effects This is the first study on EMG muscle activity in the gastrocnemius medialis and lateralis under IMTPJ dorsiflexion mobility restrictions by two different kinds of orthoses, the TMEO and the NIRO, in healthy subjects during running. TMEO has been used to treat symptoms of the first stages of OA [9–11] moving away dorsally from the contact between the proximal phalanx of the hallux and first metatarsal head surfaces. However, it is unclear if the effects on the triceps surae activity that were caused by the windlass mechanism [24] alteration through the IMTPJ caused the restriction. Some authors have shown the need for proper dorsiflexion of the IMTPJ during the push-off phase to ensure normal activity of the calcaneus–plantar system [24]. We hypothesized that TMEO would increase the EMG triceps surae activity that is induced by restriction of IMTPJ dorsiflexion. Our results showed that EMG activity of the gastrocnemius lateralis and medialis increased with TMEO-6 mm and that there is a further increase with TMEO-8 mm compared to SO (Table 3), although the last one could not be considered because of the low ICC values. Even knowing that there are no studies related to EMG activity during running with the orthopedic restriction of IMPTJ dorsiflexion, these results are consistent with other simulated running research [24,25], which showed that engaging the windlass mechanism by promoting 30◦ of IMTPJ dorsiflexion caused the arch to absorb and dissipate more elastic energy than under normal circumstances, and likely the energy of the triceps surae would be saved. In the present research, we decreased the windlass capacity through the TMEO, and followed the lack of storage and release energy in the medial longitudinal arch primary in the heel-off phase; this could have been supported by increasing gastrocnemius musculature EMG activity, as shown by our results, and by sustaining the connection between the IMTPJ and triceps surae through the windlass mechanism, according with other authors [24,25]. We hypothesized that NIRO would produce less EMG activity on triceps surae than the TMEO compared to SO. The rationale behind this approach was that its smooth edges and inverted rocker would produce a slight movement restriction of the IMTPJ; therefore, less effort would be required of the triceps surae to move the heel up. However, the present research showed surprising results, with a higher increase in EMG activity in both the gastrocnemius medialis and lateralis muscles (Table 3) with NIRO compared to TMEO, especially with NIRO-8 mm. This could be partly explained because of the soft edges of the NIRO, which yielded instability on the IMTPJ and transferred it to triceps surae in the heel-off phase. This is consistent with other studies with inverted rocker-sole shoes [40] that showed increased plantarflexion at the ankle joint and an increase in lower limb muscular activity [13]. This conclusion is not consistent with other research that showed increasing toe joint stiffness and increased ankle foot push-off work by up to 181% [41]. 4.2. Osteoarthritis OA has been defined as one of the most important and incapacitating musculoskeletal disorders in the world and OA of the IMTPJ, is the most commonly affected region on the foot [42]. This pathology can involves partial (FHL) or total (HR) rolling fail of the proximal phalanx of the hallux around first metatarsal bone in the last phase of gait [3], and there are a few treatments to relieve them, looking to avoid contact of the dorsal aspect of theses bones, such as TMEO [9–11] or classical rocker soles [12]. No studies about triceps surae EMG activity and IMTPJ OA using orthoses and/or rocker soles during running have been reported; nevertheless, our observations with simulated IMTPJ restriction through TMEO and NIRO, showed an increase of EMG activity pattern of the gastrocnemius medialis and lateralis, in contrast with a recently study [12] with IMTPJ OA and traditional rocker bottom soles, which argued that the reduction of the concentric activity of the triceps surae inferred from the forward displacements of the body center of mass was probably due to passively roll-over of the whole base of support. Sensors 2020, 20, 3205 10 of 12 4.3. Running Economy Elastic energy is stored and returned by the plantar muscles, plantar aponeurosis, and triceps surae with the Achilles tendon during the mid-stance and heel-off phases of running because of its isometric, concentric, and eccentric stretching–shortening pattern [43,44], which shows that the foot has an important role in RE. RE is related to different biomechanical parameters such as shorter ground contact times, higher stride frequency, joint stiffness, and neuromuscular response [20], specifically the pre-activation of gastrocnemius muscular group [14,17,20]. TMEO and NIRO somehow produced decreased stiffness in the IMTPJ by dorsal migration of the I metatarsal bone, and this was shown by the compensatory increase effect on the gastrocnemius musculature activity that attempts to stabilize IMTPJ instability when joined with the windlass mechanism. This would cause worse RE [20]. Our obtained values confirm the results of some studies [45,46], which showed the importance of neuromuscular pre-activation of the gastrocnemius to increase the leg stiffness, anticipating the loading forces and attenuating the effort of the foot to stabilize the joint as required, improving the energy cost and, therefore, the RE. 5. Limitations The sample size that was calculated in a previous study could not be attained because three individuals were excluded. This must be taken into account when interpreting the results. In addition, we were not able to assess the “order effect” on our sample because didn’t write the different orders of each participant’s choice, despite the fact that both groups had a similar participant number, the hypothetical order effect can take over, and we recommended future study designed to improve this aspect of the assessments. Because of the short running test duration when NIRO and TMEO were worn, the hypothetical muscular adaptations of the triceps surae could not be assessed. Longer studies in the future are needed to determine how the exertion levels can influence these muscular adaptations during running. Considering that most ±SD values obtained in the present research are higher than SEM, authors recommended to have caution in interpreting the results. 6. Conclusions NIRO and TMEO have shown a high interaction with triceps surae, increasing the gastrocnemius medialis and lateralis EMG activity during running. This may be additional evidence of the biomechanics relationship between IMTPJ and the windlass mechanism connection. Higher values of the triceps surae EMG activity wearing NIRO and TMEO during running could have a negative impact on RE; therefore, clinicians should be prescribing them with caution when they want to treat IMTPJ OA in runners. Author Contributions: Conceptualization, R.S.-G.; methodology, C.R.-M., M.K. and I.O.-S.; software, I.Z.-G. and I.O.-S.; validation, Á.G.-C. and P.A.; formal analysis, C.R.-M., B.D.-l.-C.-T. and I.Z.-G.; investigation, R.S.-G. and P.A.; resources, C.R.-M. and B.D.-l.-C.-T.; data curation, B.D.-l.-C.-T.; writing—original draft preparation; R.S.-G., C.R.-M., B.D.-l.-C.-T., I.O.-S., I.Z.-G., P.A. and M.K.; visualization, P.A.; supervision, Á.G.-C. and M.K.; project administration, B.D.-l.-C.-T. All authors have read and agreed to the published version of the manuscript Funding: This research received no external fundings. Conflicts of Interest: The authors declare no conflict of interest. References 1. Cotterill, J.M. Stiffness of the Great Toe in Adolescents. Br. Med. J. 1887, 1, 1158. Available online: http://www.ncbi.nlm.nih.gov/pubmed/20751923 (accessed on 21 February 2019). [CrossRef] [PubMed] 2. Laird, P.O. Functional Hallux Limitus. Ill. Podiatr. 1972, 9, 4. 3. Drago, J.J.; Oloff, L.; Jacobs, A.M. A comprehensive review of hallux limitus. J. Foot Surg. 1984, 23, 213–220. [PubMed] Sensors 2020, 20, 3205 11 of 12 4. Zammit, G.V.; Menz, H.B.; Munteanu, S.E. Structural factors associated with hallux limitus/rigidus: A systematic review of case control studies. J. Orthop. Sports Phys. Ther. 2009, 39, 733–742. [CrossRef] [PubMed] 5. Dananberg, H.J. Functional hallux limitus and its relationship to gait efficiency. J. Am. Podiatr. Med. Assoc. 1986, 76, 648–652. Available online: http://www.ncbi.nlm.nih.gov/pubmed/3814239 (accessed on 17 August 2015). [CrossRef] 6. Tenforde, A.S.; Yin, A.; Hunt, K.J. Foot and Ankle Injuries in Runners. Physical Medicine and Rehabilitation Clinics of North America; Saunders, W.B., Ed.; Elsevier Inc.: Amsterdam, The Netherlands, 2016; Volume 27, pp. 121–137. Available online: http://www.ncbi.nlm.nih.gov/pubmed/26616180 (accessed on 15 May 2020). 7. Polzer, H.; Polzer, S.; Brumann, M.; Mutschler, W.; Regauer, M. Hallux rigidus: Joint preserving alternatives to arthrodesis—A review of the literature. World J. Orthop. 2014, 5, 6–13. [CrossRef] 8. Shurnas, P.S. Hallux Rigidus: Etiology, Biomechanics, and Nonoperative Treatment. Foot Ankle Clin. 2009, 14, 1–8. [CrossRef] 9. Dananberg, H.J. Gait style as an etiology to chronic postural pain. Part, I. Functional hallux limitus. J. Am. Podiatr. Med. Assoc. 1993, 83, 433–441. [CrossRef] 10. Smith, R.W.; Katchis, S.D.; Ayson, L.C. Outcomes in hallux rigidus patients treated nonoperatively: A long-term follow-up study. Foot Ankle Int. 2000, 21, 906–913. [CrossRef] 11. Grady, J.F.; Axe, T.M.; Zager, E.J.; Sheldon, L.A. A retrospective analysis of 772 patients with hallux limitus. J. Am. Podiatr. Med. Assoc. 2002, 92, 102–108. [CrossRef] 12. Menz, H.B.; Auhl, M.; Tan, J.M.; Levinger, P.; Roddy, E.; Munteanu, S.E. Biomechanical Effects of Prefabricated Foot Orthoses and Rocker-Sole Footwear in Individuals with First Metatarsophalangeal Joint Osteoarthritis. Arthritis Care Res. 2016, 68, 603–611. [CrossRef] 13. Romkes, J.; Rudmann, C.; Brunner, R. Changes in gait and EMG when walking with the Masai Barefoot Technique. Clin. Biomech. 2006, 21, 75–81. Available online: https://www.sciencedirect.com/science/article/ abs/pii/S0268003305001816 (accessed on 15 May 2020). [CrossRef] [PubMed] 14. Saunders, P.U.; Pyne, D.B.; Telford, R.D.; Hawley, J.A. Reliability and variability of running economy in elite distance runners. Med. Sci. Sports Exerc. 2004, 36, 1972–1976. [CrossRef] [PubMed] 15. Hayes, P.R.; Bowen, S.J.; Davies, E.J. The relationships between local muscular endurance and kinematic changes during a run to exhaustion at vVO2MAX. J. Strength Cond. Res. 2004, 18, 898–903. 16. Roca-Dols, A.; Losa-Iglesias, M.E.; Sánchez-Gómez, R.; Becerro-de-Bengoa-Vallejo, R.; López-López, D.; Rodríguez-Sanz, D.; Martínez-Jiménez, E.M.; Calvo-Lobo, C. Effect of the cushioning running shoes in ground contact time of phases of gait. J. Mech. Behav. Biomed. Mater. 2018, 88, 196–200. [CrossRef] [PubMed] 17. Roca-Dols, A.; Elena Losa-Iglesias, M.; Sánchez-Gómez, R.; Becerro-de-Bengoa-Vallejo, R.; López-López, D.; Palomo-López, P.; Rodríguez-Sanz, D.; Calvo-Lobo, C. Electromyography activity of triceps surae and tibialis anterior muscles related to various sports shoes. J. Mech. Behav. Biomed. Mater. 2018, 86, 158–171. [CrossRef] 18. Roca-Dols, A.; Losa-Iglesias, M.E.; Sánchez-Gómez, R.; López-López, D.; Becerro-de-Bengoa-Vallejo, R.; Calvo-Lobo, C. Electromyography comparison of the effects of various footwear in the activity patterns of the peroneus longus and brevis muscles. J. Mech. Behav. Biomed. Mater. 2018, 82, 126–132. [CrossRef] 19. Blagrove, R.C.; Howatson, G.; Hayes, P.R. Effects of Strength Training on the Physiological Determinants of Middle- and Long-Distance Running Performance: A Systematic Review. Sports Medicine; Springer International Publishing: Cham, Switzerland, 2018; Volume 48, pp. 1117–1149. 20. Tam, N.; Tucker, R.; Santos-Concejero, J.; Prins, D.; Lamberts, R.P. Running Economy: Neuromuscular and Joint-Stiffness Contributions in Trained Runners. Int. J. Sports Physiol. Perform. 2019, 14, 16–22. [CrossRef] 21. Hicks, J. The mechanics of the foot. II. The plantar aponeurosis and the arch. J. Anat. 1954, 88, 25–30. 22. Stearne, S.M.; McDonald, K.A.; Alderson, J.A.; North, I.; Oxnard, C.E.; Rubenson, J. The Foot’s Arch and the Energetics of Human Locomotion. Sci. Rep. 2016, 6, 1–10. [CrossRef] 23. Ker, R.F.; Bennett, M.B.; Bibby, S.R.; Kester, R.C.; Alexander, R.M. The spring in the arch of the human foot. Nature 1987, 325, 147–149. [CrossRef] 24. Maceira, E.; Monteagudo, M. Functional hallux rigidus and the Achilles-calcaneus-plantar system. Foot Ankle Clin. 2014, 19, 669–699. [CrossRef] 25. Welte, L.; Kelly, L.A.; Lichtwark, G.A.; Rainbow, M.J. Influence of the windlass mechanism on arch-spring mechanics during dynamic foot arch deformation. J. R. Soc. Interface 2018, 15, 20180270. [CrossRef] Sensors 2020, 20, 3205 12 of 12 26. Johanson, M.A.; Allen, J.C.; Matsumoto, M.; Ueda, Y.; Wilcher, K.M. Effect of Heel Lifts on Plantarflexor and Dorsiflexor Activity During Gait. Foot Ankle Int. 2010, 31, 1014–1020. [CrossRef] [PubMed] 27. Von Elm, E.; Altman, D.G.; Egger, M.; Pocock, S.J.; Gøtzsche, P.C.; Vandenbroucke, J.P. The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Statement: Guidelines for reporting observational studies. Int. J. Surg. 2014, 12, 1495–1499. [CrossRef] [PubMed] 28. Redmond, A.C.; Crosbie, J.; Ouvrier, R.A. Development and validation of a novel rating system for scoring standing foot posture: The Foot Posture Index. Clin. Biomech. (Bristol. Avon.) 2006, 21, 89–98. [CrossRef] 29. Root, M.L.; Orien, W.P.; Weed, J.H.; Hughes, R.J. Normal and Abnormal Function of the Foot; Clinical Biomechanics Corp.: Los Angeles, CA, USA, 1977; Volume 2. 30. Sánchez-Gómez, R.; Becerro-de-Bengoa-Vallejo, R.; Losa-Iglesias, M.E.; Calvo-Lobo, C.; Navarro-Flores, E.; Palomo-López, P.; Romero-Morales, C.; López-López, D. Reliability Study of Diagnostic Tests for Functional Hallux Limitus. Foot Ankle Int. 2020, 41, 457–462. [CrossRef] [PubMed] 31. Garrow, J.S.; Webster, J. Quetelet’s index (W/H2) as a measure of fatness. Int. J. Obes. 1985, 9, 147–153. [PubMed] 32. Naess, I.; Bø, K. Can maximal voluntary pelvic floor muscle contraction reduce vaginal resting pressure and resting EMG activity? Int. Urogynecol. J. 2018, 29, 1623–1627. [CrossRef] 33. Hermens, H.J.; Freriks, B.; Disselhorst-Klug, C.; Rau, G. Development of recommendations for SEMG sensors and sensor placement procedures. J. Electromyogr. Kinesiol. 2000, 10, 361–374. [CrossRef] 34. Murley, G.S.; Buldt, A.K.; Trump, P.J.; Wickham, J.B. Tibialis posterior EMG activity during barefoot walking in people with neutral foot posture. J. Electromyogr. Kinesiol. 2009, 19, e69–e77. [CrossRef] [PubMed] 35. Shih, Y.; Lin., K.-L.; Shiang, T.-Y. Is the foot striking pattern more important than barefoot or shod conditions in running? Gait Posture 2013, 38, 490–494. [CrossRef] [PubMed] 36. Goto, K.; Abe, K. Gait characteristics in women’s safety shoes. Appl. Ergon. 2017, 65, 163–167. [CrossRef] 37. Fleming, N.; Walters, J.; Grounds, J.; Fife, L.; Finch, A. Acute response to barefoot running in habitually shod males. Hum. Mov. Sci. 2015, 42, 27–37. [CrossRef] [PubMed] 38. Landis, J.R.; Koch, G.G. The measurement of observer agreement for categorical data. Biometrics 1977, 33, 159–174. [CrossRef] [PubMed] 39. Jacobson, N.; Truax, P. Clinical significance: A statistical approach to defining meaningful change in psychotherapy research. J. Consult Clin. Psychol. 1991, 59, 12–19. [CrossRef] 40. Myers, K.A.; Long, J.T.; Klein, J.P.; Wertsch, J.J.; Janisse, D.; Harris, G.F. Biomechanical implications of the negative heel rocker sole shoe: Gait kinematics and kinetics. Gait Posture 2006, 24, 323–330. [CrossRef] 41. Honert, E.C.; Bastas, G.; Zelik, K.E. Effect of toe joint stiffness and toe shape on walking biomechanics. Bioinspir. Biomim. 2018, 13, 066007. [CrossRef] 42. Sebbag, E.; Felten, R.; Sagez, F.; Sibilia, J.; Devilliers, H.; Arnaud, L. The world-wide burden of musculoskeletal diseases: A systematic analysis of the World Health Organization Burden of Diseases Database. Ann. Rheum. Dis. 2019, 78, 844–848. [CrossRef] 43. Biewener, A.A.; Roberts, T.J. Muscle and tendon contributions to force, work, and elastic energy savings: A comparative perspective. Exerc. Sport Sci. Rev. 2000, 28, 99–107. 44. Wilson, A.; Lichtwark, G. The anatomical arrangement of muscle and tendon enhances limb versatility and locomotor performance. Philos. Trans. R. Soc. Lond. B Biol. Sci. 2011, 366, 1540–1553. [CrossRef] [PubMed] 45. Boyer, K.A.; Nigg, B.M. Muscle activity in the leg is tuned in response to impact force characteristics. J. Biomech. 2004, 37, 1583–1588. [CrossRef] [PubMed] 46. Hamner, S.R.; Seth, A.; Delp, S.L. Muscle contributions to propulsion and support during running. J. Biomech. 2010, 43, 2709–2716. [CrossRef] [PubMed] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Effects of Novel Inverted Rocker Orthoses for First Metatarsophalangeal Joint on Gastrocnemius Muscle Electromyographic Activity during Running: A Cross-Sectional Pilot Study.

06-05-2020

Sánchez-Gómez, Rubén,Romero-Morales, Carlos,Gómez-Carrión, Álvaro,De-la-Cruz-Torres, Blanca,Zaragoza-García, Ignacio,Anttila, Pekka,Kantola, Matti,Ortuño-Soriano, Ismael

eng

PMC7379642

Supplement Table 7. Change in VO2max (L·min-1 and ml·min-1·kg-1) from 1995-1997 to 2016-2017 in relation to sex and length of education. L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 Year n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change 95-97 278 2.41 (0.12) Ref 36.1 (2.42) Ref 1 719 2.44 (0.12) Ref 37.6 (2.18) Ref 398 2.59 (0.12) Ref 40.2 (2.16) Ref 98-99 324 2.44 (0.15) 1,4% 36.5 (2.65) 1,2% 1 998 2.35 (0.17) -3,9% 35.7 (2.39) -5,0% 642 2.59 (0.14) 0,1% 39.6 (2.37) -1,5% 00-01 721 2.33 (0.10) -3,2% 33.9 (1.99) -6,0% 4 172 2.47 (0.14) 1,2% 37.2 (2.33) -1,0% 1 313 2.46 (0.19) -5,0% 37.9 (2.74) -5,6% 02-03 1 154 2.12 (0.17) -12,1% 32.1 (2.54) -11,1% 8 157 2.33 (0.13) -4,7% 35.0 (2.18) -6,9% 2 547 2.45 (0.12) -5,5% 37.4 (2.20) -6,8% 04-05 1 512 2.17 (0.16) -9,9% 32.6 (2.60) -9,7% 12 670 2.34 (0.13) -4,1% 34.7 (2.30) -7,7% 5 318 2.46 (0.13) -4,9% 37.6 (2.20) -6,3% 06-07 1 546 2.18 (0.16) -9,6% 32.5 (2.62) -10,0% 12 075 2.36 (0.12) -3,3% 35.0 (2.09) -6,8% 5 093 2.49 (0.11) -3,8% 37.8 (1.89) -6,0% 08-09 1 416 2.26 (0.13) -6,1% 33.0 (2.27) -8,5% 12 591 2.37 (0.12) -3,0% 34.9 (1.98) -7,1% 6 061 2.52 (0.12) -2,6% 38.2 (2.06) -4,9% 10-11 1 131 2.22 (0.12) -7,8% 32.3 (1.98) -10,5% 10 401 2.36 (0.13) -3,2% 34.5 (2.13) -8,2% 5 769 2.52 (0.13) -2,8% 38.0 (2.24) -5,4% 12-13 1 212 2.23 (0.10) -7,4% 32.1 (1.89) -11,2% 13 049 2.36 (0.12) -3,3% 34.4 (2.03) -8,4% 9 075 2.49 (0.13) -3,7% 37.9 (2.22) -5,8% 14-15 982 2.15 (0.13) -11,0% 30.8 (2.21) -14,7% 11 722 2.33 (0.12) -4,4% 34.0 (1.94) -9,6% 8 190 2.47 (0.11) -4,8% 37.3 (1.88) -7,2% 16-17 521 2.14 (0.10) -11,1% 31.4 (1.44) -13,1% 7 351 2.32 (0.13) -4,7% 34.0 (2.04) -9,6% 5 592 2.48 (0.11) -4,3% 37.4 (1.71) -6,9% L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 Year n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change 95-97 453 2.98 (0.12) Ref 36.3 (1.44) Ref 1 497 3.19 (0.17) Ref 39.5 (2.33) Ref 180 3.21 (0.17) Ref 39.5 (2.08) Ref 98-99 556 2.75 (0.20) -7,6% 33.4 (3.12) -8,0% 2 418 3.13 (0.17) -2,0% 38.5 (2.39) -2,5% 605 3.13 (0.17) -2,6% 38.3 (2.46) -2,9% 00-01 822 2.93 (0.15) -1,6% 36.2 (2.35) -0,4% 4 226 3.05 (0.15) -4,4% 36.7 (2.33) -7,2% 1 291 3.06 (0.24) -4,8% 37.2 (3.78) -5,9% 02-03 1 418 2.66 (0.20) -10,6% 32.7 (2.81) -9,8% 7 394 2.97 (0.18) -7,0% 36.0 (2.38) -8,9% 1 959 3.14 (0.16) -2,3% 38.6 (2.23) -2,3% 04-05 2 113 2.84 (0.16) -4,8% 34.8 (2.29) -4,3% 11 642 2.97 (0.16) -6,8% 36.4 (1.93) -7,7% 4 165 3.08 (0.18) -4,1% 38.0 (2.29) -3,8% 06-07 2 363 2.87 (0.14) -3,7% 34.3 (2.08) -5,5% 13 092 2.96 (0.16) -7,4% 35.7 (1.99) -9,7% 4 350 3.10 (0.17) -3,6% 38.1 (2.14) -3,5% 08-09 2 755 2.71 (0.18) -9,1% 32.5 (2.51) -10,5% 15 466 2.99 (0.15) -6,2% 35.8 (1.93) -9,4% 5 190 3.14 (0.17) -2,3% 38.4 (2.15) -2,8% 10-11 2 495 2.80 (0.14) -6,0% 33.3 (2.06) -8,2% 14 436 2.98 (0.15) -6,6% 35.5 (1.90) -10,2% 4 945 3.18 (0.14) -0,9% 38.6 (1.83) -2,3% 12-13 3 172 2.66 (0.17) -10,7% 31.8 (2.23) -12,3% 21 789 2.95 (0.15) -7,6% 35.0 (1.94) -11,4% 8 949 3.10 (0.15) -3,6% 37.9 (2.05) -3,9% 14-15 3 071 2.69 (0.15) -9,8% 32.0 (1.93) -11,9% 23 325 2.91 (0.13) -8,7% 34.4 (1.77) -13,0% 8 294 3.04 (0.15) -5,3% 37.1 (2.09) -6,0% 16-17 1 925 2.65 (0.16) -11,1% 31.7 (2.03) -12,6% 15 990 2.91 (0.14) -8,9% 34.1 (1.78) -13,6% 5 182 3.03 (0.16) -5,5% 36.7 (2.06) -7,1% Women Men ≥12 years 10-12 years ≤9 years ≤9 years 10-12 years ≥12 years

Decline in cardiorespiratory fitness in the Swedish working force between 1995 and 2017.

11-15-2018

Ekblom-Bak, Elin,Ekblom, Örjan,Andersson, Gunnar,Wallin, Peter,Söderling, Jonas,Hemmingsson, Erik,Ekblom, Björn

eng

PMC6358870

medicina Article Pacing of Women and Men in Half-Marathon and Marathon Races Pantelis T. Nikolaidis 1,* , Ivan ´Cuk 2 and Beat Knechtle 3 1 Exercise Physiology Laboratory, 18450 Nikaia, Greece 2 Faculty of Physical Education and Sports Management, Singidunum University, 11000 Belgrade, Serbia; [email protected] 3 Institute of Primary Care, University of Zurich, 8006 Zürich, Switzerland; [email protected] * Correspondence: [email protected]; Tel.: +306977820298 Received: 4 November 2018; Accepted: 9 January 2019; Published: 14 January 2019



 Abstract: Background and objective: Half-marathon is the most popular endurance running race in terms of number of races and runners competing annually; however, no study has compared pacing strategies for this race distance with marathon. The aim of the present study was to profile pacing in half-marathon, compare half-marathon and marathon for pacing, and estimate sex differences in pacing. Materials and methods: A total of 9137 finishers in the half-marathon (n = 7258) and marathon race (n = 1853) in Ljubljana 2017 were considered for their pacing in five race segments (0–23.7%, 23.7–47.4%, 47.4–71.1%, 71.1–94.8%, and 94.8–100% of the race. Results: Half-marathon runners followed a positive pacing with every segment being slower than its previous one without the presence of an endspurt. Compared to marathon (where the average percent of change in speed (ACS) was 5.71%), a more even pacing was observed in half-marathon (ACS = 4.10%). Moreover, women (ACS = 4.11%) had similar pacing as men (ACS = 4.09%) in half-marathons. Conclusions: In summary, running a half-marathon followed a unique pattern that differentiated this race distance from marathon, with the former showing a more even pacing with an absence of endspurt, and sex difference compared to the latter. Consequently, runners should be advised to adopt a less variable pacing when competing in a half-marathon, regardless of their sex. To the best of our knowledge, the more even pacing in half-marathon, than in marathon, was a novel finding, as it was the first study to compare the two race distances for this characteristic. Keywords: aerobic exercise; endurance; marathon; performance; running 1. Introduction The half-marathon has evolved as a sport discipline of increasing popularity, documented by the annual number of finishers and races taking place worldwide [1]. In general, performance in endurance running, as well as in endurance sports of other modes of exercise (cycling, swimming, cross-country skiing), has been shown to be associated with pacing, among other physiological and psychological variables [2,3]. Despite the popularity of half-marathon, limited information about the pacing, in this sport discipline, exists [4]. Since half-marathon has been characterized by an increased woman participation compared to marathon [5], and sex differences in pacing in marathon have been observed [2,3], estimating the pacing in half-marathon would be of great practical interest, especially considering the aspect of sex. Pacing has been well studied in many endurance and ultra-endurance sports, such as cycling [6,7], swimming [8,9], and triathlon [10]. It has been observed that pacing in cycling varied depending on whether cyclists performed exercise for a given time or distance, showing a faster start when competing for distance [6]. Research on the world’s longest ultra-cycling race showed a decrease of speed across Medicina 2019, 55, 14; doi:10.3390/medicina55010014 www.mdpi.com/journal/medicina Medicina 2019, 55, 14 2 of 9 the race [7]. The variation of speed, in 800 m for women [8] and 1500 m freestyle swimming for men, followed a U shape [9]. In sprint triathlon, a comparison among three pacing strategies (positive, negative, and even, where speed either decreased, increased, or remained stable across race) of swimming indicated that a positive pacing in swimming induced a lower rate of perceived exertion than a negative pacing [10]. The abovementioned studies found different pacing strategies among endurance and ultra-endurance sports, which did not allow the generalization of their findings to other sports. Adopting a pacing strategy might depend on athletes’ decision making [11], perception of risk [12], and individual variability [13]. For instance, pacing might be influenced by internal bodily state feedback, teleoanticipation, template formation of perceived exertion, and human-environment interaction (e.g., interaction among competitors) [11]. In addition, it was shown that cyclists and ultra-marathon runners with low perception of risk started the race faster than their counterparts with high perception of risk [12]. Furthermore, inter-individual differences in the distribution of effort, across a race, might be partially explained by athletes’ motivation [13]. So far, with regards to endurance running sport disciplines of high popularity, pacing has been well studied in marathon, where it has been found that runners adopted a positive pacing, i.e., their speed decreased across race, accompanied by the presence of an end spurt, i.e., the speed increased in the final section [2,14–16]. A positive pacing has also been observed in shorter distances, such as 800 m [17], 5 km [18], and 10 km [19]. With regards to sex differences, women marathon runners adopted a more even pacing in marathon [3,20] and in 100 km, than men [21]. Perceived effort and physiological parameters have been identified as correlates of pacing, and their role might vary across a 10 km race; e.g., perceived effort influenced speed at the start of the race, whereas mainly aerobic capacity and muscle strength-to a lesser degree-influenced speed for the rest of the race [22]. Moreover, with regards to the relationship of pacing with motivation, it has been shown that men with more even pacing scored higher in psychological coping, self-esteem, life meaning, recognition, and competition, than their counterparts with more variable pacing [23]. Although many studies have been conducted on the pacing of marathon [15,16,24–26], the limited relevant information that existed for half-marathon [4] had focused on elite runners. Since these two distance races differed for performance characteristics [5], it would be reasonable to assume that knowledge from pacing in marathon could not be “transferred” to half-marathon. Knowledge about pacing in half-marathon would have both theoretical and practical interest. From a theoretical perspective, exercise physiologists would be interested in the patterns of energy distribution across a 21 km running race and on potential sex differences in these patterns. From a practical perspective, coaches and fitness trainers working with half-marathon runners would use such knowledge to assist their athletes adopting sex-tailored pacing strategies. Therefore, the aim of the present study was to (a) examine changes in speed and whether pacing would be positive, negative, or even across half-marathon and marathon races, (b) investigate whether pacing would vary by race distance, and (c) estimate sex differences in pacing for both race distances. It was hypothesized that half-marathon would present a positive pacing with an endspurt, and women would adopt a more even pacing than men. 2. Materials and Methods 2.1. Participants This study was approved by the Institutional Review Board of Kanton St. Gallen, Switzerland, with a waiver of the requirement for informed consent of the participants as the study involved the analysis of publicly available data. The study was conducted in accordance with recognized ethical standards according to the Declaration of Helsinki adopted in 1964, and revised in 2013. For the purpose of this study, valid results and split times from 1853 participants (age 41.7 ± 9.8 years, range 17–78 years; average race speed 3.03 ± 0.46 m/s, 2.13–5.47 m/s; mean ± standard deviation) of the 2017 Ljubljana marathon and 7258 participants (age 40.3 ± 10.7 years, 12–86 Medicina 2019, 55, 14 3 of 9 years; 2.99 ± 0.45 m/s, 1.61–5.15 m/s) of the 2017 Ljubljana half-marathon (total 9137 participants) were analyzed. The obtained data presents publicly available, official results from the “Ljubljana Marathon” website [27]. Participants who did not finish the race, or did not have any recorded split times, were excluded from the study. Both marathon and half-marathon were held on the same day and on the same track, whereas half-marathon race is entirely contained within marathon race. This approach assisted in eliminating the potential influence of environmental conditions [28]. Moreover, both marathon and half-marathon were rather flat, with an elevation difference of 29 m (ranging from 295–324 m). The temperature on the race day ranged from 4.2–15.4 ◦C, without strong wind or excess humidity. 2.2. Variables and Research Measures From the available data, the average race speed 0–100% (0–21.0975 km for half-marathon and 0–42.195 km for marathon) was calculated. Moreover, average running speed in five race segments was estimated for both marathon and half-marathon, that corresponded to: Segment 1-Average running speed from 0–23.7% of the race (i.e., 0–5 km for half-marathon and 0–10 km for marathon) Segment 2-Average running speed from 23.7–47.4% of the race (i.e., 5–10 km for half-marathon and 10–20 km for marathon) Segment 3-Average running speed from 47.4–71.1% of the race (i.e., 10–15 km for half-marathon and 20–30 km for marathon) Segment 4-Average running speed from 71.1–94.8% of the race (i.e., 15–20 km for half-marathon and 30–40 km for marathon) Segment 5-Average running speed from 94.8–100% of the race (i.e., 20–21.0975 km for half-marathon and 40–42.195 km for marathon) 2.3. Process Thereafter, the percentage of average change in speed for each segment (ACSS), with regards to the average race speed, was calculated. The applied equation is as follows: ACSS = 100 − (100 × average race speed/average segment speed). Finally, the average percentage of change in speed, through the 5 race segments (average change in speed (ACS)), was estimated. Note that absolute values of ACS were presented and statistically tested. The applied equation was as follows: ACS = (ACSS1 + ACSS2 + ACSS3 + ACSS4 + ACSS5)/5. Changes in speed across a running race have been previously used to study pacing in 800 m [17], 5 km [18], 10 km [19], half-marathon [4], and marathon [15]. In the present study, the speed in each segment was expressed as the percentage difference from the average race speed, in order to provide comparable data considering potential differences in average race speed between races or sexes. 2.4. Data Analysis Prior to all statistical tests, descriptive statistics were calculated as mean, standard deviation, and minimum and maximum values. Since the Kolmogorov-Smirnov test is not sensitive in large samples, data distribution normality was assessed by inspecting histograms and QQ plots. After careful examination, the obtained data showed rather normal distribution. Mixed between-within analysis of variance (ANOVA) was performed for ACSS to test differences between segments (i.e., Segments 1–5; within-subject factor), race (i.e., marathon and half-marathon; between-subjects factor), as well as their interaction (segment × race). To further assess race differences, four additional mixed between-within ANOVAs for ACSS were performed. Two ANOVAs were performed to assess differences between segments (i.e., Segments 1 to 5; within-subject factor), race (i.e., marathon and half-marathon; between-subjects factor) as well as their interaction (segment × race), separately for men and women. Another two ANOVAs were performed to assess differences between segments (i.e., Segments 1 to Medicina 2019, 55, 14 4 of 9 5; within-subject factor), sex (i.e., men and women), as well as their interaction (segment × sex), separately for marathon and half-marathon. Finally, one two-way ANOVA was performed on ACS to assess differences between races (i.e., marathon and half-marathon), sex (i.e., men and women), as well as their interaction (race × sex). For all ANOVAs, a post hoc Bonferroni test was performed. Effect size was presented via eta squared ( 2), where the values of 0.01, 0.06, and above 0.14 were considered small, medium, and large, respectively [29]. Alpha level was set at p < 0.05. All statistical tests were performed using Microsoft Office Excel 2007 (Microsoft Corporation, Redmond, WA, USA) and SPSS 20 (IBM, Armonk, NY, USA). 3. Results The average running speeds for five segments, as well as the average race speed of participants by sex and race distance, are presented in Table 1. From the descriptive data in Table 1, a gradual decrease in average speed through the race segments was observed for both sexes in both marathon and half-marathon, e.g., from 3.26 ± 0.44 m/s in segment 1 to 2.99 ± 0.50 m/s in segment 5 in men half-marathon runners. Moreover, the largest deviation of running speed was observed in marathon men, i.e., decrease by 0.38 m/s from segment 1–5, whereas the smallest deviation of running speed was observed in half-marathon women, i.e., decrease by 0.25 m/s from segment 1–5. Further examination of participants’ speed and speed change is presented in Figures 1–3. Table 1. Segments and race speed for men and women, marathon, and half-marathon runners. Segment 1 Speed (m/s) Segment 2 Speed (m/s) Segment 3 Speed (m/s) Segment 4 Speed (m/s) Segment 5 Speed (m/s) Average Race Speed (m/s) Men 42.2 km N = 1478 Median 3.24 3.18 3.06 2.80 2.88 3.04 Mean 3.29 3.22 3.09 2.82 2.91 3.08 SD 0.44 0.45 0.49 0.52 0.49 0.46 CV 0.13 0.14 0.16 0.18 0.17 0.15 Women 42.2 km N = 375 Median 2.96 2.90 2.78 2.63 2.77 2.80 Mean 3.04 2.95 2.82 2.68 2.81 2.86 SD 0.36 0.38 0.43 0.44 0.41 0.39 CV 0.12 0.13 0.15 0.16 0.15 0.14 Men 21.1 km N = 4406 Median 3.21 3.18 3.06 3.04 2.97 3.10 Mean 3.26 3.22 3.09 3.06 2.99 3.14 SD 0.44 0.44 0.46 0.49 0.50 0.45 CV 0.14 0.14 0.15 0.16 0.17 0.14 Women 21.1 km N = 2852 Median 2.90 2.83 2.69 2.69 2.64 2.76 Mean 2.91 2.83 2.70 2.69 2.66 2.77 SD 0.32 0.35 0.36 0.39 0.39 0.35 CV 0.11 0.12 0.14 0.14 0.15 0.13 SD = standard deviation; CV = coefficient of variation. Figure 1. Percentage of speed change by split section in half-marathon and marathon. Error bars present standard deviation. ** p < 0.01 for significant difference between races. Medicina 2019, 55, 14 5 of 9 Figure 2. Percentage of speed change by segment and race distance in men (a) and women (b). Error bars present standard deviation. Figure 3. Absolute average change of speed by race distance and sex. Error bars present standard deviation. ** p < 0.01 for differences between sexes; ˆˆ p < 0.01 for differences between races. In regards to marathon and half-marathon runners of both sexes (Figure 1), the significant main effects of segment (F(4, 9106) = 5959.2, 2 = 0.36, large magnitude, p < 0.01), race (F(4, 9106) = 11.7, 2 < 0.01, trivial, p = 0.17), and segment × race interaction (F(4, 9106) = 723.1, 2 = 0.04, small, p < 0.01) were observed. Specifically, in both marathon and half-marathon runners, each segment significantly differs (p < 0.01) in speed change compared to the others. When only men runners were considered (Figure 2a), results indicated the significant main effects of segment (F(4, 5879) = 4392.3, 2 = 0.39, large, p < 0.01), race (F(4, 5879) = 64.2, 2 < 0.01, trivial, p < 0.01), and segment × race interaction (F(4, 5879) = 609.4, 2 = 0.05, small, p < 0.01). Specifically, for the first two segments (2.99% and 2.16% respectively), and the last two segments (6.81% and 0.69% respectively), men marathon runners showed greater changes in speed than half-marathon runners of the same sex (p < 0.01). However, in the third segment, men marathon runners presented more stable speed than half-marathon runners by 1.81% (p < 0.01). Medicina 2019, 55, 14 6 of 9 Similar results were obtained when only women runners were observed (Figure 2b). Significant main effects of segment (F(4, 3222) = 1279.2, 2 = 0.27, large, p < 0.01), race (F(4, 3222) = 57.3, 2 < 0.01, trivial, p < 0.01), and segment × race interaction (F(4, 3222) = 100.0, 2 = 0.02, small, p < 0.01) were observed. In particular, for the first two segments (1.13% and 0.94% respectively), and the fourth segment (3.90%), women marathon runners showed a greater speed change than women half-marathon runners (p < 0.01). On the other hand, in the third (1.28%) and the fifth segment (2.53%), women marathon runners exhibited more stable speed than women half-marathon runners of the same sex (p < 0.01). When marathon runners only were observed (Figure 2a,b), the results confirmed the significant main effects of segment (F(4, 1848) = 1131.2, 2 = 0.36, large, p < 0.01), sex (F(4, 1848) = 88.0, 2 < 0.01, trivial, p < 0.01), and segment × sex interaction (F(4, 1848) = 57.0, 2 = 0.02, small, p < 0.01). Specifically, for the first two segments (0.69% and 1.38% respectively), and the last two segments (2.75% and 4.35% respectively), men marathon runners presented greater changes in speed than women marathon runners (p < 0.01). For the third and fourth segments (1.63% respectively), men marathon runners showed more stable speed than women marathon runners (p < 0.01). Considering half-marathon results only (Figure 2a,b), the significant main effects of segment (F(4, 7253) = 4772.6, 2 = 0.38, large, p < 0.01), sex (F(4, 7253) = 46.9, 2 < 0.01, trivial, p < 0.01), and segment × sex interaction (F(4, 7253) = 71.3, 2 = 0.01, trivial, p < 0.01) were observed. Particularly, in the first three segments (1.17%, 1.10%, and 0.17%, respectively), men half-marathon runners had more stable speed than women (p < 0.01), whereas in the fifth segment (1.13%) women had more stable speed (p < 0.01). Regarding fourth segment (0.16%), no significant differences between men and women half-marathon runners were observed. Finally, in men and women runners in both marathon and half-marathon, each segment significantly differences in speed change than the other (p < 0.01). The only exception is the lack of difference between the third and fifth segment in women marathon runners (p = 0.96). Regarding ACS (Figure 3), the significant main effects of race (F(3, 9107) = 300.8, 2 = 0.03, small, p < 0.01), sex (F(3, 9107) = 55.6, 2 = 0.01, trivial, p < 0.01), and race × sex interaction (F(3, 9107) = 60.5, 2 = 0.01, trivial, p < 0.01) were observed. Men marathon runners had a greater average change of speed of 2.34% than half-marathon runners (p < 0.01), whereas women marathon runners had a 0.89% greater change of speed than women half-marathon runners (p < 0.01). Moreover, the difference between men and women in marathon showed that there was a 1.41% greater average change of speed in marathon men (p < 0.01). Finally, the average speed difference between men and women in half-marathon was only 0.030% (p = 0.67). 4. Discussion The main findings of the present study were that (a) half-marathon runners followed a positive pacing with every segment being slower than its previous one without the presence of an endspurt; (b) compared to marathon, a more even pacing was observed in half-marathon; and (c) women had similar pacing as men in half-marathon. The overall distribution of energy across the race in the half-marathon followed the so-called “positive pacing” [30], that is, the speed decreased continuously across the race. This observation was in agreement with a previous research on IAAF World Half Marathon Championships where slower athletes had decreased speeds from the first segment onwards [4]. It should be highlighted that no endspurt was shown, which was in disagreement with the study of Hanley [4] who examined only elite runners. The discrepancy between the two studies might be attributed to the different performance level, as the abovementioned study focused on world championships, where the faster athletes showed larger endspurt than their slower peers. This suggested that lack of an endspurt might appear in half-marathon races with a large participation of recreational runners. To the best of our knowledge, the more even pacing in half-marathon than in marathon was a novel finding, as it was the first study to compare the two race distances for this characteristic. Medicina 2019, 55, 14 7 of 9 It might be assumed that this difference was due to the additional fatigue induced in the marathon. This assumption is supported by the observation that their difference reached its peak in the fourth segment, i.e., close to the end of the race. On the other hand, both race distances adopted a positive pacing, which was in line with the notion that the perception of effort scaled with the proportion of exercise time that remained [31]. With regards to sex differences, surprisingly, women had similar pacing as men in a half-marathon. In marathon, women had a more even pacing than men, which confirmed the previous findings in other marathon races [20,25,26]. This sex difference has been attributed to differences in physiology and decision making between women and men [32]. Both half-marathon and marathon runners adopted a variable pacing instead of maintaining a steady speed across race. Following a variable self-pacing has been shown to present certain advantages-i.e., enhancement of critical power and high-intensity exercise performance compared to constant work rate cycling exercise [33]. In addition, the rate of perceived exertion has been shown to associate with pacing [22] and might vary across race [34]. Moreover, it was acknowledged that head-to-head competition improved performance compared to running alone [35]. Although this aspect was not examined in the present study, it would be assumed that head-to-head competition exerted a similar influence in both race distances, since half-marathon and marathon races were massive events. Overall, pacing should be considered as a complex system, where individual responses interacted with environment [36]. In this context, athletes were requested to balance behavior and thinking (self-regulation) to optimize their speed across the race [37]. A limitation of the present research was that it considered a sport event (Ljubljana) with relatively small participation compared to other races [38]. Thus, the findings should be considered as preliminary, and should be verified in future studies. On the other hand, strength was the novelty as it added original information in the existing literature with regards to one of the most popular race distances. Considering that half-marathon was the most popular running race event in terms of annual number of races and participants [1,5], the findings of the present study would have practical applications for a wide range of professionals working endurance runners, e.g., coaches, fitness trainers, nutritionists, and physicians. As it has been shown that endurance runners might compete to both half-marathon and marathon [39], it would be of great practical relevance to know how these two race distances differed in pacing. With regards to sex, men and women should be advised to adopt similar pacing patterns in half-marathon, whereas women should aim to run more evenly than men in marathon. Considering the race distance, runners should be guided to regulate their pacing as less or more variable, depending on whether they intended to run a half-marathon or marathon, respectively. Since recent studies reported an association of pacing with physiological and psychological parameters in marathon [23] and 10 km run [22], future research should verify this association in half-marathon, too. 5. Conclusions In summary, both half-marathon and marathon races presented a positive pacing, i.e., speed decreased across the race; however, half-marathon runners did not show an endspurt, which was observed in marathon runners of the present study, as well as of all previous research in marathon. Furthermore, women and men adopted similar pacing pattern in half-marathon, whereas in marathon, women had more even pacing than men that was in agreement with the existing literature about marathon. Consequently, runners should be advised to adopt a less variable pacing when competing in a half-marathon, regardless of their sex. Further research would be needed to shed light on the physiological and psychological correlates of pacing in half-marathon. Author Contributions: Conceptualization, P.T.N., I.C. and B.K.; methodology, I.C.; software, I.C.; validation, P.T.N., I.C. and B.K.; formal analysis, I.C.; investigation, I.C.; resources, I.C.; data curation, I.C.; writing-original draft preparation, P.T.N., I.C. and B.K.; writing-review and editing, P.T.N., I.C. and B.K.; visualization, I.C.; supervision, P.T.N. and B.K.; project administration, P.T.N. Medicina 2019, 55, 14 8 of 9 Funding: This research received no external funding. Acknowledgments: We thank reviewers for their constructive criticism that resulted in improvement of the content of the paper. Conflicts of Interest: The authors declare no conflict of interest. References 1. Knechtle, B.; Nikolaidis, P.T.; Onywera, V.O.; Zingg, M.A.; Rosemann, T.; Rust, C.A. Male and female ethiopian and kenyan runners are the fastest and the youngest in both half and full marathon. SpringerPlus 2016, 5, 223. [CrossRef] [PubMed] 2. Nikolaidis, P.T.; Knechtle, B. Effect of age and performance on pacing of marathon runners. Open Access J. Sports Med. 2017, 8, 171–180. [CrossRef] [PubMed] 3. Breen, D.; Norris, M.; Healy, R.; Anderson, R. Marathon pace control in masters athletes. Int. J. Sports Physiol. Perform. 2018, 13, 332–338. [CrossRef] [PubMed] 4. Hanley, B. Pacing profiles and pack running at the IAAF world half marathon championships. J. Sports Sci. 2015, 33, 1189–1195. [CrossRef] 5. Knechtle, B.; Nikolaidis, P.T.; Zingg, M.A.; Rosemann, T.; Rust, C.A. Half-marathoners are younger and slower than marathoners. SpringerPlus 2016, 5, 76. [CrossRef] [PubMed] 6. Abbiss, C.R.; Thompson, K.G.; Lipski, M.; Meyer, T.; Skorski, S. Difference in pacing between time- and distance-based time trials in trained cyclists. Int. J. Sports Physiol. Perform. 2016, 11, 1018–1023. [CrossRef] [PubMed] 7. Heidenfelder, A.; Rosemann, T.; Rust, C.A.; Knechtle, B. Pacing strategies of ultracyclists in the “race across America”. Int. J. Sports Physiol. Perform. 2016, 11, 319–327. [CrossRef] 8. Lipinska, P.; Allen, S.V.; Hopkins, W.G. Modeling parameters that characterize pacing of elite female 800-m freestyle swimmers. Eur. J. Sport Sci. 2016, 16, 287–292. [CrossRef] 9. Lipinska, P.; Allen, S.V.; Hopkins, W.G. Relationships between pacing parameters and performance of elite male 1500-m swimmers. Int. J. Sports Physiol. Perform. 2016, 11, 159–163. [CrossRef] 10. Wu, S.S.; Peiffer, J.J.; Peeling, P.; Brisswalter, J.; Lau, W.Y.; Nosaka, K.; Abbiss, C.R. Improvement of sprint triathlon performance in trained athletes with positive swim pacing. Int. J. Sports Physiol. Perform. 2016, 11, 1024–1028. [CrossRef] 11. Konings, M.J.; Hettinga, F.J. Pacing decision making in sport and the effects of interpersonal competition: A critical review. Sports Med. 2018, 48, 1829–1843. [CrossRef] [PubMed] 12. Micklewright, D.; Parry, D.; Robinson, T.; Deacon, G.; Renfree, A.; St Clair Gibson, A.; Matthews, W.J. Risk perception influences athletic pacing strategy. Med. Sci. Sports Exerc. 2015, 47, 1026–1037. [CrossRef] 13. Schiphof-Godart, L.; Hettinga, F.J. Passion and pacing in endurance performance. Front. Physiol. 2017, 8, 83. [CrossRef] [PubMed] 14. Santos-Lozano, A.; Collado, P.S.; Foster, C.; Lucia, A.; Garatachea, N. Influence of sex and level on marathon pacing strategy. Insights from the New York city race. Int. J. Sports Med. 2014, 35, 933–938. [CrossRef] [PubMed] 15. Nikolaidis, P.T.; Knechtle, B. Do fast older runners pace differently from fast younger runners in the “New York city marathon”? J. Strength Cond. Res. 2017. [CrossRef] [PubMed] 16. Nikolaidis, P.T.; Knechtle, B. Pacing in age group marathoners in the “New York city marathon”. Res. Sports Med. 2018, 26, 86–99. [CrossRef] [PubMed] 17. Filipas, L.; Nerli Ballati, E.; Bonato, M.; La Torre, A.; Piacentini, M.F. Elite male and female 800-m runners’ display of different pacing strategies during season-best performances. Int. J. Sports Physiol. Perform. 2018, 13, 1344–1348. [CrossRef] 18. Deaner, R.O.; Lowen, A. Males and females pace differently in high school cross-country races. J. Strength Cond. Res. 2016, 30, 2991–2997. [CrossRef] 19. Deaner, R.O.; Addona, V.; Carter, R.E.; Joyner, M.J.; Hunter, S.K. Fast men slow more than fast women in a 10 kilometer road race. PeerJ 2016, 4, e2235. [CrossRef] 20. March, D.S.; Vanderburgh, P.M.; Titlebaum, P.J.; Hoops, M.L. Age, sex, and finish time as determinants of pacing in the marathon. J. Strength Cond. Res. 2011, 25, 386–391. [CrossRef] Medicina 2019, 55, 14 9 of 9 21. Renfree, A.; Crivoi do Carmo, E.; Martin, L. The influence of performance level, age and gender on pacing strategy during a 100-km ultramarathon. Eur. J. Sport Sci. 2016, 16, 409–415. [CrossRef] [PubMed] 22. Bertuzzi, R.; Lima-Silva, A.E.; Pires, F.O.; Damasceno, M.V.; Bueno, S.; Pasqua, L.A.; Bishop, D.J. Pacing strategy determinants during a 10-km running time trial: Contributions of perceived effort, physiological, and muscular parameters. J. Strength Cond. Res. 2014, 28, 1688–1696. [CrossRef] [PubMed] 23. Nikolaidis, P.T.; Knechtle, B. Pacing strategies in the ‘Athens classic marathon’: Physiological and psychological aspects. Front. Physiol. 2018, 9, 1539. [CrossRef] [PubMed] 24. Diaz, J.J.; Fernandez-Ozcorta, E.J.; Santos-Concejero, J. The influence of pacing strategy on marathon world records. Eur. J. Sport Sci. 2018, 18, 781–786. [CrossRef] [PubMed] 25. Hanley, B. Pacing, packing and sex-based differences in Olympic and IAAF world championship marathons. J. Sports Sci. 2016, 34, 1675–1681. [CrossRef] [PubMed] 26. Trubee, N.W.; Vanderburgh, P.M.; Diestelkamp, W.S.; Jackson, K.J. Effects of heat stress and sex on pacing in marathon runners. J. Strength Cond. Res. 2014, 28, 1673–1678. [CrossRef] [PubMed] 27. Available online: http://vw-ljubljanskimaraton.si (accessed on 15 September 2018). 28. Knechtle, B.; Di Gangi, S.; Rust, C.A.; Rosemann, T.; Nikolaidis, P.T. Men’s participation and performance in the ‘Boston marathon’ from 1897 to 2017. Int. J. Sports Med. 2018, 39, 1018–1027. [CrossRef] [PubMed] 29. Cohen, J. Statistical Power Analysis for the Behavioral Sciences, 2nd ed.; Lawrence Erlbaum Associates: Hillsdale, NJ, USA, 1988. 30. Abbiss, C.R.; Laursen, P.B. Describing and understanding pacing strategies during athletic competition. Sports Med. 2008, 38, 239–252. [CrossRef] 31. Faulkner, J.; Parfitt, G.; Eston, R. The rating of perceived exertion during competitive running scales with time. Psychophysiology 2008, 45, 977–985. [CrossRef] 32. Deaner, R.O.; Carter, R.E.; Joyner, M.J.; Hunter, S.K. Men are more likely than women to slow in the marathon. Med. Sci. Sports Exerc. 2014, 47, 607–616. [CrossRef] 33. Black, M.I.; Jones, A.M.; Bailey, S.J.; Vanhatalo, A. Self-pacing increases critical power and improves performance during severe-intensity exercise. Appl. Physiol. Nutr. Metab. 2015, 40, 662–670. [CrossRef] [PubMed] 34. Schallig, W.; Veneman, T.; Noordhof, D.A.; Rodriguez-Marroyo, J.A.; Porcari, J.P.; de Koning, J.J.; Foster, C. The role of the rating-of-perceived-exertion template in pacing. Int. J. Sports Physiol. Perform. 2018, 13, 367–373. [CrossRef] [PubMed] 35. Tomazini, F.; Pasqua, L.A.; Damasceno, M.V.; Silva-Cavalcante, M.D.; de Oliveira, F.R.; Lima-Silva, A.E.; Bertuzzi, R. Head-to-head running race simulation alters pacing strategy, performance, and mood state. Physiol. Behav. 2015, 149, 39–44. [CrossRef] [PubMed] 36. Renfree, A.; Casado, A. Athletic races represent complex systems, and pacing behavior should be viewed as an emergent phenomenon. Front. Physiol. 2018, 9, 1432. [CrossRef] [PubMed] 37. Brick, N.E.; MacIntyre, T.E.; Campbell, M.J. Thinking and action: A cognitive perspective on self-regulation during endurance performance. Front. Physiol. 2016, 7, 159. [CrossRef] 38. Knechtle, B.; Nikolaidis, P.T. Sex- and age-related differences in half-marathon performance and competitiveness in the world’s largest half-marathon-The goteborgsvarvet. Res. Sports Med. 2018, 26, 75–85. [CrossRef] 39. Salinero, J.J.; Soriano, M.L.; Lara, B.; Gallo-Salazar, C.; Areces, F.; Ruiz-Vicente, D.; Abian-Vicen, J.; Gonzalez-Millan, C.; Del Coso, J. Predicting race time in male amateur marathon runners. J. Sports Med. Phys. Fit. 2017, 57, 1169–1177. © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Pacing of Women and Men in Half-Marathon and Marathon Races.

01-14-2019

Nikolaidis, Pantelis T,Ćuk, Ivan,Knechtle, Beat

eng

PMC6720831

International Journal of Environmental Research and Public Health Article Blood Lactate Concentration Is Not Related to the Increase in Cardiorespiratory Fitness Induced by High Intensity Interval Training Todd A. Astorino *, Jamie L. DeRevere, Theodore Anderson, Erin Kellogg, Patrick Holstrom, Sebastian Ring and Nicholas Ghaseb Department of Kinesiology, California State University—San Marcos, San Marcos, CA 92096, USA * Correspondence: [email protected] Received: 9 July 2019; Accepted: 7 August 2019; Published: 9 August 2019



 Abstract: Background: There is individual responsiveness to exercise training as not all individuals experience increases in maximal oxygen uptake (VO2max), which does not benefit health status considering the association between VO2max and mortality. Approximately 50% of the training response is genetic, with the other 50% accounted for by variations in dietary intake, sleep, recovery, and the metabolic stress of training. This study examined if the blood lactate (BLa) response to high intensity interval training (HIIT) as well as habitual dietary intake and sleep duration are associated with the resultant change in VO2max (∆VO2max). Methods: Fourteen individuals (age and VO2max = 27 ± 8 years and 38 ± 4 mL/kg/min, respectively) performed nine sessions of HIIT at 130% ventilatory threshold. BLa was measured during the first and last session of training. In addition, sleep duration and energy intake were assessed. Results: Data showed that VO2max increased with HIIT (p = 0.007). No associations occurred between ∆VO2max and BLa (r = 0.44, p = 0.10), energy intake (r = 0.38, p = 0.18), or sleep duration (r = 0.14, p = 0.62). However, there was a significant association between training heart rate (HR) and ∆VO2max (r = 0.62, p = 0.02). Conclusions: When HIIT is prescribed according to a metabolic threshold, energy intake, sleep status, and BLa do not predict ∆VO2max, yet the HR response to training is associated with the ∆VO2max. Keywords: high intensity exercise; blood lactate concentration; maximal oxygen uptake; individual responsiveness to training 1. Introduction One adaptation to moderate intensity continuous training (MICT) is a significant increase in maximal oxygen uptake (VO2max) [1], which reduces mortality risk [2]. It is apparent that some individuals reveal marked increases in VO2max in response to training, whereas, others experience little to no change [3]. Approximately 50% of the change in VO2max (∆VO2max) with training is genetic [3], with the other 50% due to variations in habitual dietary intake, sleep status, physical activity, and the specifics of the training regime [4]. Ross et al. [5] showed greater increases in VO2max in response to high amount, higher intensity (75% VO2max) versus lower intensity MICT (50% VO2max). This suggests that vigorous continuous exercise may optimize ∆VO2max, which seems important as a 1 metabolic equivalent (MET) increase in VO2max is associated with a 19% reduction in all-cause mortality [6]. One additional factor that may mediate training responsiveness is the absolute metabolic stress of training [4]. Although there are various ways to monitor this, one approach is via the measurement of blood lactate concentration (BLa). Previous studies show that this measure may serve as an index of intramuscular stress [7] and may be related to the initiation of signaling pathways that regulate muscle Int. J. Environ. Res. Public Health 2019, 16, 2845; doi:10.3390/ijerph16162845 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2019, 16, 2845 2 of 8 plasticity in response to training [8]. It is apparent that the accumulation of BLa follows a threshold response in that, at intensities below the work rate coincident with the lactate threshold, there is little change in BLa, whereas, at work rates above the lactate threshold, BLa increases dramatically due to the enhanced recruitment of higher threshold motor units and greater activation of glycolysis [9]. In three groups of men (VO2max = 46–71 mL/kg/min) of similar aerobic fitness, Scharhag-Rosenberger et al. [10] showed dissimilar increases in blood lactate concentration (BLa) during prolonged cycling at 60% and 75% VO2max, which suggests discrepant metabolic strain at identical relative intensities. In addition, these authors reported that some of their participants were unable to complete the prolonged exercise bouts, which was partially explained by the onset of neuromuscular fatigue or enhanced fast-twitch motor unit recruitment, which likely varied across subjects. Recently, the BLa response to moderate intensity continuous exercise at 65 percent of peak power output (% PPO) was significantly associated with ∆VO2max seen with chronic training [11], which suggests that the metabolic response to acute exercise may be predictive of the adaptive response. Whether this relationship also occurs in response to high intensity interval training (HIIT) is unknown. Elucidating predictors of ∆VO2max with chronic exercise such as HIIT is important considering that this modality elicits superior increases in VO2max than MICT [12]. Moreover, VO2max is related to mortality [6], and better understanding predictors of the VO2max response to training may help clinicians better utilize physical activity when employing exercise-based rehabilitation. Health and fitness practitioners implement exercise training in their clientele to promote gains in fitness and body composition which likely lead to improved health status. In addition, it is likely that they try and identify specific variables that impact training responsiveness to optimize exercise programming for their clientele. In this preliminary study, we measured changes in BLa during a brief HIIT regime to determine if the adaptive response is predicted by BLa changes during training. Participants’ sleep status and dietary intake were also measured, as Hawley et al. [13] and Samuels et al. [14] suggest that they are related to training responsiveness. In the case of sleep deprivation, it is associated with fatigue and may make individuals more susceptible to overtraining [13]. In addition, Mann et al. [4] postulated that the effect of habitual dietary intake on resultant variations in the training response is unknown. Overall, it is plausible that adequate energy intake as well as sleep serve to promote recovery from individual sessions, which in turn may benefit the resultant adaptation to training. It is hypothesized that the BLa response to training is significantly related to ∆VO2max. 2. Materials and Methods Participants: Three men and eleven women (age and VO2max = 27 ± 8 years and 38 ± 4 mL/kg/min, respectively) who perform >150 min/week of exercise in the last 12 months, including resistance training, aerobic exercise, noncompetitive sport, and group exercise, participated in this study. They were recruited via convenience sampling. They completed a standard health-history questionnaire, which contained information pertaining to their current physical activity regimen. Participants also provided written informed consent to take part in the study, whose procedures were approved by the University Institutional Review Board. The study was carried out in accordance with the rules of the Declaration of Helsinki. Design: VO2max was assessed before and after nine sessions of HIIT. On days 1 and 9 of training, BLa was measured. Changes in sleep status and dietary intake were monitored during the study. Data concerning ∆VO2max and ventilatory threshold in response to this regimen were previously reported [15]. The present study summarized changes in these outcomes (BLa, sleep status, and dietary intake) that were only obtained from the experimental group who underwent interval training. Sessions occurred at the same time of day within participants, were preceded by a 24 h abstention from physical activity, and were separated by ≥24 h. Participants were instructed to record habitual physical activity during the study in a written log, and were instructed to maintain this behavior during the study. Int. J. Environ. Res. Public Health 2019, 16, 2845 3 of 8 Testing of maximal oxygen uptake: Participants performed incremental exercise on an electrically braked cycle ergometer (Velotron Dynafit Pro, RacerMate, Seattle, WA, USA). After a 2 min warm-up at 40–70 Watt (W), the work rate was increased by 20–35 W/min until volitional exhaustion, represented by cadence <50 rev/min. After 10 min of pedaling at 10% PPO, participants pedaled at 105% PPO to volitional exhaustion [16]. VO2max was identified as the average of these two values. VO2max testing was repeated at least 48 h after the last training session. During all bouts, HR was measured continuously using telemetry (Polar Electro, Beth Page, NY, USA), and gas exchange data were acquired every 15 s using indirect calorimetry (ParvoMedics True One, Sandy, UT, USA). Assessment of blood lactate concentration, sleep status, and dietary intake: During sessions 1 and 9 of HIIT, BLa was determined using a portable monitor (Lactate Plus, Nova Biomedical, Waltham, MA, USA) and lancet (Owen Mumford, Inc., Marietta, GA, USA). After a 5 min rest, the fingertip was washed with a damp paper towel and dried, and BLa was measured by using the second drop of blood, as the first one was wiped away. This procedure was repeated immediately after intervals 4 and 8. The change in BLa during the session of HIIT was calculated as the difference between the pre-exercise and interval 8 values. In addition, the change in BLa in response to training was calculated as the difference in the sum of BLa recorded after intervals 4 and 8 on day 9 versus day 1 of training. On day 1, participants recorded their dietary intake in the 24 h prior and replicated this prior to session 9 to standardize the fed state before these sessions. Dietary intake was measured for three days prior to and for three days during the last week of training. Participants recorded all food and drink ingested over this period in a log. This information was analyzed using software (http://ndb.nal.usda.gov/ndb/foods/list) to determine energy intake (in kilocalories). Sleep status was quantified every day of training, as participants reported the number of hours they slept on the night prior, and an average value was calculated. High intensity interval training: At least 48 h after baseline testing, participants initiated HIIT at a work rate equal to 130% ventilatory threshold (VT), which was equal to 70% PPO in our participants. Ventilatory threshold was estimated by two experienced reviewers independently using the methods of Caiozzo et al. [17]. If their evaluation differed, consensus was reached by consulting a third investigator. Sessions were held three days/week for three weeks and were performed on the same cycle ergometer. Subjects performed eight 1 min intervals on days 1–3 of training, nine on days 4–6, and ten on days 7–9 [18]. Sessions were preceded by a 5 min warm-up at 10% PPO, and intervals were separated by a 75 s active recovery at 10% PPO. Heart rate (HR) was measured continuously using telemetry (Polar Electro USA, Beth Page, NY, USA). The HR response to training was represented by the average HR attained at the end of each interval across all sessions of training. Data analyses: Data were expressed as mean ± standard deviation (SD) and were analyzed using SPSS Version 24.0 (IBM, Armonk, NY, USA). The Shapiro–Wilk test was used to assess normality. A two-way analysis of variance (training = pre versus post, time = three levels) with repeated measures was performed to identify differences in BLa. The Greenhouse–Geisser correction was used to account for the sphericity assumption of unequal variances across groups. If a significant F ratio occurred, Tukey’s post hoc test was used. Pearson’s pairwise correlation was used to determine relationships between variables. Statistical significance was set as p < 0.05. 3. Results Training fidelity: Training elicited 90% ± 5% PPO (79%–96% across participants), which verifies the intensity of this regime. VO2max was increased by 6% with HIIT, and this change was equal to 0.15 ± 0.13 L/min (range = −0.06–0.47 L/min) and 2.4 ± 1.8 mL/kg/min (range = −0.9–4.7 mL/kg/min) across participants. Changes in blood lactate concentration in response to training: BLa increased during HIIT (p < 0.001) yet there was no difference from day 1 to 9 of HIIT (p = 0.91) or trainingXtime interaction (p = 0.87). The change in BLa during session 1 (10.5 ± 2.2 mM) did not differ compared to session 9 (10.3 ± 2.2 mM). The overall change in BLa from pre- to post-training was equal to −0.36 ± 2.4 mM. Int. J. Environ. Res. Public Health 2019, 16, 2845 4 of 8 Four participants showed a greater than 1 mM reduction in BLa from pre- to post-training, whereas, five participants showed increases in BLa above 1 mM. Changes in sleep status and energy intake: Sleep duration was equal to 7.5 ± 0.7 h and ranged from 6.0–8.8 h per night across participants. Dietary intake did not change from pre- (2050 ± 686 kcal) to post-training (2074 ± 639 kcal) (p = 0.39), although it varied from 950–3400 kcal/d across participants. Relationship between ∆VO2max and BLa, sleep status, and dietary intake: No correlation was shown between absolute ∆VO2max and the change in BLa (r = 0.44, p = 0.10) (Figure 1a). In addition, no correlation was shown between ∆VO2max and the mean BLa on day 1 (r = 0.17, p = 0.57) or 9 of training (r = 0.48, p = 0.08) (Figure 1b,c). In addition, results showed no correlation between ∆VO2max and sleep status (r = 0.14, p = 0.62) or energy intake (r = 0.38, p = 0.18) (Figure 2a,b). Int. J. Environ. Res. Public Health 2019, 16, x FOR PEER REVIEW 4 of 8 Changes in sleep status and energy intake: Sleep duration was equal to 7.5 ± 0.7 h and ranged from 6.0–8.8 h per night across participants. Dietary intake did not change from pre- (2050 ± 686 kcal) to post-training (2074 ± 639 kcal) (p = 0.39), although it varied from 950–3400 kcal/d across participants. Relationship between ΔVO2max and BLa, sleep status, and dietary intake: No correlation was shown between absolute ΔVO2max and the change in BLa (r = 0.44, p = 0.10) (Figure 1a). In addition, no correlation was shown between ΔVO2max and the mean BLa on day 1 (r = 0.17, p = 0.57) or 9 of training (r = 0.48, p = 0.08) (Figure 1b and 1c). In addition, results showed no correlation between ΔVO2max and sleep status (r = 0.14, p = 0.62) or energy intake (r = 0.38, p = 0.18) (Figure 2a and 2b). Figure 1. Association between change in VO2max and (a) change in blood lactate concentration, (b) blood lactate concentration on day 1 of training, and (c) blood lactate concentration on day 9 of training. Relationship between ΔVO2max and training heart rate: Heart rate in response to training ranged from 88%–100% percent of maximal heart rate (HRmax), with a mean value equal to 95% ± 3% HRmax. Results revealed a significant correlation between ΔVO2max and training HR (r = 0.62, p = 0.02) (Figure 2c). Figure 1. Association between change in VO2max and (a) change in blood lactate concentration, (b) blood lactate concentration on day 1 of training, and (c) blood lactate concentration on day 9 of training. Int. J. Environ. Res. Public Health 2019, 16, 2845 5 of 8 Int. J. Environ. Res. Public Health 2019, 16, x FOR PEER REVIEW 5 of 8 Figure 2. Association between change in VO2max and (a) sleep duration, (b) dietary intake, and (c) training intensity expressed as % HRmax. 4. Discussion Previous studies document a heterogeneous BLa response to continuous exercise [10] and that the change in BLa in response to acute continuous exercise at 65% PPO predicts the resultant training response [11]. Our study examined if the BLa response to short-term HIIT is associated with ΔVO2max. The results oppose our hypothesis and they suggest that HR response to training is significantly associated with change in VO2max, yet BLa, sleep duration, and energy intake are not. These preliminary data suggest that the cardiovascular strain of HIIT prescribed according to a metabolic threshold may predict resultant responses to training. Our data showing no correlation between ΔVO2max and the BLa response to HIIT refute previous results [11]. However, there are methodological differences between studies which may explain these discrepancies. First, our sample contained men and women, while the former study included only men. It is apparent that muscle fiber type may differ between men and women [19], leading to greater reliance on glycolysis and resultant blood lactate accumulation in men who have a greater proportion of higher threshold motor units. Moreover, men exhibit higher BLa in response to HIIT and sprint interval training (SIT) versus women, which may be due to the greater work completed during training due to a propensity to maintain a higher cadence, especially during SIT [20]. Second, HIIT was performed at an intensity equal to 90% PPO, whereas, in the former study, training was performed at 65% PPO. This different composition of training led to markedly different BLa responses, as our subjects exhibited BLa ranging from 7.6–14.1 mM, which was higher than values exhibited in their study (6–8 mM). Lastly, our regimen was prescribed according to VT, which Figure 2. Association between change in VO2max and (a) sleep duration, (b) dietary intake, and (c) training intensity expressed as % HRmax. Relationship between ∆VO2max and training heart rate: Heart rate in response to training ranged from 88%–100% percent of maximal heart rate (HRmax), with a mean value equal to 95% ± 3% HRmax. Results revealed a significant correlation between ∆VO2max and training HR (r = 0.62, p = 0.02) (Figure 2c). 4. Discussion Previous studies document a heterogeneous BLa response to continuous exercise [10] and that the change in BLa in response to acute continuous exercise at 65% PPO predicts the resultant training response [11]. Our study examined if the BLa response to short-term HIIT is associated with ∆VO2max. The results oppose our hypothesis and they suggest that HR response to training is significantly associated with change in VO2max, yet BLa, sleep duration, and energy intake are not. These preliminary data suggest that the cardiovascular strain of HIIT prescribed according to a metabolic threshold may predict resultant responses to training. Our data showing no correlation between ∆VO2max and the BLa response to HIIT refute previous results [11]. However, there are methodological differences between studies which may explain these discrepancies. First, our sample contained men and women, while the former study included only men. It is apparent that muscle fiber type may differ between men and women [19], leading to greater reliance on glycolysis and resultant blood lactate accumulation in men who have a greater proportion of higher threshold motor units. Moreover, men exhibit higher BLa in response to HIIT and sprint interval training (SIT) versus women, which may be due to the greater work completed during training Int. J. Environ. Res. Public Health 2019, 16, 2845 6 of 8 due to a propensity to maintain a higher cadence, especially during SIT [20]. Second, HIIT was performed at an intensity equal to 90% PPO, whereas, in the former study, training was performed at 65% PPO. This different composition of training led to markedly different BLa responses, as our subjects exhibited BLa ranging from 7.6–14.1 mM, which was higher than values exhibited in their study (6–8 mM). Lastly, our regimen was prescribed according to VT, which may potentially reduce variability in metabolic stress versus their regimen, which may have had select participants training above or below the workload associated with VT. Further work is merited to examine if the BLa response to HIIT prescribed according to an absolute intensity is associated with ∆VO2max. Our data showed no association between ∆VO2max and sleep duration or calorie intake. Sleep deprivation may mitigate the adaptive response to training by eliciting fatigue, which may reduce performance [21]. Previous results exhibit that one (25–30 h of sleeplessness) [22] and three nights of sleep deprivation (3 h of sleep per night) [23] decreased time to exhaustion and muscular strength in athletes, whereas, in basketball players, two additional hours of sleep per night for up to seven weeks were consequent with increased speed and shooting performance [24]. Our data show that the participant with the lowest amount of sleep (6.0 h) did not exhibit increases in VO2max with HIIT. However, two additional participants who received 8 h per night of sleep also showed minimal increases in VO2max, which suggests that sleep duration by itself may not predict training responsiveness. Two of these participants also completed 7 h/week of physical activity outside the HIIT regimen, so an effect of overreaching on their lack of response may exist. Our participants’ habitual dietary intake ranged from 1.8–3.8 kcal/kg body mass. Two men and two women exhibiting substantial increases in VO2max (0.13–0.47 L/min) revealed dietary intakes approaching 3.8 kcal/kg body mass, whereas, two participants ingesting less than 2 kcal/kg revealed no change in VO2max in response to training. These individual results reveal that nutritional state may impact the training response. Nevertheless, in adults with diabetes, the addition of post-exercise protein did not affect adaptation to training [25], which may indicate that diet has little impact on training responsiveness. A significant association between HR response to training and ∆VO2max was revealed (Figure 2). Our results reveal that participants with the greatest increases in VO2max exhibited training HR above 95% HRmax. Whether this greater adaptation is due to some unique sympathetic response or alternatively, maintaining a higher cadence during HIIT, which would elicit greater work, is unknown. In sedentary men, six sessions of HIIT increased VO2max and PPO [26], with these outcomes associated with the ratio of low to high frequency power of R-R oscillation, which represents sympathovagal balance. Higher vagal activity has been shown to be directly related to ∆VO2max in response to MICT [27], and further study is merited to confirm this result in response to other HIIT regimens. One limitation of our study is that dietary intake was quantified through self-reported logs, which are prone to underreporting [28]. Studies show that the pre-exercise nutritional state may alter the molecular response to training [13], so monitoring participants’ dietary patterns before and after each session may be useful to better understand the effects of nutrition on training responsiveness. Sleep status was assessed by identifying the duration of sleep the night before each session, rather than using various questionnaires, which may be more valid to assess sleep quality [29]. Thus, the reliability of our relatively simple measure is unknown. In addition, our sample included both men and women who may show unique responses to interval training [30], yet our study was underpowered to examine this potential effect of sex. No measure of critical power was performed in our study, so it is likely that participants were training at different workloads within the heavy and/or severe intensity domain. Lastly, our training protocol was relatively brief, so data cannot be used to explain responsiveness to prolonged regimens of interval training. 5. Conclusions Overall, training HR was associated with the VO2max response to HIIT, yet dietary intake, sleep duration, and BLa accumulation were not predictive of this response. Overall, these preliminary Int. J. Environ. Res. Public Health 2019, 16, 2845 7 of 8 data suggest that the absolute cardiovascular strain may be a mediator of the adaptive response to interval training. Author Contributions: Conceptualization, T.A.A., J.L.D., and T.A.; methodology, T.A.A., J.L.D., and T.A.; formal analysis, T.A.A.; resources, T.A.A.; data curation, T.A.A., J.L.D., T.A., E.K., P.H., S.R., and N.G.; writing—original draft preparation, T.A.A.; writing—review and editing, T.A.A., J.L.D., T.A., E.K., P.H., S.R., and N.G.; supervision, T.A.A.; project administration, T.A.A. Funding: This research received no external funding. Acknowledgments: The authors thank the participants for their effort and dedication in completing the study requirements. Conflicts of Interest: The authors declare no conflict of interest. References 1. Duscha, B.D.; Slentz, C.A.; Johnson, J.L.; Houmard, J.A.; Bensimhon, D.R.; Knetzger, K.J.; Kraus, W.E. Effects of exercise training amount and intensity on peak oxygen consumption in middle-age men and women at risk for cardiovascular disease. Chest 2005, 128, 2788–2793. [CrossRef] [PubMed] 2. Lee, D.C.; Sui, X.; Artero, E.G.; Lee, I.M.; Church, T.S.; McAuley, P.A.; Stanford, F.C.; Kohl, H.W., 3rd; Blair, S.N. Long-term effects of changes in cardiorespiratory fitness and body mass index on all-cause and cardiovascular disease mortality in men: The Aerobics Center Longitudinal Study. Circulation 2011, 124, 2483–2490. [CrossRef] [PubMed] 3. Bouchard, C.; An, P.; Rice, T.; Skinner, J.S.; Wilmore, J.H.; Gagnon, J.; Pérusse, L.; Leon, A.S.; Rao, D.C. Familiar aggregation of VO2max response to exercise training: Results from the HERITAGE Family Study. J. Appl. Physiol. 1999, 87, 1003–1008. [CrossRef] [PubMed] 4. Mann, T.N.; Lamberts, R.P.; Lambert, M.I. High responders and low responders: Factors associated with individual variation in response to standardized training. Sports Med. 2014, 44, 1113–1124. [CrossRef] [PubMed] 5. Ross, R.; de Lannoy, L.; Stotz, P.J. Separate effects of intensity and amount of exercise on interindividual cardiorespiratory fitness response. Mayo Clin. Proc. 2015, 90, 1506–1514. [CrossRef] [PubMed] 6. Kodama, S.; Saito, K.; Tanaka, S.; Maki, M.; Yachi, M.; Asumi, A.; Sugawara, A.; Totsuka, K.; Shimano, H.; Ohashi, Y.; et al. Cardiorespiratory fitness as a quantitative predictor of all-cause mortality and cardiovascular events in healthy men and women: A meta-analysis. J. Am. Med. Assoc. 2009, 301, 2024–2035. [CrossRef] [PubMed] 7. Spriet, L.L.; Howlett, R.A.; Heigenhauser, G.J. An enzymatic approach to lactate production in human skeletal muscle during exercise. Med. Sci. Sports Exerc. 2000, 32, 756–763. [CrossRef] [PubMed] 8. Hood, D.A. Invited review: Contractile activity-induced mitochondrial biogenesis in skeletal muscle. J. Appl. Physiol. 2001, 90, 1137–1157. [CrossRef] 9. Brooks, G.A.; Mercier, J. Balance of carbohydrate and lipid utilization during exercise: The “crossover” concept. J. Appl. Physiol. 1994, 76, 2253–2261. [CrossRef] 10. Scharhag-Rosenberger, F.; Meyer, T.; Gabler, N.; Faude, O.; Kindermann, W. Exercise at given percentages of VO2max: Heterogeneous metabolic responses between individuals. J. Sci. Med. Sport. 2010, 13, 74–79. [CrossRef] 11. Preobrazenski, N.; Bonafiglia, J.T.; Nelms, M.W.; Lu, S.; Robins, L.; LeBlanc, C.; Gurd, B.J. Does blood lactate predict the chronic adaptive response to training: A comparison of traditional and talk test prescription methods. Appl. Physiol. Nutr. Metab. 2019, 44, 179–186. [CrossRef] [PubMed] 12. Milanovi´c, Z.; Sporiš, G.; Weston, M. Effectiveness of High-intensity interval training (HIIT) and continuous endurance training for VO2max improvements: A systematic review and meta-analysis of controlled trials. Sports Med. 2015, 45, 1469–1481. [CrossRef] [PubMed] 13. Hawley, J.A.; Tipton, K.D.; Millard-Stafford, M.L. Promoting training adaptations through nutritional interventions. J. Sports Sci. 2006, 24, 709–721. [CrossRef] [PubMed] 14. Samuels, C. Sleep, recovery, and performance: The new frontier in high-performance athletics. Phys. Med. Rehabil. Clin. N. Am. 2009, 20, 149–159. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2019, 16, 2845 8 of 8 15. Astorino, T.A.; DeRevere, J.; Anderson, T.; Kellogg, E.; Holstrom, P.; Ring, S.; Ghaseb, N. Change in VO2max and time trial performance in response to high-intensity interval training prescribed using ventilatory threshold. Eur. J. Appl. Physiol. 2018, 118, 1811–1820. [CrossRef] [PubMed] 16. Astorino, T.A.; White, A.C.; Dalleck, L.C. Supramaximal testing to confirm attainment of VO2max in sedentary men and women. Int. J. Sports Med. 2009, 30, 279–284. [CrossRef] [PubMed] 17. Caiozzo, V.J.; Davis, J.A.; Ellis, J.F.; Azus, J.L.; Vandagriff, R.; Prietto, C.A.; McMaster, W.C. A comparison of gas exchange indices used to detect the anaerobic threshold. J. Appl. Physiol. 1982, 53, 1184–1189. [CrossRef] 18. Clark, A.; De La Rosa, A.B.; DeRevere, J.L.; Astorino, T.A. Effects of various interval training regimes on changes in maximal oxygen uptake, body composition, and muscular strength in sedentary women with obesity. Eur. J. Appl. Physiol. 2019, 119, 879–888. [CrossRef] 19. Billaut, F.; Bishop, D. Muscle fatigue in males and females during multiple-sprint exercise. Sports Med. 2009, 39, 257–278. [CrossRef] 20. Astorino, T.A.; Sheard, A.C. Does sex mediate the affective response to high intensity interval exercise? Physiol. Behav. 2019, in press. [CrossRef] 21. Chennaoui, M.; Arnal, P.J.; Sauvet, F.; Léger, D. Sleep and exercise: A reciprocal issue? Sleep Med. Rev. 2015, 20, 59–72. [CrossRef] [PubMed] 22. Azboy, O.; Kaygisiz, Z. Effects of sleep deprivation on cardiorespiratory functions of the runners and optimizing sleep in elite athletes volleyball players during rest and exercise. Acta Physiol. Hung. 2009, 96, 29–36. [CrossRef] [PubMed] 23. Reilly, T.; Piercy, M. The effect of partial sleep deprivation on weightlifting performance. Ergonomics 1994, 37, 107–115. [CrossRef] [PubMed] 24. Mah, C.D.; Mah, K.E.; Kezirian, E.J.; Dement, W.C. The effects of sleep extension on the athletic performance of collegiate basketball players. Sleep 2011, 34, 943–950. [CrossRef] [PubMed] 25. Francois, M.E.; Durrer, C.; Pistawka, K.J.; Halperin, F.A.; Chang, C.; Little, J.P. Combined interval training and post-exercise nutrition in type 2 diabetes: A randomized control trial. Front. Physiol. 2017, 8, 528. [CrossRef] [PubMed] 26. Kiviniemi, A.M.; Tulppo, M.P.; Eskelinen, J.J.; Savolainen, A.M.; Kapanen, J.; Heinonen, I.H.; Hautala, A.J.; Hannukainen, J.C.; Kalliokoski, K.K. Autonomic function predicts fitness response to short-term high-intensity interval training. Int. J. Sports Med. 2015, 36, 915–921. [CrossRef] 27. Hautala, A.J.; Mäkikallio, T.H.; Kiviniemi, A.; Laukkanen, R.T.; Nissilä, S.; Huikuri, H.V.; Tulppo, M.P. Cardiovascular autonomic function correlates with the response to aerobic training in healthy sedentary subjects. Am. J. Physiol. 2003, 285, 1717–1752. [CrossRef] 28. Martin, L.J.; Su, W.F.; Jones, P.H.; Lockwood, G.A.; Tritchler, D.A.; Boyd, N.F. Comparison of energy intakes determined by food records and doubly labeled water in women participating in a dietary-intervention trial. Am. J. Clin. Nutr. 1996, 63, 483–490. [CrossRef] 29. Claudino, J.G.; Gabbet, T.J.; de Sá Souza, H.; Simim, M.; Fowler, P.; de Alcantara Borba, D.; Melo, M.; Bottino, A.; Loturco, I.; D’Almeida, V.; et al. Which parameters to use for sleep quality monitoring in team sport athletes? A systematic review and meta-analysis. BMJ Open Sport Exerc. Med. 2019, 5, e000475. [CrossRef] 30. Gillen, J.B.; Percival, M.E.; Skelly, L.E.; Martin, B.J.; Tan, R.B.; Tarnopolsky, M.A.; Gibala, M.J. Three minutes of all-out intermittent exercise per week increases skeletal muscle oxidative capacity and improves cardiometabolic health. PLoS ONE 2014, 9, e111489. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Blood Lactate Concentration Is Not Related to the Increase in Cardiorespiratory Fitness Induced by High Intensity Interval Training.

08-09-2019

Astorino, Todd A,DeRevere, Jamie L,Anderson, Theodore,Kellogg, Erin,Holstrom, Patrick,Ring, Sebastian,Ghaseb, Nicholas

eng

PMC5783408

RESEARCH ARTICLE Similarities and differences among half- marathon runners according to their performance level Ana Ogueta-Alday1*, Juan Carlos Morante1, Josue´ Go´mez-Molina2, Juan Garcı´a-Lo´pez1,3 1 Department of Physical Education and Sports, Institute of Biomedicine (IBIOMED), Faculty of Physical Activity and Sports Sciences (FCAFD), University of Leo´n, Leo´n, Spain, 2 Department of Physical Education and Sports, Faculty of Education and Sport, University of the Basque Country, UPV/EHU, Spain, 3 High Sport Performance Centre of Leo´n (CAR-Leo´n), Spanish Council of Sports (CSD), Leo´n, Spain * [email protected] Abstract This study aimed to identify the similarities and differences among half-marathon runners in relation to their performance level. Forty-eight male runners were classified into 4 groups according to their performance level in a half-marathon (min): Group 1 (n = 11, < 70 min), Group 2 (n = 13, < 80 min), Group 3 (n = 13, < 90 min), Group 4 (n = 11, < 105 min). In two separate sessions, training-related, anthropometric, physiological, foot strike pattern and spatio-temporal variables were recorded. Significant differences (p<0.05) between groups (ES = 0.55–3.16) and correlations with performance were obtained (r = 0.34–0.92) in train- ing-related (experience and running distance per week), anthropometric (mass, body mass index and sum of 6 skinfolds), physiological (VO2max, RCT and running economy), foot strike pattern and spatio-temporal variables (contact time, step rate and length). At standardized submaximal speeds (11, 13 and 15 km
h-1), no significant differences between groups were observed in step rate and length, neither in contact time when foot strike pattern was taken into account. In conclusion, apart from training-related, anthropometric and physiological variables, foot strike pattern and step length were the only biomechanical variables sensitive to half-marathon performance, which are essential to achieve high running speeds. How- ever, when foot strike pattern and running speeds were controlled (submaximal test), the spatio-temporal variables were similar. This indicates that foot strike pattern and running speed are responsible for spatio-temporal differences among runners of different perfor- mance level. Introduction The participation in long-distance running events has grown significantly in the last decade. In races between 5 km and the marathon, the total number of finishers in the USA in 2015 was about 17,114,800 runners [1]. The half-marathon was the favorite distance for male runners between 25 and 44 years of age, and finishers’ average time was around 123 min [1]. This PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 1 / 11 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Ogueta-Alday A, Morante JC, Go´mez- Molina J, Garcı´a-Lo´pez J (2018) Similarities and differences among half-marathon runners according to their performance level. PLoS ONE 13 (1): e0191688. https://doi.org/10.1371/journal. pone.0191688 Editor: Alena Grabowski, University of Colorado Boulder, UNITED STATES Received: September 11, 2017 Accepted: January 9, 2018 Published: January 24, 2018 Copyright: © 2018 Ogueta-Alday et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: This work was supported by the Spanish Sports Council (CSD) under the project 157/ UPB10/12, by a grant of the High Sport Performance Centre of Leo´n (CAR-Leo´n); and by the Basque Country Government, under a predoctoral grant number reference PRE_2013_1_1109 (J.G.). The funders had no role in study design, data collection and analysis, indicates that not only elite runners take part in these events, but so do amateur runners. It is important to understand the demands and characteristics of all types of runners (i.e. recrea- tional, moderately-trained, highly-trained), and the scientific community is interested in addressing the discipline of running from different performance-related perspectives (e.g. anthropometry, training, physiology and biomechanics). The relationship between physiological variables and running performance has been deeply investigated. A high VO2max, respiratory compensation threshold and a good running econ- omy are highly related to performance in long-distance races [2]. Some anthropometric vari- ables are also important for good running performance, as they can affect the aforementioned physiological variables [3–5]. A lower body mass [4,5], body mass index [3,5] and sum of skin- folds [5] implies a lower muscular effort to support and accelerate the body and the legs, requiring less energy expenditure [4], lower heat production and higher heat dissipation [6], and therefore allowing a better long-distance running performance. However, the influence of some biomechanical variables on long-distance running perfor- mance is quite unclear. Some studies identified the foot strike pattern (i.e. midfoot/forefoot vs rearfoot) as a key factor of performance, and found a higher percentage of midfoot/forefoot runners in the top place finishers of high-level half-marathon and marathon races [7,8]. In contrast, in low-level races this tendency was not observed [9]. On the other hand, some stud- ies have associated a shorter contact time with better performance or running economy [7,10,11], while others have not [10,12]. These discrepancies could be due to the dependence of contact time on both running speed and foot strike pattern [13]. In regards to step rate and length, some studies observed a higher step rate in highly-trained runners compared to well- trained and non-trained ones [14,15]. This seems to be a natural adaptation to obtain an ener- getically more optimal step rate [10]. However, at similar running speeds, step rate and length have not been associated with performance [12]. Therefore, the main purpose of the present study was to analyze the similarities and differ- ences between training-related, anthropometric, physiological, foot strike pattern and spatio- temporal variables in half-marathon trained runners, according to their performance level. The hypothesis was that there would be differences among runners of different level in train- ing-related, anthropometric and physiological variables, as well as in foot strike pattern, but not in spatio-temporal variables if running speed and foot strike patterns are controlled. Materials and methods Experimental design The study was approved by the University of Leo´n Ethics Committee. Forty-eight half-mara- thon runners with different performance level (from 63 to 101 min) were analyzed. Runners reported to the laboratory on two different days, with an interval of at least one week. On the first day, training-related and anthropometric characteristics were recorded, and an incremen- tal treadmill test was performed. On the second day, a submaximal test at different running speeds was performed. The submaximal running speeds were set at 11, 13 and 15 km
h-1 to assure that low- and high-level runners were between 60–90% of VO2max in one of these speeds, and therefore obtain their running economy [16]. During both tests, foot strike pat- tern, physiological (VO2, RER and HR) and spatio-temporal variables (i.e. contact and flight times, step rate and length) were simultaneously registered. Subjects Runners were recruited from national and local track and field clubs, as well as from recrea- tional running training groups. Finally, forty-eight long-distance male runners participated Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 2 / 11 decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. according to the following inclusion criteria: 1- runners had to be Caucasian from 20 to 50 years-old, 2- they must have participated in at least one self-selected half-marathon during the six-week period prior to the study, 3- their performance level must be better than 105 min, determined by the “chip time” (time from the start to the finish line after 21,097 m). Runners were divided into four groups according to their performance level: Group 1 (n = 11, < 70 min), Group 2 (n = 13, between 70 and < 80 min), Group 3 (n = 13, between 80 and < 90 min) and Group 4 (n = 11, between 90 and < 105 min). Additionally, following the criteria of Hasegawa et al. [8], runners were divided into two groups according to their foot strike pat- tern: rearfoot or midfoot/forefoot strikers, in order to 1- analyze the influence of foot strike pattern in long-distance running performance and 2- avoid the influence of foot strike pattern on spatio-temporal parameters. Written consent was obtained from the subjects and the study was approved by the University Ethics Committee. Procedures All testing sessions were conducted at the same time of day (between 10 a.m. and 1 p.m.) under similar environmental conditions (~ 800 m altitude, 20–25 ˚C, 20–35% relative humid- ity). During these days, a correct intake of carbohydrate (~ 400 gr) was recommended [17]. Participants fasted for 2 h before the submaximal test and during the tests, but they were able to drink water ad libitum to avoid dehydration. Both running tests were preceded by a stan- dardized warm-up (treadmill running at 10–12 km
h-1 for 10 min followed by 5 min of free stretching). All runners wore the same running shoes in every testing session (250–300 gr weight for each shoe) to prevent this variable from affecting running economy [18]. Running tests were performed on a treadmill (HP Cosmos Pulsar, HP Cosmos Sports & Medical GMBH, Nussdorf-Traunstein, Germany) with a 1% slope in an attempt to mimic the effects of air resistance on the metabolic cost of flat outdoor running [19]. Two fans with a wind speed between 4–8 km
h-1 (according to the preference of each runner) were placed around the treadmill (~ 50–100 cm) to cool the subjects during running [17]. Respiratory gases (Medisoft Ergocard, Medisoft Group, Sorinnes, Belgium) and heart rate (HR) (Polar Team, Polar Electro Oy, Kempele, Finland) were monitored throughout the tests. Running spatio-temporal parameters (i.e. contact and flight times, step rate and length) were recorded with a contact laser platform installed in the treadmill (SportJUMP System PRO1, DSD Inc., Leo´n, Spain) and connected to a specific software (Sport-Bio Running, DSD Inc., Leo´n, Spain). The spatio-temporal variables computed from this system were previously validated [20]. A minimum recording time of 20 s was set at each running speed to obtain at least 32 consecutive steps and thus reduce the effect of intra-individual step variability [13]. Runners’ foot strike pattern was determined using a high-speed video camera (240 Hz) (Casio Exilim Pro EX-F1, CASIO Europe GMBH, Norderstedt, Germany) placed on the right side of the treadmill (~ 1 m), perpendicular to the sagittal plane at a height of 40 cm from the ground. All runners were analyzed by the same observer, who identified their foot strike pattern (i.e. rear- foot or midfoot/forefoot) at their competitive running speed during the incremental treadmill test. This running speed was calculated from the time needed to complete the half-marathon (e.g. 18 km
h-1 for a runner with a performance of 70 min). Anthropometry Subject’s body mass, height and 6 skinfold measurements (triceps, subscapular, supra-iliac, abdominal, anterior thigh and medial calf) were recorded using standard equipment (HSB-BI, British Indicators LTD, West Sussex, UK). The total leg and lower leg (shank) lengths were also obtained (Harpender anthropometer, CMS instruments, London, UK), taking into Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 3 / 11 account the distance from the floor to the femur (greater trochanter) and to the tibia (superior point on the lateral border of the head of the tibia), respectively. Maximal thigh and shank cir- cumferences as well as minimum ankle circumference were measured (Holtain LTD, Cry- mych, UK). All measurements were made by the same researcher following the international guidelines for anthropometry [21] and the criteria of previous studies [17]. Incremental test The test started at 6 km
h-1 and treadmill speed was increased 1 km
h-1 every 1-min until voli- tional exhaustion. VO2max and HRmax were recorded as the highest values obtained in the 30 s before exhaustion [13]. The ventilatory threshold (VT) and the respiratory compensation threshold (RCT) were identified according to the criteria of Davis [22]. Spatio-temporal parameters were recorded in the last 20 s of each running speed, from 10 km
h-1 (i.e. when runners started to have flight time) until peak speed [13]. Submaximal test Subjects performed 6-min running at 11, 13 and 15 km
h-1 with a 5-min rest in between. VO2 and HR were continuously recorded during the test, considering the average of the last 3-min period of each set as representative data [17]. Running economy was determined as the VO2 cost at a given running speed, expressed in ml
kg-1
km-1 and ml
kg-0.75
km-1. This last unit was chosen to avoid the possible influence of body mass in running economy [23]. The best value between 60–90% of VO2max was chosen as running economy representative value [16]. Spatio- temporal parameters were recorded for a minimum of 20 s during the 5th minute of each set. Statistical analysis The results are expressed as mean ± SD. The Kolmogorov-Smirnov test was applied to ensure a Gaussian distribution of all results. A one-way Analysis of Variance (ANOVA) was used to analyze the differences between the four groups of runners. Additionally, the Analysis of Covariance (ANCOVA) was used to analyze the differences between the four groups of run- ners in biomechanical variables, taking into account as covariates runners’ foot strike pattern (i.e. midfoot/forefoot and rearfoot) and running speeds where physiological variables were obtained (i.e. peak, RCT and VT speeds). When a significant F value was found, the Newman- Keuls post hoc analysis was used to establish statistical differences between means. Effect sizes (ES) (Cohen’s d) were also calculated [20]. The magnitude of the difference was considered to be trivial (ES < 0.2), small (0.2  ES < 0.5), moderate (0.5  ES < 0.8) and large (ES  0.8). Pearson correlation coefficient (r) was used to obtain relationships between variables. SPSS + version 17.0 statistical software (SPSS, Inc., Chicago, IL, USA) was used. Values of p<0.05 were considered statistically significant. Results Anthropometry, training-related and physiological parameters The four groups of runners (n = 48) were not different in age (32.0 ± 7.0 years), height (176.0 ± 5.0 m), total leg length (90.0 ± 4.0 cm), lower leg (shank) length (44.0 ± 2.0 cm), and maximal thigh, shank and ankle circumferences (51.1 ± 3.1, 36.0 ± 1.0 and 22.0 ± 1.0 cm, respectively). Table 1 shows that running experience (ES = 1.62), weekly training volume (ES = 1.65), body mass (ES = 0.55), body mass index (ES = 1.42), sum of skinfolds (ES = 2.08), peak speed (ES = 3.27), VO2max expressed in ml
kg-1
min-1 (ES = 1.31) and ml
kg-0.75
min-1 (ES = 1.24), speed in both VT (ES = 1.80) and RCT (ES = 3.16), and running economy Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 4 / 11 expressed in ml
kg-1
km-1 (ES = 1.06) and ml
kg-0.75
km-1 (ES = 1.12) had a significant effect on performance level (p<0.01), and were related to running performance (p<0.05) (Table 1). Foot strike pattern Fig 1 shows that performance level had a moderate effect on foot strike pattern distribution among groups (ES = 0.72, p<0.01). The percentage of midfoot/forefoot strikers was higher in Group 1 with respect to Groups 2, 3 and 4 (73, 31, 15 and 9%, respectively). Spatio-temporal parameters during the incremental test (comparison at the same relative physiological intensities) Table 2 shows that, during the incremental test at different running speeds (i.e. peak, RCT and VT speeds), there were significant differences between groups of runners in contact time and step length (p<0.01), but not in step rate. Besides, significant correlations (p<0.05) between half-mar- athon performance (i.e. time spent) and contact time (r  0.50), step rate (r  -0.38) and length (r  -0.62) were observed. These differences and correlations disappeared taking into account the runners’ foot strike pattern and the running speed where these variables were obtained. Spatio-temporal parameters during the submaximal test (comparison at standardized running speeds) Table 3 shows that, at standardized submaximal speeds (11, 13 and 15 km
h-1), no significant differences between groups were observed in step rate and length. On the contrary, contact Table 1. Mean (± SD) training-related, anthropometric and physiological variables of the different groups of runners. Correlation (r) with running performance (time to complete a half-marathon). G1 (n = 11) G2 (n = 13) G3 (n = 13) G4 (n = 11) r Running performance (min) 66.0±2.3†# 73.0±3.4†# 85.2±2.5# 96.0±3.2 --- Running experience (years) 16.5±5.6†# 11.0±3.7†# 4.5±3.3 3.6±4.2 -0.75 Training volume (km
week-1) 118.6±30.3†# 85.8±23.3†# 51.7±21.3 43.3±15.4 -0.80 Mass (kg) 66.5±5.3†# 68.1±5.0† 73.0±5.6 73.0±8.9 0.45 Body mass index (kg
m-2) 21.4±1.4†# 21.1±0.9†# 23.3±1.3 24.1±2.4 0.64 ∑ of 6 skinfolds (mm) 37.4±9.1†# 40.4±6.3†# 58.6±13.8# 70.3±15.9 0.78 Peak speed (km
h-1) 22.1±0.8†# 20.6±1.0†# 18.8±0.4# 17.4±0.9 -0.92 VO2max (ml
kg-1
min-1) 69.2±5.0†# 64.4±5.7†# 56.9±4.5 55.9±6.2 -0.76 VO2max (ml
kg-0.75
min-1) 197.4±13.8†# 184.9±14.1†# 166.1±13.2 163.1±16.0 -0.67 RCT speed (km
h-1) 18.6±1.2†# 17.4±1.2†# 15.5±0.8# 13.8±1.1 -0.92 RCT—% VO2max 87.8±4.8 90.2±3.7 87.6±5.0 84.4±5.3 -0.33 VT speed (km
h-1) 12.7±1.2†# 11.8±1.3†# 10.2±0.5 9.8±1.3 -0.76 VT—% VO2max 58.9±4.5 61.1±7.1 59.7±6.4 62.7±7.4 0.11 RE (ml
kg-1
km-1) 196.1±18.8# 205.5±12.1 205.2±12.9 219.5±18.4 0.39 RE (ml
kg-0.75
km-1) 559.7±55.1# 590.0±35.6 600.0±41.8 640.4±52.8 0.50 RER (VCO2
VO2 -1) 0.79±05# 0.83±0.06 0.84±0.06 0.89±0.05 0.51 Note: G1, G2, G3, G4, groups of runners of different performance level (< 70, < 80, < 90 and < 105 min, respectively). ∑ of 6 skinfolds, sum of six skinfolds. VO2max, maximun oxygen uptake. RCT, respiratory compensation threshold. VT, ventilatory threshold. RE, running economy. RER, Respiratory Exchange Ratio. , significant differences with Group 2. †, significant differences with Group 3. #, significant differences with Group 4. r, significant correlations (p<0.05) in bold type. https://doi.org/10.1371/journal.pone.0191688.t001 Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 5 / 11 time was significantly shorter (p<0.01) in higher level runners (ES = 0.72, 0.74 and 0.88, respectively). These differences disappeared when the runners’ foot strike pattern was taken into account. Discussion The main outcome of this study was that there were no differences in spatio-temporal parame- ters (i.e. contact time, step rate and length) among half-marathon runners of different perfor- mance level (from 63 to 101 min) when the same foot strike pattern is used and they are running at equal submaximal speed. However, high-level runners’ group exhibited the highest percentage of midfoot/forefoot strikers (~ 73%) compared to the other three groups (~ 9–31%) (Fig 1), and therefore they showed lower contact times than rearfoot strikers (i.e. low- level runners). Anthropometry, training-related and physiological parameters Strong relationships between performance and training-related variables were found (Table 1). This is in line with previous studies that considered the excellence in long-distance running as the combination of genetic, environmental (i.e. socio-demographic) and training- related factors (i.e. deliberate practice theory) [24]. In the present study, in line with previous ones [3,4,5], higher level runners were lighter, had lower body mass index and lower fat/sum of skinfolds. In contrast, linear anthropometric variables (i.e. height, lengths or circumfer- ences) had no influence on running performance, which is in agreement with some previous studies [3,4,5]. However, other studies found the contrary, which could be due to the different ethnicities compared (e.g. Caucasian vs African) and not to the performance level itself [17,25]. Additionally, as expected, VO2max, peak speed, and speed in both VT and RCT were strongly related to half-marathon performance (Table 1), in the same line with previous Fig 1. Foot strike pattern distribution (midfoot/forefoot and rearfoot) in each group of runners. G1, G2, G3, G4, groups of runners of different performance level (< 70, < 80, < 90 and < 105 min, respectively). , significant differences with Group 1. https://doi.org/10.1371/journal.pone.0191688.g001 Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 6 / 11 findings [2,11,16,26,27]. It is noteworthy the weak relationship between performance and run- ning economy (r  0.50), coinciding with studies that did not observe any influence of this variable [12, 28]. This could be because: 1- running economy is just one factor explaining per- formance and it can be compensated by other factors [28]; 2- both economical and uneconom- ical runners have been identified at all levels of performance [29]; 3- the dependence of running economy on training status [2], as all runners in this study were well-trained; and 4- the higher percentage of midfoot/forefoot strikers in the Group 1 (Fig 1), being less economi- cal runners than rearfoot strikers [13,30]. Table 2. Mean (± SD) spatio-temporal variables of the different groups of runners during the incremental tests. Correlation (r) with running performance (time to complete a half-marathon). G1 (n = 11) G2 (n = 13) G3 (n = 13) G4 (n = 11) r PEAK Contact time (ms) 177±15†# 193±17†# 215±17 222±14 0.76 Step rate (spm) 190.7±4.7 187.6±6.3 190.6±8.0 189.7±15.5 0.01 Step length (m) 1.86±0.09†# 1.80±0.12†# 1.61±0.13 1.54±0.16 -0.73 RCT Contact time (ms) 198±23†# 219±19†# 241±19# 260±19 0.82 Step rate (spm) 181.7±6.9 177.4±7.3 178.5±8.9 172.7±9.6 -0.38 Step length (m) 1.66±0.09†# 1.58±0.11†# 1.42±0.09# 1.29±0.10 -0.87 VT Contact time (ms) 246±22†# 282±34†# 304±21 313±33 0.66 Step rate (spm) 167.5±4.8 166.2±8.0 162.6±6.2 159.6±6.2 -0.43 Step length (m) 1.22±0.09†# 1.13±0.12†# 1.03±0.06 1.05±0.08 -0.62 Note: G1, G2, G3, G4, groups of runners of different performance level (< 70, < 80, < 90 and < 105 min, respectively). PEAK, peak speed reached during the incremental test. RCT, respiratory compensation threshold. VT, ventilatory threshold. spm, steps per minute. , significant differences with Group 2. †, significant differences with Group 3. #, significant differences with Group 4. r, significant correlations (p<0.05) in bold type. https://doi.org/10.1371/journal.pone.0191688.t002 Table 3. Mean (± SD) spatio-temporal variables of the different groups of runners during the submaximal tests. Correlation (r) with running performance (time to complete a half-marathon). G1 (n = 11) G2 (n = 13) G3 (n = 13) G4 (n = 11) r 11 km
h-1 Contact time (ms) 258±19†# 279±19 290±20 295±26 0.53 Step rate (spm) 165.1±3.7 165.5±7.3 164.4±7.8 163.1±11.6 0.52 Step length (m) 1.11±0.03 1.11±0.05 1.12±0.05 1.13±0.08 0.19 13 km
h-1 Contact time (ms) 236±16†# 253±19 264±16 263±11 0.51 Step rate (spm) 169.3±3.7 168.2±6.2 173.4±9.8 171.1±11.1 0.13 Step length (m) 1.28±0.03 1.29±0.05 1.25±0.07 1.27±0.08 -0.10 15 km
h-1 Contact time (ms) 219±16†# 233±16 242±15 242±11 0.50 Step rate (spm) 174.9±3.6 172.1±6.6 180.5±10.3 178.5±13.0 0.23 Step length (m) 1.43±0.03 1.46±0.06 1.39±0.08 1.41±0.10 -0.21 Note: G1, G2, G3, G4, groups of runners of different performance level (< 70, < 80, < 90 and < 105 min, respectively). spm, steps per minute. , significant differences with Group 2. †, significant differences with Group 3. #, significant differences with Group 4. r, significant correlations (p<0.05) in bold type. https://doi.org/10.1371/journal.pone.0191688.t003 Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 7 / 11 Foot strike pattern Foot strike pattern distribution among groups found in this study is in line with previous stud- ies that compared the foot strike patterns of top and bottom place finishers of high-level half- marathon and marathon races [7,8]. Runners with a higher performance level tend to more frequently use a midfoot/forefoot strike pattern, which allows them to shorten contact time by 10% at the same running speed than rearfoot strikers [7,13,20,30–32]. This could be beneficial to reach high running speeds during training and competition (> 20 km
h-1) without compromising step rate [13,32]. Table 1 showed that peak running speed in Groups 1 and 2 was higher than 20 km
h-1, and contact time was lower than 200 ms (10% shorter than in the Groups 3 and 4), which highlights the importance of foot strike pattern to shorten contact time to achieve those high running speeds. Spatio-temporal parameters during the incremental test (comparison at relative physiological intensities) The differences in spatio-temporal variables (i.e. contact time, step rate and length) among groups and the correlations with performance during the incremental test were reasonable (Table 2). All these variables are highly dependent on running speed, and as it was previously commented, contact time is also dependent on foot strike pattern. In fact, it was observed in a previous study that an increase of 2 km
h-1 in running speed could mean an increase of ~ 7.4 steps per minute in step rate, ~ 0.284 m in step length and a decrease of ~ 20 ms in contact time, independently of the type of foot strike pattern [13]. However, during the incremental test, when foot strike pattern and running speed were considered as covariates (i.e. ANCOVA), the differences in spatio-temporal variables disappeared. This finding suggests that foot strike pattern and running speed are responsible for spatio-temporal differences between runners. At similar physiological intensities, step length was different among the groups of runners, while step rate was not (Table 2). This is in agreement with previous studies performed in vet- eran marathon runners, where shorter step length was the cause of speed reduction with age [33], possibly due to a loss of strength over the years [34]. Similarly, a strong relationship was also established between strength training and the improvement in long-distance running per- formance [35]. Nevertheless, to the best of our knowledge, none of these studies analyzed the effect of strength training programs on running spatio-temporal variables, which could be a future aim. Spatio-temporal parameters during the submaximal test (comparison at standardized running speeds) When running speed was controlled (i.e. submaximal text, Table 3), there were no differences among groups in step rate and length, in concordance with previous findings [13,20]. On the contrary, a recent study performed in collaboration with our research group and following similar experimental procedures showed differences in both step rate and length, but not in contact time when trained and untrained runners were compared [15]. Trained runners showed higher step rate and shorter step length at the same running speeds than untrained ones. This condition (i.e. higher step rate and shorter length) could be a natural adaptive mechanism to prevent some of the most common running-related injuries as it decreases the magnitude of the center of mass vertical excursion, ground reaction force, impact shock, and may ameliorate energy absorption at the hip, knee, and ankle joints impacts during running [36]. However, when experienced runners of different performance level are compared, as the present study showed, these differences in step rate and length are not observed, probably due Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 8 / 11 to the high training status of the runners (i.e. more than 40 km
week-1 of running, more than 3 years of running experience and a RCT above the 84% of VO2max) regardless their perfor- mance level. Thereby, from the results of the present and previous studies [13,15], the association between shorter contact times and better performance in long-distance runners [7,11] is quite questionable, because it depends on both foot strike pattern and running speed. When both variables are controlled, there are no differences in contact time among runners of different performance level. In other words, contact time seems to be very consistent among highly- trained runners of different performance level, which could constitute further investigation. Conclusions The present study demonstrated that runners from different performance level differed in training-related (i.e. years of experience and weekly training volume), anthropometric (i.e. body mass, body mass index and sum of skinfolds), physiological (i.e. VO2max, RCT and run- ning economy), foot strike pattern and spatio-temporal variables (i.e. contact time, step rate and length). However, when foot strike pattern and running speed were controlled (i.e. run- ning at the same absolute speed), spatio-temporal variables were similar among them. Higher level participants more frequently adopt midfoot/forefoot strike patterns and they run at higher running speeds, which implies differences in spatio-temporal variables. Nonetheless, future studies should analyze why spatio-temporal variables are so consistent when running speed and foot strike pattern are similar. Supporting information S1 Dataset. Individual dataset of the runners. (XLS) Acknowledgments The authors would like to thank the runners who participated in this study for their collaboration. Author Contributions Conceptualization: Ana Ogueta-Alday, Josue´ Go´mez-Molina, Juan Garcı´a-Lo´pez. Formal analysis: Ana Ogueta-Alday, Juan Garcı´a-Lo´pez. Funding acquisition: Juan Garcı´a-Lo´pez. Investigation: Ana Ogueta-Alday, Juan Carlos Morante, Juan Garcı´a-Lo´pez. Methodology: Ana Ogueta-Alday, Juan Carlos Morante, Josue´ Go´mez-Molina, Juan Garcı´a- Lo´pez. Resources: Juan Carlos Morante, Josue´ Go´mez-Molina. Software: Juan Carlos Morante, Juan Garcı´a-Lo´pez. Supervision: Juan Carlos Morante, Juan Garcı´a-Lo´pez. Validation: Juan Carlos Morante. Writing – original draft: Ana Ogueta-Alday, Juan Carlos Morante, Josue´ Go´mez-Molina, Juan Garcı´a-Lo´pez. Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 9 / 11 Writing – review & editing: Ana Ogueta-Alday, Josue´ Go´mez-Molina, Juan Garcı´a-Lo´pez. References 1. Running USA [Internet]. Running USA annual half-marathon report; c2015 [cited 2015 Aug 21]. http:// www.runningusa.org/annual-reports 2. Basset DR, Howley ET. Limiting factors for maximum oxygen uptake and determinants of endurance performance. Med Sci Sports Exerc. 2000; 32(1):70–84. PMID: 10647532 3. Hagan RD, Upton SJ, Duncan JJ, Gettman LR. Marathon performance in relation to maximal aerobic power and training indices in female distance runners. Br J Sports Med. 1987; 21(1):3–7. PMID: 3580726 4. Knechtle B, Duff B, Welzel U, Kohler G. Body mass and circumference of upper arm are associated with race performance in ultraendurance runners in a multistage race—the Isarrun 2006. Res Q Exerc Sport. 2009; 80(2):262–8. https://doi.org/10.1080/02701367.2009.10599561 PMID: 19650392 5. Zillmann T, Knechtle B, Ru¨st CA, Knechtle P, Rosemann T, Lepers R. Comparison of training and anthropometric characteristics between recreational male half-marathoners and marathoners. Chin J Physiol. 2013; 56(3):138–46. https://doi.org/10.4077/CJP.2013.BAB105 PMID: 23656215 6. Cheuvront SN, Haymes EM. Thermoregulation and marathon running: biological and environmental influences. Sports Med. 2001; 31(10):743–62. PMID: 11547895 7. Hasegawa H, Yamauchi T, Kramer WJ. Foot strike patterns of runners at the 15 km point during an elite-level half marathon. J Strength Cond Res. 2007; 21(3):888–93. PMID: 17685722 8. Kasmer ME, Liu XC, Roberts KG, Valadao JM. Foot-strike pattern and performance in a marathon. Int J Sports Physiol Perfom. 2013; 8(3):286–92. 9. Larson P, Higgins E, Kaminski J, Decker T, Preble J, Lyons D, et al. Foot strike patterns of recreational and sub-elite runners in a long-distance road race. J Sports Sci. 2011; 29(15):1665–73. https://doi.org/ 10.1080/02640414.2011.610347 PMID: 22092253 10. Moore I. Is there an economical running technique? A review of modifiable biomechanical factors affect- ing running economy. Sports Med. 2016; 46(6):793–807. https://doi.org/10.1007/s40279-016-0474-4 PMID: 26816209 11. Paavolainen LM, Nummela AT, Rusko HK. Neuromuscular characteristics and muscle power as deter- minants of 5-km running performance. Med Sci Sports Exerc. 1999; 31(1):124–30. PMID: 9927020 12. Støren Ø, Helgerud J, Hoff J. Running stride peak forces inversely determines running economy in elite runners. J Strength Cond Res. 2011; 25(1):117–23. https://doi.org/10.1519/JSC.0b013e3181b62c8a PMID: 20093965 13. Ogueta-Alday A, Rodrı´guez-Marroyo JA, Garcı´a-Lo´pez J. Rearfoot striking runners are more economi- cal than midfoot strikers. Med Sci Sports Exerc. 2014; 46(3):580–5. https://doi.org/10.1249/MSS. 0000000000000139 PMID: 24002340 14. Slawinski JS, Billat VL. Difference in mechanical and energy cost between highly, well, and nontrained runners. Med Sci Sports Exercise. 2004; 36(8):1440–6. 15. Go´mez-Molina J, Ogueta-Alday A, Stickley C, Ca´mara J, Cabrejas-Ugartondo J, Garcı´a-Lo´pez J. Differ- ences in spatiotemporal parameters between trained and untrained participants. J Strength Cond Res. 2017; 31(8):2169–75. https://doi.org/10.1519/JSC.0000000000001679 PMID: 28731978 16. Helgerud J, Støren Ø, Hoff J. Are there differences in running economy at different velocities for well- trained distance runners? Eur J Appl Physiol. 2010; 108(6):1099–105. https://doi.org/10.1007/s00421- 009-1218-z PMID: 20024579 17. Lucia A, Esteve-Lanao J, Olivan J, Go´mez-Gallego F, San Juan A, Santiago C, et al. Physiological char- acteristics of the best Eritrean runners-exceptional running economy. Appl Physiol Nutr Metab. 2006; 31(5):530–40. https://doi.org/10.1139/h06-029 PMID: 17111007 18. Franz JR, Wierzbinski CM, Kram R. Metabolic cost of running barefoot versus shod: is lighter better? Med Sci Sports Exerc. 2012; 44(8):1519–25. https://doi.org/10.1249/MSS.0b013e3182514a88 PMID: 22367745 19. Jones AM, Doust JH. A 1% treadmill grade most accurately reflects the energetic cost of outdoor run- ning. J Sports Sci. 1996; 14(4):321–7. https://doi.org/10.1080/02640419608727717 PMID: 8887211 20. Ogueta-Alday A, Morante JC, Rodrı´guez-Marroyo JA, Garcı´a-Lo´pez J. Validation of a new method to measure contact and flight times during treadmill running. J Strength Cond Res. 2013; 27(5):1455–62. https://doi.org/10.1519/JSC.0b013e318269f760 PMID: 22836607 21. Marfell-Jones M, Olds T, Stewart A, Carter JEL. International standards for anthropometric assess- ment. Potchefstroom, South Africa: ISAK; 2006. Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 10 / 11 22. Davis JA. Anaerobic threshold: a review of the concept and directions for future research. Med Sci Sports Exerc. 1985; 17(1):6–21. PMID: 3884961 23. Svedenhag J, Sjodin B. Body-mass-modified running economy and step length in elite male middle- and long-distance runners. Int J Sports Med. 1994; 15(6):305–10. https://doi.org/10.1055/s-2007- 1021065 PMID: 7822068 24. Tucker R, Collins M. What makes champions? A review of the relative contribution of genes and training to sporting success. Br J Sports Med. 2012; 46(8):555–61. https://doi.org/10.1136/bjsports-2011- 090548 PMID: 22535537 25. Larsen HB, Christensen DL, Nolan T, Sondergaard H. Body dimensions, exercise capacity and physical activity level of adolescent Nandi boys in western Kenya. Ann Hum Biol. 2004; 31(2):159–73. https:// doi.org/10.1080/03014460410001663416 PMID: 15204359 26. Tartaruga MP, Brisswalter J, Peyre´-Tartaruga LA, Avila AO, Alberton CL, Coertjens M, et al. The rela- tionship between running economy and biomechanical variables in distance runners. Res Q Exerc Sport. 2012; 83(3):367–75. https://doi.org/10.1080/02701367.2012.10599870 PMID: 22978185 27. Go´mez-Molina J, Ogueta-Alday A, Camara J, Stickley C, Rodrı´guez-Marroyo JA, Garcı´a-Lo´pez J. Pre- dictive variables of half-marathon performance for male runners. J Sports Sci Med. 2017; 16(2):187– 94. PMID: 28630571 28. Mooses M, Mooses K, Haile DW, Durussel J, Kaasik P, Pitsiladis YP. Dissociation between running economy and running performance in elite Kenyan distance runners. J Sports Sci. 2015; 33(2):136–44. https://doi.org/10.1080/02640414.2014.926384 PMID: 24916998 29. Morgan DW, Bransford DR, Costill DL, Daniels JT, Howley ET, Krahenbuhl GS. Variation in the aerobic demand of running among trained and untrained subjects. Med Sci Sports Exerc. 1995; 27(3):404–9. PMID: 7752868 30. Gruber AH, Umberger BR, Braun B, Hamill J. (2013). Economy and rate of carbohydrate oxidation dur- ing running with rearfoot or forefoot strike patterns. J Appl Physiol. 2013; 115(2):194–201. https://doi. org/10.1152/japplphysiol.01437.2012 PMID: 23681915 31. Di Michele R, Merni F. The concurrent effects of strike pattern and ground-contact time on running economy. J Sci Med Sport. 2014; 17(4):414–8. https://doi.org/10.1016/j.jsams.2013.05.012 PMID: 23806876 32. Hayes P, Caplan N. Foot strike patterns and ground contact times during high-calibre middle-distance races. J Sports Sci. 2012; 30(12):1275–83. https://doi.org/10.1080/02640414.2012.707326 PMID: 22857152 33. Conoboy P, Dyson R. Effect of aging on the stride pattern of veteran marathon runners. Br J Sports Med. 2006; 40(7):601–4. https://doi.org/10.1136/bjsm.2006.026252 PMID: 16687480 34. Piacentini MF, De Ioannon G, Comotto S, Spedicato A, Vernillo G, La Torre A. Concurrent strength and endurance training effects on running economy in master endurance runners. J Strength Cond Res. 2013; 27(8):2295–303. https://doi.org/10.1519/JSC.0b013e3182794485 PMID: 23207882 35. Taipale RS, Mikkola J, Vesterinen V, Nummela A, Ha¨kkinen K. Neuromuscular adaptations during com- bined strength and endurance training in endurance runners: maximal versus explosive strength train- ing or a mix of both. Eur J Appl Physiol. 2013; 113(2):325–35. https://doi.org/10.1007/s00421-012- 2440-7 PMID: 22711181 36. Schubert AG, Kempf J, Heiderscheit BC. Influence of stride frequency and length in running mechanics: a systematic review. Sports Health. 2014; 6(3):210–17. https://doi.org/10.1177/1941738113508544 PMID: 24790690 Comparison of different level runners PLOS ONE | https://doi.org/10.1371/journal.pone.0191688 January 24, 2018 11 / 11

Similarities and differences among half-marathon runners according to their performance level.

01-24-2018

Ogueta-Alday, Ana,Morante, Juan Carlos,Gómez-Molina, Josué,García-López, Juan

eng

PMC9566275

Citation: Cassirame, J.; Godin, A.; Chamoux, M.; Doucende, G.; Mourot, L. Physiological Implication of Slope Gradient during Incremental Running Test. Int. J. Environ. Res. Public Health 2022, 19, 12210. https://doi.org/10.3390/ ijerph191912210 Academic Editor: Luca Paolo Ardigo Received: 18 August 2022 Accepted: 22 September 2022 Published: 26 September 2022 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affil- iations. Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). International Journal of Environmental Research and Public Health Article Physiological Implication of Slope Gradient during Incremental Running Test Johan Cassirame 1,2,3,* , Antoine Godin 4 , Maxime Chamoux 5 , Gregory Doucende 5,† and Laurent Mourot 4,† 1 Laboratory Culture Sport Health and Society (C3S−UR 4660), Sport and Performance Department, University of Bourgogne Franche-Comte, 25000 Besançon, France 2 EA7507, Laboratoire Performance, Santé, Métrologie, Société, 51100 Reims, France 3 Mtraining, R&D Division, 25480 Ecole-Valentin, France 4 EA3920-Prognostic Markers and Regulatory Factors of Heart and Vascular Diseases, and Exercise Performance, Health, Innovation Platform, University Bourgogne Franche-Comté, 25000 Besançon, France 5 Laboratoire Interdisciplinaire Performance Santé en Environnement de Montagne (LIPSEM), UR-4604, Université de Perpignan Via Domitia, 7 Avenue Pierre de Coubertin, 66120 Font-Romeu, France * Correspondence: [email protected] † These authors contributed equally to this work. Abstract: Uphill running induces a higher physiological demand than level conditions. Although many studies have investigated this locomotion from a psychological point of view, there is no clear position on the effects of the slope on the physiological variables during an incremental running test performed on a slope condition. The existing studies have heterogeneous designs with different populations or slopes and have reported unclear results. Some studies observed an increase in oxygen consumption, whereas it remained unaffected in others. The aim of this study is to investigate the effect of a slope on the oxygen consumption, breathing frequency, ventilation and heart rate during an incremental test performed on 0, 15, 25 and 40% gradient slopes by specialist trail runners. The values are compared at the first and second ventilatory threshold and exhaustion. A one-way repeated measures ANOVA, with a Bonferroni post-hoc analysis, was used to determine the effects of a slope gradient (0, 15, 25 and 40%) on the physiological variables. Our study shows that all the variables are not affected in same way by the slopes during the incremental test. The heart rate and breathing frequency did not differ from the level condition and all the slope gradients at the ventilatory thresholds or exhaustion. At the same time, the ventilation and oxygen consumption increased concomitantly with the slope (p < 0.001) in all positions. The post-hoc analysis highlighted that the ventilation significantly increased between each successive gradient (0 to 15%, 15% to 25% and 25% to 40%), while the oxygen consumption stopped increasing at the 25% gradient. Our results show that the 25 and 40% gradient slopes allow the specialist trail runners to reach the highest oxygen consumption level. Keywords: trail running; exercise physiology; maximal oxygen consumption; performance; testing 1. Introduction The physiology of running has been largely investigated in the last century to explore human physiology and performance [1–4]. The bipedal locomotion speed is commonly used to prescribe an exercise intensity with the aim of analyzing a physiological request [5]. For the purposes of clinical and sports performance, incremental running tests have been developed to evaluate the maximal cardiorespiratory capabilities [6] to perform clinical diagnoses or prescribe physical training programs. Such evaluations have been designed mainly on treadmill machines, allowing an increase in the exercise intensity by adjusting the speed, slope or both [7,8]. Such testing procedures are massively used in sports science because they allow the subject to reach the maximum oxygen uptake ( . VO2max), which Int. J. Environ. Res. Public Health 2022, 19, 12210. https://doi.org/10.3390/ijerph191912210 https://www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2022, 19, 12210 2 of 12 is an important contributor to performance in many sports [3,9,10]. These procedures also allow the determination of physiological landmarks, such as ventilatory thresholds, that allow the setting of intensity levels during a training program [11,12]. On a level condition, the protocols designed to assess the . VO2max during running could differ, but may lead to similar . VO2max values [13–15] if the progressivity and starting intensity are appropriately set. In recent decades, the increase in popularity of trail running races [16] has attracted interest in uphill and downhill running evaluations, including the . VO2max [17–19]. Physio- logically speaking, the running economy and energy cost of uphill running has been largely described [2,20–22]. However, few studies have investigated the specific effect of a slope on the maximal physiological values, such as the . VO2max, during incremental running tests. The existing literature provides contrasting or opposite results as the studies have been conducted with different populations and/or testing protocols. The studies using a constantly increasing slope, [23] or with slopes from +7% to +25% [24–26] did not reported an alteration in the . VO2max with a positive slope, whereas others pointed out an increase in the . VO2max with a slope of up to 35% [27–29]. Hence, based on the current literature, it is not clear if a slope gradient can induce a significant alteration in the cardiopulmonary variables, especially with well-adapted athletes, such as trail runners, or if an optimal slope can be identified to reach the highest cardiorespiratory involvement, without being impacted by peripheral limitations, such as a lack of muscular force. Specifically, with trail runners, Balducci et al.’s or Schöffl et al.’s studies compared the maximal cardiorespiratory performance while running at 12.5, 16, and 25% on a treadmill and in ecological field situations [24,26]. They reported no significant change in the . VO2max but a progressive increase in the ventilation ( . VE) with the slope. Contrary to this result, Scheer et al. reported a larger . VO2max and blood lactate concentration post-exhaustion (3 min) [28]. However, the protocol used for this study included a concomitant speed and slope increment during testing (+0.5 km·h−1 and +1% per minute) and provided a final slope of only around a +10% gradient. Hence, the aim of this study was to examine the effects of different gradient slopes (from level to +40%) on the cardiorespiratory variables reached at exhaustion and the ventilatory thresholds during maximum incremental tests in specialist trail runners. Based on the previous studies, and especially on the physiological limitations of the . VO2max [30], we hypothesized that the steeper the slope, the higher the . VO2max, since a steeper slope will require a larger muscular mass to elevate the body mass [31,32]. However, we also hypothesized that this phenomenon would tend to a plateau, so that a further increase in the slope would not lead to an increase and could lead to a decrease in the . VO2, due to the peripheral limitations. 2. Methodology 2.1. Participants Fourteen young trail runners, free of injury in the last six months, were involved in this study, with a training volume (8.4 ± 3.2 h) in last two months: four females (age: 20.2 ± 2.8 years, size 1.73 ± 0.03 m, weight 61 ± 7.1 kg), and ten males (age: 20.6 ± 2.2 years, size 1.76 ± 0.04 m, weight 65.1 ± 5.2 kg). The measurement period took place in the second part of the racing season, with no competitions in last two weeks preceding the measure- ments. All the athletes have more than four years of active practice of trail running, and the training volumes for both genders are, respectively, 8.7 ± 3 h and 8.9 ± 2.5 h. All the participants were informed of the design and aim of the study and provided their written consent to participate in this study. The experiment was conducted in accordance with the Declaration of Helsinki and received the approval ID-RCB: 2019-A03012-55 from “COMITE DE PROTECTION DES PERSONNES SUD MEDITERRANEE IV”. Int. J. Environ. Res. Public Health 2022, 19, 12210 3 of 12 2.2. Experiential Design All the athletes involved in this study performed, in random order, four incremental test sessions with different constant slopes. All the sequence possibilities (24 different randomizations) were assigned to the athletes by a draw, eliminating successive sequences to avoid each athlete performing a similar sequence. The tests were completed in a period of two weeks, respecting at least three days of rest after each assessment and avoiding other strenuous activities. The protocols were designed with 0%, 15%, 25% and 40% positive slopes in ecological field conditions. The level protocol was performed on a track and field loop, whereas the slope protocols were performed on a regular track with a constant slope in a ski resort. The level protocol starts at 8 km·h−1 and increases by 0.5 km·h−1 every minute [33]; the 15% slope protocol starts at 3.37 km·h−1 and increases by 0.41 km·h−1; the 25% slope protocol starts at 2 km·h−1 and increases by 0.34 km·h−1; and the 40% protocol starts at 1.35 km·h−1 and increases by 0.27 km·h−1 (Figure 1). The uphill protocols were designed, based on previous experiments, in order to reach similar test durations whatever the slope, in line with the current recommendations [34]. These protocols start at the same ascending speed (AS) of 500 m per hour and with increments of 50 m per hour for the 15% slope condition, and of 100 m per hour for the 25% and 40% conditions. The speed control was performed using an audio soundtrack read by a mobile MP3 player. For the level condition, the pacing was done by a soundtrack reading a sound every 20 m [33]. Regarding the uphill test, the fixed pacing was done every 15 s. To maintain the right speed, the athlete must be at the flag position when the signal sounds. The test was interrupted if athletes deviated more than 5 m difference from the appropriate position during two successive intervals. Int. J. Environ. Res. Public Health 2022, 19, x 3 of 13 2.2. Experiential Design All the athletes involved in this study performed, in random order, four incremental test sessions with different constant slopes. All the sequence possibilities (24 different ran- domizations) were assigned to the athletes by a draw, eliminating successive sequences to avoid each athlete performing a similar sequence. The tests were completed in a period of two weeks, respecting at least three days of rest after each assessment and avoiding other strenuous activities. The protocols were designed with 0%, 15%, 25% and 40% pos- itive slopes in ecological field conditions. The level protocol was performed on a track and field loop, whereas the slope protocols were performed on a regular track with a constant slope in a ski resort. The level protocol starts at 8 km·h−1 and increases by 0.5 km·h−1 every minute [33]; the 15% slope protocol starts at 3.37 km·h−1 and increases by 0.41 km·h−1; the 25% slope protocol starts at 2 km·h−1 and increases by 0.34 km·h−1; and the 40% protocol starts at 1.35 km·h−1 and increases by 0.27 km·h−1 (Figure 1). The uphill protocols were designed, based on previous experiments, in order to reach similar test durations what- ever the slope, in line with the current recommendations [34]. These protocols start at the same ascending speed (AS) of 500 m per hour and with increments of 50 m per hour for the 15% slope condition, and of 100 m per hour for the 25% and 40% conditions. The speed control was performed using an audio soundtrack read by a mobile MP3 player. For the level condition, the pacing was done by a soundtrack reading a sound every 20 m [33]. Regarding the uphill test, the fixed pacing was done every 15 s. To maintain the right speed, the athlete must be at the flag position when the signal sounds. The test was inter- rupted if athletes deviated more than 5 m difference from the appropriate position during two successive intervals. Figure 1. Graphical representation of each protocol performed by participants. The grey area repre- sents the speed in km·h, while the black solid line represents the slope gradient in degrees. Protocol designs are displayed from 0° to 40° from left to right. ((A): 0°, (B): 15°, (C): 25° and (D): 40°). 2.3. Physiological Measurements During the incremental testing, the physiological parameters were measured using a portable gas exchange system, the Metamax 3B-R2 (Cortex Biophysics, Leipzig, Ger- many), previously validated by Marcfarlane et al. [35]. This system was installed on par- ticipants with a vest in a thoracic position and carefully placed on the clavicula to permit free arm movement while running. An oronasal face mask (7450 series V2 (HansRudolph, Shawnee, KS, USA)) was adjusted on each participant to install a bi-directional digital turbine. This turbine measured the respiration flow and obtained the 𝑉ሶ 𝐸 in L.min−1 and the breathing frequency (BF) in cycles.min−1. A short sample line tube (0.6 m) collected the inspired and expired air between the mask and turbine to measure the O2 and CO2 con- centrations and calculate the O2 consumption (𝑉ሶ 𝑂ଶ, L·min−1) and CO2 output (𝑉ሶ 𝐶𝑂ଶ L·min−1). For each subject, the 𝑉ሶ 𝑂ଶ was normalized and expressed in mL·min−1·kg−1. The heart rate (HR) was collected by a thoracic belt strap, the Polar H7 (Polar Electro, Kemplele, Finland), and transmitted via Bluetooth Low Energy technology to the gas ex- change measurement system. For each data collection, the tests were initialized and Figure 1. Graphical representation of each protocol performed by participants. The grey area represents the speed in km·h, while the black solid line represents the slope gradient in degrees. Protocol designs are displayed from 0◦ to 40◦ from left to right. ((A): 0◦, (B): 15◦, (C): 25◦ and (D): 40◦). 2.3. Physiological Measurements During the incremental testing, the physiological parameters were measured using a portable gas exchange system, the Metamax 3B-R2 (Cortex Biophysics, Leipzig, Germany), previously validated by Marcfarlane et al. [35]. This system was installed on participants with a vest in a thoracic position and carefully placed on the clavicula to permit free arm movement while running. An oronasal face mask (7450 series V2 (HansRudolph, Shawnee, KS, USA)) was adjusted on each participant to install a bi-directional digital turbine. This turbine measured the respiration flow and obtained the . VE in L.min−1 and the breathing frequency (BF) in cycles.min−1. A short sample line tube (0.6 m) collected the inspired and expired air between the mask and turbine to measure the O2 and CO2 concentrations and calculate the O2 consumption ( . VO2, L·min−1) and CO2 output ( . VCO2 L·min−1). For each subject, the . VO2 was normalized and expressed in mL·min−1·kg−1. The heart rate (HR) was collected by a thoracic belt strap, the Polar H7 (Polar Electro, Kemplele, Finland), and transmitted via Bluetooth Low Energy technology to the gas exchange measurement system. For each data collection, the tests were initialized and started from a computer Int. J. Environ. Res. Public Health 2022, 19, 12210 4 of 12 using MetaSoft Studio© software 5.5.1, and the data were collected into the internal memory of the portable device. Then, all the data were downloaded with the software to be stored and analyzed. Before each test, the flow sensor was calibrated with a 3L syringe and the gas sensors were calibrated with ambient air and the reference gas (15% O2, 5% CO2), as recommended by the manufacturer. For each individual test, the MetaSoft Studio© software determined automatically the maximal physiological values in the highest average of 30 s. During this period (MAX), the . VO2max in mL·min−1·kg−1, ventilation ( . VEmax) in L·min−1, breathing frequency (BFmax), heart rate (HRmax) were calculated. An experienced examinator determined the positions of the ventilatory thresholds 1 (VT1) and 2 (VT2) using Wasserman and Beaver’s method [36] to allow the system to extract following parameters: . VO2vt1, . VEvt1, BFvt1, HRvt1 and . VO2vt2, . VEvt2, BFvt2 and HRvt2. The VT1 and VT2 positions were set using the graphical interface of the MetaSoft Studio© software 5.5.1 displaying the . VE/ . VO2 and . VE/ . VCO2 curves over the time, with averaging every 10 s. The VT1 was set at the first increase of . VE/ . VO2 without an increase of . VE/ . VCO2, and the VT2 was set at a concomitant increase of . VE/ . VO2 and . VE/ . VCO2. For each tests series with a gradient (15, 25 and 40%), the ascending speed in meters per hour was calculated and identified at the VT1, VT2 and the maximum moments. 2.4. Statistic All the data exported from the MetaSoft Studio© software were merged into a Mi- crosoft Office 365 Excel spreadsheet (Microsoft, Redmond, WA, USA) and computed to be analyzed with a custom R-Studio algorithm on the desktop software version 1.4.1106 (RStudio PBC, Boston, MA, USA). The descriptive statistics are presented as the mean and SD for the physiological variables and AS. The normal distribution of all the physiological variables was confirmed through the Shapiro–Wilk test (p > 0.05) except for the . VO2, HR and . VE measurements at the maximum moment on a 0% of slope. A one-way repeated measures ANOVA with a Bonferroni post-hoc analysis was used to determine the effects of the slope gradient (0, 15, 25 and 40%) on the physiological variables (HR, BF, . VE and . VO2) at specific time points: the VT1, VT2 and the maximum (MAX). In a second time, similar procedures were performed for the AS values for the gradient slopes of 15, 25 and 40% for each moment. The physiological variables not having a normal distribution were analyzed using the Friedman test to determine the impact of the slope gradient [37]. A post-hoc test was a paired Wilcoxon signed-rank test with the Bonferroni correction [38]. These analyses were complemented by an effect size estimation using Hedges’ g (population <16, repeated measures design) [39]. Hedges’ g was also used for the variables not following a normal distribution. The magnitude thresholds for the effect size were defined as 0.20, 0.60, 1.20, 2.0 and 4.0 for small, moderate, large, very large and extremely large correlation coefficients, in accordance with previous recommendations [40]. For this study a p level inferior at 0.05 was considered as significant. 3. Results The durations of the tests were 944.0 ± 115.8 s, 907.6 ± 99.6 s, 900.2 ± 100.2 s and 904.3 ± 101.2 s for the 0, 15, 25 and 40% gradient slopes, respectively, without significant differences between the conditions. The data for each slope and time points (VT1, VT2 and MAX) are displayed in Table 1. All the individual values, as well as the mean and standard deviations are displayed as violin plot graphics in Figure 2 to observe the distribution and changes with the slope conditions. A second violin plots series was designed to show the AS values for the gradient conditions (Figure 3). Int. J. Environ. Res. Public Health 2022, 19, 12210 5 of 12 Int. J. Environ. Res. Public Health 2022, 19, x 8 of 13 Figure 2. Violin plot of dependent variables, heart rate (HR), breathing frequency (BF), ventilation (𝑉ሶ 𝐸), oxygen consumption (𝑉ሶ 𝑂ଶ) for all time points (VT1, VT2 and MAX) for slope conditions (0, Figure 2. Violin plot of dependent variables, heart rate (HR), breathing frequency (BF), ventilation ( . VE), oxygen consumption ( . VO2) for all time points (VT1, VT2 and MAX) for slope conditions (0, 15, 25 and 40%). Statistical differences observed were noted as follow; * p < 0.05, ** p < 0.01 and *** p < 0.001. Int. J. Environ. Res. Public Health 2022, 19, 12210 6 of 12 Int. J. Environ. Res. Public Health 2022, 19, x 9 of 13 15, 25 and 40%). Statistical differences observed were noted as follow; * p < 0.05, ** p < 0.01 and *** p < 0.001. Figure 3. Violin plot of ascending speed (AS) at all positions (VT1, VT2 and MAX) for all slope conditions (15, 25 and 40%). Statistical differences observed were noted as follow; * p < 0.05, ** p < 0.01 and *** p < 0.001. 4. Discussion The aim of this study was to investigate the effect of a positive gradient slope on cardiopulmonary variables during maximal incremental testing. As hypothesized, we ob- served that the steeper the slope, the higher the 𝑉ሶ 𝑂ଶ𝑚𝑎𝑥 up to +25%, without a further significant increase thereafter. However, we also noticed that not all cardiorespiratory variables are equally influenced by the slope gradient increase. As expected, our results confirmed an increase in the cardiorespiratory requirements while increasing the slope angles. However, not all cardiorespiratory variables were im- pacted equally. Indeed, we observed that the HR at VT1, VT2 and MAX remained unaf- fected by the slope gradients from 15% to 40%. This finding at MAX is in line with the previous studies [24,26,28,29]. This observation provides an interesting confirmation for training purposes and allows the transposition of the HR intensities in the various positive slopes when targeting the ventilatory threshold intensity based on the HR values. Beyond the HR, the breathing pattern was impacted by the slope gradients. For the BF, we noted that no difference was found at the VT1 and VT2 landmarks. However, a weak trend was observed at the MAX, highlighting that the greater the slope, the higher the BF (p = 0.011) associated with a small effect size between the level condition and the +25% and +40% gradient or +15% and the steeper slopes. Moreover, we can observe in Figure 2 that the individual BF values are more dispersed for the 25 and 40% slopes. This observation is also corroborated with an increased standard deviation, especially at the MAX and for 40%. This bigger dispersion of the BF at the 25 and 40% gradients could be explained by different running pattern strategies when the slope increases. Indeed, run- ners can choose to increase cadence and decrease step-length, or vice-versa, [24] and it is known that a tight locomotor–respiratory coupling exists, especially while running [41]. Moreover this coupling could be exacerbated when the upper limbs are more involved [42,43] potentially, or in a more pronounced way, at higher gradients. On the other hand, the 𝑉ሶ 𝐸 was largely positively influenced by the increase in the slope for the VT1, VT2 and MAX time points. We observed small to moderate effect sizes between the conditions, excepting between +25% and 40% at the VT2 and MAX land- marks, which remained trivial. These results differ from the literature, since Balducci et al. did not observe any differences during the incremental running test for the level con- dition vs. the 15% and the level vs. the 25% slope gradients [24], and Scheer et al. did not report any differences between the level test and the test with the constant slope increase (1% per minute) [28]. Given that the BF remained largely unchanged, this phenomenon Figure 3. Violin plot of ascending speed (AS) at all positions (VT1, VT2 and MAX) for all slope conditions (15, 25 and 40%). Statistical differences observed were noted as follow; * p < 0.05, ** p < 0.01 and *** p < 0.001. Table 1. Mean and standard deviation of heart rate (HR), breathing frequency (BF), ventilation ( . VE), oxygen consumption ( . VO2) and ascending speed (AS) for each position; ventilatory thresholds (VT1 and VT2) and at maximum (MAX) for 0, 15, 25, 40% gradient slope conditions. §, Ø and β indicate that value is significantly different than 0%, 15% and 25% gradient slope condition, respectively. 0% 15% 25% 40% VT1 HR (bpm) 148.3 ± 11.9 150.4 ± 11.4 150.1 ± 11.6 150.4 ± 11.4 BF (cycles·min−1) 44.1 ± 4.8 44.9 ± 4.5 45 ± 5.1 45.3 ± 5.4 VE (L·min−1) 90 ± 5.8 Ø β 98.9 ± 10.8 § 101.6 ± 11.9 § 105.9 ± 12.4 § Ø β VO2 (mL·min−1·kg) 42 ± 4.2 Ø β 46 ± 6 § 46.9 ± 6.2 § 46.9 ± 6.7 § AS (m·h−1) - 807.1 ± 85.1 β 1057.1 ± 126.8 1096.4 ± 135.1 Ø VT2 HR (bpm) 171.9 ± 8.2 172.4 ± 9.7 171.9 ± 9.1 172.3 ± 8.5 BF (cycles·min−1) 50.9 ± 5 52 ± 5 51.4 ± 5.4 52.6 ± 4.5 VE (L·min−1) 131.9 ± 18.2 β 137 ± 18.8 141.5 ± 17.9 § 144.1 ± 18.1 § VO2 (mL·min−1·kg) 55.1 ± 6 Ø β 58.6 ± 5 § β 59.9 ± 4.4 § 61 ± 5.6 § Ø AS (m·h−1) - 1039.2 ± 163.1 β 1478.5 ± 106.9 1503.5 ± 120 Ø Max HR (bpm) 189 ± 7.6 188.7 ± 8.2 188.2 ± 8.1 188.4 ± 7.8 BF (cycles·min−1) 57.1 ± 3.9 57.5 ± 3.6 58.4 ± 4.4 59.1 ± 4.7 VE (L·min−1) 164.5 ± 25.3 Ø β 172.6 ± 24.8 § 177.4 ± 23.6 § Ø 181.5 ± 24.5 § Ø β VO2 (mL·min−1·Kg) 63.3 ± 7.3 Ø β 66.5 ± 5.9 § β 67.9 ± 5.8 § Ø 68.1 ± 5.9 § Ø AS (m·h−1) - 1175 ± 141.1 β 1750 ± 137.2 1771.4 ± 132.6 Ø The results of the statistical analysis from the repeated one-way ANOVA and Bonfer- roni post-hoc analysis are displayed in Table 2 for all the physiological variables, whereas the ascending speed can be read in Table 3. At the VT1, VT2 and MAX time points, the . VE, . VO2 and AS significantly increased with an increase in the slope gradients (p < 0.001), whereas the HR and BF were not significantly affected in any condition. Specifically, only the HR (at VT1; significantly increased with slopes; p = 0.006) and BF (at MAX; significantly increased with slopes; p = 0.011) were affected. Moreover, the AS showed significantly slower values (p < 0.001) between 15% and the other gradients (25 and 40%) but no differ- ence was observed at any time points between the 25 and 40% slope gradients (p = 0.999). Int. J. Environ. Res. Public Health 2022, 19, 12210 7 of 12 Table 2. Results of statistical analysis including Shapiro–Wilk test for normality and p value for one-way ANOVA and Bonferroni post-hoc analysis comparing heart rate (HR), breathing frequency (BF), ventilation ( . VE) and oxygen consumption ( . VO2) at all positions (VT1, VT2 and MAX) and between slope conditions. Effect sizes are presented in right part of the table using Hedges’ g method. Anova One-Way Repeated Measures Bonferroni Post-Hoc Effect Size HR VT1 Slope p value Normality p value p value g Hedges 0% 0.074 Valid 0.006 0% 15% 25% 0% 15% 25% 15% 0.449 Valid 15% 0.22 15% 0.172 25% 0.202 Valid 25% 0.13 0.999 25% 0.142 0.024 40% 0.352 Valid 40% 0.16 0.999 0.999 40% 0.173 0 0.241 VT2 Slope p value Normality p value p value g Hedges 0% 0.182 Valid 0.920 0% 15% 25% 0% 15% 25% 15% 0.753 Valid 15% 0.999 15% 0.061 25% 0.806 Valid 25% 0.999 0.999 25% 0.001 0.059 40% 0.771 Valid 40% 0.999 0.999 0.999 40% 0.049 0.015 0.047 MAX Slope p value Normality p value p value g Hedges 0% 0.453 Failed 0.899 0% 15% 25% 0% 15% 25% 15% 0.068 Valid 15% 0.999 15% 0.035 25% 0.72 Valid 25% 0.999 0.999 25% 0.096 0.059 40% 0.287 Valid 40% 0.999 0.999 0.999 40% 0.074 0.034 0.026 BF VT1 Slope p value Normality p value p value g Hedges 0% 0.377 Valid 0.421 0% 15% 25% 0% 15% 25% 15% 0.434 Valid 15% 0.999 15% 0.163 25% 0.444 Valid 25% 0.999 0.999 25% 0.181 0.028 40% 0.3 Valid 40% 0.5 0.5 0.999 40% 0.23 0.083 0.053 VT2 Slope p value Normality p value p value g Hedges 0% 0.453 Valid 0.182 0% 15% 25% 0% 15% 25% 15% 0.49 Valid 15% 0.999 15% 0.209 25% 0.107 Valid 25% 0.999 0.999 25% 0.08 0.121 40% 0.054 Valid 40% 0.19 0.999 0.78 40% 0.336 0.117 0.329 MAX Slope p value Normality p value p value g Hedges 0% 0.08 Valid 0.011 0% 15% 25% 0% 15% 25% 15% 0.187 Valid 15% 0.999 15% 0.11 25% 0.586 Valid 25% 0.058 0.666 25% 0.301 0.207 40% 0.913 Valid 40% 0.051 0.064 0.307 40% 0.468 0.382 0.169 VE VT1 Slope p value Normality p value p value g Hedges 0% 0.643 Valid <0.001 0% 15% 25% 0% 15% 25% 15% 0.49 Valid 15% 0.01 15% 0.994 25% 0.163 Valid 25% 0.008 0.439 25% 1.2 0.232 40% 0.056 Valid 40% <0.001 <0.001 0.002 40% 1.6 0.591 0.348 VT2 Slope p value Normality p value p value g Hedges 0% 0.095 Valid < 0.001 0% 15% 25% 0% 15% 25% 15% 0.118 Valid 15% 0.569 15% 0.266 25% 0.208 Valid 25% <0.001 0.205 25% 0.515 0.238 40% 0.119 Valid 40% <0.001 0.086 0.097 40% 0.651 0.372 0.139 MAX Slope p value Normality p value p value g Hedges 0% 0.186 Failed <0.001 0% 15% 25% 0% 15% 25% 15% 0.097 Valid 15% <0.001 15% 0.316 25% 0.193 Valid 25% <0.001 0.01 25% 0.511 0.189 40% 0.689 Valid 40% <0.001 0.003 0.036 40% 0.663 0.348 0.167 Int. J. Environ. Res. Public Health 2022, 19, 12210 8 of 12 Table 2. Cont. Anova One-Way Repeated Measures Bonferroni Post-Hoc Effect Size VO2 VT1 Slope p value Normality p value p value g Hedges 0% 0.398 Valid <0.001 0% 15% 25% 0% 15% 25% 15% 0.614 Valid 15% <0.001 15% 0.754 25% 0.409 Valid 25% <0.001 0.201 25% 0.886 0.136 40% 0.848 Valid 40% <0.001 0.505 0.999 40% 0.86 0.143 0.010 VT2 Slope p value Normality p value p value g Hedges 0% 0.22 Valid <0.001 0% 15% 25% 0% 15% 25% 15% 0.81 Valid 15% <0.001 15% 0.588 25% 0.184 Valid 25% <0.001 0.048 25% 0.898 0.264 40% 0.204 Valid 40% <0.001 0.001 0.739 40% 0.994 0.43 0.205 MAX Slope p value Normality p value p value g Hedges 0% 0.034 Failed <0.001 0% 15% 25% 0% 15% 25% 15% 0.749 Valid 15% <0.001 15% 0.427 25% 0.889 Valid 25% <0.001 0.042 25% 0.680 0.237 40% 0.482 Valid 40% <0.001 0.039 0.999 40% 0.712 0.271 0.035 Table 3. Results of statistical analysis including Shapiro–Wilk test for normality and p value for one-way ANOVA and Bonferroni post-hoc analysis comparing ascending speed (AS) at all positions (VT1, VT2 and MAX) and between slope conditions 15, 25 and 40%. Effect sizes are presented in right part of the table using Hedges’ g method. Anova One-Way Repeated Measures Bonferroni Post-Hoc Effect Size AS VT1 Slope p value Normality p value p value g Hedges 15% 0.655 Valid <0.001 15% 25% 15% 25% 25% 0.046 Valid 25% <0.001 25% 2.25 40% 0.036 Valid 40% <0.001 0.999 40% 2.49 0.291 VT2 Slope p value Normality p value p value g Hedges 15% 0.013 Valid <0.001 15% 25% 15% 25% 25% 0.65 Valid 25% <0.001 25% 3.09 40% 0.275 Valid 40% <0.001 0.999 40% 3.15 0.214 Max Slope p value Normality p value p value g Hedges 15% 0.094 Valid <0.001 15% 25% 15% 25% 25% 0.518 Valid 25% <0.001 25% 4.01 40% 0.056 Valid 40% <0.001 0.999 40% 4.23 0.154 4. Discussion The aim of this study was to investigate the effect of a positive gradient slope on cardiopulmonary variables during maximal incremental testing. As hypothesized, we observed that the steeper the slope, the higher the . VO2max up to +25%, without a further significant increase thereafter. However, we also noticed that not all cardiorespiratory variables are equally influenced by the slope gradient increase. As expected, our results confirmed an increase in the cardiorespiratory requirements while increasing the slope angles. However, not all cardiorespiratory variables were impacted equally. Indeed, we observed that the HR at VT1, VT2 and MAX remained unaffected by the slope gradients from 15% to 40%. This finding at MAX is in line with the previous studies [24,26,28,29]. This observation provides an interesting confirmation for training purposes and allows the transposition of the HR intensities in the various positive slopes when targeting the ventilatory threshold intensity based on the HR values. Beyond the HR, the breathing pattern was impacted by the slope gradients. For the BF, we noted that no difference was found at the VT1 and VT2 landmarks. However, a weak trend was observed at the MAX, highlighting that the greater the slope, the higher the BF Int. J. Environ. Res. Public Health 2022, 19, 12210 9 of 12 (p = 0.011) associated with a small effect size between the level condition and the +25% and +40% gradient or +15% and the steeper slopes. Moreover, we can observe in Figure 2 that the individual BF values are more dispersed for the 25 and 40% slopes. This observation is also corroborated with an increased standard deviation, especially at the MAX and for 40%. This bigger dispersion of the BF at the 25 and 40% gradients could be explained by different running pattern strategies when the slope increases. Indeed, runners can choose to increase cadence and decrease step-length, or vice-versa, [24] and it is known that a tight locomotor–respiratory coupling exists, especially while running [41]. Moreover this coupling could be exacerbated when the upper limbs are more involved [42,43] potentially, or in a more pronounced way, at higher gradients. On the other hand, the . VE was largely positively influenced by the increase in the slope for the VT1, VT2 and MAX time points. We observed small to moderate effect sizes between the conditions, excepting between +25% and 40% at the VT2 and MAX landmarks, which remained trivial. These results differ from the literature, since Balducci et al. did not observe any differences during the incremental running test for the level condition vs. the 15% and the level vs. the 25% slope gradients [24], and Scheer et al. did not report any differences between the level test and the test with the constant slope increase (1% per minute) [28]. Given that the BF remained largely unchanged, this phenomenon can only be explained by an increase in the tidal volume with an increasing slope to increase the air volume exchange at each breathing cycle. Similar finding have been reported by Lemire et al., comparing uphill and downhill to an incremental level running test [18]. In the same way, these authors also observed a higher metabolic coupling of ventilation when running uphill at 15% [18]. Moreover, the . VE and tidal volume have already been demonstrated to be higher at the maximal intensity (>80% maximum) during an incremental test with a slope gradient compared to the level condition [27], or in a field test performed at 16% compared with the treadmill test at 1% [26]. Here again, the hypothesis of locomotor–respiratory coupling could potentially explain the mechanisms. Potentially, the slope gradient led to a reduction of the step frequency [24,44] at a similar intensity and could allow the use of a deeper breathing pattern with increased tidal volume. Finally, we observed that the . VO2 was higher during uphill running compared to the slope gradients for the VT1, VT2 and MAX time points. Amongst the different positive slopes, we noticed that, at VT1, no significant differences were found between the 15% and 25% (p = 0.201) and between the 25% and 40% gradients (p = 0.999). Both the VT2 and MAX time points provided the same pattern, with a significant increase in the . VO2 between 15% and 25% (p < 0.001 with a small effect size), while no differences were found between 25% and 40% (p = 0.739 and 0.999) These results corroborated the previous findings comparing the . VO2max between the level condition and the various slopes conditions [27–29,31,45,46]. Nevertheless, it is important to notice that these previous studies focused on a smaller slope gradient (7%). Moreover, differences in the studied populations could be reported as the previous studies were conducted without specialist trail runners [25] or with older and heavier athletes [24,26]. In these studies, the athletes’ ages were 38.5 ± 6.4 years and 42.8 ± 14.6 years, respectively, whereas the were 20.4 ± 2.3 years in our study, and the body weights were 69.8 ± 8.6 kg and 75.8 ± 10.2 kg vs. 63.9 ± 5.8 kg in our study. The main mechanism that could explain an increase in the . VO2 while running uphill is likely the higher muscular mass involved in this condition, especially for the vastus, soleus, gluteus, biceps femoris and gastrocnemius muscles [31,32,47–49]. In addition, uphill running over a 15% gradient eliminates the bouncing mechanism and the use of elastic energy, which is helpful for displacement [22,50,51]. The locomotion related to uphill running induces stronger muscular contractions and mechanical works [52]. To produce this additional mechanical energy, Robert and Belliveau concluded the involvement of a greater contribution of ankle, knee and hip extensor muscles in line with Sloniger et al. or Swanson and Caldwell’s work [31,32]. In addition, a greater hip extensors’ contribution has an additional effect on the metabolic expenditure, because theses muscles are not able to Int. J. Environ. Res. Public Health 2022, 19, 12210 10 of 12 produce force economically with a low contribution of elastic storage and recovery [53,54]. Gottschall and Kram reported an increase of 75% in the propulsive force during treadmill running at a 9% slope compared to a level condition [55]. Basically, more important propulsive work induces additional concentric muscle contraction levels, which are most costly physiologically [21]. Furthermore, even if our result indicates a positive influence of the slope on the . VO2 at the VT1, VT2 and MAX time points, we can also underline that the 40% gradient does not produce significantly greater . VO2 max values compared to the 25%, while the . VEmax is superior on this gradient. Based on this finding, we can exclude ventilatory limitation in this condition to produce a higher oxygen consumption. We can hypothesis a higher O2 extraction, transportation or utilization in muscles as limiting factors at steep grades [30]. In addition, this result provides an interesting insight to evaluate trail runner specialists. Indeed, based on this result, we can recommend the evaluation of the . VO2max capacity of trail runners on slope conditions at or over 25% to obtain their real maximal capacities, which has been highly correlated with performance levels for short and middle-distance trail-running competitions [56]. Finally, our study investigated the maximal AS for the VT1, VT2 and MAX time points and for different slope conditions (15, 25 and 40%). Our results provided similar findings for all the time points; the 15% gradient provided a systematically lower AS than 25% (p < 0.001) and 40% (p < 0.001), with a very large effect size. We also noted that the AS did not differ between 25 and 40%, whatever the time points considered (p = 0.999). This observation is very interesting for training purposes and may allow using ascending speeds as intensity indexes during sessions. Moreover, this information provides new evidence for athletes intending to compete in maximal elevation challenges, where the goal is to reach the highest elevation in an allocated duration (from 4 to 24 h). In our study, the slopes between 25 and 40% allowed the achievement of a greater ascending speed (approx. 1750 m per hour) than the 15% gradient (approx. 1175 m per hour) and could be selected for such contests based on the individual’s capabilities. 5. Conclusions The aim of this study was to investigate the effect of different gradient slopes (from level to +40%) on the cardiorespiratory variables reached at exhaustion and ventilatory thresholds during maximum incremental tests in specialist trail runners. First, our study clearly indicates that slope conditions over a 15% gradient allow reaching higher . VO2 and . VE levels at the VT1, VT2 and MAX time points, whereas the BF and HR remain unchanged for the specialist trail runners. Our study also pointed out a limitation of the slope influence on the physiological variables over the 25% gradient, whereas the . VO2 does not further increase at the VT1, VT2 and MAX time points. Finally, an analysis of the speed provides similar findings. Our study pointed out that the maximal AS can be reached for slopes of 25% and 40% equally. Based on these results, we recommend trail runner testing on slope conditions between 25 and 40% to stress and reach their real maximal cardiorespiratory capacities and obtain their maximal AS for training purposes. Author Contributions: Conceptualization, G.D. and J.C.; methodology, G.D., L.M. and J.C.; software, A.G.; investigation, G.D. and M.C.; writing—original draft preparation, J.C.; writing—review and editing, J.C., M.C., A.G., G.D. and L.M. supervision, G.D., J.C. and L.M.; project administration, L.M. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: The experiment was conducted in accordance with the Declaration of Helsinki and received the approval ID-RCB: 2019-A03012-55 from “COMITE DE PROTECTION DES PERSONNES SUD MEDITERRANEE IV”. Conflicts of Interest: The authors declare no conflict of interest. Int. J. Environ. Res. Public Health 2022, 19, 12210 11 of 12 References 1. Billat, L.V.; Koralsztein, J.P. Significance of the Velocity at VO2max and Time to Exhaustion at This Velocity. Sports Med. Auckl. NZ 1996, 22, 90–108. [CrossRef] [PubMed] 2. di Prampero, P.E.; Atchou, G.; Brückner, J.C.; Moia, C. The Energetics of Endurance Running. Eur. J. Appl. Physiol. 1986, 55, 259–266. [CrossRef] [PubMed] 3. Joyner, M.J. Physiological Limiting Factors and Distance Running: Influence of Gender and Age on Record Performances. Exerc. Sport Sci. Rev. 1993, 21, 103–133. [CrossRef] 4. Margaria, R.; Cerretelli, P.; Aghemo, P.; Sassi, G. Energy Cost of Running. J. Appl. Physiol. 1963, 18, 367–370. [CrossRef] 5. Jamnick, N.A.; Pettitt, R.W.; Granata, C.; Pyne, D.B.; Bishop, D.J. An Examination and Critique of Current Methods to Determine Exercise Intensity. Sports Med. 2020, 50, 1729–1756. [CrossRef] 6. Beltz, N.M.; Gibson, A.L.; Janot, J.M.; Kravitz, L.; Mermier, C.M.; Dalleck, L.C. Graded Exercise Testing Protocols for the Determination of VO2max: Historical Perspectives, Progress, and Future Considerations. J. Sports Med. Hindawi Publ. Corp. 2016, 2016, 3968393. [CrossRef] 7. Shephard, R.J. The Relative Merits of the Step Test, Bicycle Ergometer, and Treadmill in the Assessment of Cardio-Respiratory Fitness. Int. Z. Angew. Physiol. 1966, 23, 219–230. [CrossRef] [PubMed] 8. Volkov, N.I.; Shirkovets, E.A.; Borilkevich, V.E. Assessment of Aerobic and Anaerobic Capacity of Athletes in Treadmill Running Tests. Eur. J. Appl. Physiol. 1975, 34, 121–130. [CrossRef] 9. Borges, T.O.; Dascombe, B.; Bullock, N.; Coutts, A.J. Physiological Characteristics of Well-Trained Junior Sprint Kayak Athletes. Int. J. Sports Physiol. Perform. 2015, 10, 593–599. [CrossRef] 10. Butts, N.K.; Henry, B.A.; Mclean, D. Correlations between VO2max and Performance Times of Recreational Triathletes. J. Sports Med. Phys. Fitness 1991, 31, 339–344. 11. Baquet, G.; van Praagh, E.; Berthoin, S. Endurance Training and Aerobic Fitness in Young People. Sports Med. 2003, 33, 1127–1143. [CrossRef] 12. Midgley, A.W.; Mc Naughton, L.R. Time at or near VO2max during Continuous and Intermittent Running. A Review with Special Reference to Considerations for the Optimisation of Training Protocols to Elicit the Longest Time at or near VO2max. J. Sports Med. Phys. Fit. 2006, 46, 1–14. 13. Berthoin, S.; Gerbeaux, M.; Turpin, E.; Guerrin, F.; Lensel-Corbeil, G.; Vandendorpe, F. Comparison of Two Field Tests to Estimate Maximum Aerobic Speed. J. Sports Sci. 1994, 12, 355–362. [CrossRef] [PubMed] 14. Kuipers, H.; Rietjens, G.; Verstappen, F.; Schoenmakers, H.; Hofman, G. Effects of Stage Duration in Incremental Running Tests on Physiological Variables. Int. J. Sports Med. 2003, 24, 486–491. [CrossRef] [PubMed] 15. Meyer, T.; Welter, J.-P.; Scharhag, J.; Kindermann, W. Maximal Oxygen Uptake during Field Running Does Not Exceed That Measured during Treadmill Exercise. Eur. J. Appl. Physiol. 2003, 88, 387–389. [CrossRef] 16. Hoffman, M.D.; Ong, J.C.; Wang, G. Historical Analysis of Participation in 161 Km Ultramarathons in North America. Int. J. Hist. Sport 2010, 27, 1877–1891. [CrossRef] [PubMed] 17. Breiner, T.J.; Ortiz, A.L.R.; Kram, R. Level, Uphill and Downhill Running Economy Values Are Strongly Inter-Correlated. Eur. J. Appl. Physiol. 2019, 119, 257–264. [CrossRef] [PubMed] 18. Lemire, M.; Hureau, T.J.; Favret, F.; Geny, B.; Kouassi, B.Y.L.; Boukhari, M.; Lonsdorfer, E.; Remetter, R.; Dufour, S.P. Physiological Factors Determining Downhill vs Uphill Running Endurance Performance. J. Sci. Med. Sport 2021, 24, 85–91. [CrossRef] 19. Vernillo, G.; Giandolini, M.; Edwards, W.B.; Morin, J.-B.; Samozino, P.; Horvais, N.; Millet, G.Y. Biomechanics and Physiology of Uphill and Downhill Running. Sports Med. 2017, 47, 615–629. [CrossRef] 20. di Prampero, P.E.; Salvadego, D.; Fusi, S.; Grassi, B. A Simple Method for Assessing the Energy Cost of Running during Incremental Tests. J. Appl. Physiol. 2009, 107, 1068–1075. [CrossRef] 21. Minetti, A.E.; Ardigò, L.P.; Saibene, F. Mechanical Determinants of the Minimum Energy Cost of Gradient Running in Humans. J. Exp. Biol. 1994, 195, 211–225. [CrossRef] [PubMed] 22. Minetti, A.E.; Moia, C.; Roi, G.S.; Susta, D.; Ferretti, G. Energy Cost of Walking and Running at Extreme Uphill and Downhill Slopes. J. Appl. Physiol. 2002, 93, 1039–1046. [CrossRef] [PubMed] 23. Kasch, F.W.; Wallace, J.P.; Huhn, R.R.; Krogh, L.A.; Hurl, P.M. VO2max during Horizontal and Inclined Treadmill Running. J. Appl. Physiol. 1976, 40, 982–983. [CrossRef] [PubMed] 24. Balducci, P.; Clémençon, M.; Morel, B.; Quiniou, G.; Saboul, D.; Hautier, C.A. Comparison of Level and Graded Treadmill Tests to Evaluate Endurance Mountain Runners. J. Sports Sci. Med. 2016, 15, 239–246. [PubMed] 25. Paavolainen, L.; Nummela, A.; Rusko, H. Muscle Power Factors and VO2max as Determinants of Horizontal and Uphill Running Performance. Scand. J. Med. Sci. Sports 2000, 10, 286–291. [CrossRef] 26. Schöffl, I.; Jasinski, D.; Ehrlich, B.; Dittrich, S.; Schöffl, V. Outdoor Uphill Exercise Testing for Trail Runners, a More Suitable Method? J. Hum. Kinet. 2021, 79, 123–133. [CrossRef] 27. Pokan, R.; Schwaberger, G.; Hofmann, P.; Eber, B.; Toplak, H.; Gasser, R.; Fruhwald, F.M.; Pessenhofer, H.; Klein, W. Effects of Treadmill Exercise Protocol with Constant and Ascending Grade on Levelling-off O2 Uptake and VO2 Max. Int. J. Sports Med. 1995, 16, 238–242. [CrossRef] 28. Scheer, V.; Ramme, K.; Reinsberger, C.; Heitkamp, H.-C. VO2max Testing in Trail Runners: Is There a Specific Exercise Test Protocol? Int. J. Sports Med. 2018, 39, 456–461. [CrossRef] Int. J. Environ. Res. Public Health 2022, 19, 12210 12 of 12 29. Taylor, H.L.; Buskirk, E.; Henschel, A. Maximal Oxygen Intake as an Objective Measure of Cardio-Respiratory Performance. J. Appl. Physiol. 1955, 8, 73–80. [CrossRef] 30. di Prampero, P.E. An Analysis of the Factors Limiting Maximal Oxygen Consumption in Healthy Subjects. Chest 1992, 101, 188S–191S. [CrossRef] 31. Sloniger, M.A.; Cureton, K.J.; Prior, B.M.; Evans, E.M. Lower Extremity Muscle Activation during Horizontal and Uphill Running. J. Appl. Physiol. 1997, 83, 2073–2079. [CrossRef] [PubMed] 32. Swanson, S.C.; Caldwell, G.E. An Integrated Biomechanical Analysis of High Speed Incline and Level Treadmill Running. Med. Sci. Sports Exerc. 2000, 32, 1146–1155. [CrossRef] [PubMed] 33. Mendez-Villanueva, A.; Buchheit, M.; Kuitunen, S.; Poon, T.K.; Simpson, B.; Peltola, E. Is the Relationship between Sprinting and Maximal Aerobic Speeds in Young Soccer Players Affected by Maturation? Pediatr. Exerc. Sci. 2010, 22, 497–510. [CrossRef] [PubMed] 34. Midgley, A.W.; Bentley, D.J.; Luttikholt, H.; McNaughton, L.R.; Millet, G.P. Challenging a Dogma of Exercise Physiology: Does an Incremental Exercise Test for Valid VO 2 Max Determination Really Need to Last between 8 and 12 Minutes? Sports Med. 2008, 38, 441–447. [CrossRef] 35. Macfarlane, D.J.; Wong, P. Validity, Reliability and Stability of the Portable Cortex Metamax 3B Gas Analysis System. Eur. J. Appl. Physiol. 2012, 112, 2539–2547. [CrossRef] 36. Beaver, W.L.; Wasserman, K.; Whipp, B.J. A New Method for Detecting Anaerobic Threshold by Gas Exchange. J. Appl. Physiol. 1986, 60, 2020–2027. [CrossRef] 37. Pereira, D.G.; Afonso, A.; Medeiros, F.M. Overview of Friedman’s Test and Post-Hoc Analysis. Commun. Stat.-Simul. Comput. 2015, 44, 2636–2653. [CrossRef] 38. Taheri, S.M.; Hesamian, G. A Generalization of the Wilcoxon Signed-Rank Test and Its Applications. Stat. Pap. 2013, 54, 457–470. [CrossRef] 39. Goulet-Pelletier, J.-C.; Cousineau, D. A Review of Effect Sizes and Their Confidence Intervals, Part I: The Cohen’s d Family. Quant. Methods Psychol. 2018, 14, 242–265. [CrossRef] 40. Hopkins, W.G.; Marshall, S.W.; Batterham, A.M.; Hanin, J. Progressive Statistics for Studies in Sports Medicine and Exercise Science. Med. Sci. Sports Exerc. 2009, 41, 3–13. [CrossRef] 41. O’Halloran, J.; Hamill, J.; McDermott, W.J.; Remelius, J.G.; Van Emmerik, R.E.A. Locomotor-Respiratory Coupling Patterns and Oxygen Consumption during Walking above and below Preferred Stride Frequency. Eur. J. Appl. Physiol. 2012, 112, 929–940. [CrossRef] [PubMed] 42. Takano, N. Phase Relation and Breathing Pattern during Locomotor/Respiratory Coupling in Uphill and Downhill Running. Jpn. J. Physiol. 1995, 45, 47–58. [CrossRef] 43. Tiller, N.B.; Price, M.J.; Campbell, I.G.; Romer, L.M. Effect of Cadence on Locomotor–Respiratory Coupling during Upper-Body Exercise. Eur. J. Appl. Physiol. 2017, 117, 279–287. [CrossRef] [PubMed] 44. Giovanelli, N.; Ortiz, A.L.R.; Henninger, K.; Kram, R. Energetics of Vertical Kilometer Foot Races; Is Steeper Cheaper? J. Appl. Physiol. 2016, 120, 370–375. [CrossRef] 45. Lemire, M.; Falbriard, M.; Aminian, K.; Millet, G.P.; Meyer, F. Level, Uphill, and Downhill Running Economy Values Are Correlated Except on Steep Slopes. Front. Physiol. 2021, 12, 697315. [CrossRef] [PubMed] 46. Pringle, J.S.; Carter, H.; Doust, J.H.; Jones, A.M. Oxygen Uptake Kinetics during Horizontal and Uphill Treadmill Running in Humans. Eur. J. Appl. Physiol. 2002, 88, 163–169. [CrossRef] [PubMed] 47. Sloniger, M.A.; Cureton, K.J.; Prior, B.M.; Evans, E.M. Anaerobic Capacity and Muscle Activation during Horizontal and Uphill Running. J. Appl. Physiol. 1997, 83, 262–269. [CrossRef] 48. Wall-Scheffler, C.M.; Chumanov, E.; Steudel-Numbers, K.; Heiderscheit, B. Electromyography Activity across Gait and Incline: The Impact of Muscular Activity on Human Morphology. Am. J. Phys. Anthropol. 2010, 143, 601–611. [CrossRef] 49. Yokozawa, T.; Fujii, N.; Ae, M. Muscle Activities of the Lower Limb during Level and Uphill Running. J. Biomech. 2007, 40, 3467–3475. [CrossRef] 50. Dewolf, A.H.; Peñailillo, L.E.; Willems, P.A. The Rebound of the Body during Uphill and Downhill Running at Different Speeds. J. Exp. Biol. 2016, 219, 2276–2288. [CrossRef] 51. Snyder, K.L.; Kram, R.; Gottschall, J.S. The Role of Elastic Energy Storage and Recovery in Downhill and Uphill Running. J. Exp. Biol. 2012, 215, 2283–2287. [CrossRef] [PubMed] 52. Roberts, T.J.; Belliveau, R.A. Sources of Mechanical Power for Uphill Running in Humans. J. Exp. Biol. 2005, 208, 1963–1970. [CrossRef] [PubMed] 53. Biewener, A.A.; Roberts, T.J. Muscle and Tendon Contributions to Force, Work, and Elastic Energy Savings: A Comparative Perspective. Exerc. Sport Sci. Rev. 2000, 28, 99–107. 54. Cavagna, G.A.; Saibene, F.P.; Margaria, R. Mechanical Work in Running. J. Appl. Physiol. 1964, 19, 249–256. [CrossRef] [PubMed] 55. Gottschall, J.S.; Kram, R. Ground Reaction Forces during Downhill and Uphill Running. J. Biomech. 2005, 38, 445–452. [CrossRef] 56. Pastor, F.S.; Besson, T.; Varesco, G.; Parent, A.; Fanget, M.; Koral, J.; Foschia, C.; Rupp, T.; Rimaud, D.; Féasson, L.; et al. Performance Determinants in Trail-Running Races of Different Distances. Int. J. Sports Physiol. Perform. 2022, 17, 844–851. [CrossRef]

Physiological Implication of Slope Gradient during Incremental Running Test.

09-26-2022

Cassirame, Johan,Godin, Antoine,Chamoux, Maxime,Doucende, Gregory,Mourot, Laurent

eng

PMC4326429

RESEARCH Open Access Contributions of lower extremity kinematics to trunk accelerations during moderate treadmill running Timothy R Lindsay1*, James A Yaggie2 and Stephen J McGregor1 Abstract Background: Trunk accelerations during running provide useful information about movement economy and injury risk. However, there is a lack of data regarding the key biomechanical contributors to these accelerations. The purpose was to establish the biomechanical variables associated with root mean square (RMS) accelerations of the trunk. Methods: Eighteen healthy males (24.0 ± 4.2 yr; 1.78 ± 0.07 m; 79.7 ± 14.8 kg) performed treadmill running with high resolution accelerometer measurement at the lumbar spine and full-body optical motion capture. We collected 60 sec of data at three speeds (2.22, 2.78, 3.33 m∙s−1). RMS was calculated for medio-lateral (ML), anterio-posterior (AP), vertical (VT), and the resultant Euclidean scalar (RES) acceleration. From motion capture, we calculated 14 kinematic variables, including mean sagittal plane joint angles at foot contact, mid-stance, and toe-off. Principal components analysis (PCA) was used to form independent components comprised of combinations of the original variables. Stepwise regressions were performed on the original variables and the components to determine contributions to RMS acceleration in each axis. Results: Significant speed effects were found for RMS-accelerations in all axes (p < 0.05). Regressions of the original variables indicated from 4 to 5 variables associated with accelerations in each axis (R2 = 0.71 to 0.82, p < 0.001). The most prominent contributing variables were associated with the late flight and early stance phase. PCA reduced the data into four components. Component 1 included all hip angles before mid-stance and component 2 was primarily associated with propulsion. Regressions indicated key contributions from components 1 and 2 to ML, VT, and RES acceleration (p < 0.05). Conclusions: The variables with highest contribution were prior to mid-stance and mechanically relate to shock absorption and attenuation of peak forces. Trunk acceleration magnitude is associated with global running variables, ranging from energy expenditure to forces lending to the mechanics of injury. These data begin to delineate running gait events and offer relationships of running mechanics to those structures more proximal in the kinetic chain. These relationships may provide insight for technique modification to maximize running economy or prevent injury. Keywords: High resolution accelerometers, Root mean square, Principal components analysis, Running, Economy, Injury, Stiffness * Correspondence: [email protected] 1School of Health Promotion & Human Performance, Eastern Michigan University, Ypsilanti, MI, USA Full list of author information is available at the end of the article JNER JOURNAL OF NEUROENGINEERING AND REHABILITATION © 2014 Lindsay et al.; licensee BioMed Central. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. Lindsay et al. Journal of NeuroEngineering and Rehabilitation 2014, 11:162 http://www.jneuroengrehab.com/content/11/1/162 Background Running is an increasingly popular sport that provides substantial health benefits at minimal expense. Estimates from 2011 are that 38.7 million Americans participate in running or jogging 6 or more days per year (up from 24.5 million in 2001), with 9.2 million doing so 110 or more days per year (up from 6.8 million in 2001) [1]. Offsetting the numerous health benefits of exercise is the relatively high incidence of injury, which according to one system- atic review, ranges from 19-79% [2]. Even at the lower end of this range, the high participation rate means that injury is a substantial concern. Since most running injuries are chronic rather than acute [3,4], the tolerable level of accu- mulated stress is an important consideration. This stress depends on multiple factors including the training dose, anatomical structure, and movement mechanics [4-7]. We focus on mechanics in this paper. Mechanically speaking, running involves the application of force to the ground to generate the resultant ground re- action force (GRF) necessary for forward propulsion and support against gravity. This places stress on soft tissue and bone via force transmission through the kinetic chain, which may lead to future injury if the exercise dosage ex- ceeds regenerative capacity. A comprehensive description of forces requires a complicated model, but the acceler- ation of the center of mass (COM) can provide a simple quantification of net force. Continuous COM data may then be expressed as a root mean square (RMS) value to represent the overall magnitude of acceleration over many strides [8]. RMS provides a measure of dispersion similar to standard deviation, only relative to zero rather than the mean [9]. The presence of more extreme values in the sig- nal (i.e., high acceleration or deceleration) increases the RMS value. Acceleration at any anatomical location de- pends on the level of attenuation through tissue de- formation and joint excursion at all points distal. The attenuation of force and acceleration can be modified with lower limb stiffness and may alter the likelihood of running-related injuries [10,11]. High stiffness may aid performance and economy but also may increase the risk of injury to structural components. In contrast, stiffness that is too low may be metabolically costly and increase the risk of soft tissue injury [8,10-12]. Stiffness depends on the intrinsic properties of bone and soft tis- sue (muscle, tendon, ligament, and cartilage) [13], but also may be modified via kinematic changes. For ex- ample, in subjects instructed to perform a soft drop landing, there was greater knee joint excursion [14,15]. As well, Derrick [11] has argued that runners generally run with extended knees prior to impact, but are able to increase knee flexion in order to reduce vertical ac- celerations. Similarly, subjects who were instructed to adopt a “Groucho running” style had longer strides (believed to be associated with decreased stiffness) and decreased stiffness, as directly measured [16]. Inter- ventions such as gait retraining to pursue this objective are promising and demonstrate that kinematics are modifiable [17,18]. There has not yet been a direct investigation into the relationship between running mechanics and RMS accel- eration. The measurement of acceleration requires little equipment, can be done in the field, and real-time feed- back is possible. Since the major movements of running are in the sagittal plane, we focused on the flexion/ex- tension behavior of the hip, knee, and ankle joints during various gait events, as well as some other key variables that are readily modifiable. The purpose of this study was to determine the biomechanical factors contributing to global axial accelerations in active healthy males. In previ- ous work [8], we observed greater accelerations in healthy untrained runners compared to trained collegiate runners. In the current study we selected a sample that was rela- tively heterogeneous with regard to chosen mode of phys- ical activity and indicative of those from the general population who might take up running as a recreational activity for health benefits. These individuals would be more likely to exhibit mechanics that would make them more susceptible to injury due to relatively high accelera- tions [8]. To accomplish our objectives, we used a multiple regression approach to determine the variables that best fit a least squares model generated for RMS acceleration in each axis. Additionally, principal components analysis (PCA) was used to establish potentially hidden interac- tions between individual variables that can be combined to form separate components. These components may be then assessed for their contribution to axial accelerations. Thus, with a view to performance and injury management, this study will provide a description of modifiable bio- mechanical factors and their relationship with RMS trunk accelerations. Methods Subjects and experimental procedure Eighteen healthy, active, college-age males volunteered to participate. Subjects participated 2–7 times per week in various forms of physical activity such as individual endur- ance sports (including running for 6 subjects), strength training, team sports, and/or combat sports. The proce- dures of this study were approved by the Human Subjects Review Committee of Eastern Michigan University College of Health and Human Services. All subjects provided writ- ten informed consent. We analyzed 60 sec of data from three randomly-ordered treadmill run trials run at 2.22, 2.78, and 3.33 m∙s−1. Sub- jects were given as much rest between trials as they desired (typically 60–180 s). Subject characteristics are presented in Table 1. Lindsay et al. Journal of NeuroEngineering and Rehabilitation 2014, 11:162 Page 2 of 8 http://www.jneuroengrehab.com/content/11/1/162 Instrumentation We placed one triaxial high resolution accelerometer (G-Link ADXL 210, Microstrain, Inc., Williston, VT) on the dorsal mid-line, at the level of the iliac crest (ap- proximately at the L4/L5 spinous process). Accelerome- ters mounted at this anatomical location can provide valid estimates of oxygen consumption during running and can distinguish mechanics between trained and un- trained individuals [8]. Although the legs primarily move in the sagittal plane during running, this is not the case for the spine and pelvis. Because the accelerometer is mounted in that region, there is significant non-sagittal movement requiring measurement in three and not just two dimensions. The accelerometer (mass = 47 g) con- sists of internal circuitry enclosed in a 58 × 43 × 21 mm casing, plus an antenna extending a bit outside the di- mensions and adding 18 mm to the thickness. The ac- celerometer was mounted to a semi-rigid strap, and secured with elastic wrap to minimize extraneous move- ment of the device. Kinematic data was collected with a 3-D optical mo- tion capture system (Vicon MX, Vicon, Centennial, CO). We employed a 39-marker full body gait model (Plug- In-Gait, Vicon, Los Angeles, CA) consisting of 15 seg- ments including the head, thorax, pelvis, upper arm, forearm, hand, thigh, shank, and foot. Seven cameras (Vicon T40 and T40 S) were placed roughly equidistant to the subject on the treadmill. Mean values for fourteen kinematic variables were calculated (mean value for left and right leg). Foot contact was defined as the point of lowest vertical displacement of the heel marker [19]. Mid-stance was defined as the lowest point of the software-estimated COM. Toe-off was defined as the point of maximum knee extension [19]. Lower limb joint angles were calculated according to the parameters of the software and model. Variables are listed and defined in Table 2. Data capture and analysis Data were collected in the medio-lateral (ML), anterio- posterior (AP), and vertical (VT) axes. Trajectories were sampled at 200 Hz and then filtered with a 4th order Butterworth filter with a low pass cutoff at 10 Hz. Accelerometer data were streamed wirelessly at 617 Hz to Agilelink software (Microstrain, Williston, VT), subsequently re-sampled at 200 Hz, and filtered similarly to correspond with motion capture data. During running, the device is not perfectly aligned relative to the room (i.e., the global coordinate system, as opposed to the body coordinate system). Corrections were made for the tilt of the accelerometer, based on the method of Moe-Nilssen [9]. We provide a brief description of the calculations, but we encourage the reader to study the details provided in that paper [9]. Correction is possible because the mean vector angles in the ML and AP sensing axes may be estimated while the participant is running (see Appendix for calculations). The RMS of the vertical (VTRMS), medio-lateral (MLRMS), and anterior-posterior (APRMS) axes was then calculated for the epochs in each trial: xRMS ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1 N X N i xi2 v u u t ð1Þ where x is the given plane and N is the total number of samples in 60 sec (at 200 Hz, N = 12,000). The result- ant Euclidian scalar variable (RES) was also calculated Table 1 Subject characteristics Mean SD Min Max Age (yr) 24.0 4.2 19 32 Height (m) 1.78 0.07 1.66 1.89 Mass (kg) 79.7 14.8 59.1 107.3 BMI (kg∙m−2) 25.2 3.6 20.8 31.7 Table 2 Biomechanical variable definitions Abbreviation Explanation Measurement convention Hip-max Maximum hip angle (before foot-strike). Positive = flexion Hip-FS Hip angle at foot-strike. Hip-MS Hip angle at mid-stance. Hip-TO Hip angle at toe-off. Knee-FS Knee angle at foot-strike. Positive = flexion Knee-MS Knee angle at mid-stance. Knee-TO Knee angle at toe-off. Ankle-FS Ankle angle at foot-strike. Positive = dorsiflexion Ankle-TO Ankle angle at toe-off. PR Mean range of pelvis rotation in the transverse plane for each gait cycle. Scalar FA Foot advance; sagittal plane distance between the heel and COM at foot contact, relative to mean leg length. Scalar DROP Vertical displacement of COM from foot contact to mid-stance, relative to mean leg length. Scalar RISE Vertical displacement of COM from mid-stance to toe-off, relative to mean leg length. Scalar SR Step rate. Steps per min Lindsay et al. Journal of NeuroEngineering and Rehabilitation 2014, 11:162 Page 3 of 8 http://www.jneuroengrehab.com/content/11/1/162 for the determination of the magnitude of the overall body acceleration: RESRMS ¼ ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi VT RMS2 þ MLRMS2 þ APRMS2 p ð2Þ The above processing and analysis of data was done using custom designed code in a Matlab environment (Matlab R2013b, Mathworks, Natick, MA). Statistical tests Correlations were first performed to assess the relation- ship between anthropometric variables and acceleration. Analysis of variance (ANOVA) was used to determine the effect of speed on the four acceleration and fourteen bio- mechanical variables. A stepwise regression was then used to determine the significant kinematic contributions to ac- celeration in each axis. We also performed principal com- ponent analysis (PCA) to reduce the dimensionality of the data into significant components using a varimax rotation and Kaiser normalization. A stepwise regression was then performed using these components as predictors of acceler- ation in each axis. Post hoc power analyses were conducted for all ANOVA and regression analyses. A Bonferroni test was used for multiple comparisons, where appropriate. Statistical significance was set at p < 0.05. Statistical analysis was done using SPSS software (version 21, IBM Corpor- ation, Armonk, NY). Results Significant speed effects were found for RMS-accelerations for ML, AP, and RES (p < 0.05, Table 3). Of the biomechan- ical variables, only maximum hip angle showed a significant speed effect (p < 0.05, Table 3). Height, mass, and BMI were not significantly correlated with acceleration in any axis (p < 0.05). Regression indicated 4 to 5 significant variables associ- ated with acceleration, depending on the axis (Table 4). We encourage the reader to take notice of the sign of the beta coefficients (Table 4) and the angle definitions (Table 2) to understand the direction of change that is associated with an increase in acceleration. The combin- ation of significant variables was different for each axis. Explained variance (R2) ranged from 0.71 to 0.82. A plot of predicted versus measured RMS acceleration for each axis is provided in Figure 1. PCA indicated 4 significant kinematic components (Table 5), explaining 79.1% of total variance. Component 1 (λ = 4.9, 37.4% of variance) was comprised of variables predominantly associated with hip flexion in late flight and early stance phase (hip-MS, hip-FS, knee-MS, hip- max). Component 2 (λ = 2.8, 21.2% of variance) was as- sociated with the propulsive phase of the gait cycle (ankle-TO, knee-TO, RISE, PR). Component 3 (λ = 1.6, 12.5% of variance) included variables associated with cushioning during the early stance phase (knee-FS, DROP, ankle-FS). Regressions (Table 6) indicated that components 1 and 2 significantly predicted ML, VT, and RES acceleration (R2 from 0.32 to 0.40, p < 0.001). Com- ponent 3 significantly predicted AP acceleration (R2 = 0.041, p = 0.041). Discussion The purpose of this study was to determine the bio- mechanical contributors to global axial RMS accelera- tions during running. We found significant relationships where explained variance using regressions on the ori- ginal variables was 0.71 for ML, 0.53 for AP, 0.74 for VT, and 0.43 for RES. PCA did identify hidden relationships that explained 79% of the variance of the original vari- ables and that were not evident using only multiple re- gression. When regressions were performed using the PCA component variables, though, explained variance was lower than with the original biomechanical variables alone. Reducing the numerous variables into a few prin- cipal components therefore does explain much of the variance in a simplified manner, but the predictive value of this simplified relationship is not as strong as using a traditional regression with a non-reduced variable set. Table 3 Mean (SD) acceleration and biomechanical variables for each speed Variable Speed (m/s) Observed power 2.22 2.78 3.33 MLRMS (g)* 0.35 (0.05) 0.41 (0.06)† 0.46 (0.07)† 1.00 APRMS (g)* 0.36 (0.06) 0.43 (0.10)† 0.50 (0.10)† 0.99 VTRMS (g) 1.09 (0.13) 1.18 (0.11) 1.19 (0.10) 0.72 RESRMS (g)* 1.21 (0.12) 1.33 (0.12)† 1.38 (0.11)† 0.99 Hip-max (deg)* 36.2 (6.8) 42.1 (7.2)† 48.3 (7.4)† 1.00 Hip-FS (deg) 28.9 (6.5) 31.5 (5.8) 34.6 (6.2) 0.68 Hip-MS (deg) 22.6 (7.6) 24.9 (6.9) 27.6 (7.4) 0.41 Hip-TO (deg) −5.4 (5.6) −8.4 (5.6) −10.5 (5.6) 0.68 Knee-FS (deg) 13.0 (8.2) 12.3 (6.6) 13.7 (6.6) 0.08 Knee-MS (deg) 37.9 (7.1) 39.3 (6.8) 40.6 (6.7) 0.16 Knee-TO (deg) 10.5 (6.5) 8.6 (5.8) 8.0 (6.1) 0.19 Ankle-FS (deg) 9.4 (5.0) 8.9 (4.8) 9.2 (4.8) 0.06 Ankle-TO (deg) −11.5 (7.5) −16.0 (5.4) −17.6 (5.7) 0.76 PR (deg) 3.9 (1.6) 4.7 (2.5) 5.6 (3.7) 0.34 FA (% mean leg length) 5.3 (2.9) 6.9 (3.0) 9.2 (3.8) 0.90 DROP (% mean leg length) 6.2 (1.7) 6.3 (1.5) 5.9 (1.4) 0.10 RISE (% mean leg length) 8.6 (1.7) 8.9 (1.5) 8.7 (1.5) 0.09 SR (steps/min) 155.5 (9.5) 158.0 (9.0) 163.1 (10.5) 0.54 *Significant speed effect at p < 0.05, †significantly different from 2.22 m∙s−1. Lindsay et al. Journal of NeuroEngineering and Rehabilitation 2014, 11:162 Page 4 of 8 http://www.jneuroengrehab.com/content/11/1/162 Accelerations measured at the lumbar spine originate from the GRF, which is transmitted through the foot, shank, thigh, and pelvis. GRF at the shank is typically bi- phasic and is significantly attenuated at proximal body segments [20,21]. The two GRF peaks are associated with impact and propulsion [22,23], with resultant body segment acceleration depending on GRF magnitude and damping effects [24]. The magnitude of force applied to the ground depends, in part, on the stiffness of the lower extremities, as does the acceleration resulting from the GRF. According to the regressions, increased RMS accelera- tions were associated with different combinations of the following kinematic characteristics during early stance phase, depending on the axis: increased hip flexion, de- creased knee flexion, and decreased foot advance. Most studies demonstrate that a combination of increased hip flexion and decreased ankle dorsiflexion at foot contact is associated with alterations in GRF during foot contact in various settings, providing evidence for the role of both the quadriceps and the ankle dorsiflexor muscles in shock absorption [25-30]. Our data did not demonstrate the importance of ankle dynamics in shock attenuation, but did highlight the role of hip angle in positioning the quadriceps for shock attenuation during running [21,31]. Indeed, decreased FA was associated with increased AP, VT, and RES acceleration. In contrast, a greater FA leads to a flatter angle of attack (i.e., angle between segment and the ground), which results in high lengthening rates and decreased rate and magnitude of loading [32]. During the late stance and propulsion phase, a greater hip extension and ankle plantar flexion at toe off were as- sociated with greater RMS acceleration. According to modeling by Hamner et al. [30], the soleus and gastrocne- mius provide the biggest contributions to the propulsion phase. Data from the present study supports the import- ant role of the ankle plantar flexors in propulsion. Kinematic observations in the current study are similar to changes observed by McMahon et al. [16] when per- forming a “Groucho running” intervention. In that study, reductions in leg stiffness were associated with reduced GRF and increased metabolic cost and are accomplished by increased knee flexion. In contrast, in the present study this appears to be facilitated by increased hip joint excur- sion and a decreased foot advance. We note that Groucho running is an exaggerated style for the purpose of estab- lishing a relationship, and not intended for exercise and performance purposes. The subjects in the present study used a freely-chosen technique and were not given any in- struction to modify their form. Still, the kinematic descrip- tions we provide would seem to be subject to modification with skill training [17,18]. There is also evidence that the level of acceleration may be modified with training. We have previously shown that the vertical accelerations of trained collegiate runners are lower than untrained individuals but greater than triathletes with similar fitness and training volume [33]. This may represent an optimization of the different performance requirements and injury risk between the different groups because the optimal magnitude of verti- cal accelerations for performance may be different than what is optimal for minimizing risk of injury, and both may be different from sport to sport. Acceleration mag- nitude and stiffness may reflect several aspects of phys- ical function during running such as energy expenditure, impact forces relating to stress and injury, and perform- ance [10,11]. Often, one aspect must be compromised if another is to be maximized. For example, high impact forces accompanying high limb stiffness may increase energy return and performance according to the spring- mass model but may require more energy and occur at the expense of an overuse injury [32]. In the current study, we employ a simple approach to modeling, including kinematic descriptions and a single acceleration quantity for each axis (representing accelera- tions over the entire gait cycle); this work represents an easily accessible method with the potential for real-time output. Although it is not possible to fully account for the myriad of interactions between force, acceleration, stiff- ness, effective segmental mass, performance, and injury, Table 4 Regression results for original variables Dependent Adjusted R2 Independent Beta p Observed power MLRMS 0.714a Hip-max 0.009 < 0.001 1.00 Hip-TO −0.005 < 0.001 Knee-FS −0.003 0.002 Knee-MS −0.003 0.020 APRMS 0.718b Hip-max 0.016 < 0.001 1.00 Hip-MS −0.010 < 0.001 Hip-TO −0.008 < 0.001 FA −0.010 < 0.001 RISE −0.020 0.001 VTRMS 0.795c Hip-FS 0.034 < 0.001 1.00 FA −0.027 < 0.001 Hip-MS −0.030 < 0.001 Hip-max 0.010 < 0.001 Ankle-TO −0.014 < 0.001 RESRMS 0.822d Hip-Max 0.017 < 0.001 1.00 FA −0.028 < 0.001 Hip-FS 0.034 < 0.001 Hip-MS −0.036 < 0.001 Ankle-TO −0.014 < 0.001 aF(4,51) = 35.362, p < 0.001; bF(5,50) = 28.960, p < 0.001; cF(5,50) = 45.542, p < 0.001. dRES: F(5,50) = 51.967, p < 0.001. Lindsay et al. Journal of NeuroEngineering and Rehabilitation 2014, 11:162 Page 5 of 8 http://www.jneuroengrehab.com/content/11/1/162 there have been several reports of the benefits of interven- tions using acceleration as an outcome variable [17,18,34]. Data reduction via PCA facilitates the tracking of such characteristics because the number of features becomes relatively smaller [35]. Indeed, satisfactory descriptions of walking gait using PCA (~80-90% explained variance) Table 5 Rotated component matrix from principal component analysis Component 1 2 3 4 Hip-MS 0.910* −0.058 0.081 −0.127 Hip-FS 0.906* 0.022 −0.016 0.134 Knee-MS 0.897* 0.034 0.239 0.213 Hip-max 0.852* −0.146 0.157 0.105 Ankle-TO −0.114 0.888* −0.033 −0.295 Knee-TO 0.345 0.812* 0.076 0.069 RISE 0.599 −0.658* −0.063 −0.009 PR 0.172 −0.517* −0.054 −0.122 Knee-FS 0.342 −0.004 0.902* −0.074 DROP 0.600 −0.006 −0.667* 0.243 Ankle-FS 0.505 0.305 0.514* 0.117 FA 0.352 0.062 −0.172 0.810* Hip-TO 0.401 0.588 −0.112 −0.591* Bold font and *indicates grouping for each component. Figure 1 Predicted RMS acceleration versus measured RMS acceleration values. Graphs indicate: (a) ML, b) AP, c) VT, and d) RES. Each graph includes data from all three speeds. Most biomechanical variables did not show a speed effect. Table 6 Regression results for principal components Dependent Adjusted R2 Predictors Beta p Observed power MLRMS 0.322a Component 1 0.418 < 0.001 1.00 Component 2 −0.415 < 0.001 APRMS 0.058b Component 3 0.273 0.041 1.00 VTRMS 0.401c Component 2 −0.529 < 0.001 1.00 Component 1 0.378 0.001 RESRMS 0.380d Component 2 −0.513 < 0.001 1.00 Component 1 0.374 0.001 aF(2,53) = 14.066, p < 0.001; bF(1,54) = 4.364, p = 0.041; cF(2,53) = 19.412, p < 0.001. dF(2,53) = 17.875, p < 0.001. Lindsay et al. Journal of NeuroEngineering and Rehabilitation 2014, 11:162 Page 6 of 8 http://www.jneuroengrehab.com/content/11/1/162 applied to continuous waveforms of joint markers or joint ankles have been obtained with only the first three or four principal components [35-39]. However, sometimes only one of many principal components is significantly differ- ent between subject groups (fallers vs. non-fallers, over- weight vs. normal weight) or experimental conditions (loaded vs. non-loaded) in walking tasks [40,41]. The re- duction of the kinematic variables into four principal components may aid conceptualization of the key gait characteristics that contribute to the magnitude of ac- celerations. That it was possible to form components from the different biomechanical variables is likely indi- cative of movement synergies employed by the individ- ual as a motor strategy [37]. To the extent that this strategy can be altered, this presents an opportunity to modify force production and impact absorption. The discrete values used in the present study represent an a priori reduction from continuous waveform data, and may be seen as a limitation, but the maximum and minimum values found in a waveform can often be the regions of most significant difference [40], and would thus likely be captured at points during the gait cycle that we examined. As well, the complete dataset of bio- mechanical variables displayed greater explained vari- ance than the principal components. This may indicate that the reduction of the complete dataset results in the loss of important information that is explanatory with regard to gait dynamics. However, this does not neces- sarily diminish the value in identifying otherwise hidden synergistic relationships perhaps indicative of a neuro- muscular strategy. Another limitation is the small num- ber of biomechanical variables chosen for analysis. While the selection of a few readily modifiable variables provides a simple preliminary analysis, there are other variables that have not been included that potentially affect RMS trunk acceleration. Indeed, our measure- ments focused on movement in the sagittal plane, but this neglects frontal plane dynamics that may influence medio-lateral acceleration. Because the accelerometer only approximates COM movement, the findings are limited if an explanation of COM per se is desired. However, if the goal is to investigate what contributes to measured accel- erations, and explain previous findings (c.f. McGregor [8]) then the factors highlighted in this paper provide a basis for future investigations. Conclusions This study helps to establish the use of lumbar- mounted accelerometers to demonstrate effects related to stiffness, impact, and the attenuation of acceleration. Previous work has demonstrated the connection be- tween RMS accelerations and energy expenditure [8]. Our present data provides a more mechanistic explan- ation of how various kinematic configurations may influence the multi-segmental force cascade from the foot-ground interface to the lumbar vertebrae where ac- celerations are measured. Specifically, we have identi- fied the role of hip and knee angles in shock absorption and the role of the hip and ankle in propulsion. In addition to establishing these key biomechanical con- tributors to acceleration, we showed how many of these variables change in concert. Wherever these variables are modifiable, the acceleration signal may be a useful way to monitor movement with a view to performance and injury management. Our findings pertain to young, healthy, and active men and women, but the relation- ships found here can form the basis from which more specific subject groups may be studied in the future. Appendix Statistical tests The following steps and equations are based on Moe- Nilssen [9]. For a dataset of large N, the acceleration vector approaches the sine of the angle of that vector: lim
aML ¼ sinθML ð3Þ lim
aAP ¼ sinθAP ð4Þ The coordinate system definitions must be strictly maintained so that the positive/negative signs are correct and the relationships hold true. The following correc- tions were applied to each of the axes: aAPcorr ¼ aAPmeas⋅ cosθAP−aVTmeas⋅ sinθAP ð5Þ aVTprov ¼ aAPmeas⋅ sinθAP þ aVTmeas⋅ cosθAP ð6Þ aMLcorr ¼ aMLmeas⋅ cosθML−aVTprov⋅ sinθML ð7Þ aVTcorr ¼ aMLmeas⋅ sinθML þ aVTprov⋅ cosθML þ 1 ð8Þ where corr, meas, and prov refer to corrected, measured, and provisional terms, respectively. The static compo- nent of gravity was also corrected for the VT axis, leav- ing only the dynamic acceleration component. Competing interests The authors declare that they have no competing interests. Authors’ contributions TL participated in study design, carried out data collection, performed data processing, and drafted the manuscript. JY participated in interpretation of data and drafting the manuscript. SM conceived of the study, performed statistical analysis, participated in interpretation of data and drafting of the manuscript. All authors read and approved the final manuscript. Acknowledgements The authors are thankful to the subjects and for data collection assistance from Aaron Stickel, Lucas Wall, Zach Maino, Ken Hayes, and Ann Brennan. Author details 1School of Health Promotion & Human Performance, Eastern Michigan University, Ypsilanti, MI, USA. 2School of Health Sciences & Human Performance, Ithaca College, Ithaca, NY, USA. Lindsay et al. Journal of NeuroEngineering and Rehabilitation 2014, 11:162 Page 7 of 8 http://www.jneuroengrehab.com/content/11/1/162 Received: 2 May 2013 Accepted: 20 November 2014 Published: 12 December 2014 References 1. Sport Business Research Network. [http://www.sbrnet.com/research.asp? subRID=512] 2. Van Gent RN, Siem D, van Middelkoop M, van Os AG, Bierma-Zeinstra SMA, Koes BW: Incidence and determinants of lower extremity running injuries in long distance runners: a systematic review. Br J Sports Med 2007, 41:469–480. 3. Ghani Zadeh Hesar N, Van Ginckel A, Cools A, Peersman W, Roosen P, De Clercq D, Witvrouw E: A prospective study on gait-related intrinsic risk factors for lower leg overuse injuries. Br J Sports Med 2009, 43:1057–1061. 4. Hreljac A: Etiology, prevention, and early intervention of overuse injuries in runners: a biomechanical perspective. Phys Med Rehabil Clin N Am 2005, 16:651–667. 5. Messier SP, Legault C, Schoenlank CR, Newman JJ, Martin DF, DeVita P: Risk factors and mechanisms of knee injury in runners. Med Sci Sports Exerc 2008, 40:1873–1879. 6. Niemuth PE, Johnson RJ, Myers MJ, Thieman TJ: Hip muscle weakness and overuse injuries in recreational runners. Clin J Sport Med 2005, 15:14–21. 7. Almeida SA, Williams KM, Shaffer RA, Brodine SK, Sandra A, Maxwell K, Richard A, Stephanie K: Epidemiological patterns of musculoskeletal injuries and physical training. Med Sci Sports Exerc 1999, 31:1176–1182. 8. McGregor SJ, Busa MA, Yaggie JA, Bollt EM: High resolution MEMS accelerometers to estimate VO2 and compare running mechanics between highly trained inter-collegiate and untrained runners. PLoS ONE 2009, 4:e7355. 9. Moe-Nilssen R: A new method for evaluating motor control in gait under real-life environmental conditions. Part 1: the instrument. Clin Biomech 1998, 13:320–327. 10. Butler RJ, Crowell HP, Davis IM: Lower extremity stiffness: implications for performance and injury. Clin Biomech 2003, 18:511–517. 11. Derrick TR: The effects of knee contact angle on impact forces and accelerations. Med Sci Sports Exerc 2004, 36:832–837. 12. Busa MA, Muth T, Herman C, Hornyak J, Bollt EM, Mcgregor SJ: High resolution accelerometry can identify changes in running dynamics across an inter-collegiate cross country season [abstract]. Med Sci Sports Exerc 2010, 42(Suppl 5):91. 13. Latash ML, Zatsiorsky VM: Joint stiffness: myth or reality? Hum Mov Sci 1993, 12:653–692. 14. Zhang S, Bates BT, Dufek JS: Contributions of lower extremity joints to energy dissipation during landings. Med Sci Sports Exerc 2000, 32:812–819. 15. Devita P, Skelly WA: Effect of landing stiffness on joint kinetics and energetics in the lower extremity. Med Sci Sports Exerc 1992, 24:108–115. 16. McMahon TA, Valiant G, Frederick EC: Groucho running. J Appl Physiol 1987, 62:2326–2337. 17. Crowell HP, Davis IS: Gait retraining to reduce lower extremity loading in runners. Clin Biomech 2011, 26:78–83. 18. Crowell HP, Milner CE, Hamill J, Davis IS: Reducing impact loading during running with the use of real-time visual feedback. J Orthop Sports Phys Ther 2010, 40:206–213. 19. Fellin RE, Rose WC, Royer TD, Davis IS: Comparison of methods for kinematic identification of footstrike and toe-off during overground and treadmill running. J Sci Med Sport 2010, 13:646–650. 20. Mercer JA, Vance J, Hreljac A, Hamill J: Relationship between shock attenuation and stride length during running at different velocities. Eur J Appl Physiol 2002, 87:403–408. 21. Derrick TR, Hamill J, Caldwell GE, Graham E: Energy absorption of impacts during running at various stride lengths. Med Sci Sports Exerc 1998, 30:128–135. 22. Novacheck T: The biomechanics of running. Gait Posture 1998, 7:77–95. 23. Munro CF, Miller DI, Fuglevand AJ: Ground reaction forces in running: a reexamination. J Biomech 1987, 20:147–155. 24. Lafortune MA, Lake MJ, Hennig E: Transfer function between tibial acceleration and ground reaction force. J Biomech 1995, 28:113–117. 25. Madigan ML, Pidcoe PE: Changes in landing biomechanics during a fatiguing landing activity. J Electromyogr Kinesiol 2003, 13:491–498. 26. Christina K, White S, Gilchrist L: Effect of localized muscle fatigue on vertical ground reaction forces and ankle joint motion during running. Hum Mov Sci 2001, 20:257–276. 27. Gerritsen GM, van den Bogert AJ, Nigg BM, Gerritsen KG: Direct dynamics simulation of the impact phase in heel-toe running. J Biomech 1995, 28:661–668. 28. Kim W, Voloshin AS, Johnson SH: Modeling of heel strike transients during running. Hum Mov Sci 1994, 13:221–244. 29. Bobbert MF, Schamhardt HC, Nigg BM: Calculation of vertical ground reaction force estimates during running from positional data. J Biomech 1991, 24:1095–1105. 30. Hamner SR, Seth A, Delp SL: Muscle contributions to propulsion and support during running. J Biomech 2010, 43:2709–2716. 31. Keller TS, Weisberger AM, Ray JL, Hasan SS, Shiavi RG, Spengler DM: Relationship between vertical ground reaction force and speed during walking, slow jogging, and running. Clin Biomech 1996, 11:253–259. 32. Seyfarth A, Geyer H, Günther M, Blickhan R: A movement criterion for running. J Biomech 2002, 35:649–655. 33. Maino ZJ, Gordon JP, Workman AD, Daoud A, McGregor SJ: Differences in running mechanics and economy between collegiate runners and triathletes identified using high-resolution accelerometers [abstract]. Med Sci Sports Exerc 2011, 43(Suppl 5):105. 34. Davis IS: Gait retraining in runners. Orthop Pract 2005, 17:8–13. 35. Daffertshofer A, Lamoth CJC, Meijer OG, Beek PJ: PCA in studying coordination and variability: a tutorial. Clin Biomech 2004, 19:415–428. 36. Milovanović I, Popović DB: Principal component analysis of gait kinematics data in acute and chronic stroke patients. Comput Math Methods Med 2012, 2012:649743. 37. Mah CD, Hulliger M, Lee RG, O’Callaghan IS: Quantitative analysis of human movement synergies: constructive pattern analysis for gait. J Mot Behav 1994, 26:83–102. 38. Dillmann U, Holzhoffer C, Johann Y, Bechtel S, Gräber S, Massing C, Spiegel J, Behnke S, Bürmann J, Louis AK: Principal component analysis of gait in Parkinson’s disease: relevance of gait velocity. Gait Posture 2014, 39:882–887. 39. Troje NF: Decomposing biological motion: a framework for analysis and synthesis of human gait patterns. J Vis 2002, 2:371–387. 40. Lee M, Roan M, Smith B: An application of principal component analysis for lower body kinematics between loaded and unloaded walking. J Biomech 2009, 42:2226–2230. 41. Kobayashi Y, Hobara H, Matsushita S, Mochimaru M: Key joint kinematic characteristics of the gait of fallers identified by principal component analysis. J Biomech 2014, 47:2424–2429. doi:10.1186/1743-0003-11-162 Cite this article as: Lindsay et al.: Contributions of lower extremity kinematics to trunk accelerations during moderate treadmill running. Journal of NeuroEngineering and Rehabilitation 2014 11:162. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Lindsay et al. Journal of NeuroEngineering and Rehabilitation 2014, 11:162 Page 8 of 8 http://www.jneuroengrehab.com/content/11/1/162

Contributions of lower extremity kinematics to trunk accelerations during moderate treadmill running.

12-12-2014

Lindsay, Timothy R,Yaggie, James A,McGregor, Stephen J

eng

PMC4390232

RESEARCH ARTICLE Can Persistence Hunting Signal Male Quality? A Test Considering Digit Ratio in Endurance Athletes Daniel Longman1*, Jonathan C. K. Wells2, Jay T. Stock1 1 Department of Archaeology and Anthropology, University of Cambridge, Cambridge, United Kingdom, 2 Childhood Nutrition Research Centre, UCL Institute of Child Health, London, United Kingdom * [email protected] Abstract Various theories have been posed to explain the fitness payoffs of hunting success among hunter-gatherers. ‘Having’ theories refer to the acquisition of resources, and include the di- rect provisioning hypothesis. In contrast, ‘getting’ theories concern the signalling of male re- sourcefulness and other desirable traits, such as athleticism and intelligence, via hunting prowess. We investigated the association between androgenisation and endurance running ability as a potential signalling mechanism, whereby running prowess, vital for persistence hunting, might be used as a reliable signal of male reproductive fitness by females. Digit ratio (2D:4D) was used as a proxy for prenatal androgenisation in 439 males and 103 fe- males, while a half marathon race (21km), representing a distance/duration comparable with that of persistence hunting, was used to assess running ability. Digit ratio was signifi- cantly and positively correlated with half-marathon time in males (right hand: r = 0.45, p<0.001; left hand: r = 0.42, p<0.001) and females (right hand: r = 0.26, p<0.01; left hand: r = 0.23, p = 0.02). Sex-interaction analysis showed that this correlation was significantly stronger in males than females, suggesting that androgenisation may have experienced stronger selective pressure from endurance running in males. As digit ratio has previously been shown to predict reproductive success, our results are consistent with the hypothesis that endurance running ability may signal reproductive potential in males, through its associ- ation with prenatal androgen exposure. However, further work is required to establish whether and how females respond to this signalling for fitness. Introduction Hunting and reproductive success The high value placed by females on male ability to acquire resources has been well docu- mented [1]. This is evident in pre-industrial human societies, where males exhibit a positive re- lationship between status and number of surviving offspring. Such observations have been made across the continents of Africa, South America and Asia [2–4]. It has been suggested that PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 1 / 12 a11111 OPEN ACCESS Citation: Longman D, Wells JCK, Stock JT (2015) Can Persistence Hunting Signal Male Quality? A Test Considering Digit Ratio in Endurance Athletes. PLoS ONE 10(4): e0121560. doi:10.1371/journal. pone.0121560 Academic Editor: Bernhard Fink, University of Goettingen, GERMANY Received: May 17, 2014 Accepted: February 11, 2015 Published: April 8, 2015 Copyright: © 2015 Longman et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: These authors have no support or funding to report. Competing Interests: The authors have declared that no competing interests exist. the same is true of contemporary Western society; increasing income has been shown to have a significant effect on male reproductive success and desirability as a marriage partner [1,5–7]. Prior to agriculture, hunting may have represented an important means by which male re- sourcefulness could be demonstrated. Food is inexorably linked to status in many cultures around the world [8], and there is evidence that superior hunters enjoy social prestige within the community [9]. Indeed, successful hunters have been shown to enjoy higher reproductive success [10,11]. Hunting may therefore be motivated by male-male competition [12]. Smith [4] has reviewed quantitative links between hunting ability and fitness-related factors such as fertility, offspring survivorship and number of mates, and qualitative links between hunting success and knowledge of hunting ability within communities. However, it is not clear whether it is the high social status of hunters (which reflects their ability to acquire resources), or the actual resources they obtain, that provides the mechanism linking hunting success with reproductive fitness. Theories posed to explain the fitness payoffs of hunting success include the "direct provi- sioning" hypothesis, which asserts that successful hunters are more able to share food with their mate and children, thereby enhancing fertility and offspring survivorship through physio- logical mechanisms. In accordance with this theory, the offspring of successful hunters may in- deed be better nourished [13]. Recent endocrine data from subsistence populations has revealed that while testosterone levels increase upon a successful kill, this increase is not associ- ated with the number and size of kills or the presence of an audience beyond the hunter's fami- ly [14]. These findings are consistent with the direct provisioning model; meat provisioning enhances reproductive success either directly [11], or indirectly through political alliances and other benefits stemming from the community's desire to retain a successful hunter as a neigh- bour [15–18]. Although this theory has intrinsic appeal, the egalitarian organisation of many forager societies ensures that this nepotistic distribution of food is not consistently observed [3]. While some male hunters are more effective than others, meat is widely distributed throughout the group. As a result most of the food consumed is caught by a man outside one's own nuclear family, such that the hunter's own family may receive no more meat than anyone else [3]. Such sharing has been attributed to a means of maintaining social identity, and has been considered a defining characteristic of hunter-gatherer behaviour [19]. Food sharing also serves as a mechanism to deal with fluctuations in food availability [20]. A hunter may ex- change a short-term food surplus for receipt of food in the future should his own hunting ef- forts fail; a phenomenon known as the reciprocal altruism [21]. When food resources are scarce and hunting success unpredictable, food sharing is prevalent as a type of culinary insur- ance policy, as seen in the Inuit of Akalivik [22]. However, food donations may not always be reciprocated [23,24]. Smith [4] is doubtful as to the efficacy of this proposal due to the lack of a system of conditional reciprocity linked to meat provisioning. Thus, perhaps counter- intuitively, there is very little clear evidence that the meat produced by hunting is the primary mechanism underpinning the reproductive fitness of hunters. The 'having' theories do not pro- vide adequate explanation. A third hypothesis, based on Zahavi's "handicap principle" [25], is that hunting success acts as a reliable signal of enviable underlying traits such as athleticism, intelligence or altruism. Successful hunters benefit the community through the provisioning of public goods, an act which enhances their reputation for generosity. The pursuant social standing thereby attained may be attractive to women due to benefits of association such as protection [26]. As such, hunting returns may be exchanged for the future fitness-enhancing benefits of increased social capital or status [27–29]. This theory differs from those previously mentioned in that it need not necessarily be the resource acquisition itself which promotes the reproductive fitness ('hav- ing'), but the signal such resourcefulness conveys of underlying male quality ('getting'). The Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 2 / 12 distribution of acquired meat beyond one's own family unit would also ensure that any such signal was more widely observed [30]. Indeed, hunters are aware that others will know who ac- quired the meat [31,32]. Male hunting may not then be motivated purely by consequent meat consumption [12,33]. Such signalling would be reliable if it were costly to the signaller, such that an individual lacking in a certain trait would be unable to afford to mimic the signal. The hypothesis that hunting may act as a signal has been addressed by several authors [30,34,35], but it is as yet unclear what the underlying mechanism may be. In other words, what makes the signal a reliable index of fitness. We first review the physical traits that contrib- ute to hunting success, and then consider potential mechanisms whereby these traits might sig- nal reproductive potential. Hunting success and physical skills The hunting of mid-sized animals using hunting technologies such as spears favours a method of disadvantaging or weakening the animal before approaching for the kill. This is because the killing range of spears is very limited, and being close enough to kill a mid-sized animal would be to risk injury to the hunter. Persistence hunting offers such a method, and may be a very ef- fective means of food acquisition under certain conditions. Persistence hunting is a technique by which hunters track and chase prey to the point of prey exhaustion or hyperthermia, often during the hottest part of the day. This technique has been observed in the Kalahari in Africa, and the Tarahumara tribe of Northern Mexico, who have been reported chasing deer until they collapse before strangling them by hand [36,37]. Other animals targeted in this way include steenbok, gemsbok, duiker, caracal, cheetah, kudu and eland [38,39]. During the hunt, chases of up to 35km have been documented. Persistence hunting is useful from the point of view of the early hunter for three reasons; it is relatively low risk, easy for a fit human with animal tracking skills, and requires low metabolic cost relative to potential pay-off [40]. Endurance running may therefore represent a valuable component of hunting success. The evolution of human endurance running ability has attracted attention previously [41,42]. This followed observations that, in comparison with other cursorial animals, humans perform very well over long distances. We are unique amongst primates in possessing the ability to run dis- tances of several kilometres using aerobic metabolism [41], with many amateur human runners able to sustain speeds of 5m/s [40]. This is fast compared with specialised quadrupedal cursors; a dog with a similar mass to a human (65kg) has a trot-gallop transition speed of 3.8m/s, and can then only sustain a gallop for a maximum of 15 minutes under ideal conditions [43]. The same is true of horses, which is surprising as they have been selectively bred for running ability [44]. Human runners can easily cover distances exceeding 10km a day; comparable with hunt- ing dogs and hyenas which run to scavenge and hunt [40]. Consequently, the physical capacity for endurance running appears have been selected for in our genus. Hunting may well have provided the primary selective pressure; evidence for hominin carnivory dates back to approxi- mately 2.5Ma [45]. The ability to run long distances to either hunt or scavenge may have im- proved the chances of acquiring meat [40,46]. But as described above, if meat was widely shared, how might hunting ability translate into fitness payoffs? Endurance running might benefit male fitness if it acted as a reliable signal of reproductive potential. Since testosterone is widely associated with reproductive success [47,48], an associa- tion between testosterone and endurance running would make running prowess a reliable sig- nal of male reproductive potential. Recent work has reported associations between sporting ability and a marker for foetal testosterone exposure, the 2D:4D ratio [49,50]. This develop- mental association between testosterone and physical abilities makes it a viable candidate as a signalling mechanism. Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 3 / 12 The ontogenetic development of digit ratio variability, and significance for reproductive success The sexually dimorphic digit ratio (2D:4D), first noted by Baker [51], is linked to prenatal androgenisation [52]. The presence of significant sexual dimorphism in the digit ratio of de- ceased human foetuses suggests 2D:4D ratio is established early in prenatal development [53]. Analysis of amniotic fluid provided direct evidence for a relationship between foetal hormones and digit ratio [54]. This followed indirect reports stemming from homeobox genes [55,56] and congenital adrenal hyperplasia [57,58]. Although the relationship between high digit ratio and prenatal testosterone exposure has been questioned [59], and may be confounded by sexu- al dimorphism in body size [60], the majority of evidence suggests that digit ratio is a suitable instrument for investigating effects of prenatal androgen exposure on subsequent phenotype [52]. As a result digit ratio is increasingly used as a proxy for foetal hormone environment—to investigate early life predictors of later phenotype, and sex differences therein. Digit ratio predicts reproductive success in both men and women once the effects of age and population have been removed [61]. Mechanistically, low digit ratios are associated with higher sperm counts and testosterone levels in men, and, conversely, high digit ratios are linked with high oestrogen and luteinising hormone in women [62]. Furthermore, males exhibit a negative relationship between digit ratio and other correlates of reproductive success such as preferred number of children, strength of sex-drive, ease of achieving sexual excitement [47] and sperm numbers per ejaculate [62]. Thus, digit ratio variability emerging early in life has significant im- plications for fitness. Digit ratio is negatively correlated with physical prowess across a range of sporting disci- plines from slalom skiing [63] to football [64]. Digit ratio has also been tentatively correlated to middle-distance running ability, albeit with a small sample size at distances of less than 4 miles [65]. This study by Manning et al was the first to discuss the relationship between digit ratio and running within the context of persistence hunting. Digit ratio was found to explain a larger portion of the variance in endurance running than expected (up to 25%); explaining more vari- ation than other sports [66] and running over shorter distances [67,68]. As such, the literature supports the view that while 2D:4D does predict running speed, the predictive power increases from sprinting events (1–2%) to events of up to 4 miles in length (20–25%). We therefore sought to test the hypothesis that physical prowess at endurance running is as- sociated with this putative marker of testosterone exposure. This was performed with a larger sample and distance than previously reported. A female sample was included, as the relation- ship between female digit ratio and running ability is as yet unknown. Since men undertake the majority of hunting [69], we further predicted that the association between digit ratio and en- durance running prowess would be stronger in males than females. Such a sex-difference would suggest that the selective pressure of endurance running has shaped running prowess to be a stronger signal of reproductive potential in males than females. Through such preferences, males would then convert their physical prowess into fitness gains. Our study therefore tests a potential mechanism whereby ‘getting’ meat could translate into male fitness returns. Materials and Methods Athletes (N = 542; m = 439, f = 103) were recruited at the 2013 Robin Hood Half Marathon (21km), held on September 29th. Athletes received an email explaining the study prior to race day, and were invited to take part following the race. Weather conditions were clear skies, with a mean day-time temperature 14°C. The race is a high-profile event, with course records of 61:38 and 73:32 for men and women respectively. Participants ranged between the ages of 19 and 35, and were all Caucasian. The half-marathon distance was chosen due to its Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 4 / 12 appropriateness to evolutionary hunting-associated running [36]. Race time over 21km was the performance indicator. While this is a measure of speed, the half-marathon distance is considered to be very different from pure speed events such as the 100m sprint in terms of the energy systems utilised. While the 100m sprint utilises the anaerobic system to supply ATP at high rates, prolonged endurance races such as the half-marathon are dependent upon the more sustainable aerobic energy system and lipid metabolism [70–74]. Half-marathons are often used to investigate the effects of endurance exercise on physiology [75–78], so we are confident the event may be used to consider endurance running ability. All participants participating in the same race, ensuring standardisation of not only dis- tance, but also of time-affecting factors such as weather conditions, time of day and race-course elevation profile. All competitors wore small electronic chips which uniquely identified them as they crossed electronic mats at the start and finish lines, guaranteeing accurate race timings. Photocopies were taken of athletes' hands upon finishing the race, and measurements of digit ratio were made at a later date. It is appreciated that ratios obtained from photocopies and from direct measurements should not be combined within one study [79], so all measurements were taken from photocopies, with the same machine being used for all copies. This method was chosen due to ease of use and speed in facilitating a large sample size. Digit ratio was mea- sured using Mitutoyo electronic callipers, reading to 0.01mm. Measurements were taken twice from each photocopy to check repeatability. Digit ratios from the first measurement were strongly correlated with the lengths recorded from the second measurement for all individual participants (all r > 0.95). In addition, the means of individual right hand and left hand digit ratios were significantly correlated (male r = 0.88, P < 0.001; female r = 0.762, P < 0.001). The precision of each measurement was found using the method of Bland & Altman [80]; male right 2D = 2.52; male right 4D = 2.21; male left 2D = 2.36; male left 4D = 2.07; female right 2D = 2.24; female right 4D = 2.04; female left 2D = 1.93; female left 4D = 2.42, all measure- ments to 2 decimal places. These associations gave confidence in the reliability of the hand-spe- cific measurements of digit ratio, and therefore in the conclusions subsequently drawn. Ethical approval for the project was granted by the University of Cambridge Human Biology Ethics Committee. Statistics Crude associations between digit ratio and race time were explored in each sex using correla- tion analysis. The difference between the male and female correlation coefficients was analysed using the Fisher r-to-z transformation and subsequent comparison of z-scores [81,82]. Multi- ple linear regression was used to analyse the sex difference in the relationship between digit ratio and race time. Sex was coded as 0 (female) versus 1 (male), and an ‘Interaction’ term was constructed by multiplying sex code by digit ratio. Each of sex, ‘interaction’ and digit ratio were then used as predictors of race time. Results Descriptive statistics A description of the male and females samples is given in Table 1. Males were older, and had significantly lower right and left digit ratios than females. Among the male subsample there was a significant positive correlation between right and left hand 2D:4D ratio and half-marathon time (right r = 0.45, p < 0.001; left r = 0.42, p < 0.001). When age was controlled for the correlation strengthened (right r = 0.47, p < 0.001; left r = 0.43, p < 0.001). The same was true among the female subsample (right hand r = 0.26, p < 0.01; left hand r = 0.23, p = 0.02), with the correlation strengthening when Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 5 / 12 age was controlled for (right r = 0.29, p < 0.01; left r = 0.26, p < 0.01). These positive correla- tions can be seen below in Fig. 1. Note the steeper regression line in the male subsample com- pared to the female subsample. Regression analysis revealed that men performed better than women, higher levels of per- formance were associated with lower digit ratio, and digit ratio increases performance more significantly in men than women (p<0.05). Age was not a significant contributor in the regres- sion model, so was removed. See Table 2 for a summary of the results. Discussion It was hypothesised that 2D:4D ratio correlates with endurance running performance over the half-marathon distance. This hypothesis has been supported. The male effect sizes are similar to those of Manning et al. [65], which is consistent with the theory that digit ratio explains more variation in endurance running than it does in shorter running events or other sports. The link between digit ratio and maximal oxygen uptake may well relate to these observations [83]. A marker of testosterone exposure is therefore associated with running ability, which eth- nographic evidence has shown to be an important attribute for hunting [37]. As testosterone (including the 2D:4D ratio) has been consistently associated with reproductive success [47], this relationship between endurance running ability and digit ratio provides mechanistic evi- dence in support of the hypothesis that running prowess could act as a reliable signal for male reproductive potential. A conceptual diagram outlining this potential mechanism by which Table 1. Descriptive characteristics of the samples. Males (n = 439) Females (n = 103) Mean SD Mean SD Age (y) 31.7 4.93 28.8 4.58 2D:4D ratio right 0.97 0.033 0.98 0.028 2D:4D ratio left 0.94 0.036 1.01 0.035 Race time (s) 6946 1313 7002 926 doi:10.1371/journal.pone.0121560.t001 Fig 1. Scatter plot of male and female right hand 2D:4D ratio versus half marathon performance (s). The steeper male gradient is visible. doi:10.1371/journal.pone.0121560.g001 Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 6 / 12 hunting success and running performance could act as a signal of reproductive fitness is given below in Fig. 2. As a positive relationship between hunting ability and male reproductive success has been reported [2,3], it may be that women are attracted to men with the capacity to ‘get’, rather than those who ‘have’. This may be a consequence of the egalitarian organisation of many forager Table 2. Regression of half-marathon time in seconds on digit ratio (right in model 1, left in model 2), sex and sex-ratio interaction (right in model 1, left in model 2). Step Predictor β t p Model 1 1 Age -.057 -1.317 .188 Δ R2 = .003, F Change (1,540) = 1.736, p = .188 2 Age -.007 -.186 .853 Sex Code -2.692 -2.058 .040 Right Interaction 2.725 2.104 .036 Right ratio .249 2.410 .016 Δ R2 = .203, F Change (3,537) = 45.692, p < .001 Model summary: F(4,537) = 34.810, p < 0.001, R2 = .206 Model 2 1 Age -.057 -1.317 .188 Δ R2 = .003, F Change (1,540) = 1.736, p = .188 2 Age -.030 -.734 .463 Sex Code -2.419 -2.174 .030 Left Interaction 2.561 2.427 .016 Left ratio .248 2.201 .028 Δ R2 = .172, F Change (3,537) = 37.239, p < .001 Model summary: F(4,537) = 28.451, p < 0.001, R2 = .175 doi:10.1371/journal.pone.0121560.t002 Fig 2. Conceptual diagram outlining a potential mechanism by which hunting success and running performance act as a signal of reproductive fitness. doi:10.1371/journal.pone.0121560.g002 Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 7 / 12 societies, which ensures that meat is widely distributed throughout the group [3]. It is acknowl- edged, however, that this study has not demonstrated that women are in fact differentially at- tracted to faster male runners, or that this is the only mechanism potentially linking hunting success with reproductive success. Of course, hunting ability and reproductive success may both be correlated with another independent variable which is itself a cause of higher reproduc- tive success, independent of the male's hunting abilities [4]. Although both sexes exhibited a statistically significant positive correlation between right and left hand 2D:4D ratio and time taken to complete a half-marathon running race, the rela- tionship exhibited by males was significantly stronger than that of females. As such, the sec- ondary hypothesis was also supported, suggesting that the ability of running ability to signal reproductive potential has been under stronger selection in males than in females. Testosterone is not only a link between digit ratio and endurance running ability, but also an important mediator of the male reproductive effort. Behaviourally, testosterone plays an im- portant role in producing sex drive [84] and mediating confidence and assertiveness in social situations [85] – qualities deemed beneficial in the male mating effort. Physiologically, testos- terone serves to promote muscle growth [86,87], providing an advantage in male-male combat situations [48,88]. Muscularity is also a sexually attractive trait; more muscular men to report a greater number of sexual partners and younger age at first conception [89,90] and more off- spring sired [91]. Muscularity per se, however, could be disadvantageous as a signal of male fitness in a hot environment, as the link between muscle mass and metabolic rate means that a larger muscle mass could induce overheating. Additionally, the high metabolic cost of muscle tissue would not be well suited to environments prone to famine. Running ability may therefore represent a signal of reproductive fitness that is better suited to a hot stochastic environment, where it is furthermore compatible with immediate benefits, given that hunting promotes food acquisition. Comparative analyses of Olympic winning times and world records has lead some authors to conclude that women will outcompete men in the marathon by the end of the 20th century and over 100m in the mid 22nd century [92,93]. However, caution is required as the initial greater rate of improvement in female performances is most likely due to historical consider- ations such as the later social acceptance and inclusion of female distance running in major events such as the Olympic Games [94]. The gender difference is now believed to have reached a plateau [95], with the remaining disparity explicable by biological differences be- tween males and females [96]. While it has been suggested that the gender difference in per- formance disappears as distances increase beyond that of the marathon [97,98], several studies have reported that such distances exhibit no change in relative ability [99]. Indeed, performances between the sexes may even diverge [100,101]. Although it is possible that fe- males improve relative to males with distance, such considerations are not deemed to be rele- vant to this study as ultra-marathon distances are not applicable to the evolutionary pressures applied by persistence hunting. To conclude, this investigation has shown that both male and female digit ratios are signifi- cantly and negatively correlated to ability in endurance running of a distance/duration compa- rable with that of persistence hunting (a positive correlation between digit ratio and half marathon time). The relationship was stronger for males than females, suggesting that males faced greater evolutionary pressure to develop endurance running capabilities. We suggest that hunting ability might therefore act as a reliable signal of male fitness, in addition to its function of calorie acquisition. Consequently, the ability to 'get' meat may contribute to the positive rela- tionship between hunting ability and male reproductive success. Our work has used sporting Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 8 / 12 ability as a proxy for hunting prowess; further work is now required to test this hypothesis in hunting societies. Author Contributions Conceived and designed the experiments: DL JTS JCKW. Performed the experiments: DL. An- alyzed the data: DL JTS JCKW. Contributed reagents/materials/analysis tools: DL JTS JCKW. Wrote the paper: DL JTS JCKW. References 1. Buss D. Sex differences in human mate preferences: Evolutionary hypotheses tested in 37 cultures. Behav Brain Sci. 1989; 12(1):1–49. 2. Pennington R, Harpending H. The structure of an African pastoralist community: demography, history, and ecology of the Ngamiland Herero. New York: Oxford Unversity Press; 1993. 3. Kaplan H, Hill K. Food sharing among Ache foragers: Tests of Explanatory Hypotheses. Curr Anthro- pol. 1985; 26(2):223–46. 4. Smith EA. Why do good hunters have higher reproductive success? Hum Nat. 2004; 15(4):343–64. 5. Pollet T, Nettle D. Driving a hard bargain: sex ratio and male marriage success in a historical US popu- lation. Biol Lett. 2008; 4(1):31–3. PMID: 18055406 6. Buunk BP, Dijkstra P, Fetchenhauer D, Kenrick DT. Age and gender differences in mate selection cri- teria for various involvement levels. Pers Relatsh. 2002; 9(3):271–8. 7. Nettle D, Pollet TV. Natural selection on male wealth in humans. Am Nat. 2008; 172(5):658–66. doi: 10.1086/591690 PMID: 18808363 8. Weissner P, Schiefenövel W, editors. Food and the status quest: an interdisciplinary perspective. Providence, RI: Berghahn Books; 1996. 9. Altman J, Peterson N. Rights to game and rights to cash among contemporary Australian hunter gath- erers. 1988. 10. Hill K, Kaplan H. On why male foragers hunt and share food. Curr Anthropol. 1993; 34(5):701–10. 11. Kaplan H, Hill K. Hunting ability success among male Ache foragers: Preliminary research conclu- sions. Curr Anthropol. 1985; 26(1):131–3. 12. Hawkes K, O’Connell JF, Blurton Jones N. Hunting and nuclear families some lessons from the Hadza about men’s work. Curr Anthropol. 2001; 42(5):681–709. 13. Hawkes K. On why male foragers hunt and share food: reply to Hill and Kaplan. Curr Anthropol. 1993; 345:706–10. 14. Trumble BC, Smith EA, Connor KAO, Kaplan HS, Gurven MD. Successful hunting increases testos- terone and cortisol in a subsistence population. Proc R Soc B Biol Sci. 2013; 281(1776). 15. Sugiyama L, Chacon R. Effects of illness and injury on foraging among the Yora and Shiwiar: patholo- gy risk as adaptive problem. Human behavior and adaptation: an anthropological perspective. 2000. p. 371–95. 16. Patton J. Coalitional effects on reciprocal fairness in the ultimatum game: A case from the Ecuadorian Amazon. Foundations of human sociality: Economic experiments and ethnographic evidence from fif- teen smallscale societies. 2004. p. 96–124. 17. Patton J. Meat sharing for coalitional support. Evol Hum Behav. 2005; 26(2):137–57. 18. Hawkes K, Altman J, Beckerman S, Grinker R, Jeske R, Peterson N, et al. Why gatherers work; An ancient version of the problem of public goods. Curr Anthropol. 1993; 34(4):341–61. 19. Lee R, DeVore I. Problems in the study of hunters and gatherers. Man the Hunter. 1968. p. 3–12. 20. Enloe JG. Food sharing past and present: archaeological evidence for economic and social interac- tions. Before Farming. 2003; 1.1:1–23. 21. Trivers R. The evolution of reciprocal altruism. Q Rev Biol. 1971;35–57. 22. Kishigami N. Contemporary Inuit food sharing and hunter support program of Nunavik, Canada. Senri Ethnol Stud. 2000; 53:171–92. 23. Tucker VA. The energetic cost of moving about: walking and running are extremely inefficient forms of locmotion. Much greater efficiency is achieved by birds, fish-and bicyclists. Am Sci. 1975; 63(4):413– 9. PMID: 1137237 Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 9 / 12 24. Bliege Bird R, Bird DW, Smith EA, Kushnick GC. Risk and reciprocity in Meriam food sharing. Evol Hum Behav. 2002; 23(4):297–321. 25. Zahavi A. Mate selection—a selection for a handicap. J Theor Biol. 1975; 53:205–14. PMID: 1195756 26. Blurton Jones N, Marlowe F, Hawkes K, O’Connell JF. Hunter-gatherer divorce rates and the paternal provisioning theory of human monogamy. In: Cronk L, Chagnon N, Irons W, editors. Adaptation and human behaviour: an anthropological perspective. New York, NY: Aldine de Gruyter; 2000. p. 65–84. 27. Bliege Bird R, Smith EA. Signaling theory, strategic interaction, and symbolic capital. Curr Anthropol. 2005; 46(2):221–48. 28. Gurven M, Allen-Arave W, Hill K, Hurtado M. “It’s a wonderful life”. signaling generosity among the Ache of Paraguay. Evol Hum Behav. 2000; 21(4):263–82. PMID: 10899478 29. Bliege Bird R, Scelza B, Bird DW, Smith EA. The hierarchy of virtue: mutualism, altruism and signaling in Martu women’s cooperative hunting. Evol Hum Behav. Elsevier Inc.; 2012; 33(1):64–78. 30. Smith EA, Bliege Bird R. Turtle hunting and tombstone opening: public generosity as costly signaling. Evol Hum Behav. 2000; 21(4):245–61. PMID: 10899477 31. Hawkes K. Showing off: tests of a hypothesis about men’s foraging goals. Ethol Sociobiol. 1991; 12- (1):29–54. 32. Hawkes K. Sharing and collective action. Evolutionary ecology and human behavior. In: Smith E, Win- terhalder B, editors. Evolutionary Ecology and human behaviour. Hawthorne, NY: de Gruyter; 1992. p. 269–300. 33. Bliege Bird R, Smith E, Bird D. The hunting handicap: costly signaling in human foraging strategies. Behav Ecol Sociobiol. 2001; 50(1):9–19. 34. Hawkes K, Bliege Bird R. Showing off, handicap signaling, and the evolution of men’s work. Evol Anthropol Issues, News, Rev. 2002; 11(2):58–67. 35. Hawkes K. Why do men hunt? Some benefits for risky strategies. In: Cashdan E, editor. Risk and un- certainty in tribal and peasant economies. Boulder: Westview Press; 1990. p. 145–66. 36. Pennington C. The Tarahumara of Mexico. Salt Lake City: University of Utah Press; 1963. 37. Liebenberg L. Persistence hunting by modern hunter gatherers. Curr Anthropol. 2006; 47(6):1017– 26. 38. Steyn H. Southern Kalahari San subsistence ecology: A reconstruction. South African Archaeol Bull. 1984;117–24. 39. Schapera I. The Khoisan Peoples of South Africa: Bushmen and Hottentots. London: Routledge and Kegan Paul; 1930. 40. Lieberman DE, Bramble DM, Raichlen DA, Shea JJ. Brains, brawn, and the evolution of human endur- ance running capabilities. In: Grine, Fleagle, Richard, Leakey, editors. The First Humans—Origin and Early Evolution of the Genus Homo. Springer; 2006. 41. Carrier DR. The energetic paradox of human running and hominid evolution. Curr Anthropol. 1984; 25 (4):483–95. 42. McDougall C. Born to Run: A hidden tribe, superathletes, and the greatest race the world has ever seen. Knopf; 2009. 43. Heglund N, Taylor C. Speed, stride frequency and energy cost per stride: how do they change with body size and gait? J Exp Biol. 1988; 138:301–18. PMID: 3193059 44. Minetti AE. Efficiency of equine express postal systems. Nature. 2003; 426:785–6. PMID: 14685222 45. Semaw S, Renne P, Harris J, Feibel C, Bernor R, Fesseha N, et al. 2.5-million-year-old stone tools from Gona, Ethiopia. Nature. 1997; 385:333–6. PMID: 9002516 46. Bunn H. Hunting, power scavenging, and butchering by Hadza foragers and by Plio-Pleistocene Homo. Meat-eating and human evolution. 2001. 47. Manning JT, Fink B. Digit ratio (2D:4D), dominance, reproductive success, asymmetry, and socio- sexuality in the BBC Internet Study. Am J Hum Biol. 2008; 20(4):451–61. doi: 10.1002/ajhb.20767 PMID: 18433004 48. Bribiescas R. Reproductive physiology of the human male. Reproductive ecology and human evolu- tion. 2001. p. 107–33. 49. Pokrywka L, Rachoń D, Suchecka-Rachoń K, Bitel L. The second to fourth digit ratio in elite and non- elite female athletes. Am J Hum Biol. 2005; 17(6):796–800. PMID: 16254897 50. Paul S, Kato B, Hunkin J, Vivekanandan S, Spector T. The big finger: the second to fourth digit ratio is a predictor of sporting ability in women. Br J Sports Med. 2006; 40(12):981–3. PMID: 17008344 51. Baker F. Anthropological Notes on the Human. Am Anthropol. 1888; 1(1):51–76. Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 10 / 12 52. Hönekopp J, Bartholdt L, Beier L, Liebert A. Second to fourth digit length ratio (2D:4D) and adult sex hormone levels: new data and a meta-analytic review. Psychoneuroendocrinology. 2007; 32(4):313– 21. PMID: 17400395 53. Galis F, Ten Broek CMA, Van Dongen S, Wijnaendts LCD. Sexual dimorphism in the prenatal digit ratio (2D:4D). Arch Sex Behav. 2010; 39(1):57–62. doi: 10.1007/s10508-009-9485-7 PMID: 19301112 54. Lutchmaya S, Baron-Cohen S, Raggatt P, Knickmeyer R, Manning JT. 2nd To 4th digit ratios, fetal testosterone and estradiol. Early Hum Dev. 2004 Apr; 77(1–2):23–8. PMID: 15113633 55. Kondo T, Zákány J, Innis JW, Duboule D. Of fingers, toes and penises. Nature. 1997; 390(6655):29. PMID: 9363887 56. Mortlock D, Innis J. Mutation of HOXA13 in hand-foot-genital syndrome. Nat Genet. 1997; 15(2):179– 80. PMID: 9020844 57. Brown WM, Hines M, Fane BA, Breedlove SM. Masculinized finger length patterns in human males and females with congenital adrenal hyperplasia. Horm Behav. 2002; 42(4):380–6. PMID: 12488105 58. Okten A, Kalyoncu M, Yariş N. The ratio of second- and fourth-digit lengths and congenital adrenal hy- perplasia due to 21-hydroxylase deficiency. Early Hum Dev. 2002; 70(1–2):47–54. PMID: 12441205 59. Berenbaum SA, Bryk KK, Nowak N, Quigley CA, Moffat S. Fingers as a marker of prenatal androgen exposure. Endocrinology. 2009; 150(11):5119–24. doi: 10.1210/en.2009-0774 PMID: 19819951 60. Kratochvíl L, Flegr J. Differences in the 2nd to 4th digit length ratio in humans reflect shifts along the common allometric line. Biol Lett. 2009; 5(5):643–6. doi: 10.1098/rsbl.2009.0346 PMID: 19553247 61. Manning JT, Barley L, Walton J, Lewis-Jones D, Trivers RL. The 2nd: 4th digit ratio, sexual dimor- phism, population differences, and reproductive success: evidence for sexually antagonistic genes? Evol Hum Behav. 2000; 21:163–83. PMID: 10828555 62. Manning J, Scutt D, Wilson J, Lewis-Jones D. The ratio of 2nd to 4th digit length: a predictor of sperm numbers and concentrations of testosterone, luteinizing hormone and oestrogen. Hum Reprod. 1998; 13(11):3000–4. PMID: 9853845 63. Manning J. The ratio of the 2nd to 4th digit length and performance in skiing. J Sports Med Phys Fit- ness. 2002; 42(4):446. PMID: 12391439 64. Manning JT, Taylor RP. Second to fourth digit ratio and male ability in sport: implications for sexual se- lection in humans. Evol Hum Behav. 2001; 22(1):61–9. PMID: 11182575 65. Manning J, Morris L, Caswell N. Endurance running and digit ratio (2D: 4D): Implications for fetal tes- tosterone effects on running speed and vascular health. Am J Hum Biol. 2007; 421:416–21. 66. Hönekopp J, Schuster M. A meta-analysis on 2D:4D and athletic prowess: Substantial relationships but neither hand out-predicts the other. Pers Individ Dif. 2010 Jan; 48(1):4–10. 67. Manning J, Hill M. Digit ratio (2D:4D) and sprinting speed in boys. Am J Hum Biol. 2009; 21(2):210–3. doi: 10.1002/ajhb.20855 PMID: 19107924 68. Trivers R, Hopp R, Manning J. A longitudinal study of digit ratio (2D:4D) and its relationship with adult running speed in Jamaicans. Hum Biol. 2013; 85(4):623–6. PMID: 25019194 69. Biesele M, Barclay S. Ju/’huan women's tracking knowledge and its contribution to their husbands' hunting success. African Study Monogr. 2001; 26:67–84. 70. Duffield R, Dawson B, Goodman C. Energy system contribution to 100-m and 200-m track running events. J Sci Med Sport. 2004; 7(3):302–13. PMID: 15518295 71. Hagerman F. Energy metabolism and fuel utilization. Med Sci Sport Exerc. 1992; 24(9 Supple):S309– 14. 72. Wilmore J, Costill D. Physiology of sport and exercise. 2nd ed. Human Kinetics Publishers; 1999. 73. Brooks GA, Mercier J. Balance of carbohydrate and lipid utilization during exercise: the “crossover” concept. J Appl Physiol. 1994; 76(6):2253–61. PMID: 7928844 74. McArdle W, Katch F, Katch V. Exercise physiology: energy, nutrition and human performance. Phila- delphia, PA: Lippincott, Williams & Wilkins; 2001. 75. Dalla Vecchia L, Traversi E, Porta A, Lucini D, Pagani M. On site assessment of cardiac function and neural regulation in amateur half marathon runners. Open Hear. 2014; 1(1):e000005. 76. Oomah S, Mousavi N, Bhullar N, Kumar K, Walker J, Lytwyn M, et al. (2011). The role of three- dimensional echocardiography in the assessment of right ventricular dysfunction after a half mara- thon: comparison with cardiac magnetic resonance imaging. J Am Soc Echocardiogr. 2011; 24 (2):207–13. doi: 10.1016/j.echo.2010.10.012 PMID: 21281911 77. Lippi G, Longo UG, Maffulli N. Genetics and sports. Br Med Bull [Internet]. 2010; 93:27–47. Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 11 / 12 78. Dion T, Savoie F, Asselin A, Gariepy C, Goulet E. Half-marathon running performance is not improved by a rate of fluid intake above that dictated by thirst sensation in trained distance runners. Eur J Appl Physiol. 2013; 113(12):3011–20. doi: 10.1007/s00421-013-2730-8 PMID: 24085484 79. Manning JT, Fink B, Neave N, Caswell N. Photocopies yield lower digit ratios (2D:4D) than direct fin- ger measurements. Arch Sex Behav [Internet]. 2005; 34(3):329–33. PMID: 15971015 80. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986; 307:1–9. 81. Preacher K. Calculation for the test of the difference between two independent correlation coefficients [Computer software]. 82. Cohen J, Cohen P. Applied multiple regression/correlation analysis for the behavioral sciences. Hills- dale, New Jersey: Erlbaum; 1983. 83. Hill R, Simpson B, Millet G, Manning J, Kilduff L. Right-left digit ratio (2D:4D) and maximal oxygen up- take. J Sports Sci. 2011; 30(2):129–34. doi: 10.1080/02640414.2011.637947 PMID: 22141747 84. Baumeister RF, Catanese KR, Vohs KD. Is there a gender difference in strength of sex drive? theoreti- cal views, conceptual distinctions, and a review of relevant evidence. Personal Soc Psychol Rev. 2001 Aug; 5(3):242–73. 85. Ellison P. Energetics and reproductive effort. Am J Hum Biol. 2003; 15:342–51. PMID: 12704710 86. Griggs R, Kingston W, Jozefowicz R, Herr B, Forbes G, Halliday D. Effect of testosterone on muscle mass and muscle protein synthesis. J Appl Physiol. 1989; 66:498–503. PMID: 2917954 87. Kadi F. Cellular and molecular mechanisms responsible for the action of testosterone on human skel- etal muscle. A basis for illegal performance enhancement. Br J Pharmacol. 2008; 154(3):522–8. doi: 10.1038/bjp.2008.118 PMID: 18414389 88. Dijkstra P, Buunk BP. Sex differences in the jealousy-evoking effect of rival characteristics. Eur J Soc Psychol. 2002; 32(6):829–52. 89. Frederick DA, Haselton MG. Why is muscularity sexy? Tests of the fitness indicator hypothesis. Pers Soc Psychol Bull. 2007; 33(8):1167–83. PMID: 17578932 90. Gallup AC, White DD, Gallup GG. Handgrip strength predicts sexual behavior, body morphology, and aggression in male college students. Evol Hum Behav. 2007; 28(6):423–9. 91. Wells J, Charoensiriwath S, Treleaven P. Reproduction, aging, and body shape by three-dimensional photonic scanning in Thai men and women. Am J Hum Biol. 2011; 23(3):291–8. doi: 10.1002/ajhb. 21151 PMID: 21387458 92. Whipp B, Ward S. Will women soon outrun men? Nature. 1992; 355(6355):25. PMID: 1731197 93. Tatem A, Guerra C, Atkinson P, Hay S. Momentous sprint at the 2156 Olympics? Nature. 2004; 431 (7008):525–6. PMID: 15457249 94. Noakes T. Lore of running. Champagn, Illinois: Leisure Press; 2003. 95. Sparling P, O’Donnell E, Snow T. The gender difference in distance running performances has pla- teaued: an analysis of world rankings from 1980 to 1996. Med Sci Sport Exerc. 1998; 30(12):1725–9. PMID: 9861606 96. Cheuvront SN, Carter R, Deruisseau KC, Moffatt RJ. Running performance differences between men and women:an update. Sports Med. 2005; 35(12):1017–24. PMID: 16336006 97. Bam J, Noakes T, Juritz J, Dennis S. Could women outrun men in ultramarathon races? Med Sci Sport Exerc. 1997; 29(2):244–7. PMID: 9044230 98. Speechly D, Taylor S, Rogers G. Differences in ultra-endurance exercise in performance-matched male and female runners. Med Sci Sport Exerc. 1996; 28(3):359–65. PMID: 8776224 99. Hoffman M. Ultramarathon trail running comparison of performance-matched men andwomen. Med Sci Sport Exerc. 2008; 40(9):2008. 100. Coast J, Blevins J, Wilson B. Do gender differences in running performance disappear with distance? Can J Appl Physiol. 2004; 29(2):139–245. PMID: 15064423 101. Eichenberger E, Knechtle B, Rüst CA, Rosemann T, Lepers R. Age and sex interactions in mountain ultramarathon running—the Swiss Alpine Marathon. Open access J Sport Med. 2012; 3:73–80. doi: 10.2147/OAJSM.S33836 PMID: 24198590 Persistence Hunting and the Signalling of Male Quality PLOS ONE | DOI:10.1371/journal.pone.0121560 April 8, 2015 12 / 12

Can persistence hunting signal male quality? A test considering digit ratio in endurance athletes.

04-08-2015

Longman, Daniel,Wells, Jonathan C K,Stock, Jay T

eng

PMC7143158

International Journal of Environmental Research and Public Health Article Interval Hypoxic Training Enhances Athletic Performance and Does Not Adversely Affect Immune Function in Middle- and Long-Distance Runners Won-Sang Jung 1 , Sung-Woo Kim 1 and Hun-Young Park 1,2,* 1 Physical Activity and Performance Institute, Konkuk University, 120 Neungdong-ro, Gwangjin-gu, Seoul 05029, Korea; [email protected] (W.-S.J.); [email protected] (S.-W.K.) 2 Department of Sports Medicine and Science, Graduate School, Konkuk University, 120 Neungdong-ro, Gwangjin-gu, Seoul 05029, Korea * Correspondence: [email protected]; Tel.: +(82)-2-2049-6035 Received: 12 February 2020; Accepted: 15 March 2020; Published: 16 March 2020



 Abstract: This study evaluated the effects of intermittent interval training in hypoxic conditions for six weeks compared with normoxic conditions, on hemodynamic function, autonomic nervous system (ANS) function, immune function, and athletic performance in middle- and long-distance runners. Twenty athletes were divided into normoxic training (normoxic training group (NTG); n = 10; residing and training at sea level) and hypoxic training (hypoxic training group (HTG); n = 10; residing at sea level but training in 526-mmHg hypobaric hypoxia) groups. All dependent variables were measured before, and after, training. The training frequency was 90 min, 3 d per week for six weeks. Body composition showed no significant difference between the two groups. However, the HTG showed more significantly improved athletic performance (e.g., maximal oxygen uptake). The hemodynamic function (e.g., oxygen uptake, oxygen pulse, and cardiac output) during submaximal exercise and ANS function (e.g., standard deviation and root mean square of successive differences, high frequency, and low/high frequency) improved more in the HTG. Immune function parameters were stable within the normal range before and after training in both groups. Therefore, hypoxic training was more effective in enhancing athletic performance, and improving hemodynamic and ANS function; further, it did not adversely affect immune function in competitive runners. Keywords: interval hypoxic training; hemodynamic function; autonomic nervous system balance; exercise performance; immune function; competitive middle- and long-distance runners 1. Introduction Endurance exercise performance is related to various factors that can be altered by diverse hypoxic training methods, including erythropoiesis, exercise economy, capillary density, hemodynamic function, and acid-base response in the skeletal muscle [1,2]. Enhancing these factors, which are related to endurance exercise performance, increases the efficiency of aerobic energy production and consequently enhances maximum oxygen uptake (VO2max). It also enhances athletic performance by improving time until fatigue and increasing exercise intensity [3–6]. In particular, endurance exercise performance is reported to be the most affected by hemodynamic function, which is an indicator of oxygen transport and utilization ability [7]. Currently, altitude/hypoxic training is a common and popular practice for enhancing athletic performance in normoxic conditions among various athletes [8]. The most typical altitude/hypoxic training regimens proposed include living high-training high (LHTH), living high-training low (LHTL), and living low-training high (LLTH) methods. The LHTH method involves living and training at Int. J. Environ. Res. Public Health 2020, 17, 1934; doi:10.3390/ijerph17061934 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2020, 17, 1934 2 of 15 1500–4000 m in natural altitude environments, while the LHTL method involves living at or near sea level but training under a natural or simulated altitude condition of 2000–3000 m [3,4,6]. The LLTH method may be of particular interest to athletes because this training commonly involves shorter hypoxic exposure (approximately two to five sessions per week of <3 h), lower cost, lesser effort, and lesser time than the LHTH and LHTL methods [1]. Further, the LLTH method, includes interval hypoxic training (IHT), repeated sprint training in hypoxia, and resistance training in hypoxic conditions; it has become an increasingly popular altitude/hypoxic practice, where athletes live at or near sea level but train in 2000–4500 m simulated hypobaric or normobaric hypoxic conditions [9–11]. Among the various LLTH methods, IHT consists of repeated exposures to 5–7 min of steady or progressive hypoxia, interrupted by equal periods of recovery; it can modify oxygen transport and energy utilization and induce permanent modifications in cardiac function [12]. Short-term repeated exposure to hypoxic conditions with high-intensity exercise enhances athletic performance via the metabolic and oxygen utilizing capacity [11,13–15]. However, some studies have not supported the enhancing effect of high-intensity training in hypoxic conditions on athletic performance [16–19]. These conflicting results are attributed to the fact that the enhancement of athletic performance was not verified on the basis of changes in hemodynamic function. Heart rate (HR) variability (HRV) is a widely used marker reflecting cardiac modulation by sympathetic and vagal components and autonomic nervous system (ANS) activity. Dynamic adjustments in cardiac and peripheral vascular control, including their regulation by the ANS, occur in response to rapid changes in the HR [20,21]. Change in HRV with exercise training have often been interpreted as increase in vagal activity or ANS balance function, which is related to athletic performance [21]. Herzig et al. [22] reported that HRV markers of vagal activity are moderately associated with athletic performance variables, such as 10-mi race time. Dong [23] explained that HRV was becoming one of the most useful tools for tracking the time course of exercise training adaptation of athletes and for setting the optimal exercise intensity that leads to enhanced athletic performance. Therefore, it is essential to examine the effectiveness of exercise training in hypoxic conditions with changes in HRV, which is useful in enhancing the athletic performance. Exercise in hypoxic conditions acts as a stressor to yield greater physiological and metabolic functions than exercise in normoxic conditions, thereby causing changes in the neuroendocrine system and affecting immune function [24,25]. Further, exposure to hypoxic conditions stimulates the release of epinephrine in the adrenal medulla, increases the sympathetic nervous system activity, and increases the concentration of cortisol and adrenal cortical hormone in the blood [26,27]. The most representative changes in immune function following exposure to hypoxic conditions include decreased CD4+ T cell count; decreased T cell activation and proliferation; lymphocytosis; neutropenia; and inflammatory upregulation of cytokines, such as interleukin (IL)-6, IL-1, C-reactive protein, and tumor necrosis factor (TNF)-α [25,28–31]. As described above, exposure to hypoxic conditions results in a change in immune function based on various changes in the physiological, metabolic, and neuroendocrine systems. However, studies on changes in immune function following exercise training in hypoxic conditions are scarce. Considering that various hypoxic training regimens are commonly used to enhance athletic performance in normoxic conditions based on hematological and non-hematological changes, it is important to examine the effects on immune function in terms of health and conditioning. Moreover, the World Anti-Doping Agency is concerned that various hypoxic training regimens can have a potentially negative effect on health [32]. Thus, an essential task for elite athletes is to examine how exercise training in hypoxic conditions affects their immune function, and establish the efficacy and stability of hypoxic training. Therefore, this study aimed to investigate the effects of intermittent interval training on hemodynamic function, ANS function, immune function, and athletic performance of competitive middle- and long-distance runners in a hypoxic condition versus that in a normoxic condition. We hypothesized that intermittent interval training in a hypoxic condition would enhance Int. J. Environ. Res. Public Health 2020, 17, 1934 3 of 15 hemodynamic function, ANS function, and athletic performance more than in a normoxic condition, and would not adversely affect immune function in competitive middle- and long-distance runners. 2. Materials and Methods 2.1. Subjects Subjects, whose characteristics are presented in Table 1, were men and were competitive, moderately trained, middle- and long-distance runners (n = 20) registered with the Korea Association of Athletics Federations. They were assigned equally to the normoxic (NTG) and hypoxic (HTG) training group based on their body composition and athletic performance. We explained the experiment and possible adverse effects before the start of the study to participants and obtained their signed informed consent to participate in this study. This study was approved by the institutional review board of Konkuk University (7001355-2020002-HR-359) and was conducted in accordance with the provisions of the Declaration of Helsinki. Table 1. Characteristics of the athletes. Variables NTG HTG t-Value p-Value Number (n) n = 10 n = 10 - - Environmental condition (mmHg) Sea level (760 mmHg) 3000-m simulated altitude (526 mmHg) - - Age (year) 25.9 ± 1.2 26.3 ± 1.5 −0.499 0.624 Height (cm) 176.9 ± 7.6 178.2 ± 3.5 −0.514 0.616 Weight (kg) 70.8 ± 5.8 71.2 ± 6.3 −0.490 0.630 BMI (kg/m2) 23.1 ± 1.5 22.8 ± 0.9 −1.554 0.138 FFM (kg) 51.1 ± 4.4 52.1 ± 4.8 −0.490 0.630 Percent body fat (%) 17.5 ± 2.7 18.4 ± 1.8 −0.882 0.389 Values are expressed as means ± standard deviations. NTG = normoxic training group, HTG = hypoxic training group, BMI = body mass index, FFM = free fat mass. 2.2. Study Design The study design is shown in Figure 1. Twenty athletes were equally divided into the NTG (n = 10; intermittent interval training in a normoxic condition; 760 mmHg) and HTG (n = 10; intermittent interval training in a hypoxic condition; 526 mmHg; simulated altitude of 3000 m). All testing and training were performed in a 9-m (width) × 7-m (length) × 3-m (height) chamber with a temperature of 22 ± 1 ◦C and humidity of 50 ± 5% regulated by an environmental control chamber (NCTC-1, Nara control, Seoul, Korea). The present study comprised a 5-day pre-test period (i.e., 3 testing days and 1 rest day between the testing days), 6-week training period under each environmental condition, and 5-day post-test period. The post-test period began 3 d after the final training session. On the first pre- and post-testing days, blood samples were collected between 8:00 and 10:00 a.m. after 12 h of fasting for the analysis of blood variables related to immune function in the normoxic condition. Thereafter, body composition and ANS function were measured. Subsequently, the VO2max was measured to evaluate exercise performance in the afternoon. On the second pre- and post-testing days, hemodynamic function parameters were measured during a 30-min bout of submaximal cycle ergometer exercise. The exercise intensity was set at individual cycle ergometer exercise load values corresponding to 80% maximal HR (HRmax) obtained during the pre-test period. On the third testing day, a 3000-m time trial record was measured on an authorized track stadium at sea level. Int. J. Environ. Res. Public Health 2020, 17, 1934 4 of 15 Figure 1. Study design of the present study. All athletes performed the following in 90-min sessions: warm-up, interval training, and cool- down. The training frequency was 90 min, 3 d per week for 6 weeks. Warm-up and cool-down were set at 50% HRmax for each participant for 5 min, which was then increased by 10% HRmax every 5 min and performed for 15 min. The interval training sessions consisted of 10 repetitions of interval running exercise (5 min of exercise corresponding to 90–95% HRmax and 1 min of rest) on a treadmill. All exercise training sessions in the hypoxic conditions were supervised by directors, coaches, and the researchers. 2.3. Blood Composition Body composition parameters, such as weight, free fat mass, and percentage body fat were analyzed using Inbody 770 (Inbody, Seoul, Republic of Korea). 2.4. Hemodynamic Function Hemodynamic function was measured before and after training while the participants performed submaximal cycle ergometer exercise corresponding to 80% HRmax obtained during the pre-test period for 30 min at sea level [1,11]. The oxygen uptake (VO2) was measured using the K5 auto metabolism analyzer (Cosmed, Rome, Italy) and a breathing valve in the form of a facemask. The HR, stroke volume index (SVi), and cardiac output index (COi) were assessed non-invasively using a thoracic bioelectrical impedance device (PhysioFlow PF-05, Paris, France). The oxygen pulse (O2 pulse) was calculated as VO2/HR. 2.5. ANS Function ANS function was assessed by measuring HRV. After approximately 10 min of rest, four pads were placed on the wrists and ankles using an HRV meter (LAXTHA; CANS-3000, Daejeon, Republic of Korea), and participants’ HRV was measured in the resting condition. The following parameters were evaluated: standard deviation (SD) of successive differences (SDNN) and root mean square of successive differences (RMSSD) for the time domain methods and low frequency (LF) band, high frequency (HF) band, and LH/HF band ratio for the frequency domain methods [33]. Figure 1. Study design of the present study. All athletes performed the following in 90-min sessions: Warm-up, interval training, and cool-down. The training frequency was 90 min, 3 d per week for 6 weeks. Warm-up and cool-down were set at 50% HRmax for each participant for 5 min, which was then increased by 10% HRmax every 5 min and performed for 15 min. The interval training sessions consisted of 10 repetitions of interval running exercise (5 min of exercise corresponding to 90–95% HRmax and 1 min of rest) on a treadmill. All exercise training sessions in the hypoxic conditions were supervised by directors, coaches, and the researchers. 2.3. Blood Composition Body composition parameters, such as weight, free fat mass, and percentage body fat were analyzed using Inbody 770 (Inbody, Seoul, Korea). 2.4. Hemodynamic Function Hemodynamic function was measured before and after training while the participants performed submaximal cycle ergometer exercise corresponding to 80% HRmax obtained during the pre-test period for 30 min at sea level [1,11]. The oxygen uptake (VO2) was measured using the K5 auto metabolism analyzer (Cosmed, Rome, Italy) and a breathing valve in the form of a facemask. The HR, stroke volume index (SVi), and cardiac output index (COi) were assessed non-invasively using a thoracic bioelectrical impedance device (PhysioFlow PF-05, Paris, France). The oxygen pulse (O2 pulse) was calculated as VO2/HR. 2.5. ANS Function ANS function was assessed by measuring HRV. After approximately 10 min of rest, four pads were placed on the wrists and ankles using an HRV meter (LAXTHA; CANS-3000, Daejeon, Korea), and participants’ HRV was measured in the resting condition. The following parameters were evaluated: Standard deviation (SD) of successive differences (SDNN) and root mean square of successive differences Int. J. Environ. Res. Public Health 2020, 17, 1934 5 of 15 (RMSSD) for the time domain methods and low frequency (LF) band, high frequency (HF) band, and LH/HF band ratio for the frequency domain methods [33]. 2.6. Immune Function To assess immune function, white blood cell (WBC), eosinophil, neutrophil, basophil, natural killer (NK) cell, B cell and T cell counts were measured before and after the intervention. Three milliliters of blood were collected between 8:00 and 10:00 a.m. after 12 h of fasting. All blood samples were placed in an anticoagulant heparin tube and centrifuged at 3500 rpm for 10 min, and the serum was collected and rapidly frozen at −70 ◦C. Thereafter, the frozen or refrigerated serum was commissioned by the Clinical Laboratory of Green Cross Medical Foundation and analyzed using the method described below. In detail, WBC, neutrophil, eosinophil, and basophil counts were measured via flow cytometry using a cellpack kit (Sysmex, Kobe, Japan). NK cell, B cell, and T cell counts were analyzed using FC500 (Beckman Counter, CA, USA) and measured via flow cytometry using an NK cell kit (Beckman Coulter, Paris, France), a CD19-PE kit (Beckman Coulter, Paris, France), and a CD3-PC5 kit (Beckman Coulter, Paris, France), respectively. 2.7. Athletic Performance To evaluate athletic performance, VO2max was measured before and after the intervention with the modified BRUCE protocol for graded exercise testing on a treadmill (S25TX, SFET, Seoul, Korea) using a K5 breath by breath auto metabolism analyzer (K5, Cosmed, Rome, Italy). The graded exercise test was completed when the following three criteria were satisfied: (1) VO2 plateau: No further increase in oxygen use per minute even with an increase in work performed, (2) HR within 10 beats of the age-predicted HRmax: This is the basis for using participants’ HRmax as a surrogate for the VO2max when designing personal training programs, and (3) plasma (blood) lactate concentrations of >7 mmol/L. The 3000-m time trial records were measured on a 400-m track at sea level in Seoul between 9:00 and 10:00 a.m. (temperature = 22–24 ◦C; humidity = 60–80%; wind = 0–10 km/h). To avoid the effect of racing strategies, all starts were staggered by at least 2 min. 2.8. Statistical Analysis Means and SDs were calculated for each primary dependent variable. Normality of distribution of all outcome variables was verified using the Sharpiro-Wilk test. The two-way analysis (time × group) of variance with repeated measures of the “time” factor was used to analyze the effects of the training methods on each dependent variable. Partial eta-squared (η2) values were calculated as measures of the effect size. When a significant interaction effect was found, the Bonferroni post-hoc test was used to identify within-group changes over time. Additionally, the paired t-test was used to compare between the post- and pre-training values of the dependent variables in each group separately. A priori power analysis was performed with G-power for the energy metabolic parameter (VO2 during 30-min of submaximal exercise) based on previous research [1], indicating that a sample size of 14 participants (7 subjects per group) would be required to provide 80% power at an α-level of 0.05. We anticipated a dropout rate of >10% and aimed for a starting population of 20. All analyses were performed using the Statistical Package for the Social Sciences version 24.0 (IBM Corp., Armonk, NY, USA). The level of significance was set at 0.05 (a priori). 3. Results 3.1. Body Composition Data on the body composition in both groups before and after training are shown in Table 2. No significant interaction was observed in all body composition parameters, i.e., body composition did not affect the change in the other dependent variables. Int. J. Environ. Res. Public Health 2020, 17, 1934 6 of 15 Table 2. Changes in the body composition of the competitive runners before and after the tests. Variables NTG (n = 10) p-Value HTG (n = 10) p-Value p-Value (η2) Pre Post Pre Post Group Time Interaction Weight (kg) 70.8 ± 5.8 71.1 ± 5.8 0.146 71.2 ± 6.3 71.1 ± 7.3 0.945 0.555 (0.020) 0.233 (0.078) 0.269 (0.067) BMI (kg/m2) 23.1 ± 1.5 23.1 ± 1.6 0.105 22.8 ± 0.9 22.6 ± 0.9 0.390 0.136 (0.119) 0.138 (0.118) 0.919 (0.001) FFM (kg) 51.1 ± 4.4 50.6 ± 4.4 0.147 52.1 ± 4.8 52.1 ± 5.5 0.952 0.555 (0.020) 0.233 (0.078) 0.269 (0.067) Percent body fat (%) 17.5 ± 2.7 17.3 ± 2.9 0.274 18.4 ± 1.8 17.8 ± 2.0 0.275 0.477 (0.028) 0.144 (0.115) 0.549 (0.020) Values are expressed as means ± standard deviations. NTG = normoxic training group, HTG = hypoxic training group, BMI = body mass index, FFM = free fat mass. 3.2. Athletic Performance Figure 2 depicts the pre- and post-training data on athletic performance in both groups. There was a significant interaction for the VO2max (η2 = 0.686, p < 0.001). The post-hoc analysis revealed significant enhancements in both groups (NTG: p < 0.01, HTG: p < 0.001), and the improvement in VO2max was greater in the HTG than in the NTG (NTG: 1.5%, HTG: 6.3%). No significant interaction was observed for the 3000-m time trial records. Table 2. Changes in the body composition of the competitive runners before and after the tests. Variables NTG (n = 10) p-Value HTG (n = 10) p-Value p-Value (ɳ2) Pre Post Pre Post Group Time Interaction Weight (kg) 70.8 ± 5.8 71.1 ± 5.8 0.146 71.2 ± 6.3 71.1 ± 7.3 0.945 0.555 (0.020) 0.233 (0.078) 0.269 (0.067) BMI (kg/m2) 23.1 ± 1.5 23.1 ± 1.6 0.105 22.8 ± 0.9 22.6 ± 0.9 0.390 0.136 (0.119) 0.138 (0.118) 0.919 (0.001) FFM (kg) 51.1 ± 4.4 50.6 ± 4.4 0.147 52.1 ± 4.8 52.1 ± 5.5 0.952 0.555 (0.020) 0.233 (0.078) 0.269 (0.067) Percent body fat (%) 17.5 ± 2.7 17.3 ± 2.9 0.274 18.4 ± 1.8 17.8 ± 2.0 0.275 0.477 (0.028) 0.144 (0.115) 0.549 (0.020) Values are expressed as means ± standard deviations. NTG = normoxic training group, HTG = hypoxic training group, BMI = body mass index, FFM = free fat mass 3.2. Athletic Performance Figure 2 depicts the pre- and post-training data on athletic performance in both groups. There was a significant interaction for the VO2max (η2 = 0.686, p < 0.001). The post-hoc analysis revealed significant enhancements in both groups (NTG: p < 0.01, HTG: p < 0.001), and the improvement in VO2max was greater in the HTG than in the NTG (NTG: 1.5%, HTG: 6.3%). No significant interaction was observed for the 3000-m time trial records. Figure 2. Athletic performance parameters before and after the 6-week exercise training. (A) Change in the VO2max after exercise training in each environmental condition. (B) Change in the 3-km time trial record after exercise training in each environmental condition. VO2max = maximum oxygen uptake, NTG = normoxic training group, HTG = hypoxic training group. * Significant difference between the pre- and post-tests, ** p < 0.01, *** p < 0.001. 3.3. Hemodynamic Function As shown in Table 3, there was a significant interaction for the VO2 (η2 = 0.251, p < 0.05), O2 pulse (η2 = 0.588, p < 0.001), and COi (η2 = 0.575, p < 0.001). Compared with the NTG, the HTG showed a significant decrease in the VO2 (p < 0.001) and a significant increase in the COi (p < 0.001) during submaximal exercise for 30 min. The O2 pulse (NTG: p < 0.01, HTG: p < 0.01) during submaximal exercise for 30 min significantly increased in both groups, and the improvement in O2 pulse was greater in the HTG than in the NTG (NTG: 17.7%, HTG: 24.7%). No significant interaction was observed for the HR and SVi. Figure 2. Athletic performance parameters before and after the 6-week exercise training. (A) Change in the VO2max after exercise training in each environmental condition. (B) Change in the 3-km time trial record after exercise training in each environmental condition. VO2max = maximum oxygen uptake, NTG = normoxic training group, HTG = hypoxic training group. * Significant difference between the pre- and post-tests, ** p < 0.01, *** p < 0.001. 3.3. Hemodynamic Function As shown in Table 3, there was a significant interaction for the VO2 (η2 = 0.251, p < 0.05), O2 pulse (η2 = 0.588, p < 0.001), and COi (η2 = 0.575, p < 0.001). Compared with the NTG, the HTG showed a significant decrease in the VO2 (p < 0.001) and a significant increase in the COi (p < 0.001) during submaximal exercise for 30 min. The O2 pulse (NTG: p < 0.01, HTG: p < 0.01) during submaximal exercise for 30 min significantly increased in both groups, and the improvement in O2 pulse was greater in the HTG than in the NTG (NTG: 17.7%, HTG: 24.7%). No significant interaction was observed for the HR and SVi. Int. J. Environ. Res. Public Health 2020, 17, 1934 7 of 15 Table 3. Changes in hemodynamic function during submaximal cycle ergometer exercise for 30 min among the competitive runners before and after the tests. Variables NTG (n = 10) p-Value HTG (n = 10) p-Value p-Value (η 2) Pre Post Pre Post Group Time Interaction HR (beat/30 min) 5375.7 ± 431.4 4946.1 ± 287.8 0.003 5207.4 ± 308.7 4830.5 ± 267.6 0.001 0.296 (0.060) 0.000 (0.674) ††† 0.694 (0.009) VO2 (mL/30 min) 1355.6 ± 114.4 1322.9 ± 118.9 0.094 1194.1 ± 139.5 1118.1 ± 141.8 <0.001 *** 0.005 (0.364) †† 0.000 (0.678) ††† 0.024 (0.251) † O2 pulse (mL/beat/30 min) 553.5 ± 111.2 651.5 ± 49.2 0.004 ** 551.4 ± 129.1 687.7± 35.9 0.007 ** 0.614 (0.014) 0.418 (0.037) 0.000 (0.588) ††† SVi (mL/beat/30 min) 1691.5 ± 99.4 1660.7 ± 102.5 0.018 1520.4 ± 117.4 1518.2 ± 188.1 0.971 0.007 (0.342) †† 0.579 (0.017) 0.629 (0.013) COi (L/30 min) 244.7 ± 20.8 253.7 ± 15.4 0.280 281.7 ± 22.1 227.2 ± 31.4 <0.001 *** 0.526 (0.023) 0.002 (0.411) †† 0.000 (0.575) ††† Values are expressed as means ± standard deviations. NTG = normoxic training group, HTG = hypoxic training group, HR = heart rate, VO2 = oxygen consumption, O2 pulse = oxygen pulse, SVi = stroke volume index, COi = cardiac output index. Significant interaction or main effect: † p < 0.05, †† p < 0.01, ††† p < 0.001; Significant difference between the pre- and post-tests: ** p < 0.01, *** p < 0.001. Int. J. Environ. Res. Public Health 2020, 17, 1934 8 of 15 3.4. ANS Function Table 4 depicts the pre- and post-training data on HRV in both groups. Significant interaction was seen for all HRV parameters, including SDNN (η2 = 0.732, p < 0.001), RMSSD (η2 = 0.777, p < 0.001), LF band (η2 = 0.616, p < 0.001), HF band (η2 = 0.693, p < 0.001), and LF/HF band ratio (η2 = 0.420, p < 0.01). In the post-hoc analysis, the NTG showed a significant decrease in the average of all R wave to R wave intervals (mean RR; p < 0.001), SDNN (p < 0.01), RMSSD (p < 0.001), LF band (p < 0.01), and HF band (p < 0.001). In contrast, the HTG presented a significant increase in the mean RR (p < 0.01), SDNN (p < 0.01), RMSSD (p < 0.01), total power (p < 0.05), HF band (p < 0.01), and LF/HF band ratio (p < 0.01). Table 4. Changes in ANS function among the competitive runners before and after the tests. Variables NTG (n = 10) p-Value HTG (n = 10) p-Value p-Value (η 2) Pre Post Pre Post Group Time Interaction SDNN (ms) 61.0 ± 4.5 50.1 ± 6.8 0.001 ** 56.6 ± 14.4 66.9 ± 11.0 0.001 ** 0.157 (0.108) 0.827 (0.003) 0.000 (0.732) ††† RMSSD (ms) 37.6 ± 7.7 23.7 ± 4.9 <0.001 *** 32.5 ± 10.4 44.8 ± 12.9 0.002 ** 0.055 (0.190) 0.646 (0.012) 0.000 (0.777) ††† LF band (ms2) 7.1 ± 0.6 6.7 ± 0.4 0.001 ** 7.0 ± 0.7 7.2 ± 0.6 0.052 0.399 (0.040) 0.013 (0.294) † 0.000 (0.616) ††† HF band (ms2) 6.5 ± 0.4 5.9 ± 0.3 <0.001 *** 6.0 ± 1.1 6.9 ± 1.0 0.003 ** 0.455 (0.031) 0.353 (0.048) 0.000 (0.693) ††† LF/HF band ratio 1.2 ± 0.1 1.3 ± 0.2 0.207 1.5 ± 0.5 1.2 ± 0.3 0.008 ** 0.447 (0.032) 0.012 (0.304) † 0.002 (0.420) †† Values are expressed as mean ± standard deviation. ANS = autonomic nervous system, NTG = normoxic training group, HTG = hypoxic training group, SDNN = standard deviation of the NN interval, RMSSD = root mean square of successive differences, LF = low frequency, HF = high frequency. Significant interaction or main effect: † p < 0.05, †† p < 0.01, ††† p < 0.001; Significant difference between the pre- and post-tests: ** p < 0.01, *** p < 0.001. 3.5. Immune Function As shown in Table 5, there was a significant interaction for the WBC (η2 = 0.293, p < 0.05), neutrophil (η2 = 0.416, p < 0.01), monocyte (η2 = 0.580, p < 0.001), and B cell (η2 = 0.258, p < 0.05) counts. Compared with the NTG, the HTG showed a significant increase in the WBC (p < 0.05) and neutrophil (p < 0.01) counts and a significant decrease in the monocyte count (p < 0.001). Conversely, the B cell count significantly decreased (p < 0.05) in the NTG compared to that in the HTG. No significant interaction was observed for the eosinophil, basophil, NK cell, and T cell counts. Table 5. Changes in immune function among the competitive runners before and after the tests. Variables NTG (n = 10) p-Value HTG (n = 10) p-Value p-Value (η 2) Pre Post Pre Post Group Time Interaction WBC count (103/µL) 5.4 ± 0.5 5.6 ± 0.6 0.124 4.6 ± 0.8 6.4 ± 2.1 0.013 * 0.977 (0.000) 0.003 (0.387) †† 0.014 (0.293) † Eosinophil count (%) 3.8 ± 1.5 3.7 ± 1.6 0.623 3.4 ± 1.3 3.1 ± 1.4 0.153 0.459 (0.031) 0.126 (0.125) 0.286 (0.063) Neutrophil count (%) 48.5 ± 5.8 46.4 ± 4.8 0.312 44.6 ± 4.6 53.0 ± 11.1 0.004 ** 0.635 (0.013) 0.047 (0.201) † 0.002 (0.416) †† Basophil count (%) 1.3 ± 0.5 1.2 ± 0.6 0.434 0.8 ± 0.1 0.6 ± 0.1 <0.001 0.004 (0.385) †† 0.018 (0.273) † 0.343 (0.050) Monocyte count (%) 9.2 ± 2.6 9.8 ± 2.0 0.141 9.1 ± 0.7 7.3 ± 0.7 <0.001 *** 0.090 (0.151) 0.023 (0.255) † 0.000 (0.580) ††† NK cell count (%) 22.0 ± 4.9 21.3 ± 4.9 0.603 19.6 ± 5.1 19.0 ± 5.6 0.622 0.280 (0.064) 0.467 (0.030) 0.988 (0.000) B cell count (%) 14.6 ± 2.9 13.1 ± 3.8 0.048 * 16.7 ± 2.0 17.6 ± 1.6 0.229 0.008 (0.331) †† 0.548 (0.020) 0.022 (0.258) † T cell count (%) 66.6 ± 8.0 65.9 ± 10.3 0.633 70.1 ± 6.9 71.1 ± 8.1 0.126 0.249 (0.073) 0.780 (0.004) 0.256 (0.071) Values are expressed as means ± standard deviations. NTG = normoxic training group, HTG = hypoxic training group, WBC = white blood cell, NK= natural killer. Significant interaction or main effect: † p < 0.05, †† p < 0.01, ††† p < 0.001; Significant difference between the pre- and post-tests: * p < 0.05, ** p < 0.01, *** p < 0.001. 4. Discussion In the present study, we hypothesized that intermittent interval training in a hypoxic condition (simulated 3000-m, 526-mmHg hypobaric hypoxia) versus that in a normoxic condition would enhance athletic performance and improve hemodynamic function and ANS function and would not adversely affect immune function in competitive middle- and long-distance runners. Our findings were consistent with these hypotheses. 4.1. Athletic Performance Our study confirmed that intermittent interval training in a hypoxic condition improved VO2max more than that in a normoxic condition. Int. J. Environ. Res. Public Health 2020, 17, 1934 9 of 15 The oxygen transport capacity in systemic conditions is most often evaluated using VO2max [8]. Exercise training in a hypoxic condition may increase exercise performance by inducing various biochemical and structural adaptive changes in the skeletal and cardiac muscles, which favor the oxidative process and can enhance non-hematological parameters, such as exercise economy, acid-base balance, and metabolic response during submaximal exercise, ultimately leading to improved oxygen delivery and utilization capacity [1,8–11,13]. Among the various LLTH methods, IHT consists of repeated exposures to 5–7 min of steady or progressive hypoxia, interrupted by equal periods of recovery; it can modify oxygen transport and energy utilization and induce permanent modifications in the cardiac function [12]. The short-term repeated exposure to hypoxic conditions, with high-intensity exercise, enhances athletic performance via the metabolic and oxygen utilizing capacity [11,13–15]. However, research findings on IHT as an effective hypoxic training method for enhancing athletic performance at sea level are inconclusive [8]. In various previous studies, the difference in the enhancement of athletic performance, via the IHT method, was attributed to the intensity of the exercise performed in the hypoxic conditions [8,9,34]. There are also some differences in the type of exercise. Park and Lim [8] evaluated the effects of six weeks of hypoxic training on exercise performance in moderately trained competitive swimmers and reported that a moderate intensity of continuous and interval training in a hypoxic condition for six weeks resulted in an unclear change in the aerobic and anaerobic performance compared to normoxic training. They also argued that the unclear improvement after hypoxic training was attributed to the relatively low exercise intensity. Conversely, Czuba et al. [14] evaluated the effects of a three-week continuous hypoxic training with a relatively high exercise intensity corresponding to 95% lactate threshold workload on athletic performance in well-trained cyclists and reported a significant increase in exercise performance (e.g., VO2max, VO2 at the lactate threshold, maximal work load, and work load at the lactate threshold). Further, Roels et al. [35] observed a significant increase in the VO2max after seven weeks of high-intensity IHT compared to that after normoxic training. Regarding the inconsistency of previous research results, McLean et al. [34] suggested that the greater athletic performance (e.g., VO2max and time trial records) with IHT was more likely achieved if exercise training in hypoxic conditions is performed with high-intensity interval workouts. Our study also showed that intermittent interval training in a hypoxic condition was effective in enhancing athletic performance by increasing the VO2max of competitive runners compared to training in a normoxic condition. We believe that these positive results were attributable to IHT performed with high-intensity interval workouts. 4.2. Hemodynamic Function Our study verified that intermittent interval training in a hypoxic condition improved VO2, O2 pulse, and COi during submaximal exercise more than that in a normoxic condition. Hemodynamic function represents the dynamics of blood flow in systemic conditions, and the hemodynamic system continuously monitors and adjusts to the conditions in the body and its environment. Further, athletic performance is highly related to hemodynamic function, which serves as an indicator of oxygen transport and utilization capacity [13]. As mentioned earlier, various previous studies have reported that IHT effectively enhanced athletic performance by improving the metabolic responses (e.g., blood lactate level, glycolytic enzyme and glucose transport, and acid-base balance regulation) and oxygen utilization capacity [1,11,14,15]. IHT may increase exercise performance by inducing various biochemical and structural adaptive changes in the skeletal and cardiac muscles that favor the oxidative process and can enhance non-hematological parameters, such as exercise economy, acid-base balance, and metabolic response during submaximal exercise, ultimately leading to improved oxygen delivery and utilization capacity [1,8–11,13]. However, there is a lack of studies demonstrating that IHT enhanced athletic performance based on changes in hemodynamic function, which indicates oxygen delivery and utilization capacity in systemic conditions. Therefore, we examined the enhancement effect of intermittent interval training in the hypoxic condition on athletic performance in relation to Int. J. Environ. Res. Public Health 2020, 17, 1934 10 of 15 hemodynamic function and found that compared to training in a normoxic condition, intermittent interval training in a hypoxic condition effectively improved hemodynamic function by decreasing the VO2 and COi and increasing the O2 pulse during submaximal cycle ergometer exercise for 30 min among competitive runners. In the present study, the significantly improved hemodynamic function (significant decrease in the VO2 and COi and significant increase in the O2 pulse) indicates enhanced exercise economy, which is defined as the amount of energy per unit distance [4,36–38]. Exercise economy and VO2max are widely known as determinant factors of athletic performance [13,36–38]. IHT has been shown to enhance ATP re-synthesis (per 1 mole O2) and decrease ATP levels at a given exercise intensity (e.g., speed and workload) [36,38,39]. Exercise training enhances athletic performance by increasing the efficiency of oxygen transport and utilization and increases energy availability, which consequently improves the invigoration of the parasympathetic nervous system via the activation of β-adrenergic receptors in the cardiac muscles and efficiently changes cardiac function by increasing venous return [13,39]. Thus, the reduced COi and increased VO2 and O2 pulse during submaximal exercise in our study indicate improvement in the function of the heart as a pump for delivering oxygen and oxygen utilization and delivery to the muscle tissue [13,39]. Thus, the IHT in our study enhanced the VO2max and exercise economy, which are two important factors of athletic performance; this enhancement was probably affected by improvements in hemodynamic function. 4.3. ANS Function Our study proved that intermittent interval training in a hypoxic condition improved successive differences (SDNN), root mean square of successive differences (RMSSD), high frequency (HF) band, and LF/HF band ratio. HRV is a widely used marker reflecting cardiac modulation by sympathetic and vagal components and ANS activity; it is the most sensitive and reproducible marker among those obtained from tests for measuring changes in the ANS [20,40,41]. Further, HRV is caused by the interaction between the sympathetic and parasympathetic nervous system on the sinoatrial nodes. Therefore, the HRV test is generally used in the field of mental health examination and health science [20,40]. Clinical application of HRV is mainly associated with the prediction of sudden cardiac infarction and assessment of progression of cardiovascular and metabolic illness [23]. Recently, the sports science field has been using the HRV test for monitoring exercise training effects and recovery [22,23,42]. Decreases and increases in the vagal-derived indices of HRV have been shown to indicate negative, and positive adaptations, respectively, to exercise training [42]. In elite athletes, HRV changes are highly related to the efficiency of exercise training, and positive adaptations, such as increased cardiovascular fitness after exercise training, have been reported to be associated with HRV [23,36,42]. Moreover, some previous studies reported that athletic performance (e.g., VO2max) correlated with improvement in HRV [23,43]. Therefore, our study examined the effect of intermittent interval training on athletic performance along HRV parameters in a hypoxic condition versus that in a normoxic condition. We found that all HRV parameters (e.g., SDNN, RMSSD, LF band, HF band, and LF/HF band ratio) showed a significant interaction between the NTG and HTG; the SDNN, RMSSD, HF band, and LF/HF band ratio indicated greater improvements in the HTG than in the NTG. Among the HRV parameters, SDNN is an indicator of comprehensive HRV, has a high correlation with the HF band, and mainly reflects parasympathetic nervous system activity [44]. The RMSSD is an estimate of the short-term components of HRV, and the larger the value, the more physiologically healthy and relaxed it is [44,45]. The HF band mainly reflects the activity of the vagal nerve supplying the heart and is representative of parasympathetic nervous system activity [46]. Conversely, the LF band correlates with stress and reflects the sympathetic nervous system activity [47]. As the LF band increases, the overall HRV decreases and heart instability increases. The LF/HF band ratio reflects the overall balance of the ANS. A higher LF/HF band ratio indicates that the sympathetic nervous system is relatively activated or the parasympathetic nervous system activity is suppressed [21]. In the present Int. J. Environ. Res. Public Health 2020, 17, 1934 11 of 15 study, the HTG showed more significant improvements in most HRV parameters (e.g., SDNN, RMSSD, HF band, and LF/HF band ratio) than did the NTG; this positive HRV adaptation may have influenced the enhancement of exercise performance. Further, our study showed that IHT could improve the training and recovery quality, and yield more efficient exercise training effect by improving HRV. 4.4. Immune Function Our study confirmed that intermittent interval training in a hypoxic condition showed significant changes in WBC, neutrophil, monocyte, and B cell count, however, all parameters of immune function were altered after the two-week IHT was clinically within the normal range. As mentioned above, the IHT regimen in our study enhanced the athletic performance with improvement in hemodynamic function and HRV; it is critical to examine the impact of the regimen on immune function in terms of the health and conditioning of athletes. The World Anti-Doping Agency is concerned that various hypoxic training regimens can have a potentially negative effect on health [32]. Therefore, it is an imperative task for elite athletes to examine how exercise training in hypoxic conditions affects immune function and to establish the efficacy and stability of hypoxic training. Exercise in hypoxic conditions acts as a stressor to yield greater physiological and metabolic function than that in normoxic conditions, causing changes in the neuroendocrine system and consequently affecting immune function [24,25]. Representative changes in the nervous and endocrine systems by acute exposure to hypoxic conditions correspond to stimulating the release of epinephrine in the adrenal medulla, increasing the sympathetic nervous system activity, and increasing the concentration of cortisol and adrenal cortical hormones in the blood [26,27]. Changes in the neuroendocrine system by exposure to hypoxic conditions induce changes in immune function, such as decreased T cell count; decreased T cell activation and proliferation; increased neutrophil count; and upregulation of inflammatory markers, including IL-6, IL-1, C-reactive protein, and TNF-α [28,48]. As described above, exposure to hypoxic conditions has been reported to result in a change in immune function based on changes in the physiological, metabolic, and neuroendocrine systems. However, studies on changes in immune function following exercise training in hypoxic conditions are scarce. Although there are differences in the regimen of hypoxic training, Tiollier et al. [49] investigated the impact of an 18-day LHTL training camp on secretory immunoglobulin A (sIgA) levels in 11 (six female and five male) elite cross-country skiers. There was a downward trend in the sIgA levels, which reached significance in the LHTL group but not in the control group. Further, the salivary IgA levels were still lower at baseline than those post-operatively. They strongly suggested a cumulative negative effect of physical exercise and hypoxia on the sIgA levels during LHTL training. Brugniaux et al. [50] examined the effect of LHTL training performed for 13–18 d through leukocyte count evaluation in 41 athletes from 3 federations (cross-country skiers, n = 11; swimmers, n = 18; runners, n = 12) and found that the leukocyte count was not affected, except at 3500 m. Park et al. [51] recently reported a case study in which an LLTL regimen was used to evaluate the effects of a 2-week hypoxic training on immune function in Korean national cycling athletes with disabilities. They found that all immune function parameters were in the normal range even after two weeks of hypoxic training. In the present study, the HTG showed a more significant increase in the WBC and neutrophil counts and a significant decrease in the monocyte count than did the NTG. Conversely, the B cell count significantly decreased in the NTG compared to that in the HTG. However, all immune function parameters that were altered after the two-week IHT were clinically within the normal range. Thus, the six-week IHT in this study did not negatively affect the immune function of the competitive runners, which is consistent with the results of a previous study [51]. 5. Limitation of the Study Some limitations of our study should be considered when interpreting our results. Although the present study was designed systematically with equally controlled experiments, small sample sizes were a limitation to check the effects of an intermittent interval training in a hypoxic condition Int. J. Environ. Res. Public Health 2020, 17, 1934 12 of 15 versus that in a normoxic condition on athletic performance, hemodynamic function, ANS function, and immune function in middle- and long-distance competitive runners. Thus, larger samples are warranted in future studies to access sports field practice. Furthermore, the athlete’s dietary intake and conditioning were not investigated. 6. Conclusions Our study confirmed that intermittent interval training in a hypoxic condition for six weeks would enhance the athletic performance and improve hemodynamic and ANS function; further, it did not adversely affect the immune function of competitive runners compared to that of runners training in a normoxic condition. Author Contributions: Study conception and design, W.-S.J. and H.-Y.P.; statistical analysis, S.-W.K. and W.-S.J.; investigation, W.-S.J., S.-W.K., and H.-Y.P.; data interpretation, W.-S.J., S.-W.K., and H.-Y.P.; writing-original draft preparation, W.-S.J. and H.-Y.P.; writing-review and editing, W.-S.J., S.-W.K., and H.-Y.P.; supervision, W.-S.J., S.-W.K., and H.-Y.P. All authors have read and agreed to the published version of the manuscript. Funding: This work was supported by the Ministry of Education of Korea and the National Research Foundation of Korea (NRF-2019S1A5A8032648). Acknowledgments: This study was supported by the Konkuk University Research Professor Program. Conflicts of Interest: The authors declare no conflicts of interest. References 1. Park, H.Y.; Shin, C.; Lim, K. Intermittent hypoxic training for 6 weeks in 3000 m hypobaric hypoxia conditions enhances exercise economy and aerobic exercise performance in moderately trained swimmers. Biol. Sport 2018, 35, 49–56. [CrossRef] 2. Im, J.Y.; Bang, H.S.; Seo, D.Y. The Effects of 12 Weeks of a Combined Exercise Program on Physical Function and Hormonal Status in Elderly Korean Women. Int. J. Environ. Res. Public Health 2019, 16. [CrossRef] 3. Park, H.Y.; Hwang, H.; Park, J.; Lee, S.; Lim, K. The effects of altitude/hypoxic training on oxygen delivery capacity of the blood and aerobic exercise capacity in elite athletes—A meta-analysis. J. Exerc. Nutr. Biochem. 2016, 20, 15–22. [CrossRef] [PubMed] 4. Millet, G.P.; Faiss, R.; Brocherie, F.; Girard, O. Hypoxic training and team sports: A challenge to traditional methods? Br. J. Sports Med. 2013, 47, i6–i7. [CrossRef] [PubMed] 5. Kim, S.W.; Jung, W.S.; Park, W. Twelve Weeks of Combined Resistance and Aerobic Exercise Improves Cardiometabolic Biomarkers and Enhances Red Blood Cell Hemorheological Function in Obese Older Men: A Randomized Controlled Trial. Int. J. Environ. Res. Public Health 2019, 16. [CrossRef] [PubMed] 6. Jung, W.S.; Hwang, H.; Kim, J.; Park, H.Y.; Lim, K. Effect of interval exercise versus continuous exercise on excess post-exercise oxygen consumption during energy-homogenized exercise on a cycle ergometer. J. Exerc. Nutr. Biochem. 2019, 23, 45–50. [CrossRef] 7. Nystoriak, M.A.; Bhatnagar, A. Cardiovascular Effects and Benefits of Exercise. Front. Cardiovasc. Med. 2018, 5, 135. [CrossRef] 8. Park, H.Y.; Lim, K. Effects of Hypoxic Training versus Normoxic Training on Exercise Performance in Competitive Swimmers. J. Sports Sci. Med. 2017, 16, 480–488. 9. Faiss, R.; Girard, O.; Millet, G.P. Advancing hypoxic training in team sports: From intermittent hypoxic training to repeated sprint training in hypoxia. Br. J. Sports Med. 2013, 47, i45–i50. [CrossRef] 10. Girard, O.; Brocherie, F.; Millet, G.P. Effects of Altitude/Hypoxia on Single- and Multiple-Sprint Performance: A Comprehensive Review. Sports Med. 2017, 47, 1931–1949. [CrossRef] 11. Hamlin, M.J.; Marshall, H.C.; Hellemans, J.; Ainslie, P.N.; Anglem, N. Effect of intermittent hypoxic training on 20 km time trial and 30 s anaerobic performance. Scand. J. Med. Sci. Sports 2010, 20, 651–661. [CrossRef] [PubMed] 12. Bernardi, L. Interval hypoxic training. Adv. Exp. Med. Biol. 2001, 502, 377–399. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2020, 17, 1934 13 of 15 13. Park, H.Y.; Jung, W.S.; Kim, J.; Hwang, H.; Lim, K. Efficacy of intermittent hypoxic training on hemodynamic function and exercise performance in competitive swimmers. J. Exerc. Nutr. Biochem. 2018, 22, 32–38. [CrossRef] 14. Czuba, M.; Waskiewicz, Z.; Zajac, A.; Poprzecki, S.; Cholewa, J.; Roczniok, R. The effects of intermittent hypoxic training on aerobic capacity and endurance performance in cyclists. J. Sports Sci. Med. 2011, 10, 175–183. 15. Geiser, J.; Vogt, M.; Billeter, R.; Zuleger, C.; Belforti, F.; Hoppeler, H. Training high-living low: Changes of aerobic performance and muscle structure with training at simulated altitude. Int. J. Sports Med. 2001, 22, 579–585. [CrossRef] 16. Beidleman, B.A.; Muza, S.R.; Fulco, C.S.; Jones, J.E.; Lammi, E.; Staab, J.E.; Cymerman, A. Intermittent hypoxic exposure does not improve endurance performance at altitude. Med. Sci. Sports Exerc. 2009, 41, 1317–1325. [CrossRef] [PubMed] 17. Katayama, K.; Sato, K.; Matsuo, H.; Ishida, K.; Iwasaki, K.; Miyamura, M. Effect of intermittent hypoxia on oxygen uptake during submaximal exercise in endurance athletes. Eur. J. Appl. Physiol. 2004, 92, 75–83. [CrossRef] 18. Rodriguez, F.A.; Truijens, M.J.; Townsend, N.E.; Stray-Gundersen, J.; Gore, C.J.; Levine, B.D. Performance of runners and swimmers after four weeks of intermittent hypobaric hypoxic exposure plus sea level training. J. Appl. Physiol. 2007, 103, 1523–1535. [CrossRef] 19. Roels, B.; Bentley, D.J.; Coste, O.; Mercier, J.; Millet, G.P. Effects of intermittent hypoxic training on cycling performance in well-trained athletes. Eur. J. Appl. Physiol. 2007, 101, 359–368. [CrossRef] 20. Park, H.Y.; Jung, W.S.; Kim, J.; Lim, K. Twelve weeks of exercise modality in hypoxia enhances health-related function in obese older Korean men: A randomized controlled trial. Geriatr. Gerontol. Int. 2019, 19, 311–316. [CrossRef] 21. Vitale, J.A.; Bonato, M. Heart Rate Variability in Sport Performance: Do Time of Day and Chronotype Play A Role? J. Clin. Med. 2019, 8. [CrossRef] [PubMed] 22. Herzig, D.; Asatryan, B.; Brugger, N.; Eser, P.; Wilhelm, M. The Association Between Endurance Training and Heart Rate Variability: The Confounding Role of Heart Rate. Front. Physiol. 2018, 9, 756. [CrossRef] [PubMed] 23. Dong, J.G. The role of heart rate variability in sports physiology. Exp. Ther. Med. 2016, 11, 1531–1536. [CrossRef] [PubMed] 24. Mazzeo, R.S.; Donovan, D.; Fleshner, M.; Butterfield, G.E.; Zamudio, S.; Wolfel, E.E.; Moore, L.G. Interleukin-6 response to exercise and high-altitude exposure: Influence of alpha-adrenergic blockade. J. Appl. Physiol. 2001, 91, 2143–2149. [CrossRef] [PubMed] 25. Mazzeo, R.S. Physiological responses to exercise at altitude: An update. Sports Med. 2008, 38, 1–8. [CrossRef] [PubMed] 26. Beidleman, B.A.; Muza, S.R.; Fulco, C.S.; Cymerman, A.; Staab, J.E.; Sawka, M.N.; Lewis, S.F.; Skrinar, G.S. White blood cell and hormonal responses to 4300 m altitude before and after intermittent altitude exposure. Clin. Sci. 2006, 111, 163–169. [CrossRef] 27. Mazzeo, R.S.; Child, A.; Butterfield, G.E.; Braun, B.; Rock, P.B.; Wolfel, E.E.; Zamudio, S.; Moore, L.G. Sympathoadrenal responses to submaximal exercise in women after acclimatization to 4300 meters. Metab. Clin. Exp. 2000, 49, 1036–1042. [CrossRef] 28. Hartmann, G.; Tschop, M.; Fischer, R.; Bidlingmaier, C.; Riepl, R.; Tschop, K.; Hautmann, H.; Endres, S.; Toepfer, M. High altitude increases circulating interleukin-6, interleukin-1 receptor antagonist and C-reactive protein. Cytokine 2000, 12, 246–252. [CrossRef] 29. Pedersen, B.K.; Steensberg, A. Exercise and hypoxia: Effects on leukocytes and interleukin-6-shared mechanisms? Med. Sci. Sports Exerc. 2002, 34, 2004–2013. [CrossRef] 30. Thake, C.D.; Mian, T.; Garnham, A.W.; Mian, R. Leukocyte counts and neutrophil activity during 4 h of hypocapnic hypoxia equivalent to 4000 m. Aviat. Space Environ. Med. 2004, 75, 811–817. 31. Lee, Y.I.; Leem, Y.H. Acid sphingomyelinase inhibition alleviates muscle damage in gastrocnemius after acute strenuous exercise. J. Exerc. Nutr. Biochem. 2019, 23, 1–6. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2020, 17, 1934 14 of 15 32. Wilber, R.L. Application of altitude/hypoxic training by elite athletes. Med. Sci. Sports Exerc. 2007, 39, 1610–1624. [CrossRef] [PubMed] 33. Park, H.Y.; Kim, S.; Kim, Y.; Park, S.; Nam, S.S. Effects of exercise training at lactate threshold and detraining for 12 weeks on body composition, aerobic performance, and stress related variables in obese women. J. Exerc. Nutr. Biochem. 2019, 23, 22–28. [CrossRef] [PubMed] 34. McLean, B.D.; Gore, C.J.; Kemp, J. Application of ’live low-train high’ for enhancing normoxic exercise performance in team sport athletes. Sports Med. 2014, 44, 1275–1287. [CrossRef] [PubMed] 35. Roels, B.; Millet, G.P.; Marcoux, C.J.; Coste, O.; Bentley, D.J.; Candau, R.B. Effects of hypoxic interval training on cycling performance. Med. Sci. Sports Exerc. 2005, 37, 138–146. [CrossRef] 36. Green, H.J.; Roy, B.; Grant, S.; Hughson, R.; Burnett, M.; Otto, C.; Pipe, A.; McKenzie, D.; Johnson, M. Increases in submaximal cycling efficiency mediated by altitude acclimatization. J. Appl. Physiol. 2000, 89, 1189–1197. [CrossRef] 37. Saunders, P.U.; Telford, R.D.; Pyne, D.B.; Cunningham, R.B.; Gore, C.J.; Hahn, A.G.; Hawley, J.A. Improved running economy in elite runners after 20 days of simulated moderate-altitude exposure. J. Appl. Physiol. 2004, 96, 931–937. [CrossRef] 38. Gore, C.J.; Hahn, A.G.; Aughey, R.J.; Martin, D.T.; Ashenden, M.J.; Clark, S.A.; Garnham, A.P.; Roberts, A.D.; Slater, G.J.; McKenna, M.J. Live high:Train low increases muscle buffer capacity and submaximal cycling efficiency. Acta Physiol. Scand. 2001, 173, 275–286. [CrossRef] 39. Park, H.Y.; Park, W.; Lim, K. Living High-Training Low for 21 Days Enhances Exercise Economy, Hemodynamic Function, and Exercise Performance of Competitive Runners. J. Sports Sci. Med. 2019, 18, 427–437. 40. Kim, J.; Park, H.Y.; Lim, K. Effects of 12 Weeks of Combined Exercise on Heart Rate Variability and Dynamic Pulmonary Function in Obese and Elderly Korean Women. Iran. J. Public Health 2018, 47, 74–81. 41. Alansare, A.; Alford, K.; Lee, S.; Church, T.; Jung, H.C. The Effects of High-Intensity Interval Training vs. Moderate-Intensity Continuous Training on Heart Rate Variability in Physically Inactive Adults. Int. J. Environ. Res. Public Health 2018, 15. [CrossRef] [PubMed] 42. Plews, D.J.; Laursen, P.B.; Stanley, J.; Kilding, A.E.; Buchheit, M. Training adaptation and heart rate variability in elite endurance athletes: Opening the door to effective monitoring. Sports Med. 2013, 43, 773–781. [CrossRef] [PubMed] 43. Hautala, A.; Tulppo, M.P.; Makikallio, T.H.; Laukkanen, R.; Nissila, S.; Huikuri, H.V. Changes in cardiac autonomic regulation after prolonged maximal exercise. Clin. Physiol. 2001, 21, 238–245. [CrossRef] [PubMed] 44. Lucini, D.; Norbiato, G.; Clerici, M.; Pagani, M. Hemodynamic and autonomic adjustments to real life stress conditions in humans. Hypertension 2002, 39, 184–188. [CrossRef] [PubMed] 45. Licht, C.M.; de Geus, E.J.; van Dyck, R.; Penninx, B.W. Association between anxiety disorders and heart rate variability in The Netherlands Study of Depression and Anxiety (NESDA). Psychosom. Med. 2009, 71, 508–518. [CrossRef] [PubMed] 46. Kemp, A.H.; Quintana, D.S.; Felmingham, K.L.; Matthews, S.; Jelinek, H.F. Depression, comorbid anxiety disorders, and heart rate variability in physically healthy, unmedicated patients: Implications for cardiovascular risk. PLoS ONE 2012, 7, e30777. [CrossRef] 47. Francis, J.L.; Weinstein, A.A.; Krantz, D.S.; Haigney, M.C.; Stein, P.K.; Stone, P.H.; Gottdiener, J.S.; Kop, W.J. Association between symptoms of depression and anxiety with heart rate variability in patients with implantable cardioverter defibrillators. Psychosom. Med. 2009, 71, 821–827. [CrossRef] 48. Pyne, D.V.; McDonald, W.A.; Morton, D.S.; Swigget, J.P.; Foster, M.; Sonnenfeld, G.; Smith, J.A. Inhibition of interferon, cytokine, and lymphocyte proliferative responses in elite swimmers with altitude exposure. J. Interferon Cytokine Res. 2000, 20, 411–418. [CrossRef] 49. Tiollier, E.; Schmitt, L.; Burnat, P.; Fouillot, J.P.; Robach, P.; Filaire, E.; Guezennec, C.; Richalet, J.P. Living high-training low altitude training: Effects on mucosal immunity. Eur. J. Appl. Physiol. 2005, 94, 298–304. [CrossRef] Int. J. Environ. Res. Public Health 2020, 17, 1934 15 of 15 50. Brugniaux, J.V.; Schmitt, L.; Robach, P.; Jeanvoine, H.; Zimmermann, H.; Nicolet, G.; Duvallet, A.; Fouillot, J.P.; Richalet, J.P. Living high-training low: Tolerance and acclimatization in elite endurance athletes. Eur. J. Appl. Physiol. 2006, 96, 66–77. [CrossRef] 51. Park, H.Y.; Jung, W.S.; Kim, J.; Hwang, H.; Kim, S.W.; An, Y.; Lee, H.; Jeon, S.; Lim, K. Effects of 2-Week Exercise Training in Hypobaric Hypoxic Conditions on Exercise Performance and Immune Function in Korean National Cycling Athletes with Disabilities: A Case Report. Int. J. Environ. Res. Public Health 2020, 17. [CrossRef] [PubMed] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Interval Hypoxic Training Enhances Athletic Performance and Does Not Adversely Affect Immune Function in Middle- and Long-Distance Runners.

03-16-2020

Jung, Won-Sang,Kim, Sung-Woo,Park, Hun-Young

eng

PMC10356634

Vol.:(0123456789) Sports Medicine (2023) 53:1595–1607 https://doi.org/10.1007/s40279-023-01850-z SYSTEMATIC REVIEW Long‑Term Development of Training Characteristics and Performance‑Determining Factors in Elite/International and World‑Class Endurance Athletes: A Scoping Review Hanne C. Staff1  · Guro Strøm Solli1,2 · John O. Osborne1  · Øyvind Sandbakk1,3 Accepted: 31 March 2023 / Published online: 13 May 2023 © The Author(s) 2023 Abstract Objective In this scoping review, we aimed to 1) identify and evaluate existing research that describes the long-term develop- ment of training characteristics and performance-determining factors in male and female endurance athletes reaching an elite/ international (Tier 4) or world-class level (Tier 5), 2) summarize the available evidence and 3) point out existing knowledge gaps and provide methodological guidelines for future research in this field. Methods This review was conducted following the Joanna Briggs Institute methodology for scoping reviews. Results Out of 16772 screened items across a 22-year period (1990-2022), a total of 17 peer-reviewed journal articles met the inclusion criteria and were considered for further analysis. These 17 studies described athletes from seven different sports and seven different countries, with 11 (69%) of the studies being published during the last decade. Of the 109 athletes included in this scoping review, one quarter were women (27%), and three quarters were men (73%). Ten studies included information about the long-term development of training volume and training intensity distribution. A non-linear, year-to-year increase in training volume was found for most athletes, resulting in a subsequent plateau. Furthermore, 11 studies described the development of performance determining factors. Here, most of the studies showed improvements in submaximal variables (e.g., lactate/anaerobic threshold and work economy/efficiency) and maximal performance-indices (e.g., peak speed/watt during performance testing). Conversely, the development of VO2max was inconsistent across studies. No evidence was found regarding possible sex differences in development of training or performance-determining factors among endurance athletes. Conclusion Overall, a low number of studies describing the long-term development of training and performance-determining factors is available. This suggests that existing talent development practices in endurance sports are built upon limited sci- entific evidence. Overall, there is an urgent need for additional long-term studies based on systematic monitoring of athletes from a young age utilizing high-precision, reproducible measurements of training and performance-determining factors. * Hanne C. Staff [email protected] Guro Strøm Solli [email protected] John O. Osborne [email protected] Øyvind Sandbakk [email protected] 1 School of Sport Sciences, UiT The Arctic University of Norway, Campus Tromsø, Hansine Hansens veg 18, 9019 Tromsø, Norway 2 Department of Sports Science and Physical Education, Nord University, Bodø, Norway 3 Department of Neuromedicine and Movement Science, Centre for Elite Sports Research, Norwegian University of Science and Technology, Trondheim, Norway 1596 H. C. Staff et al. Key Points Only 17 studies described the long-term development of training characteristics and performance-determining factors of elite/international and world-class athletes, with 16 studies using a retrospective study design, 11 studies being case studies, and the majority of partici- pants being male. A non-linear year-to-year increase in training volume, mainly driven by increases in low-intensity training until reaching a subsequent plateau at elite/international and world-class level was found for most of the included endurance athletes. Consistent improvement in maximal performance tests and submaximal performance indices were found for most athletes, while the developments in maximal oxy- gen uptake were inconsistent across studies. This scoping review highlights an urgent need for addi- tional long-term studies based on systematic monitoring of athletes and suggests that a common framework is required for comparing results across different long-term studies in endurance sports. 1 Introduction Long-term performance development in endurance sports is determined by a multifaceted interaction of manifold vari- ables. Extensive sport-specific practice, including optimal progression of training volume, frequency, and intensity distribution, is required to stimulate sport-specific adap- tive responses. This process normally requires a relatively long period (10–15 years) of dedicated training, although recent studies report considerable variation within and across sports in the amount of training and the time needed to reach elite and super-elite levels [1–4]. In addition to the obvious role of the genetic potential, the realization of ath- letes’ potential is also influenced by motivation, skillset and experience of the athlete and coach, training peers, support- ing staff, training environment and facilities, well-being, and life balance [5, 6]. The training characteristics among elite/international and world-class athletes in endurance sports have been widely described in retrospective studies [7–10]. The outcomes from this research have emphasized the importance of high- endurance training volumes (TV) with sport-specific dif- ferences owing to variations in muscular loads and injury risks across exercise modalities [10]. Furthermore, there is an established consensus that a relatively long period of dedicated training is required to tolerate these TV and reach an elite level [4, 11–13]. Accordingly, gradual progression in TV is required to tolerate and respond positively to the overall training load. However, training load can also be manipulated by changing the intensity and/or frequency of training, although limited evidence exists on how the pro- gressive increase in these factors interacts to provide the best possible training stimulus and to avoid setbacks, thereby ensuring continuity to optimize the development of physi- ological factors and performance [9, 14]. Describing and comparing the intensity distribution of endurance training (TID) across different studies and ath- letes necessitates a standardized intensity scale. Here, a three-zone model is often used, with the zones referred to as: low-intensity training (LIT), moderate-intensity train- ing (MIT), and high-intensity training (HIT). Although both conceptual and practical challenges are associated with the division of intensity zones, the separation of each zone using reproducible blood lactate anchor points, combined with corresponding heart rate and ratings of perceived exertion, is arguably the most effective available method [9, 15]. Other methods that are used to determine intensity zones include ventilatory thresholds or critical power [16, 17]. Although there are differences in the methods for quantifying training intensity, there seems to be similarities in the basic TID pat- terns selected by successful endurance athletes [9]. Previous studies report that the training of successful endurance ath- letes include 70–90% LIT, with the remaining 10–30% per- formed as MIT and HIT [9, 18, 19]. This variation in TID is likely caused by differences in the demands of the examined sports, individual development areas, and the methodology used to determine LIT, MIT, and HIT [10, 20, 21]. Still, it is unclear if the same TID should be employed in all stages of the development process in an endurance athlete’s career. Successful endurance performance is characterized by high levels of maximal oxygen uptake (VO2max), anaerobic threshold or lactate threshold, and work economy or effi- ciency [22]. However, the long-term development of these performance-determining factors is influenced by vari- ous aspects such as training, psychophysiological matura- tion, and sex, resulting in different developmental patterns throughout an athlete’s career [23]. Therefore, an overview of the studies including information about the long-term development of training characteristics and performance- determining factors of elite/international and world-class athletes would provide a starting point for better under- standing the long-term development process of endurance athletes. Accordingly, this scoping review aimed to (1) iden- tify and evaluate existing research that has focused on the long-term development of training characteristics and 1597 Long-Term Development of Training Characteristics and Performance-Determining Factors performance-determining factors in male and female endur- ance athletes reaching an elite/international or world-class level, (2) summarize the available evidence, and (3) point out existing knowledge gaps and provide methodological guidelines for future research in this field. 2 Methods This scoping review was conducted following the Joanna Briggs Institute methodology for scoping reviews [24]. The review protocol and search results for each step of the review are available on the Open Science Framework (https:// osf. io/ b3fwu/). The Preferred Reporting Items for Systematic reviews and Meta-Analyses extension for Scoping Reviews Checklist (PRISMA-ScR) was followed step by step [25]. An initial limited search of PubMed was undertaken to identify potentially relevant articles. The words contained in the titles and abstracts of relevant articles, and the index terms used to describe the articles, were then utilized to develop a full search strategy. Broad inclusion criteria were initially employed to increase the probability of mapping the existing literature of interest and obtaining a compre- hensive list of articles. The search strategy (Table 1), includ- ing all identified keywords and index terms, was adapted for use across four major databases: PubMed, PsychINFO, SPORTDiscus, and Web of Science. Boolean search terms were used to link nested concepts. Once the search strategy was completed, search results were collated and exported to EndNote referencing soft- ware (version X9.3.3; Clarivate Analytics, Philadelphia, PA, USA). Duplicates were removed using the duplication detection tool of the Endnote software, before all remaining unique records were made available to reviewers for further processing (i.e., study screening and selection). In addition to the systematic search of the four primary databases, an additional search was performed using Google Scholar, with the first 200 results exported for further screening. The ini- tial database search output can be viewed at https:// osf. io/ b3fwu/. The types of publications included in the first stage of the literature review were: peer-reviewed journal papers (pub- lished between the period 1 January, 1990 and 8 December, 2022, written in English and involving human participants), reviews, and meta-analyses; while non-peer reviewed arti- cles published in magazines, unpublished doctoral disser- tations, and masters’ theses were excluded. Both quantita- tive, qualitative, and mixed-method studies were included to consider different aspects of the development process. To chart data related to long-term development, the stud- ies were included if training or physiological characteristics were reported for ≥ 2 years. The participant classification Table 1 Search strategy, including all identified keywords and index terms MeSH Medical Subject Headings, RPE ratings of perceived exhaustion An asterisk (*) indicates a Boolean operator for truncation searching from the word stem, while a question mark (?) represents a wild card replacement of a single letter MeSH terms Concept 1 Concept 2 Concept 3 Concept 4 Athletic level Population Sex differences Training characteristics Athletes Sports Elite Professional Medalist Olympic “High performance” “World class” “World champion” “Highly trained” Endurance Aerobic Cycli* Skier* “Cross country” Skiing Runn* Triat* Biath* Swim* Rowing Rower Orienteer* “Long distance” Marathon Athletics Skating Biking Female Woman/Women Girl Male Man/Men Boy Sex Gender Training Endurance Load TRIMP Intensity Speed Velocity Frequency Volume Distance Distribution Time RPE Mode Modality Movement Activity Terrain Periodi?ation Tapering Peaking Altitude Progression Longitudinal 1598 H. C. Staff et al. framework of McKay et al. [26] was used and only stud- ies with participants classified as Tier 4 (elite/international level) or Tier 5 (world-class level) were included. The review process consisted of three levels of screening: (1) an initial title screening; (2) an abstract review; and (3) a full-text review. Two investigators (HS and JOO) inde- pendently screened all articles against the forementioned inclusion and exclusion criteria and then compared results. Where consensus was not reached, it was resolved by means of consolidation with a third independent researcher (GSS). Reasons for the exclusion of any full-text source are reported in the scoping review report. The search results are presented in a PRISMA flow diagram (Fig. 1) [27, 28]. Following the final full-text review screening step, an expert panel (n = 6) of experienced academics in exercise physiology and athlete development was assembled to review the included stud- ies and suggest any additional relevant articles that could be considered for inclusion. Snowball searching was also employed on the reference lists of the included studies, to identify any other relevant sources. A data extraction form was developed and key infor- mation on the selected articles, population, concept, and context was collected. This form was reviewed and tested by all research team members before implementation, to ensure that the form accurately captured the necessary data. Extracted study variables included: primary author, year of publication, athletes’ country, study aim/purpose, sample description and size, participant details, study methodology, body composition, training characteristics (TV, TID), physi- ological characteristics (VO2max, submaximal responses, per- formance indicators), and performance. The charting process was an iterative process with three researchers (HS, JOO, and GSS) extracting the data. 3 Results 3.1 Study Characteristics A total of 17 peer-reviewed journal articles were included. Sixteen of these studies used a retrospective study design, with a mean duration of ~ 7 years (range 2–17 years). Out of the 17 studies, ten included men exclusively, five included only women, and two included a mix of men and women. Cumulatively, the studies included a total of 109 partici- pants, with approximately a quarter (n = 29; 27%) being women. The two studies that included both sexes represented two-thirds (n = 73; 67%) of the total participants, with a total of 24 women and 49 men. The five women-only studies were all individual case studies, accounting for just 5% of the 109 Records identified from databases (n = 16,772) • PubMed (n = 2,728) • SPORTDiscus (n = 5,653) • Web of Science (n = 7,656) • PsycInfo (n = 735) Duplicates removed before screening (n = 4,411) Records title screened (n = 12,361) Records excluded at title screening (n = 11,390) Full-text articles assessed for eligibility (n = 132) • Databases (n = 121) • Google Scholar (n = 11) Full-text articles excluded: • Intervention duration too short (n = 33) • Out of scope (n = 26) • Not elite-level athletes (n = 24) • Review paper (n = 18) • Age (n = 10) • Not endurance exercise (n = 4) • No development (n = 2) • Not English language (n = 1) • Cannot access full-text (n = 1) Records identified from Google Scholar (n = 200) Additional records identified from: • Expert reference group (n = 6) • Citation searching (n = 6) Full-text articles excluded: • Previous excluded (n = 1) • Out of scope (n = 2) • Not elite-level athletes (n = 3) • Not endurance exercise (n = 2) Studies included (n = 17) • Database (n = 13) • Alternative sources (n = 4) Identification of studies via databases Identification of studies via other methods Identification Screening Included Records removed before screening (n = 78) • Duplicate records (n = 7) • Duplicate to database search (n = 71) Records abstract screened (n = 122) Records excluded at abstract screening (n = 111) Records abstract screened (n = 971) Records excluded at abstract screening (n = 850) Eligibility Fig. 1 Preferred Reporting Items for Systematic reviews and Meta- Analyses (PRISMA) flow diagram showing the flow of informa- tion through the review process [28]. From Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting system- atic reviews. BMJ 2021;372:n71. https:// doi. org/ 10. 1136/ bmj. n71 1599 Long-Term Development of Training Characteristics and Performance-Determining Factors total participants, while men-only studies represented 28% (n = 31). A total of 11 different studies were individual case reports. Athletes from seven Olympic endurance sports were represented in the study, middle- and long-distance run- ning (n = 41); swimming (n = 41); cycling (n = 13); rowing (n = 6); triathlon (n = 6); biathlon (n = 1); and cross-country skiing (n = 1), while only one athlete represented the Para- lympic disciplines (swimming). The majority of included studies (n = 11; 65%) were published after 2010. Athletes from seven countries were included, with the majority of athletes (85%) from Spain (n = 52) and Australia (n = 40), and the remaining from Norway (n = 7), Croatia (n = 4), UK (n = 3), France (n = 2), and Belgium (n = 1). 3.2 Training Characteristics The ten studies that provided information about the long- term development of training characteristics are presented in Table 2. Nine of the studies were individual case studies that were conducted on athletes in cross-country skiing, biathlon, running, cycling, rowing, and para-swimming. Six studies described training data that ranged from 6 to 17 years dura- tion. No information about training before a junior age was reported by any of the studies. Table 2 includes a summary of TV and TID from the included studies. Other important training characteristics such as training frequency, strength, speed, and altitude training were rarely described and are not included in Table 2. Specifically, four studies included information about training frequency [29–32], three stud- ies reported strength and speed training [30, 33, 34], and four studies included information about the use of altitude training.[30, 33–35]. One study had a detailed description of the altitude training during the 5 most successful years (30–35 years of age) but no information about altitude train- ing from earlier years was presented [30]. The other studies only briefly described that altitude training was employed, without providing any detailed data. 3.2.1 Training Volume In total, eight studies reported a progressive non-linear increase in TV [29, 30, 32–37]. Two female world-class athletes, from cross-country skiing and marathon running, had relatively low TV at a junior age, and increased their TV by 80–500% over a 10- to 12-year period, from 18 to 20 years of age until the age of peak performance [30, 35]. A similar pattern was seen in two male athletes, from row- ing and cycling, with a 50–80% increase in TV from the age of 18–23 years [29, 36], in one female para-swimmer with an almost 70% increase in TV from the age of 23–26 years [34] and in two male middle-distance runners from the age of 17–21 years with TV increases of approximately 50% and 66% [32]. In contrast, a much lower increase in TV (30%) was reported in a world-class male biathlete from the age of 21–31 years [33]. Three studies reported a plateau in TV (500–900 h·year−1) between the ages of 26–30 years [30, 31, 33]. Particularly large increases in TV were observed to occur relatively early in the development process, such as a 60% increase in TV from the age of 20–24 years in a world- class female cross-country skier and a 60% TV increase from the age of 18–20 years for a male Spanish cyclist [30, 36]. 3.2.2 Training Intensity Distribution Training intensity distribution was described in six indi- vidual case studies [29, 30, 33, 34, 37, 38]. One of these studies reported increased LIT and MIT, and an associ- ated decrease in the amount of HIT, at a later stage in the career of a female world-class cross-country skier [30]. Two studies showed a change towards a higher volume of both LIT and HIT, but reduced volume in MIT, for male rowers (number of kilometers rowed per week) and long-distance runners (relative distribution) [29, 38]. In contrast, a middle- distance runner reported an increase in the number of kilo- meters run per week at both LIT, MIT, and HIT from the age of 17–22 years [37]. Finally, a relatively stable TID was reported over 10 years in a world-class male biathlete and over 4 years in a world-class para-swimmer [33, 34]. 3.3 Performance‑Determining Factors The 11 studies that describe the development of physiologi- cal parameters are presented in Table 2. Five of the studies were individual case studies and described world-class ath- letes in cross-country skiing, rowing, and running. Only two studies included both male and female athletes. An increase in VO2max was reported in four studies [29, 38–40]. The relative (i.e., body mass normalized) VO2max of a male rower increased by 4% from the age of 25 years until he retired at 32 years [39]. Two other studies on male rowers found a 29% absolute to 26% relative increase in VO2max from 16 to 20 years [29, 40]. In one of these stud- ies, a further 13% increase was observed from the age of 20–27 years, before stabilizing at 28 years [29]. An 11% increase in relative VO2max was also reported in a male mid- dle-distance runner who altered his TID by increasing the proportion of LIT and HIT, but decreasing MIT, over two consecutive seasons [38]. Five studies found no change in relative values of VO2max of elite/international and world-class level athletes in long- distance running, triathlon, cycling, and cross-country skiing [30, 35, 41–43]. Six studies described improvements in sub- maximal performance-determining variables (e.g., lactate/ anaerobic threshold and/or economy/efficiency) [29, 30, 35, 38, 40, 44] and six studies showed improvements in per- formance indicators (e.g., maximal speed, maximal power 1600 H. C. Staff et al. Table 2 Overview of the development of training characteristics and physiological-determining factors References Participants Nation and sport n (sex) Age (y) Time period (y) Training characteristics TV development TID development Performance-determining factors VO2max Submaximal responses Maximal performance indicators Female population  [30] Norwegian XC skier; Tier 5; (♀ = 1); 18–35 y 17 Age 20–35 y.: ↑TV 522–940 h·y−1 (+ 80%) ↑LIT ~ 430 to ~ 800 h·y−1 Age 20–23 and 29–35 y: MIT + HIT ~ 60 h·y−1 Age 23–28 y: MIT + HIT ~ 80 h·y−1 Age 20–27 y: LIT/MIT/ HIT ~ 88/2/10% Age 28–35 y: LIT/MIT/HIT ~ 92/3/5% Age 30–35 y (five most successful seasons): ↔ VO2max = 67.7 ± 1.7 mL·kg−1·min−1 ↔ vAT = 10.7 ± 0.4 km·h−1 (running @10.5% inclination) ↔ AT = ~ 89% of VO2max  [35] British marathon runner; Tier 5; (♀ = 1); 19–30 y 12 Age 18–29 y: ↑TV from 25–30 miles·wk−1 to 120–160 miles·wk−1 (380–433%) ↑TV from 40–48 km·wk−1 to 193–258 km·wk−1 (382–438%) Age 18–29 y: VO2max = ~ 70 mL·kg−1·min−1 (range 65–80 mL·kg−1·min−1) ↑vVO2max = 20.5–23.5 km·h−1 (+ 15%) ↓VO2 @16 km·h−1 = 205– 175 mL·kg−1·km−1 (− 15%) ↑LT = ~ 15.5 km·h−1 to ~ 17.5– 18.5 km·h−1 (+ 13–19%)  [44] British marathon runner; Tier 5; (♀ = 1); 18–22 y 5 Age 18–22 y: ↓VO2max = 72.8–66.7 mL·kg−1·min−1 (− 8%) ↑LT = 15–18 km·h−1 (+ 20%) ↑vVO2max = 19.5–22 km·h−1 (+ 13%) ↓VO2@16 km·h−1 = 53–48 mL·kg−1·m in−1 (+ 9%)  [34] Norwegian Paralympic swimmer; Tier 5; (♀ = 1); 23–26 y 4 Age 23–26 y: ↑TV 388–656 h·y−1 (+ 69%), 1126– 1993 km·y−1 (+ 77%) ↔ LIT/MIT/HIT ˃90/2–4/3–6% (of total training)  [31] Norwegian marathon runner; Tier 5; (♀ = 1); 25–26 and 29–30 y 2 Age 25–26 y (track focus): TV = 123 km·wk−1 (119– 132 km·week−1) Age 29–30 y (marathon focus): TV = 121 km·wk−1 (first 26 weeks of the year) ↓TV than current marathon runners Similar TV when changing from track races to marathon races Male population  [29] Belgium rower; Tier 5; (♂ = 1); 16–30 y 15 From junior (18 y) to senior (23 y): ↑TV 4372–6091 km·y−1 (+ 39%), 11.3–17.2 h·wk−1 (+ 52%) ↑LIT = 4021–5664 km·y−1 (+ 40%) ↓MIT = 218–121 km·y−1 (-44%) ↑HIT = 87–280 km·y−1 (+ 221%) Age 16–20 y: ↑VO2max = 4.1–5.3 L·min−1 (+ 29%) Age 20–27 y: ↑VO2max = 5.3–6.0 L·min−1 (+ 13%) Age 27–30 y: ↔ VO2max = 6.0 L·min−1 Age 16–27 y: ↑POLa4 = 200–404 W (+ 101%) Age 27–30 y: ↔ POLa4 = 396 W Age 16–25 y: ↑POmax = 330–536 W (+ 62%) Age 25–30 y: ↔ POmax = 536 W 1601 Long-Term Development of Training Characteristics and Performance-Determining Factors Table 2 (continued) References Participants Nation and sport n (sex) Age (y) Time period (y) Training characteristics TV development TID development Performance-determining factors VO2max Submaximal responses Maximal performance indicators  [33] French biathlete; Tier 5; (♂ = 1); 19–31 y 11 Age 19–31 y: ↑TV = ~ 530–700 h·y−1 (+ 32%) Age 19–30 y: ↔ LIT/MIT/HIT ~ 86/3/4% (of total training) Age 30–31 y: ↑MIT = 7.4% (of total training)  [39] French rower; Tier 5; (♂ = 1); 26–36 y 10 Age 26–32 y: ↑VO2max = 67.6–70.7 mL·kg−1·min−1 (+ 5%) ↑POmax = 455–461 W (+ 1%)  [40] Croatian rowers; Tier 5; (♂ = 4); 16–21 y 6 Age 16–20 y: ↑VO2max = 61.5– 69.7 ml·kg−1·min−1 (+ 26%) Age 20–21 y: ↔ VO2max = 69.7 mL·kg −1·min−1 Age 16–21 y: ↑POAT = 297–359 W (+ 21%) ↑POVO2max = 400–481 W (+ 20%)  [36] Spanish cyclist; Tier 4; (♂ = 1); 18–23 y 6 Age 18–23 y: ↑TV = 526–943 h·y−1 (+ 79.2%), 14.733–29.383 km·y−1 (+ 100%) Large increase (60–62%) before becoming professional, but smaller increases afterwards  [32] Norwegian MD runners; Tier 5; (♂ = 3; HI, FI, and JI); 17–28 y 6 HI age 17–21 y: ↑TV = 100–110 to 156 km·wk−1 (~ + 50%) FI age 17–20 y: ↑TV = 70–80 to 120–130 km·wk−1 (~ + 66%) JI age 17–18 y: TV = 130– 140 km·wk−1 All had a similar TV = 150– 160 km·wk−1 in the 2019 preparation period (HI 28 y, FI 26 y, JI 19 y)  [37] Norwegian MD runner; Tier 5; (♂ = 1); 17–21 y 5 Age 17–21 y: ↑TV = 100–110 to 145–160 km·wk−1 (~ + 50%) ↑LIT = ~ 80 to ~ 110 km·wk−1 (~ + 37%) ↑MIT = ~ 10 to ~ 20 km·wk−1 (~ + 100%) ↑HIT = ~ 2 to ~ 3 km·wk−1 (~ + 50%) Training recorded 10 weeks. January- March  [43] Spanish cyclists; Tier 5; (♂ = 12); 22–27 y 5 Age 22–27 y: ↔ VO2max = range 75.5– 77.3 mL·kg−1·min−1 ↑DE = 24–27%  [42] Spanish triathletes; Tier 4; (♂ = 6); 24–25 y 2 Age 24–25 y: ↔ VO2max = range 77.8– 77.4 mL·kg−1·min−1 ↔ POmax = range 5.7–5.9 W·kg−1) 3-km TT (as running and running after cycling) did not change 1602 H. C. Staff et al. output, and speed at VO2max) [29, 35, 38–40, 44] over dura- tions of 2–17 years in world-class runners, cross-country skiers, and rowers. 4 Discussion This scoping review aimed to (1) identify and evaluate exist- ing research that has focused on the long-term development of training characteristics and performance-determining fac- tors in male and female endurance athletes reaching an elite/ international or world-class level, (2) summarize the avail- able evidence, and (3) point out existing knowledge gaps and provide methodological guidelines for future research in this field. In total, 17 studies were included in the review, with all but one using a retrospective study design and the major- ity of participants being male. A non-linear year-to-year increase in TV was reported for most athletes, resulting in a plateau at the elite/international and world-class levels. Only six case studies reported details about the development of TID, with all showing an increased volume of LIT while the long-term changes in MIT and HIT distribution varied across studies. Improvements in submaximal performance- determining factors (e.g., lactate/anaerobic threshold and work economy/efficiency) and various performance indices (e.g., peak speed/watt during performance testing) were reported for seven of the studies, with inconsistent find- ings in the ten studies reporting long-term development of VO2max. No evidence regarding possible sex differences in the development of training or performance-determining variables among endurance athletes reaching an elite/inter- national or world-class level was described for any of the included studies. 4.1 Study Characteristics Only studies with elite/international or world-class level athletes (i.e., performance level Tier 4 and 5) as classi- fied according to the definition by McKay et al. [26] were included in the review. Accordingly, this criterion decreased the pool of potentially relevant research, and of the included studies, the majority had small sample sizes (n < 5). A pos- sible explanation for the limited number of relevant stud- ies is the lack of systematic monitoring of elite/world-class athletes and/or restrictions on publishing unique data from such individuals. It is understandable that athletes may wish minimal distractions during their sporting careers, and that national federations likely want to maintain secrecy to gain a competitive advantage in the short-term perspective. How- ever, we believe that systematic monitoring and publishing of long-term athletic data would benefit the sporting com- munity at large, by contributing to the body of literature regarding elite-level training and athletic development. The majority of the included research in this scoping review were case studies, which are considered the weakest form of scientific evidence and limit the possibility for gen- eralization of the findings. Still, the case studies provide rich Table 2 (continued) References Participants Nation and sport n (sex) Age (y) Time period (y) Training characteristics TV development TID development Performance-determining factors VO2max Submaximal responses Maximal performance indicators  [38] British MD runner; Tier 5; (♂ = 1); 26–27 y 2 Age 26–27 y: ↔ TV = range 112–114 km·wk−1 ↑LIT, ↓MIT, ↑HIT (absolute numbers not reported) Age 26–27 y: ↑ VO2max 70.5–78.,5 mL·kg−1·min−1 (+ 11%) ↑LT = 16–18 km·h−1 (+ 13%) ↑vVO2max = 10.4–23.1 km·h−1 (+ 13%) Mixed population  [45] Australian swimmers; Tier 4; (♀ = 16, ♂ = 24); 19–25 y (♂); 18–24 y (♀) 6 Age 19–25 y (♂) and 18–24 (♀) y: ↑LT = 1.2% annual increase (♀) ↑maximal speed = 0.6–1.0% annual increase (♂/♀)  [41] Spanish MD and LD runners; Tier 4; (♀ = 8, ♂ = 25); 23–26 y (♂); 26–29 y (♀) 3 Age 23–26 (♂) and 26–29 (♀) y: ↔ VO2max = ♂ ~ 76 mL·kg−1·min−1 and ♀ ~ 70 mL·kg−1·min−1 Running performance (800-mara- thon): + 1.8% (♂) and + 0.7% (♀) AT anaerobic threshold, DE delta efficiency, HIT high-intensity training, LD long-distance, LIT low-intensity training, LT lactate threshold, MD middle-distance, MIT moderate-intensity training, PO power output, TID training intensity distribution, TT time trial, TV training volume, v velocity, VO2max maximal oxygen uptake, XC cross-country, y years, ↑ increase, ↓ decrease, ↔ stabilized, ♀ female, ♂ male 1603 Long-Term Development of Training Characteristics and Performance-Determining Factors in-depth material on unique world-class level athletes such as Grete Waitz, Paula Radcliffe, Marit Bjørgen, Martin Four- cade, Henrik, Filip, and Jacob Ingebrigtsen, Tim Maeyens, Sarah Louise Rung, and Mo Farah. While studies including more athletes would improve the ability to generalize find- ings, another possibility would be merging data from several individual case studies of world-class athletes, to produce stronger evidence. However, such assimilation would require a common framework for the reporting of high-quality long- term training data in elite athletes. Overall, implementation of such a policy would require collaboration between sports federations and research institutions, resulting in national and international projects with a concurrent focus on helping today’s athletes optimize their abilities, while the long-term data would enhance the performance of future generations of athletes. Furthermore, the finding that no information about training before a junior age was reported by any of the stud- ies in this review demonstrates the importance of systemati- cally monitoring athletes from a younger age. Over the past decade, there has been a burgeoning aware- ness and discussion regarding the lack of female-specific sports science research [46]. The present systematic scoping review highlights that female participants are considerably under-represented, and these findings align with other recent studies that emphasize the continued paucity of research on women in sport [47]. Out of the 17 studies included in this review, only 5% of the participants were from female-only studies. Similarly, Cowley et al. [47] reported that only 6% of randomly sampled sport and exercise studies, published between 2014 and 2020, were on women. Furthermore, the data in this review showed an under-representation of female participants and Paralympic athletes, a small num- ber of unique sports, and a clear predominance of athletes from Western Europe. This restricts the generalizability of the existing scientific evidence and limits the possibility to inform sport practices and policies [48]. 4.2 Training Characteristics Although scientific evidence is lacking, long-term dedi- cated training is crucial for reaching a world-class level in endurance sports. In our results, only seven of the studies included detailed information about the long-term progres- sion in training; from a junior age or beginning of a sen- ior age (18–20 years), until reaching elite/international or world-class level (i.e., 23–29 years). Interestingly, none of these studies included information about training before the age of 18 years, which could be a topic to investigate in future studies. The studies demonstrate a non-linear increase in TV, varying from 30 to 500% over periods that range from 2 to 17 years. Such large overall increases in TV required a considerable elevation in TV for specific years. However, more information is needed to understand the observed increases in TV, and if larger increases are associated with a more rapid performance and physiological development, or conversely, a greater risk of stagnation. Three studies documented a plateau in TV occur- ring close to peak performance, from ~ 650 to 900 h·yr−1 depending on the type of endurance sport and individual needs. This is not unexpected for the long-term develop- ment process, as a TV plateau is often observed around the same time an athlete reaches their peak performance level. However, we observed a decrease in TV, although performance level was maintained in the final years of a world-class female cross-country skier [30]. The find- ings of a gradual TV increase prior to reaching a plateau support the guidelines provided by sporting bodies, but additional research on how training progression can be further optimized is required. The effectiveness of utilizing TID concepts to maxi- mize endurance adaptations and performance is a “hot topic” in the scientific literature [19, 20, 49]. However, little research has investigated the long-term development of TID in elite/international or world-class endurance athletes. In this scoping review, six case studies detailed athletic TID development, with all studies reporting an increased LIT volume. Two of the studies showed a stable portion of MIT and HIT over time [33, 34]; one study observed a change towards a higher volume for both MIT and HIT [37], while another study showed a small rela- tive increase in MIT and a corresponding decrease in HIT [30]. The remaining two studies described a reduction in MIT and increased HIT [29, 38]. It is therefore difficult to draw any conclusions from this summary. In addition, six studies used different methodologies to determine TID, included athletes from different sports, and detailed dif- ferent timespans. For example, one study compared only 2 years of training [38], while another study described training changes over 12 years [30]. In addition, the differ- ent methodology for logging intensity zones [21] and the complexity of the long-term development process, make it challenging to form generalizations about TID. However, increased LIT was associated with progression in the train- ing load for all studies, and as such, this factor appears to be a critical cornerstone of any successful endurance training program. Accordingly, the proportion or volume of MIT and HIT is a crux of the training debate that has been previously described [20]. Still, an optimal endur- ance training program should provide the necessary total TV, whilst balancing the appropriate proportion of MIT and HIT for each individual athlete. The current scientific understanding of how TID should be divided over a long duration is limited and more information regarding the long-term development of TID during different stages of an athlete’s career is needed. 1604 H. C. Staff et al. 4.3 Performance‑Determining Factors The description of a world-class athlete implies a positive performance development across multiple years, and seven of the included studies also reported positive developments of performance and/or maximal performance indicators [29, 35, 38, 40, 44, 45, 50]. However, a compilation of the results is challenging because of testing in different periods of the season, and the fact that these performance determi- nants appear particularly sensitive to seasonal variations in training. While high VO2max values have been measured in world- class athletes for most endurance sports [51], less data are available on the long-term development of VO2max. In this scoping review, VO2max was reported in ten studies [29, 30, 35, 38–44], and suggests a considerable individual varia- tion in the development of VO2max of elite athletes during their athletic careers. These cumulative data indicate that for some athletes, VO2max may develop and become optimized in the early stages of their career, while other performance- determining factors then drive subsequent improvements. In contrast, other athletes were able to further develop their VO2max at later stages in their careers. The causative rea- son behind this divergent response may be due to training pattern changes that stress complementary VO2max-limiting factors during this period. However, this theory should be considered speculative and additional research is required to further investigate this concept. For example, changes in body mass or body composition could change the relative VO2max values. While VO2max showed different development patterns in world-class athletes, performance-determining factors that were based on submaximal responses demonstrated consid- erably more consistent developments, both with and without improvements in VO2max [29, 30, 35, 38, 40, 44, 50]. This result provides further support for the concept that endur- ance performance improvements after the age of 18–20 years are primarily related to other factors than VO2max, such as improved fractional utilization of VO2max and work econ- omy/efficiency. This is exemplified in the studies of Paula Radcliff [35, 44] who already reached a high value of VO2max at the age of 18 years, while improvements in running econ- omy and running performance continued to develop gradu- ally over years. 4.4 Existing Knowledge Gaps The low number of peer-reviewed articles that have pre- sented data on the long-term development of athletes reach- ing elite/international or world-class level, in combination with varying data quality and lack of important details, high- lights the urgent need for more long-term studies to support evidence-based talent development in sport. As more than half of the included studies were case studies, and most of the data were collected retrospectively, prospective studies would be of particularly interest. The low number of studies in women also confirms their current under-representation in the scientific literature. Although participation and professionalization in Para- lympic sports are increasing, it is problematic that only one study with Paralympic athletes met the inclusion criteria in this systematic scoping review. The same applies for the small number of unique sports and the clear predominance of athletes from Western Europe, which highlights the need for further examinations of different sports, cultures, and ethnicities. Finally, only four of the 17 studies reported concurrent data of training and performance-determining variables, lim- iting the ability to identify potential associations between relevant variables of interest. In this context, future long- term development studies should follow a common frame- work, enabling the possibility to compare data across studies and the performance of future meta-analyses. 4.5 Methodological Guidelines for Future Research The findings in this scoping review demonstrate that a com- mon methodological framework to permit a detailed com- parison between different studies is needed. Based on the findings in this study, we have devised the following guide- lines regarding the type of information to include, and the standardization required, for all future studies that wish to report on long-term training development and performance- determining factors in endurance sports (see Table 3). We hope that these guidelines can assist future studies to stand- ardize the collection and presentation of training data, and we encourage other researchers to further develop and vali- date this proposed framework. 5 Conclusions The current review found that only a handful of previous studies have reported the long-term development of train- ing characteristics and performance-determining factors in male and female endurance athletes reaching an elite/inter- national or world-class level. There are particularly limited data on women, and athletes aged younger than 18 years. No evidence was found for possible sex differences. Although 17 studies were included in this systematic scoping review, athletes from only a small number of countries and sports are described. Accordingly, current long-term talent devel- opment practices in endurance sports have insufficient sci- entific evidence. The training characteristics described a non-linear year- to-year increase in TV for most world-class endurance 1605 Long-Term Development of Training Characteristics and Performance-Determining Factors athletes, subsequently resulting in a plateau. However, the progression of TID showed individual patterns. While it is likely that a gradual progression in TV, with most of the increase stemming from more LIT, is required to reach a high level in endurance sports, no pattern was identified for the optimal development of MIT and HIT. The few stud- ies reporting the development of performance-determining variables indicated a consistent improvement in maximal performance tests and submaximal performance indicators for most athletes. Conversely, VO2max development was observed to be inconsistent. Overall, there is an urgent need for additional research that describes the long-term development of world-class athletes. Specifically, the implementation of systematic monitoring of athletes from a young age, employing high- precision reproducible measurements of training and per- formance-determining variables would enable prospective and high-quality retrospective study designs of considerable Table 3 Methodological guidelines for future research focusing on the long-term development of endurance athletes AT anaerobic threshold, CP critical power, HIT high-intensity training, HR heart rate, LIT low-intensity training, LT lactate threshold, max maxi- mum, MIT moderate-intensity training, PO power output, TID training intensity distribution, TV training volume, VE minute ventilation, VO oxygen uptake, VO2max maximal oxygen uptake, VT ventilatory threshold Topics Information and standardization Time frame Years without using training diary/logs: qualitatively describe the training/activity background until the start of systematically logging Years with the use of training diary/logs: record daily/weekly training from the year they started logging training until the end of their career Performance development Logging of all competitions (type, duration) Logging of results from major events (national and international championship and World Cup as junior and senior) Training characteristics Training volume Training frequency Training form (endurance, strength, and speed) Exercise mode (modality) TID 3-zone model (LIT, MIT, HIT) Session design (continuous or interval and choice of terrain) Mobility Qualitative descriptions of methodology used to record TID, TV, and pauses between intermitted training methods (interval training) Recovery parameters Rest days Sleeping time and quality Nutrition Qualitative registrations of other loading factors:  Work/school or other cognitive stress  Traveling (including time-zone changes)  Environmental (heat, cold, humidity, altitude)  Traumatic challenging emotional events/situations Health parameters Illness and injury days Menstrual or hormonal contraceptive cycle Periodization phases Annual General preparation Specific preparation Competition period Altitude Tapering Anthropometric and physi- ological parameters Body height (cm) Body mass (kg) Lean body mass (kg) Total body fat (%) Systematic measurements of  VO2max (L·min−1, mL·kg−1·min−1)    speed@VO2max, HR@VO2max, VE@VO2max  Performance indices (peak/average speed and power)   Peak/max HR  Threshold concepts (LT, VT, CP)    speed@AT, watt@AT, HR@AT, lactat@AT, VO@AT, PO@AT  Work economy or efficiency  Relevant speed and strength measurements if possible 1606 H. C. Staff et al. scientific and practical value. In addition, the use of a com- mon methodological framework is also necessary to permit a detailed comparison between different studies and allow for future meta-analyses. Acknowledgements The authors thank Tromsø Research Foundation and the FENDURA project team for their enthusiasm and contribution in this study. Declarations Funding This study was funded by the Tromsø Research Foundation (Project-ID: 19_FENDURA_BW) and UiT The Arctic University of Norway. Open access funding provided by UiT The Arctic University of Norway (incl University Hospital of North Norway). Conflicts of Interest/Competing Interests Hanne Staff, Guro Strøm Solli, John O. Osborne, and Øyvind Sandbakk have no conflicts of interest that are directly relevant to the content of this article. Ethics Approval Not applicable. Consent to Participate Not applicable. Consent for Publication Not applicable. Availability of Data and Material Not applicable. Code Availability Not applicable. Authors’ Contributions HS, GS, JOO, and ØS designed the study; JOO performed the database search; HS, GS, and JOO performed the screen- ing process; HS, GS, JOO, and ØS contributed to interpretation of the results; HS, GS, and ØS wrote the draft manuscript; and HS, GS, JOO, and ØS contributed to the final manuscript. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. References 1. Rees T, Hardy L, Güllich A, Abernethy B, Côté J, Woodman T, et al. The Great British Medalists Project: a review of current knowledge on the development of the world’s best sporting talent. Sports Med. 2016;46(8):1041–58. 2. Ericsson KA, Krampe RT, Tesch-Romer C. The role of deliberate practice in the acquisition of expert performance. Psychol Rev. 1993;100(3):363–406. 3. Tucker R, Collins M. What makes champions? A review of the relative contribution of genes and training to sporting success. Br J Sports Med. 2012;46(8):555–61. 4. Ericsson KA. Towards a science of the acquisition of expert performance in sports: clarifying the differences between deliberate practice and other types of practice. J Sports Sci. 2020;38(2):159–76. 5. Issurin VB. Evidence-based prerequisites and precursors of ath- letic talent: a review. Sports Med. 2017;47(10):1993–2010. 6. Vaeyens R, Güllich A, Warr CR, Philippaerts R. Talent identifica- tion and promotion programmes of Olympic athletes. J Sports Sci. 2009;27(13):1367–80. 7. Casado A, González-Mohíno F, González-Ravé JM, Foster C. Training periodization, methods, intensity distribution, and vol- ume in highly trained and elite distance runners: a systematic review. Int J Sports Physiol Perform. 2022;17(6):820–33. 8. González-Ravé JM, Hermosilla F, González-Mohíno F, Casado A, Pyne DB. Training intensity distribution, training volume, and periodization models in elite swimmers: a systematic review. Int J Sports Physiol Perform. 2021;16(7):913–26. 9. Seiler S. What is best practice for training intensity and duration distribution in endurance athletes? Int J Sports Physiol Perform. 2010;5(3):276–91. 10. Sandbakk O, Haugen T, Ettema G. The influence of exercise modality on training load management. Int J Sports Physiol Perform. 2021;16(4):605–8. 11. Bloom BS. Developing talent in young people. New York: Bal- lantine Books; 1985. 12. Helsen WF, Starkes JL, Van Winckel J. The influence of relative age on success and dropout in male soccer players. Am J Hum Biol. 1998;10(6):791–8. 13. Rusko H. Physiology of cross country skiing. In: Rusko H (Ed) Handbook of Sports Medicine and Science - Cross Country Skiing. Oxford, UK: Blackwell Science; 2003. pp. 1–31 14. Foster C, Barroso R, Beneke R, Bok D, Boullosa D, Casado A, et al. How to succeed as an athlete: what we know, what we need to know. Int J Sports Physiol Perform. 2022;17(3):333–4. 15. Haugen T, Sandbakk Ø, Seiler S, Tønnessen E. The training characteristics of world-class distance runners: an integration of scientific literature and results: proven practice. Sports Med Open. 2022;8(1):1–18. 16. MacIntosh BR, Murias JM, Keir DA, Weir JM. What is mod- erate to vigorous exercise intensity? Front Physiol. 2021;12: 682233. 17. Ruiz-Alias SA, Olaya-Cuartero J, Ñancupil-Andrade AA, García- Pinillos F. 9/3-Minute running critical power test: mechani- cal threshold location with respect to ventilatory thresholds and maximum oxygen uptake. Int J Sports Physiol Perform. 2022;17(7):1111–8. 18. Seiler S, Tønnessen E. Intervals, thresholds, and long slow dis- tance: the role of intensity and duration in endurance training. Sportscience. 2009;13:32–53. 19. Stöggl T, Sperlich B. Polarized training has greater impact on key endurance variables than threshold, high intensity, or high volume training. Front Physiol. 2014;5:33. 20. Burnley M, Bearden SE, Jones AM. Polarized training is not optimal for endurance athletes. Med Sci Sports Exerc. 2022;54(6):1032–4. 21. Sylta O, Tonnessen E, Seiler S. From heart-rate data to training quantification: a comparison of 3 methods of training-intensity analysis. Int J Sports Physiol Perform. 2014;9(1):100–7. 22. Joyner MJ, Coyle EF. Endurance exercise performance: the physi- ology of champions. J Physiol. 2008;586(1):35–44. 23. Lloyd RS, Oliver JL, Faigenbaum AD, Howard R, De Ste Croix MB, Williams CA, et al. Long-term athletic development: part 1: a pathway for all youth. J Strength Cond Res. 2015;29(5):1439–50. 24. Peters MD, Marnie C, Tricco AC, Pollock D, Munn Z, Alexan- der L, et al. Updated methodological guidance for the conduct of scoping reviews. JBI Evid Synth. 2020;18(10):2119–26. 1607 Long-Term Development of Training Characteristics and Performance-Determining Factors 25. Tricco AC, Lillie E, Zarin W, O’Brien KK, Colquhoun H, Levac D, et al. PRISMA extension for scoping reviews (PRISMA-ScR): checklist and explanation. Ann Intern Med. 2018;169(7):467–73. 26. McKay AK, Stellingwerff T, Smith ES, Martin DT, Mujika I, Goosey-Tolfrey VL, et al. Defining training and performance caliber: a participant classification framework. Int J Sports Physiol Perform. 2022;17(2):317–31. 27. Moher D, Liberati A, Tetzlaff J, Altman DG, Group P. Reprint: preferred reporting items for systematic reviews and meta-analy- ses: the PRISMA statement. Phys Ther. 2009;89(9):873–80. 28. Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Int J Surgery. 2021;88: 105906. 29. Bourgois J, Steyaert A, Boone J. Physiological and anthropometric progression in an international oarsman: a 15-year case study. Int J Sports Physiol Perform. 2014;9(4):723–6. 30. Solli GS, Tonnessen E, Sandbakk O. The training characteristics of the world's most successful female cross-country skier. Front Physiol. 2017;8:1069. 31. Tjelta LI, Tonnessen E, Enoksen E. A case study of the training of nine limes New York Marathon winner Grete Waitz. Int J Sports Sci Coach. 2014;9(1):139–57. 32. Tjelta LI. Three Norwegian brothers all European 1500 m champions: what is the secret? Int J Sports Sci Coach. 2019;14(5):694–700. 33. Schmitt L, Bouthiaux S, Millet GP. Eleven years’ monitoring of the world’s most successful male biathlete of the last decade. Int J Sports Physiol Perform. 2021;16(6):900–5. 34. Baumgart JK, Tønnessen E, Eklund M, Sandbakk Ø. Training distribution during a paralympic cycle for a multiple swimming champion with paraplegia: a case report. Int J Sports Physiol Per- form. 2021;16(12):1888–94. 35. Jones AM. The physiology of the world record holder for the women’s marathon. Int J Sports Sci Coach. 2006;1(2):101–16. 36. Pinot J, Grappe F. A six-year monitoring case study of a top-10 cycling Grand Tour finisher. J Sports Sci. 2015;33(9):907–14. 37. Tjelta LI. A longitudinal case study of the training of the 2012 European 1500 m track champion. Int J Appl Sports Sci. 2013;25:11–8. 38. Ingham SA, Fudge BW, Pringle JS. Training distribution, physi- ological profile, and performance for a male international 1500-m runner. Int J Sports Physiol Perf. 2012;7(2):193–5. 39. Lacour J-R, Messonnier L, Bourdin M, Lacour J-R. Physiological correlates of performance. Case study of a world-class rower. Eur J Appl Physiol. 2009;106(3):407–13. 40. Mikulic P. Maturation to elite status: a six-year physiological case study of a world champion rowing crew. Eur J Appl Physiol. 2011;111(9):2363–8. 41. Arrese AL, Ostariz ES, Mallen JAC, Munguia ID. The changes in running performance and maximal oxygen uptake after long-term training in elite athletes. J Sports Med Phys Fitness. 2005;45(4):435–40. 42. Díaz V, Peinado AB, Vleck VE, Alvarez-Sánchez M, Benito PJ, Alves FB, et al. Longitudinal changes in response to a cycle-run field test of young male national “talent identification” and senior elite triathlon squads. J Strength Cond Res. 2012;26(8):2209–19. 43. Santalla A, Naranjo J, Terrados N. Muscle efficiency improves over time in world-class cyclists. Med Sci Sports Exerc. 2009;41(5):1096–101. 44. Jones AM. A five year physiological case study of an Olympic runner. Br J Sports Med. 1998;32(1):39–43. 45. Anderson ME, Hopkins WG, Roberts AD, Pyne DB. Monitoring seasonal and long-term changes in test performance in elite swim- mers. Eur J Sport Sci. 2006;6(3):145–54. 46. Costello JT, Bieuzen F, Bleakley CM. Where are all the female participants in sports and exercise medicine research? Eur J Sport Sci. 2014;14(8):847–51. 47. Cowley ES, Olenick AA, McNulty KL, Ross EZ. “Invisible sports- women”: the sex data gap in sport and exercise science research. Women Sport Phys Activity J. 2021;29(2):146–51. 48. Curran O, MacNamara A, Passmore D. What about the girls? Exploring the gender data gap in talent development. Frontier Sports Active Living. 2019;1:3. 49. Foster C, Casado A, Esteve-Lanao J, Haugen T, Seiler S. Polarized training is optimal for endurance athletes. Med Sci Sports Exerc. 2022;54(6):1028–31. 50. Mujika I, Orbañanos J, Salazar H. Physiology and training of a world-champion paratriathlete. Int J Sports Physiol Perform. 2015;10(7):927–30. 51. Haugen T, Paulsen G, Seiler S, Sandbakk Ø. New records in human power. Int J Sports Physiol Perform. 2018;13(6):678–86.

Long-Term Development of Training Characteristics and Performance-Determining Factors in Elite/International and World-Class Endurance Athletes: A Scoping Review.

05-13-2023

Staff, Hanne C,Solli, Guro Strøm,Osborne, John O,Sandbakk, Øyvind

eng

PMC9864860

Citation: Roberts, J.D.; Lillis, J.B.; Pinto, J.M.; Chichger, H.; López-Samanes, Á.; Coso, J.D.; Zacca, R.; Willmott, A.G.B. The Effect of a Hydroxytyrosol-Rich, Olive-Derived Phytocomplex on Aerobic Exercise and Acute Recovery. Nutrients 2023, 15, 421. https://doi.org/10.3390/ nu15020421 Academic Editor: David C. Nieman Received: 14 December 2022 Revised: 6 January 2023 Accepted: 10 January 2023 Published: 13 January 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). nutrients Article The Effect of a Hydroxytyrosol-Rich, Olive-Derived Phytocomplex on Aerobic Exercise and Acute Recovery Justin D. Roberts 1,* , Joseph B. Lillis 1, Jorge Marques Pinto 1 , Havovi Chichger 2 , Álvaro López-Samanes 3 , Juan Del Coso 4 , Rodrigo Zacca 5,6 and Ashley G. B. Willmott 1 1 Cambridge Centre for Sport and Exercise Sciences (CCSES), School of Psychology and Sport Science, Anglia Ruskin University, Cambridge CB1 1PT, UK 2 School of Life Sciences, Anglia Ruskin University, Cambridge CB1 1PT, UK 3 Exercise Physiology Group, Faculty of Health Sciences, Universidad Francisco de Vitoria, 28223 Madrid, Spain 4 Centre for Sport Studies, Rey Juan Carlos University, 28943 Fuenlabrada, Spain 5 Research Center in Physical Activity, Health and Leisure (CIAFEL), Faculty of Sports, University of Porto (FADEUP), 4200-450 Porto, Portugal 6 Laboratory for Integrative and Translational Research in Population Health (ITR), 4050-600 Porto, Portugal * Correspondence: [email protected]; Tel.: +44-845-196-5154 Abstract: There is current scientific interest in naturally sourced phenolic compounds and their potential benefits to health, as well as the effective role polyphenols may provide in an exercise setting. This study investigated the chronic effects of supplementation with a biodynamic and organic olive fruit water phytocomplex (OliPhenolia® [OliP]), rich in hydroxytyrosol (HT), on submaximal and exhaustive exercise performance and respiratory markers of recovery. Twenty-nine recreationally active participants (42 ± 2 yrs; 71.1 ± 2.1 kg; 1.76 ± 0.02 m) consumed 2 × 28 mL·d−1 of OliP or a taste- and appearance-matched placebo (PL) over 16 consecutive days. Participants completed a demanding, aerobic exercise protocol at ~75% maximal oxygen uptake ( . VO2max) for 65 min 24 h before sub- and maximal performance exercise tests prior to and following the 16-day consumption period. OliP reduced the time constant (τ) (p = 0.005) at the onset of exercise, running economy (p = 0.015) at lactate threshold 1 (LT1), as well as the rating of perceived exertion (p = 0.003) at lactate turnpoint (LT2). Additionally, OliP led to modest improvements in acute recovery based upon a shorter time to achieve 50% of the end of exercise . VO2 value (p = 0.02). Whilst OliP increased time to exhaustion (+4.1 ± 1.8%), this was not significantly different to PL (p > 0.05). Phenolic compounds present in OliP, including HT and related metabolites, may provide benefits for aerobic exercise and acute recovery in recreationally active individuals. Further research is needed to determine whether dose-response or adjunct use of OliP alongside longer-term training programs can further modulate exercise-associated adaptations in recreationally active individuals, or indeed support athletic performance. Keywords: polyphenols; OliPhenolia®; hydroxytyrosol; exercise; oxygen uptake kinetics; lactate threshold; running economy 1. Introduction Nutritional strategies to enhance exercise performance and recovery are of current scientific interest to individuals who regularly undertake physical activity, competitive athletes, military workers, as well as the general population. Recent approaches which have gained popularity in an attempt to attenuate exercise-induced muscle damage (EIMD) and oxidative stress include the supplementation of naturally occurring phytochemicals (i.e., polyphenols) from sources such as pomegranate, cocoa, or cherries [1–3]. The average adult consumption of polyphenols is suggested to be ~1 g·d−1 [4], with primary sources from fruits, vegetables, beverages such as tea and coffee, wine, and chocolate [5]. With antioxidant properties [6], nutritional polyphenols may act as radical scavengers and metal Nutrients 2023, 15, 421. https://doi.org/10.3390/nu15020421 https://www.mdpi.com/journal/nutrients Nutrients 2023, 15, 421 2 of 20 chelators, regulating metabolism, body mass, chronic disease, and cell proliferation [7]. Free radicals and reactive oxygen and nitrogen species (RONS) are the primary oxidizing agents produced in cellular biochemical reactions for aerobic energy production [5]. Aerobic exercise is characterized by increased total energy expenditure [8], where the availability of endogenous substrates and aerobic metabolism are crucial for overall performance [9,10]. The increased oxygen (O2) demand by skeletal muscles during exercise results in greater free radical production and an increase in RONS [11]. Whilst viewed as detrimental to the cell for many years, recent evidence shows that RONS are crucial physiological activators and regulators of various intracellular signaling pathways in response to stress, enhanc- ing defense, improving cell adaptation, and upregulating the expression of endogenous antioxidant enzymes [12,13]. Furthermore, exercise adaptations are dependent, at least partially, on an acute ox- idative stress response. When exercise intensity is matched, individuals expressing lower levels of RONS have demonstrated inferior training adaptations compared to those with moderate or higher levels of RONS [14]. However, during excessive and demanding exer- cise, an imbalance between RONS and endogenous antioxidants induces oxidative damage, potentially impacting at a mitochondrial or DNA level [15], reducing vasodilatory capac- ity [16] and contractile force within the muscle through impaired calcium sensitivity [17]. This can have inferences for repetitive training sessions or longer-term adaptations and may, therefore, impair exercise performance and/or the recovery process. In sports where arterial blood flow and maximum cardiac output are determinants of performance (i.e., endurance and team-based sports), acute ingestion (<3 h before competition) or chronic supplementation of polyphenols (~7-days) could improve time to exhaustion at 70% maximum oxygen uptake ( . VO2max) by +9.7% [18] and intermittent high-intensity running distance by +10% [19]. The mechanisms by which polyphenols may facilitate ergogenic effects reportedly occur via nitric oxide synthase production [20] as well as the activation of sirtuin 1 (SIRT1) [21,22]. SIRT1 deacetylates several transcription factors such as forkhead (FOXO) proteins and peroxisome proliferator-activated receptor gamma coactivator 1-alpha (PGC- 1α) [23]. This can facilitate mitochondrial biogenesis, endothelial function, cell proliferation and differentiation, metabolic efficiency, resistance to stress, and improve inflammatory and immune function [24–26]. The supplementation of phenolic compounds, and gut-derived metabolites, may therefore provide adjunct or indirect ergogenic effects on physical perfor- mance by way of potentially reducing the O2 cost of exercise (i.e., economy), enhancing . VO2max or exercise tolerance, and/or improving substrate utilization efficiency. Previ- ous findings have highlighted that polyphenol nutrients (e.g., resveratrol) may support mitochondrial function [27] and may therefore modulate essential biological functions (including thermogenesis, mitochondrial biogenesis and adenosine triphosphate produc- tion) [28]. These functions are pivotal for trained, recreationally active and untrained exercising individuals, ensuring that substrate supply kinetics and waste product removal match the requirements of the specific exercise bout [29]. Furthermore, it could be inferred that due to the anti-inflammatory and immuno- modulatory effects of phenolic compounds, an increase in polyphenol consumption (from food or supplementation) may be pertinent to exercise recovery. A reduction in physio- logical stressors that negatively impact exercise training [30] may support fast and slow phases of recovery, influencing performance in both prolonged or repeated bouts of exercise. Evidence for enhanced functional recovery from foods/supplements high in polyphenol compounds (e.g., Montmorency cherries > 5-days) have been exhibited in both trained and untrained individuals in a multitude of general exercise settings [31–33]. However, further research is warranted to investigate other polyphenols or novel food products, to assess markers of exercise recovery and identify the potential impact of phenolic compounds in specific exercise settings. This is the first study to undertake an investigation into a commercially available polyphenol-rich olive fruit water, OliPhenolia® (OliP), which has not been assessed in an exercise domain. Originating during the olive picking season, this polyphenol-rich drink Nutrients 2023, 15, 421 3 of 20 is extracted via concentration, reverse osmosis, and ceramic membrane technology at the aqueous part of the olive fruit. Whilst OliP contains a variety of phenolic compounds, it is particularly rich in hydroxytyrosol (HT). Abundant in olives in the form of pure HT, HT glycosides and oleuropein, HT is an effective antioxidant, with studies highlighting protection against oxidative stress in vascular tissue [34,35], low-density lipoprotein oxidation [36–38], and a reduction in oxidative damage in intestinal epithelial cells [39], hepatocytes, and erythrocytes [40]. However, OliP has yet to be considered within an exercise and/or recovery domain and thus, requires investigation. Therefore, this study investigated the effect of OliP on submaximal and exhaustive exercise, as well as respiratory markers of acute recovery, in recreationally active volunteers. Understanding the efficacy of OliP may inform future nutritional strategies pertinent to exercise training and recovery. 2. Materials and Methods 2.1. Ethical Approval and Trial Registration This study was registered with clinical-trials.gov (ID: NCT04959006) with ethical approval obtained from the Faculty of Science and Engineering Research Ethics Panel, Anglia Ruskin University (Ethical approval no. FSE/FREP/20/946). Following a priori power calculation (G*power3, Dusseldorf, Germany [41]) using α = 0.05 and 1-β = 0.80, from previous reports of a time trial run and following recovery (plasma free radicals, post run pain and time to recovery [h]) [42], a minimum sample size of eight per intervention group was estimated. 2.2. Participant Characteristics Eligibility for the study required participants to be recreationally active (undertaking ~3 exercise sessions a week), with a . VO2max of >25 mL·kg−1·min−1 determined at the first visit. All participants were >21 yrs, with no known metabolic disorders, viruses, or infec- tions; were not self-administering any polyphenol or antioxidant-rich supplementation or adhering to specific diets that could conflict with study parameters. A total of 32 healthy participants volunteered and engaged with the study. However, following a review of indi- vidual protocol adherence and analysis of outliers, 3 participants’ datasets were removed. General characteristics of the remaining 29 participants satisfactorily completing the study are displayed in Table 1. Table 1. Mean ± standard error (SE) participant characteristics overall and for OliPhenolia® (OliP) and placebo (PL) groups respectively. Variable Overall OliP PL (n = 29; 20 M, 9 F) (n = 15; 11 M, 4 F) (n = 14; 9 M, 5 F) Age (yrs) 42 ± 2 42 ± 3 42 ± 3 Height (m) 1.76 ± 0.02 1.77 ± 0.03 1.75 ± 0.03 Body mass (kg) 71.08 ± 2.14 73.57 ± 2.44 68.41 ± 3.52 Fat free mass (kg) 57.67 ± 2.31 59.33 ± 3.05 55.89 ± 3.56 Body mass index (kg·m2) 22.9 ± 0.4 23.5 ± 0.4 22.3 ± 0.7 Body fat (%) 18.7 ± 1.8 19.5 ± 2.2 17.8 ± 3.0 . VO2max (L·min−1) 3.53 ± 0.16 3.56 ± 0.22 3.49 ± 0.24 . VO2max (mL·kg−1·min−1) 49.6 ± 1.7 48.3 ± 2.5 51.0 ± 2.2 M = male; F = female; . VO2max = maximal oxygen uptake. No statistical differences were reported between groups (p > 0.05). 2.3. Experimental Design Using a randomized number generator process (www.randomizer.org; accessed on 10 May 2021), participants were allocated into two supplement intervention groups (OliP or PL) in a double-blind manner. All participants reported to the Cambridge Centre for Sport and Exercise Sciences (CCSES), Anglia Ruskin University, on five separate occasions, the first of which involved an initial familiarization session [43,44]. All laboratory visits were Nutrients 2023, 15, 421 4 of 20 conducted at the same time of day following an overnight fast (~10 h), with participants arriving in a euhydrated state. Participants were instructed to avoid strenuous and/or excessive exercise for the 24 h prior to testing visits, as well as adhere to all dietary instruc- tions for the 3-days pre-testing (Section 2.4). For the duration of the supplementation period, participants were asked to continue habitual exercise and diet regimes, ensuring each week was matched to the previous in terms of training load, caloric and macronutrient intake. 2.4. Dietary and Exercise Activity Monitoring Dietary and hydration intake was tracked via the use of a mobile based application (MyFitnessPal, Inc., San Francisco, CA, USA). Participants were provided with a personal login and guidance instructions to support detailed dietary tracking. Participants were required to record consumption of all food items and liquids for 3-days leading into each exercise test as part of this study, as well as across the 16-day intervention period [45] and were checked regularly by the same researcher for consistency throughout the intervention. A list of polyphenol-rich ‘foods to avoid’ was also provided for participants to adhere to in the 3-days leading into each laboratory visit (see Supplementary Materials, Table S1). Participants were also required to complete a standardized, daily exercise activity diary for the 3-days prior to exercise trials and the duration of the 16 consecutive days supplementa- tion period, ensuring they were rested for the 24 h prior to each visit. Participants were requested to maintain habitual lifestyle and exercise patterns across the study, ensuring consistency across the 16-day period throughout the course of supplementation. Session type, mean session heart rate, exercise duration and perceived session exertion (using a standard 0–10 visual analogue scale) were recorded following the completion of each training session as reported elsewhere [45]. 2.5. Laboratory Procedures All tests took place under controlled environmental conditions (temperature: 19.6 ± 0.3 ◦C; barometric pressure: 1005.6 ± 1.2 mBar; and relative humidity: 48.4 ± 2.2%). Upon arrival, participants rested for 10 min in a seated position before assessment of blood pressure (Omron 750CP, Kyoto, Japan), body mass (electronic scale, Seca, Hamburg, Germany), and height (Seca stadiometer, Hamburg, Germany). At rest (baseline) and throughout exercise, 20 µL capilliarized fingertip blood samples were collected for the assessment of blood lactate and glucose (Biosen C Line EKF-diagnostic analyzer, Cardiff, UK). Heart rate (HR) data were recorded in 5 s intervals using a short-range telemetric monitor (Polar 810s, Polar T34 strap, Kempele, Finland). For the initial familiarization trial, body composition was also recorded using bioelectrical impedance for the indirect assessment of body fat percentage, fat-free mass, and fat mass (Tanita SC-330ST, Amsterdam, The Netherlands). Breath-by-breath pulmonary gas variables (volume of O2 [ . VO2], volume of carbon dioxide [ . VCO2], minute ventilation [ . VE], respiratory exchange ratio [RER], breathing frequency [BF] and tidal volume [TV]) were measured continuously via a metabolic cart (MetaLyzer 3B-R2, Cortex Ltd., Leipzig, Germany) using a suitable facemask for each participant (7600 face mask with headgear, Hans Rudolph, Shawnee, Kansas, USA). Prior to each test, the MetaLyzer was calibrated as per manufacturers’ specifications. All exercise testing was completed on a Quasar Med Treadmill (HP Cosmos, Nussdorf, Germany). 2.5.1. Experimental Protocols—Visit 1, 3 and 5 Exercise intensities were calculated using lactate profiles from the familiarization trial (visit 1) and remained consistent in visit 3 and 5. Visits 1, 3 and 5 consisted of a two-part graded exercise test [46,47] including: (1) a submaximal incremental protocol, with a 10 min recovery period; and (2) a maximal test to volitional exhaustion (Figure 1). Nutrients 2023, 15, 421 5 of 20 3 min 10 min Ramp to volitional exhaustion Recovery B B B B 5 min 50 min 1 min 60% ∆ LT1−LT2 10% above LT2 LT1 speed 1 h 3 min 10 min Ramp to volitional exhaustion Recovery 24 h Visit 5: Repeat of visit 3 Visit 4: Repeat of visit 2 Visit 1: Familiarization session Visit 2: Demanding aerobic session Visit 3: Submaximal and performance test session 16-day intervention 24 h ≥3-days End of study Figure 1. Schematic of study protocol outlining the familiarization (visit 1), demanding aerobic session (visit 2) and submaximal and performance test session (visit 3). B = blood sample; LT1 = lactate threshold; LT2 = lactate turnpoint. 2.5.2. Submaximal and Performance Test Protocol The speed for the submaximal protocol was selected at a pre-defined level and in- creased by 1 km·h−1 every 4 min, with 3 min of running at a constant speed [46,47] followed by a 1 min break for capilliarized fingertip blood sample collection. The gradient was main- tained at 1% [48] with rating of perceived exertion (RPE; 0−10 scale) and HR assessed in the final 30 s of each running stage. For the . VO2max performance test, speed was held con- sistent with gradient increasing by 1% per min, with RPE and blood lactate concentration (B[La]) obtained at the end of the test. Participants ran until volitional exhaustion (which determined time to exhaustion [TTE]), with standardized verbal encouragement provided towards the end of the test. 2.5.3. Determination of Physiological Parameters and Respiratory Kinetics Lactate threshold (LT1) was determined by an initial rise in B[La] above baseline [49], and lactate turnpoint (LT2) was determined by a sudden and sustained increase in B[La] [50]. Nutrients 2023, 15, 421 6 of 20 Mean and standard deviations (SD) of . VO2 for the last 30 breaths of each increment were calculated. Values ±4 SD were removed as outliers with all remaining breaths averaged [44]. Running economy was calculated in mL·kg−1·km−1 [51], described in Equation (1) below: Economy = . VO2  mL·kg−1·min−1 Speed (km·h −1)/60 (1) The . VO2 kinetics for exercise (on-kinetics) and recovery periods (off-kinetics) were modelled and calculated using validated software VO2FITTING [52]. Errant breaths were omitted by only including those within . VO2 local mean ± 4 SD. Subsequently, the individ- ual on-transient breath-by-breath . VO2 responses were modelled using a mono-exponential model [52] described in Equation (2): . VO2(t) = . VO2baseline + A  1 − e− t τ  (2) where . VO2(t) represents the relative . VO2 at the time t, A and τ are the amplitude and time constant (τ) of the fast . VO2 component. The individual off-transient breath-by-breath . VO2 responses were modelled using the a mono-exponential model [52] described in Equation (3): . VO2(t) = EE . VO2 − A  1 − e− t τ  (3) where EE . VO2 represents the relative end-exercise . VO2 during the on-transient kinetics phase. During exercise (on-kinetics), O2 deficit, . VO2 demand and τ were estimated [53]. The acute recovery period in this study reflected the 10 min following the submaximal exercise protocol. Within this period, time to 50% (T50%) was determined by the recording of consistent breaths under 50% of the . VO2max value reached [19]. . VO2max was determined from the highest . VO2 values recorded over a 15-breath rolling average [54]. 2.5.4. Demanding Aerobic Session—Visit 2 and 4 Visit 2 and 4 involved 65 min of exercise, with an overall target exercise intensity of ~75% . VO2max, designed to elicit muscular oxidative stress [45]. Participants completed a 5 min warm-up at a speed corresponding to LT1. Exercise intensity then increased to speeds that corresponded with 60% of the difference between LT1 and LT2 (∆LT1-LT2) for 50 min, before completing a maximum of 5 × 1 min intervals at a speed 10% above LT2, interspersed with 1 min active recovery at 60% ∆LT1-LT2. B[La], HR and RPE were measured at rest, and at 10, 30 and 48 min, and following the last interval. Exercise intensity was consistent between visit 2 and 4. Additionally, as a means to quantify whether the nutritional intervention influenced plasma HT (as the main polyphenol in OliP), resting whole blood measures were undertaken prior to both visit 2 and 4 (as part of a larger study reported elsewhere [45]). For this, whole blood samples were collected into 4 mL Vacuette™ K2EDTA tubes (Greiner Bio-One GmbH, Kremsmunster, Austria), centrifuged at 2000 rcf for 10 min, with extracted plasma stored at −80 ◦C until analysis for HT. Plasma HT was assessed using a liquid–liquid extraction method following acidic hydrolysis, with gas chromatography–mass spectrometry (GC-MS) analysis (Agilent 7820A GC, Santa Clara, CA, USA [45]). 2.6. Nutritional Intervention Nutritional supplementation was distributed in a double-blinded manner to partici- pants upon completion of visit 3. Product dose and timeframe were based on commercial product supply and company recommendations to consume 1 box (32 jars) of OliP as an acute intervention period. Therefore, participants were provided 32 jars in identically Nutrients 2023, 15, 421 7 of 20 labelled boxes and requested to consume 2 jars per day (56 mL total) separated by ~6 h between meals across 16-days. Each jar contained ~28 mL of either OliP (sweetened version, Batch 14, Fattoria La Vialla, Castiglion Fibocchi, Arezzo, Italy [see Supplementary Mate- rials, Table S2 for independent product analysis]) or taste- and appearance-matched PL (equal ratio of: prune juice [Sunsweet California Prune Juice, Tesco, Welwyn Garden City, UK], diet cola [Tesco Cola, Tesco, Welwyn Garden City, UK] and tonic water [low-calorie Indian tonic water, Tesco, Welwyn Garden City, UK]). To monitor adherence, participants returned all jars at the end of the trial, including any remaining full jars for confirmation of intervention adherence. 2.7. Dietary and Exercise Activity Analysis Dietary analysis occurred via Nutritics Professional Dietary Analysis software (Nu- tritics Ltd., Co., Dublin, Ireland). Three-day dietary intake was analyzed prior to each laboratory visit to ensure study guideline protocols were adhered to (Table 2). Estimation of dietary HT intake was also assessed (excluding supplementation) using the US Department of Agriculture and the Phenol-Explorer databases. Exercise data allowed for assessment of training load, monotony and strain as previously reported [45,55] (Table 3) to quantify relative consistency between cohorts. Table 2. Energy and macronutrient intake for both groups within the 3-day control period prior to visit 2 (pre-intervention) and 4 (post-intervention). Variable Time OliP PL Kcal·d−1 Pre 2134.9 ± 139.7 2149.6 ± 112.3 Post 2172.1 ± 135.5 2456.8 ± 151.2 CHO (g·d−1) Pre 232.9 ± 17.4 259.7 ± 15.9 Post 240.0 ± 16.1 273.8 ± 22.8 CHOrelative (g·kg−1·d−1) Pre 3.2 ± 0.3 3.9 ± 0.2 Post 3.3 ± 0.2 4.1 ± 0.3 FAT (g·d−1) Pre 85.9 ± 6.2 76.7 ± 4.8 Post 87.0 ± 6.4 97.9 ± 6.2 FATrelative (g·kg−1·d−1) Pre 1.2 ± 0.1 1.2 ± 0.1 Post 1.1 ± 0.1 1.5 ± 0.1 PRO (g·d−1) Pre 103.0 ± 8.6 104.2 ± 7.6 Post 103.4 ± 7.7 115.6 ± 9.1 PROrelative (g·kg−1·d−1) Pre 1.4 ± 0.1 1.5 ± 0.1 Post 1.4 ± 0.1 1.7 ± 0.2 No statistical differences were reported between groups (p > 0.05). Units: Kcal·d−1 = kilocalories per day; g· d−1 = grams per day; g·kg−1·d−1 = grams per kilogram body mass per day. 2.8. Statistical Analysis Statistical analysis was performed using SPSS (V.28, IBM Corporation, Armonk, New York, USA), with statistical significance determined as p ≤ 0.05. All data were assessed for homogeneity using Levene’s test and normality through a Shapiro-Wilk’s test [56]. A two- way repeated measures ANOVA was used for the main analysis with Greenhouse-Geisser corrections applied if sphericity could not be assumed. For plasma HT analysis, a mixed design ANOVA was also undertaken. Bonferroni post-hoc comparisons were employed where applicable, with effect sizes (partial eta squared ηp2) also reported (small = 0.02, medium = 0.13, large = 0.26). An independent samples t-test was also adopted to compare relevant data between groups (i.e., participant characteristics, dietary intake and training records), whereby Cohen’s d effect sizes were utilized (trivial ≤ 0.19, small = 0.20–0.49, medium = 0.50–0.70, large ≥ 0.80). Data are presented as mean ± SE. Nutrients 2023, 15, 421 8 of 20 Table 3. Mean habitual exercise activity for both groups within the 3-day control period prior to visit 2 (pre-intervention) and 4 (post-intervention), as well as collated mean training parameters across the intervention period. Control Period (3-Days) Prior to Laboratory Visits Variable Time OliP PL Training load (AU) Pre 641 ± 106 646 ± 116 Post 722 ± 80 802 ± 150 Training monotony (AU) Pre 0.9 ± 0.1 1.0 ± 0.1 Post 1.0 ± 0.1 1.2 ± 0.1 Training strain (AU) Pre 730 ± 182 583 ± 118 Post 753 ± 143 1112 ± 272 Across the 16-day intervention Exercise sessions completed 13 ± 1 15 ± 1 Session duration (min) 58.6 ± 4.2 67.3 ± 4.9 HR (b·min−1) 135 ± 4 140 ± 4 Session perceived exertion 5.2 ± 0.3 4.8 ± 0.3 AU = arbitrary units. No statistical differences were reported between groups (p > 0.05). 3. Results 3.1. Dietary Intake, Supplement Adherence, and Exercise Monitoring No significant differences were reported within or between groups in the 3-day period leading into exercise testing for energy (kcal·d−1), carbohydrate (CHO), fat (FAT), and protein (PRO) intake (Table 2). HT intake during the 3-day dietary control periods prior to each visit demonstrated no differences within or between groups (p > 0.05). Mean intake of dietary HT during control periods prior to visit 2 and 4 were 0.06 ± 0.11 and 0.04 ± 0.10 mg for OliP, and 0.08 ± 0.13 and 0.13 ± 0.16 mg for PL respectively. Estimation of general dietary HT throughout the intervention period (excluding OliP intake) indicated a mean total intake of 0.14 ± 0.07 mg·kg−1·d−1 for OliP compared to 0.18 ± 0.11 mg·kg−1·d−1 for PL (p > 0.05) and was considered low. Within groups, supplementation adherence rates were 99.0% OliP and 98.1% PL, with no between group differences reported (p > 0.05). No differences were reported within or between groups during the 3-day control period for training load, monotony, or strain (p > 0.05, Table 3). Additionally, no differ- ences were reported between groups for habitual exercise activity throughout the 16-day intervention (p > 0.05; [45]). 3.2. Demanding Aerobic Session—Visit 2 and 4 No differences within or between groups were found for the aerobic test (p > 0.05, Table 4) demonstrating relative consistency. Overall, mean target exercise intensity of ~75% . VO2max was achieved and sustained in both exercise sessions (76.08 ± 1.04 and 75.42 ± 0.98%, visit 2 and 4, respectively). Exercise intensity was consistent between groups for visit 2 and 4 in the OliP (11.4 ± 1.7 km·h−1) and PL group (11.5 ± 1.8 km·h−1; both p > 0.05). Plasma HT was not detected at baseline (visit 2, pre-supplementation) or following PL, but significantly increased from 0.0 ± 0.0 to 6.3 ± 1.6 ng·mL−1 following OliP (F = 14.28, p = 0.001, ηp2 = 0.43). 3.3. Submaximal Exercise—Visit 3 and 5 Onset of exercise: For time constant (τ), there was a significant effect for time (F = 5.23, p = 0.031, ηp2 = 0.17) and group (F = 4.44, p = 0.045, ηp2 = 0.15). A significant difference between groups pre-intervention (visit 3) was found (F = 4.36, p = 0.047, ηp2 = 0.15). A significant reduction from visit 3 and 5 was found in τ within OliP (F = 9.51, p = 0.005, ηp2 = 0.28, Figure 2, Table 5A) only. No differences were found for the PL group (p > 0.05). Nutrients 2023, 15, 421 9 of 20 Table 4. Physiological responses for both groups during the demanding aerobic session at visit 2 (pre-intervention) and 4 (post-intervention). Variable OliP PL Pre Post Pre Post . VO2 (mL·kg−1·min−1) 36.6 ± 1.5 36.7 ± 1.5 37.7 ± 1.6 37.4 ± 1.4 % of baseline . VO2max (%) 76.6 ± 1.5 76.7 ± 1.2 75.0 ± 1.8 74.6 ± 1.8 . VO2 (L·min−1) 2.70 ± 0.14 2.70 ± 0.15 2.56 ± 0.17 2.54 ± 0.15 . VCO2 (L·min−1) 2.49 ± 0.14 2.49 ± 0.13 2.36 ± 0.16 2.34 ± 0.15 . VE (L·min−1) 85.12 ± 4.59 84.87 ± 4.68 80.99 ± 4.66 80.49 ± 4.11 . VE/ . VO2 29.44 ± 0.46 29.33 ± 0.48 29.82 ± 1.01 30.01 ± 0.98 . VE / . VCO2 31.88 ± 0.45 31.75 ± 0.46 32.34 ± 1.11 32.55 ± 1.11 RER 0.92 ± 0.01 0.92 ± 0.01 0.92 ± 0.01 0.92 ± 0.01 Economy (mL·kg−1·km−1) 193.6 ± 3.5 194.0 ± 3.4 198.4 ± 4.7 197.2 ± 4.0 B[La] (mmol·L−1) 1.32 ± 0.11 1.31 ± 0.10 1.33 ± 0.11 1.33 ± 0.09 . VO2max = maximal oxygen uptake; . VO2 = volume of oxygen; . VCO2 = volume of carbon dioxide; . VE = minute ventilation; RER = respiratory exchange ratio; B[La] = blood lactate concentration. No differences reported within or between groups (p > 0.05). ab e . ysio ogica espo ses o bo g oups u i g e e a i g ae obic sessio a isi (pre-intervention) and 4 (post-intervention). Variable OliP PL Pre Post Pre Post V̇ O2 (mL·kg−1·min−1) 36.6 ± 1.5 36.7 ± 1.5 37.7 ± 1.6 37.4 ± 1.4 % of baseline V̇ O2max (%) 76.6 ± 1.5 76.7 ± 1.2 75.0 ± 1.8 74.6 ± 1.8 V̇ O2 (L·min−1) 2.70 ± 0.14 2.70 ± 0.15 2.56 ± 0.17 2.54 ± 0.15 V̇ CO2 (L·min−1) 2.49 ± 0.14 2.49 ± 0.13 2.36 ± 0.16 2.34 ± 0.15 V̇ E (L·min−1) 85.12 ± 4.59 84.87 ± 4.68 80.99 ± 4.66 80.49 ± 4.11 V̇ E/V̇ O2 29.44 ± 0.46 29.33 ± 0.48 29.82 ± 1.01 30.01 ± 0.98 V̇ E /V̇ CO2 31.88 ± 0.45 31.75 ± 0.46 32.34 ± 1.11 32.55 ± 1.11 RER 0.92 ± 0.01 0.92 ± 0.01 0.92 ± 0.01 0.92 ± 0.01 Economy (mL·kg−1·km−1) 193.6 ± 3.5 194.0 ± 3.4 198.4 ± 4.7 197.2 ± 4.0 B[La] (mmol·L−1) 1.32 ± 0.11 1.31 ± 0.10 1.33 ± 0.11 1.33 ± 0.09 V̇ O2max = maximal oxygen uptake; V̇ O2 = volume of oxygen; V̇ CO2 = volume of carbon dioxide; V̇ E = minute ventilation; RER = respiratory exchange ratio; B[La] = blood lactate concentration. No differ- ences reported within or between groups (p > 0.05). 3.3. Submaximal Exercise—Visit 3 and 5 Onset of exercise: For time constant (τ), there was a significant effect for time (F = 5.23, p = 0.031, ηp2 = 0.17) and group (F = 4.44, p = 0.045, ηp2 = 0.15). A significant difference between groups pre-intervention (visit 3) was found (F = 4.36, p = 0.047, ηp2 = 0.15). A sig- nificant reduction from visit 3 and 5 was found in τ within OliP (F = 9.51, p = 0.005, ηp2 = 0.28, Figure 2, Table 5A) only. No differences were found for the PL group (p > 0.05). Figure 2. Time constant (τ) at the onset of exercise pre- and post-intervention for OliP and PL groups. * denotes significant difference between groups pre intervention (p = 0.047), ** denotes sig- nificant difference between time points for OliP (p = 0.005).Lactate threshold 1 (LT1): Respiratory parameters, exercise economy, B[La] and RPE are shown in Table 5A pre- to post-intervention for both OliP and PL. There was a significant interaction effect for relative V̇ O2 (time x group: F = 4.66, p = 0.039, ηp2 = 0.16, Figure 3A), where a reduction in relative V̇ O2 was found in the OliP group between visit 3 and 5 (F = 7.09, p = 0.013, ηp2 = 0.22). Whilst no differences were found post interven- tion between OliP and PL, when expressed as relative change, OliP demonstrated a −2.7 ± 1.2% reduction in V̇ O2 compared with the PL group at LT1 intensity (−0.7 ± 1.0%; t = 2.13, p = 0.043, d = 0.82, 95% confidence interval [CI] range 0.05 to 1.64). This corresponded with a significant reduction in the % of V̇ O2max from baseline (73.7 ± 1.8% in visit 3 to 71.2 ±1.6% in visit 5) (F = 7.72, p = 0.01, ηp2 = 0.24, Figure 3B) for OliP only. No differences were reported in the PL group (p > 0.05), or post ✱✱ ✱ Figure 2. Time constant (τ) at the onset of exercise pre- and post-intervention for OliP and PL groups. * denotes significant difference between groups pre intervention (p = 0.047), ** denotes significant difference between time points for OliP (p = 0.005). Lactate threshold 1 (LT1): Respiratory parameters, exercise economy, B[La] and RPE are shown in Table 5A pre- to post-intervention for both OliP and PL. There was a significant interaction effect for relative . VO2 (time x group: F = 4.66, p = 0.039, ηp2 = 0.16, Figure 3A), where a reduction in relative . VO2 was found in the OliP group between visit 3 and 5 (F = 7.09, p = 0.013, ηp2 = 0.22). Whilst no differences were found post intervention between OliP and PL, when expressed as relative change, OliP demonstrated a −2.7 ± 1.2% reduction in . VO2 compared with the PL group at LT1 intensity (−0.7 ± 1.0%; t = 2.13, p = 0.043, d = 0.82, 95% confidence interval [CI] range 0.05 to 1.64). This corresponded with a significant reduction in the % of . VO2max from baseline (73.7 ± 1.8% in visit 3 to 71.2 ±1.6% in visit 5) (F = 7.72, p = 0.01, ηp2 = 0.24, Figure 3B) for OliP only. No differences were reported in the PL group (p > 0.05), or post intervention in comparison to OliP. A significant interaction effect was also observed for running economy (time x group: F = 5.22, p = 0.031, ηp2 = 0.17, Figure 3C), with a significant improvement demonstrated between visit 3 and 5 for OliP only (F = 6.82, p = 0.015, ηp2 = 0.21, 95% CI range 189.80 to 207.11). Finally, it was also noted that when expressed as relative change, there was a pre- to post-intervention reduction in . VCO2 within OliP (−1.6 ± 0.9%) compared to PL (+1.5 ± 0.9%; t = 2.33, p = 0.028, d = 0.90). Nutrients 2023, 15, 421 10 of 20 Table 5. Respiratory, exercise economy, perceived exertion, and blood lactate parameters at the onset of exercise and lactate threshold (A), and lactate turnpoint (B) at visit 3 (pre-intervention) and 5 (post-intervention) interspersed with 16 consecutive days of either OliP or PL. (A) OliP PL Onset of Exercise Pre Post Pre Post . VO2 at baseline (mL·kg−1·min−1) 8.48 ± 1.13 6.88 ± 0.55 6.43 ± 0.81 6.25 ± 0.60 . VO2 demand at 60 s (L) 1.69 ± 0.15 1.74 ± 0.12 1.80 ± 0.16 1.78 ± 0.14 . VO2 demand at 120 s (L) 3.39 ± 0.29 3.47 ± 0.23 3.60 ± 0.31 3.55 ± 0.28 . VO2 demand at 180 s (L) 4.80 ± 0.43 4.94 ± 0.32 5.08 ± 0.43 5.03 ± 0.40 O2 deficit (L) 0.97 ± 0.08 0.85 ± 0.07 0.84 ± 0.07 0.81 ± 0.07 τ (s) 40.5 ± 4.6 30.5 ± 1.5 * 29.2 ± 1.9 28.1 ± 1.1 Mean . VO2 in the last 60 s (mL·kg−1·min−1) 30.7 ± 1.5 30.0 ± 1.5 30.3 ± 2.2 31.1 ± 1.6 Lactate threshold (LT1) % . VO2max of baseline (%) 73.0 ± 1.8 70.9 ± 1.6 * 68.7 ± 1.4 69.2 ± 1.6 . VO2 (L·min−1) 2.56 ± 0.13 2.50 ± 0.13 * 2.45 ± 0.15 2.46 ± 0.13 . VCO2 (L·min−1) 2.30 ± 0.12 2.26 ± 0.12 2.16 ± 0.15 2.19 ± 0.13 . VE (L·min−1) 71.81 ± 3.55 71.36 ± 3.74 67.62 ± 4.09 68.56 ± 4.19 . VE/ . VO2 26.17 ± 0.36 26.65 ± 0.42 25.83 ± 0.78 25.94 ± 0.88 . VE/ . VCO2 29.22 ± 0.45 29.41 ± 0.44 29.30 ± 0.91 29.21 ± 1.04 RER 0.90 ± 0.01 0.91 ± 0.01 0.88 ± 0.01 0.89 ± 0.01 Economy (mL·kg−1·km−1) 201.3 ± 3.6 195.6 ± 4.0 * 199.4 ± 5.1 201.2 ± 5.9 RPE 2.9 ± 0.3 2.7 ± 0.3 3.2 ± 0.5 3.3 ± 0.1 B[La] (mmol·L−1) 1.32 ± 0.11 1.31 ± 0.10 1.33 ± 0.11 1.33 ± 0.09 (B) OliP PL Lactate turnpoint (LT2) Pre Post Pre Post % . VO2max (%) 84.1 ± 1.8 83.5 ± 1.4 81.2 ± 2.0 82.8 ± 2.1 . VO2 (L·min−1) 2.96 ± 0.16 2.95 ± 0.17 2.91 ± 0.21 2.96 ± 0.16 . VCO2 (L·min−1) 2.81 ± 0.15 2.81 ± 0.15 2.73 ± 0.20 2.79 ± 0.19 . VE (L·min−1) 92.73 ± 4.87 92.28 ± 5.16 88.00 ± 5.48 89.80 ± 5.89 . VE / . VO2 29.46 ± 0.50 29.41 ± 0.52 28.72 ± 0.98 28.60 ± 1.11 . VE / . VCO2 31.00 ± 0.48 30.80 ± 0.47 30.69 ± 1.04 30.25 ± 1.13 RER 0.95 ± 0.01 0.95 ±0.01 0.94 ± 0.01 0.94 ± 0.01 Economy (mL·kg−1·km−1) 192.7 ± 4.5 191.1 ± 3.5 193.7 ± 4.7 197.7 ± 4.2 RPE 6.0 ± 0.34 5.4 ± 0.4 * 5.6 ± 0.4 5.4 ± 0.4 B[La] (mmol·L−1) 2.31 ± 0.12 2.18 ± 0.11 2.24 ± 0.12 2.22 ± 0.12 τ = time constant; . VO2max = maximal oxygen uptake; . VO2 = volume of oxygen; . VCO2 = volume of carbon dioxide; . VE = minute ventilation; RER = respiratory exchange ratio; RPE = rating of perceived exertion; B[La] = blood lactate concentration. * denotes a significant within group difference (p < 0.05). intervention in comparison to OliP. A significant interaction effect was also observed for running economy (time x group: F = 5.22, p = 0.031, ηp2 = 0.17, Figure 3C), with a significant improvement demonstrated between visit 3 and 5 for OliP only (F = 6.82, p = 0.015, ηp2 = 0.21, 95% CI range 189.80 to 207.11). Finally, it was also noted that when expressed as relative change, there was a pre- to post- intervention reduction in V̇ CO2 within OliP (−1.6 ± 0.9%) compared to PL (+1.5 ± 0.9%; t = 2.33, p = 0.028, d = 0.90). Table 5. Respiratory, exercise economy, perceived exertion, and blood lactate parameters at the on- set of exercise and lactate threshold (A), and lactate turnpoint (B) at visit 3 (pre-intervention) and 5 (post-intervention) interspersed with 16 consecutive days of either OliP or PL. (A) OliP PL Onset of Exercise Pre Post Pre Post V̇ O2 at baseline (mL·kg−1·min−1) 8.48 ± 1.13 6.88 ± 0.55 6.43 ± 0.81 6.25 ± 0.60 V̇ O2 demand at 60 s (L) 1.69 ± 0.15 1.74 ± 0.12 1.80 ± 0.16 1.78 ± 0.14 V̇ O2 demand at 120 s (L) 3.39 ± 0.29 3.47 ± 0.23 3.60 ± 0.31 3.55 ± 0.28 V̇ O2 demand at 180 s (L) 4.80 ± 0.43 4.94 ± 0.32 5.08 ± 0.43 5.03 ± 0.40 O2 deficit (L) 0.97 ± 0.08 0.85 ± 0.07 0.84 ± 0.07 0.81 ± 0.07 τ (s) 40.5 ± 4.6 30.5 ± 1.5 * 29.2 ± 1.9 28.1 ± 1.1 Mean V̇ O2 in the last 60 s (mL·kg−1·min−1) 30.7 ± 1.5 30.0 ± 1.5 30.3 ± 2.2 31.1 ± 1.6 Lactate threshold (LT1) %V̇ O2max of baseline (%) 73.0 ± 1.8 70.9 ± 1.6 * 68.7 ± 1.4 69.2 ± 1.6 V̇ O2 (L·min−1) 2.56 ± 0.13 2.50 ± 0.13 * 2.45 ± 0.15 2.46 ± 0.13 V̇ CO2 (L·min−1) 2.30 ± 0.12 2.26 ± 0.12 2.16 ± 0.15 2.19 ± 0.13 V̇ E (L·min−1) 71.81 ± 3.55 71.36 ± 3.74 67.62 ± 4.09 68.56 ± 4.19 V̇ E /V̇ O2 26.17 ± 0.36 26.65 ± 0.42 25.83 ± 0.78 25.94 ± 0.88 V̇ E /V̇ CO2 29.22 ± 0.45 29.41 ± 0.44 29.30 ± 0.91 29.21 ± 1.04 RER 0.90 ± 0.01 0.91 ± 0.01 0.88 ± 0.01 0.89 ± 0.01 Economy (mL·kg−1·km−1) 201.3 ± 3.6 195.6 ± 4.0 * 199.4 ± 5.1 201.2 ± 5.9 RPE 2.9 ± 0.3 2.7 ± 0.3 3.2 ± 0.5 3.3 ± 0.1 B[La] (mmol·L−1) 1.32 ± 0.11 1.31 ± 0.10 1.33 ± 0.11 1.33 ± 0.09 (B) OliP PL Lactate turnpoint (LT2) Pre Post Pre Post %V̇ O2max (%) 84.1 ± 1.8 83.5 ± 1.4 81.2 ± 2.0 82.8 ± 2.1 V̇ O2 (L·min−1) 2.96 ± 0.16 2.95 ± 0.17 2.91 ± 0.21 2.96 ± 0.16 V̇ CO2 (L·min−1) 2.81 ± 0.15 2.81 ± 0.15 2.73 ± 0.20 2.79 ± 0.19 V̇ E (L·min−1) 92.73 ± 4.87 92.28 ± 5.16 88.00 ± 5.48 89.80 ± 5.89 V̇ E /V̇ O2 29.46 ± 0.50 29.41 ± 0.52 28.72 ± 0.98 28.60 ± 1.11 V̇ E /V̇ CO2 31.00 ± 0.48 30.80 ± 0.47 30.69 ± 1.04 30.25 ± 1.13 RER 0.95 ± 0.01 0.95 ±0.01 0.94 ± 0.01 0.94 ± 0.01 Economy (mL·kg−1·km−1) 192.7 ± 4.5 191.1 ± 3.5 193.7 ± 4.7 197.7 ± 4.2 RPE 6.0 ± 0.34 5.4 ± 0.4 * 5.6 ± 0.4 5.4 ± 0.4 B[La] (mmol·L−1) 2.31 ± 0.12 2.18 ± 0.11 2.24 ± 0.12 2.22 ± 0.12 τ = time constant; V̇ O2max = maximal oxygen uptake; V̇ O2 = volume of oxygen; V̇ CO2 = volume of carbon dioxide; V̇ E = minute ventilation; RER = respiratory exchange ratio; RPE = rating of perceived exertion; B[La] = blood lactate concentration. * denotes a significant within group difference (p < 0.05). V̇ O2 (mL·kg−1·min−1) ✱ V̇ O2max (%) ✱ Economy (mL·kg−1·km−1) ✱ Figure 3. Summary respiratory and economy differences at lactate threshold (LT1) intensity pre and post 16 consecutive days consumption of either OliP or PL for (A) . VO2; (B) . VO2max % of baseline and (C) running economy. * denotes significance between time points in the OliP group (p < 0.05). Nutrients 2023, 15, 421 11 of 20 Lactate turnpoint (LT2): There was a significant interaction reported for RPE at LT2 (time x group: F = 7.99, p = 0.009, ηp2 = 0.24), where a reduction in RPE was found in the OliP group between visit 3 and 5 only (F = 11.01, p = 0.003, ηp2 = 0.30). No other differences were found within or between groups at this exercise intensity (p > 0.05, Table 5B). 3.4. Recovery from Submaximal Exercise There was a significant interaction effect for T50% in acute recovery responses (time × group: F = 7.72, p = 0.010, ηp2 = 0.24), where post-hoc assessment indicated a reduction in T50% for the OliP group only between visit 3 and 5 (F = 5.67, p = 0.026, ηp2 = 0.19, Figure 4). No other changes were observed for respiratory variables assessed (Table 6). Nutrients 2023, 15, x FOR PEER REVIEW 11 of 20 Figure 3. Summary respiratory and economy differences at lactate threshold (LT1) intensity pre and post 16 consecutive days consumption of either OliP or PL for (A) V̇ O2; (B) V̇ O2max % of baseline and (C) running economy. * denotes significance between time points in the OliP group (p < 0.05). Lactate turnpoint (LT2): There was a significant interaction reported for RPE at LT2 (time x group: F = 7.99, p = 0.009, ηp2 = 0.24), where a reduction in RPE was found in the OliP group between visit 3 and 5 only (F = 11.01, p = 0.003, ηp2 = 0.30). No other differences were found within or between groups at this exercise intensity (p > 0.05, Table 5B). 3.4. Recovery from Submaximal Exercise There was a significant interaction effect for T50% in acute recovery responses (time × group: F = 7.72, p = 0.010, ηp2 = 0.24), where post-hoc assessment indicated a reduction in T50% for the OliP group only between visit 3 and 5 (F = 5.67, p = 0.026, ηp2 = 0.19, Figure 4). No other changes were observed for respiratory variables assessed (Table 6). Figure 4. Recovery to T50% at pre- and post-intervention time points for the OliP and PL groups. * denotes significance between time points in the OliP group (p = 0.026). Table 6. Recovery from submaximal exercise by intervention group (visit 3 and visit 5). Variable OliP PL Pre Post Pre Post EEV̇ O2 (mL·kg−1·min−1) 44.3 ± 2.1 43.8 ± 2.2 46.7 ± 2.4 47.0 ± 2.0 %V̇ O2max (%) 92.8 ± 1.4 91.6 ± 1.8 91.6 ± 7.6 92.5 ± 5.9 Amplitude (mL·kg−1·min−1) 36.9 ± 1.8 36.59 ± 2.1 39.16 ± 2.0 38.5 ± 2.1 τ (s) 51.6 ± 2.6 52.3 ± 9.4 47.9 ± 2.2 49.6 ± 2.7 T50% (s) 55.1 ± 2.2 50.4 ± 3.4 * 53.9 ± 2.0 50.6 ± 1.5 EEV̇ O2 = End of exercise volume of oxygen; V̇ O2max = maximum oxygen uptake; τ = time constant; T50% = 50% of end of exercise volume of oxygen. * denotes a significant within group difference (p = 0.026). 3.5. Time to Exhaustion and V̇ O2max A significant effect was found for time during TTE (F = 11.49, p = 0.002, ηp2 = 0.32) which increased post-intervention for both OliP (+4.1 ± 1.8%) and PL (+5.8 ± 2.6%), with no differences reported between groups for final run speed (12.6 ± 0.5 km∙h−1 for OliP, 12.9 ± 0.7 km∙h−1 for PL, p > 0.05, Table 7). A significant interaction effect was reported for V̇ O2max (time x group: F = 16.79, p = 0.033, ηp2 = 0.17), where V̇ O2max increased post-inter- vention for PL (F = 7.17, p = 0.013, ηp2 = 0.22, 95% CI range 44.37 to 55.24), but not OliP (F = 0.16, p = 0.693, ηp2 = 0.01, 95% CI range 347.39 to 414.41). A significant interaction effect was reported for V̇ CO2max (time x group: F = 18.69, p = 0.018, ηp2 = 0.20), with a post-inter- vention increase in V̇ CO2max reported for the PL (F = 13.77, p = 0.001, ηp2 = 0.36, 95% CI ✱ Figure 4. Recovery to T50% at pre- and post-intervention time points for the OliP and PL groups. * denotes significance between time points in the OliP group (p = 0.026). Table 6. Recovery from submaximal exercise by intervention group (visit 3 and visit 5). Variable OliP PL Pre Post Pre Post EE . VO2 (mL·kg−1·min−1) 44.3 ± 2.1 43.8 ± 2.2 46.7 ± 2.4 47.0 ± 2.0 % . VO2max (%) 92.8 ± 1.4 91.6 ± 1.8 91.6 ± 7.6 92.5 ± 5.9 Amplitude (mL·kg−1·min−1) 36.9 ± 1.8 36.59 ± 2.1 39.16 ± 2.0 38.5 ± 2.1 τ (s) 51.6 ± 2.6 52.3 ± 9.4 47.9 ± 2.2 49.6 ± 2.7 T50% (s) 55.1 ± 2.2 50.4 ± 3.4 * 53.9 ± 2.0 50.6 ± 1.5 EE . VO2 = End of exercise volume of oxygen; . VO2max = maximum oxygen uptake; τ = time constant; T50% = 50% of end of exercise volume of oxygen. * denotes a significant within group difference (p = 0.026). 3.5. Time to Exhaustion and . VO2max A significant effect was found for time during TTE (F = 11.49, p = 0.002, ηp2 = 0.32) which increased post-intervention for both OliP (+4.1 ± 1.8%) and PL (+5.8 ± 2.6%), with no differences reported between groups for final run speed (12.6 ± 0.5 km·h−1 for OliP, 12.9 ± 0.7 km·h−1 for PL, p > 0.05, Table 7). A significant interaction effect was reported for . VO2max (time x group: F = 16.79, p = 0.033, ηp2 = 0.17), where . VO2max increased post- intervention for PL (F = 7.17, p = 0.013, ηp2 = 0.22, 95% CI range 44.37 to 55.24), but not OliP (F = 0.16, p = 0.693, ηp2 = 0.01, 95% CI range 347.39 to 414.41). A significant interaction effect was reported for . VCO2max (time x group: F = 18.69, p = 0.018, ηp2 = 0.20), with a post-intervention increase in . VCO2max reported for the PL (F = 13.77, p = 0.001, ηp2 = 0.36, 95% CI range 47.56 to 60.76) but not the OliP group (F = 0.61, p = 0.444, ηp2 = 0.24, 95% CI range 47.74 to 59.54). Nutrients 2023, 15, 421 12 of 20 Table 7. Time to exhaustion, respiratory, exercise economy and perceived exertion parameters during maximal intensity exercise at visit 3 (pre-intervention) and 5 (post-intervention) for OliP and PL groups. Variable OliP PL Pre Post Pre Post TTE (s) 378.7 ± 13.5 393.1 ± 13.4 * 357.0 ± 15.7 377.5 ± 18.3 * . VO2max (mL·kg−1·min−1) 48.2 ± 2.6 48.0 ± 2.5 48.8 ± 2.6 50.8 ± 2.3 * . VCO2max (mL·kg−1·min−1) 53.4 ± 3.0 53.9 ± 2.8 52.7 ± 3.2 55.6 ± 3.2 * . VE (L·min−1) 141.29 ± 8.24 142.40 ± 8.86 136.76 ± 11.21 142.23 ± 10.70 hRER 1.13 ± 0.01 1.14 ± 0.02 1.11 ± 0.02 1.13 ± 0.02 Speed at . VO2max (km·h−1) 14.7 ± 0.7 14.9 ± 0.7 15.0 ± 0.9 15.4 ± 0.8 RPE 9.8 ± 0.1 9.6 ± 0.2 9.6 ± 0.1 9.8 ± 0.1 TTE = time to exhaustion; . VO2max = maximal oxygen uptake; . VO2 = volume of oxygen; . VCO2 = volume of carbon dioxide; . VE = minute ventilation; RER = respiratory exchange ratio; RPE = rating of perceived exertion. * denotes a significant within group difference (p ≤ 0.031). 4. Discussion To the authors’ knowledge, this is the first study to undertake research focusing on OliP in an exercise domain and aligns with concurrent research pertinent to olive-derived phytonutrients [45]. The key findings from this study demonstrate that 16 consecutive days consumption of OliP resulted in positive effects on several key markers of running performance. Of particular interest, OliP consumption significantly improved respiratory parameters at the onset of exercise within condition (i.e., τ), and oxygen consumption and running economy at LT1 (particularly when expressed as relative change in comparison to PL). Whilst respiratory parameters at LT2 were largely unaffected by OliP, perceived exertion was improved with the phytocomplex beverage. Acute recovery (T50%) following incremental exercise was also notably improved with OliP. Whilst maximal effort and TTE measures were not different between OliP and PL, an elevated . VCO2max was reported for PL only. Furthermore, it was noted that both groups improved TTE following the intervention. Importantly, no adverse effects were reported throughout the intervention. Regarding methodological approaches to the demanding aerobic session, steady- state moderate intensity exercise (60–70% . VO2max) for 30–60 min followed by arduous (90% . VO2max) [57] or performance efforts [58] have been shown to provoke a heightened oxidative stress response and elicit peripheral fatigue. Accordingly, the demanding aerobic sessions employed in the current study resulted in an intensity of ~75% . VO2max, with no differences within or between groups. It can therefore be assumed that an equal degree of physiological strain was achieved between cohorts prior to the main performance tests. As dietary and exercise habits were maintained across the intervention, it is feasible that physiological adaptations observed, may therefore be partly attributed to the phenolic compounds within OliP. As a naturally derived phytocomplex, OliP is notably rich in HT, which is a key polyphenol of interest and may support endogenous antioxidant mechanisms pertinent to mitochondrial respiratory capacity and/or efficiency, such as upregulation of PGC-1α [28,59–62]. Consumption of OliP may therefore be of relevance to individuals who engage in regular aerobic exercise, considering the negligible dietary HT content in both the pre-visit control period and habitual diet assessments for both cohorts. Plasma HT concentrations were not detected at baseline (pre-supplementation), or post PL, but were significantly elevated in response to the OliP intervention. Therefore, any impact on aerobic exercise may be associated with increased systemic HT concentrations, or gut-derived metabolites. At present, however, there is a paucity of scientific research surrounding HT and exercise performance. Additionally, there does not appear to be any existing research evidencing the effects of HT on aerobic running performance in humans. Plant-based polyphenols have peaked scientific interest in recent years [29,63], in particular HT, due to its potential to impact multiple physiological pathways. In an exercise domain, recent animal studies Nutrients 2023, 15, 421 13 of 20 have demonstrated the ability of HT to enhance endurance capacity [59], prevent exercise induced fatigue, muscle damage and immunosuppression [64,65] and improve mitochon- drial function in both trained and sedentary rodents [65]. However, these findings need to be corroborated in human models as well as within an exercise domain to ascertain the efficacy of HT-rich supplements. It is also important to outline the current debate surrounding the efficacy of antiox- idant and polyphenol supplementation as an exercise or training aid. Adaptations from exercise are dependent, at least partially, on individual oxidative stress responses [66]. One perspective highlights the potential inhibition of natural training adaptations through lim- iting the upregulation of endogenous antioxidant enzymes, and therefore diminishing the hormetic response to moderate exercise [13]. However, the counterargument highlights that the subsequent reduction in oxidative stress following antioxidant and/or polyphenol sup- plementation may positively influence recovery kinetics, development in contractile force, calcium handling, and therefore the ability to exercise and/or recover more ‘economically’. This may facilitate adaptations to exercise training and/or athletic performance [17]. Findings from this study demonstrated a ~17% improvement in τ at the onset of exercise for OliP. τ reflects the speed at which the steady-state is achieved [53], and in turn the size of the O2 deficit [67]. However, these results were only significant within condition and should therefore be interpreted with caution. In addition, it was noted that non-significant differences were observed between conditions prior to the nutritional inter- vention based on random participant allocation, which may in part impact the observed findings. Contrary to these findings, Breese et al. [68] reported no differences in . VO2 phase II time constant, from unloaded to moderate exercise after 6-days supplementation with beetroot juice (BTJ; ~8 mmol nitrate (NO3−)). Although mitochondrial respiratory capacity was not assessed, it is known that the speed of the O2 uptake response during the onset of moderate exercise intensity is associated with the respiratory capacity of mitochondrial complex II and the capacity of the mitochondrial electron transport system [69,70]. As HT has been shown to improve the expression of mitochondrial complex I/II/IV, this is of particular interest in an exercise domain as complex I is recognized to be the primary complex for the electron transport chain [71]. Moreover, HT has been reported to promote the congregation of complex I (CI) into supercomplexes (SCs) [65], therefore decreasing the diffusion distance for transfer of electrons between complexes, and improving the efficiency of the mitochondrial electron transfer between complexes [59,72]. More research is required to ascertain the above stated mechanisms in humans, particularly in relation to OliP consumption. Consumption of OliP in the dose provided also resulted a significant decrease (−2.7%) in . VO2 consumption at LT1 compared with PL. This aligns with existing research into both high [73]- and low [74]-dose BTJ supplementation whereby a ~5% reduction . VO2 consumption was reported with no changes in . VE, RER or HR [74]. This modest change could be partially attributed to the increase in mitochondrial function and increased ex- pression of PGC-1α following supplementation of OliP. In vitro, HT administration has been shown to upregulate nuclear respiratory factors 1 and 2, mitochondrial transcription factor A, and peroxisome proliferator active receptor γ (PPAR γ) in response to increased phosphorylation of adenosine monophosphate kinase (AMPK) [61]. The role HT may play in enhancing mitochondrial respiratory capacity could also provide a rationale for the reduced oxygen consumption observed during sub-maximal exercise at low to moderate intensities (LT1). In vitro, HT has been shown to improve mitochondrial biogenesis, O2 and fatty acid utilization in adipocyte cells [61,75,76]. Although not measured in this study, this may support the proposed benefits of OliP in a submaximal exercise domain, however, more research is required in humans to confirm such mechanisms. It is also viable that other phenolic compounds [38] (i.e., oleuropein aglycone) and HT derivatives (i.e., HT glucosides) found in OliP may also support antioxidant pathways that may influence aerobic performance [59,77]. Indeed, olive-derived phenolic compounds are not entirely absorbed during digestion and are extensively transformed into different metabolites by Nutrients 2023, 15, 421 14 of 20 the gut microbiota [78]. For instance, whilst oleuropein transformation by gut bacteria can increase HT yield [79], HT is further transformed into homovanillin derivatives [80,81] and glutathione conjugates [79] which may have pertinent antioxidant properties [82]. These metabolites may exert further physiological effects [83] potentially explaining findings from the current study. Furthermore, HT-derived metabolite variability and quantity are also de- pendent on the phenolic composition of the product consumed [81]. Such complexities should also be considered when determining the physiological impact of combined polyphenols. A relevant parameter of aerobic performance is the efficiency of movement, i.e., ex- ercise economy [84]. This reflects the amount of O2 required to generate a constant sub- maximal running speed and therefore, is directly associated with the efficiency of aerobic fuel metabolism and the sparing of glycogen reserves [85]. Mitochondria are crucial for aerobic energy generation in exercise [86]. Improvement in mitochondrial respiratory capacity and functional efficiency following HT supplementation in animal studies has been established [87] and is associated with the constitution of supramolecular entities, the mitochondrial SCs, including respiratory complex I, III, and IV [88]. Administration of HT for 10-weeks in rodents (20 vs. 300 mg·kg−1·d−1) and exercise (up to 65 min a day at 75% of maximal velocity) compared to exercise alone improved mitochondrial function and antioxidant capacity induced by exercise [65]. However, when the HT dose was increased to 300 mg·kg−1·d−1, pro-oxidant effects were evident [65], which appeared to negatively influence SCs assembly, aligning with existing published literature [64]. Collectively, these results indicate that whilst exercise induces the formation of mitochondrial SCs [89], low- dose HT consumption may support or enhance this process [65] whilst a high dose of HT may provoke pro-oxidant mechanisms, disrupting the mitochondria and potentially limiting or diminishing SC adaptation [64]. In the current study, a relatively low HT dose was employed as part of the olive-derived phytocomplex (~0.8 mg·kg−1·d−1) in healthy volunteers. Whilst mitochondrial function was not directly assessed, it is feasible that HT and related gut-derived phenolic metabolites may have supported SC assembly and facilitated improved oxygen cost responses observed at the onset of exercise and during low to moderate exercise (LT1). Furthermore, the low HT dose employed in the current study may also explain why exercise performance (TTE) was not significantly different between cohorts in line with previous research [64]. Despite OliP presenting no significant impact on respiratory mechanisms at LT2, a poignant finding was the observed significant reduction in RPE at this intensity. Mecha- nisms for this are unclear, however it is feasible that there may be a link to a reduction in brain oxygenation that is present during intensive exercise and directly associated with an increase of fatigue (subjectively quantified as perceived exertion) [90]. Alternatively, mechanisms potentially occurring at a mitochondrial level and the effect upon SCI and SCII, may indicate that beneficial responses to OliP are more likely to be present at lower intensities only. Further research is required to accurately ascertain potential mechanisms involved in subjective measures associated with exercise. Similarly, recovery was largely unaffected based upon off-kinetic modeling; however, current findings did present a −9.4% decrease in T50% for OliP compared with a −5.6% decrease in PL, during the initial recovery period from sub-maximal exercise. This in itself warrants further investigation considering that previous findings utilizing a similar exercise intensity (70% maximum aerobic power) did not find a benefit to . VO2 half-recovery time following the supplementation of mixed polyphenols (250 mg Vinitrox™ for 7-days [19]). In the current study it is feasible that the HT content in OliP (and related gut-dervied metabolites) may be influencing recovery indirectly, and may therefore have applications following repeated bouts of exercise. However, results should be interpreted with caution and further research should be undertaken to corroborate findings. Finally, although improvements in TTE were evident in both groups (+4.1% OliP and +5.8% PL), the overall change in exercise performance was not different between OliP and PL. In the current study, exercise performance was based upon physical tolerance to sustained near-maximal exercise. Based upon findings at LT1 intensity, it could be prudent to assess whether OliP Nutrients 2023, 15, 421 15 of 20 is more effective when determining performance employing other measures such as an extended time trial (i.e., 5 km run time) or total work completed in a fixed time period as opposed to an acute near-maximal TTE bout. Study Limitations and Future Directions It is important to note that there were several limitations to the current study. Firstly, improvements were found in specific, but not all parameters assessed. As example, change in time constant at the onset of exercise was noted within-group only for OliP and therefore should be interpreted with caution. Likewise, during acute recovery, whilst improvements were observed for T50%, other parameters using respiratory off-kinetics were not deemed significant, and again results should be interpreted carefully. However, where main in- teraction effects were identified (including relative changes in oxygen consumption and running economy at LT1 compared with PL) important adaptations following the inclusion of dietary OliP may be evident. It should, however, be noted that differences were not observed post-intervention between conditions which should be taken into consideration. Although improvements were observed particularly at LT1 with healthy, recreationally active volunteers, we did not specifically distinguish whether such effects were pertinent to gender, training status, or the type/intensity of habitual exercise. Further research may therefore be relevant to determine the potential applications of OliP in various cohorts. Additionally, the protocol used in the current study was designed to standardize the demanding aerobic run prior to the following day exercise performance session for all participants [45]. It is important to recognize the translation of controlled laboratory findings to real-world exercise applications [91], and future research should investigate the adjunct use of OliP in applied and field-based settings (e.g., single exercise sessions, events that require repeated bouts, or multi-day events). It should also be noted that existing literature has outlined the potential variability in polyphenol products [92]. Whilst 16-days of OliP consumption (Batch 14) positively influenced aerobic exercise parameters and acute recovery, results may differ between batches and additional investigation is needed to corroborate current findings. Indeed, as previously noted, the intestinal microbiota plays an important role influencing gut-derived phenolic metabolites, which are additionally dependent on the phenolic composition of dietary products. As this was the first study to assess the use of OliP in an exercise domain, a parallel co- hort design was employed to ascertain the influence of a single course of the phytocomplex (16 consecutive days) whilst minimizing potential for longer term training effects. Further research should investigate whether time course (>16-days), dose-response (>56 mL· d−1) and/or dose-frequency (>2 serves·d−1) can influence sustained exercise training adap- tations or accumulated recovery, i.e., during marathon training. Additional exploration into alternate recovery periods (i.e., respiratory measures up to 1 h post exercise, and inflammatory or muscle soreness measures 1, 12, 24, and 48 h+ following exhaustive ex- ercise) is also warranted. Finally, based upon current findings, including effects of OliP on exercise-induced oxidative stress presented elsewhere [45], it would be beneficial to assess the potential impact of this olive-derived phytocomplex on inflammatory markers associated with EIMD (particularly within other populations, e.g., trained athletes), or within clinical applications where functional movement may be impacted (e.g., arthritis, fibromyalgia). 5. Conclusions This is the first study to investigate the use of OliP in an exercise domain. Findings demonstrated that 16-days supplementation of OliP positively influenced parameters of aerobic exercise, most notably at submaximal levels. Reduced oxygen cost and improved running economy at exercise intensities corresponding with LT1, as well as improvements in acute recovery may have implications for recreationally active individuals undertaking demanding or repeated aerobic exercise training. Further research is warranted to cor- roborate these findings and explore potential applications (time-course, dose-response) to prolonged training periods and/or repetitive bouts of exercise. Nutrients 2023, 15, 421 16 of 20 Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nu15020421/s1, Table S1: List of food sources requested to be avoided by participants for the 3-days prior to main laboratory testing sessions; Table S2: Independent product analysis. Author Contributions: Conceptualization, J.D.R.; methodology, J.D.R., J.B.L., J.M.P., A.G.B.W. and H.C.; formal analysis, J.D.R., J.B.L., J.M.P. and R.Z.; investigation, J.D.R., J.B.L., J.M.P. and H.C.; resources, J.D.R., J.B.L. and J.M.P.; data curation, J.D.R., J.B.L. and J.M.P.; writing—original draft preparation, J.D.R., J.B.L. and J.M.P.; writing—review and editing, J.D.R., A.G.B.W., J.M.P., J.B.L., Á.L.-S., J.D.C., R.Z. and H.C.; supervision, J.D.R.; project administration, J.D.R. and H.C.; funding acquisition, J.D.R. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Fattoria La Vialla, Castiglion Fibocchi, Arezzo, Italy for product and related consumables, and research/analytical costs (Number: R9039). This study was undertaken independently of the funding company. Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and approved by Faculty of Science and Engineering Research Ethics Panel, Anglia Ruskin University (approval number: FSE/FREP/20/946; date: 13 October 2020). Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: The data presented in this study is available upon request from the corresponding authors. The data is not publicly available due to ethical considerations, in accordance with participant consent on the use of confidential data. Acknowledgments: The authors would like to acknowledge the team at Fattoria la Vialla, Castiglion Fibocchi, Arezzo, Italy, in particular Rosa Briamonte, for their open communication and support in this study. Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript or in the decision to publish results. References 1. Hutchison, A.T.; Flieller, E.B.; Dillon, K.J.; Leverett, B.D. Black Currant Nectar Reduces Muscle Damage and Inflammation Following a Bout of High-Intensity Eccentric Contractions. J. Diet. Suppl. 2014, 13, 1–15. [CrossRef] 2. Kelley, D.S.; Rasooly, R.; Jacob, R.A.; Kader, A.A.; Mackey, B.E. Consumption of Bing Sweet Cherries Lowers Circulating Concentrations of Inflammation Markers in Healthy Men and Women. J. Nutr. 2006, 136, 981–986. [CrossRef] 3. Martín, M.A.; Ramos, S. Cocoa polyphenols in oxidative stress: Potential health implications. J. Funct. Foods 2016, 27, 570–588. [CrossRef] 4. Scalbert, A.; Williamson, G. Dietary Intake and Bioavailability of Polyphenols. J. Nutr. 2000, 130, 2073S–2085S. [CrossRef] [PubMed] 5. Scalbert, A.; Morand, C.; Manach, C.; Rémésy, C. Absorption and metabolism of polyphenols in the gut and impact on health. Biomed. Pharmacother. 2002, 56, 276–282. [CrossRef] 6. Tsao, R. Chemistry and Biochemistry of Dietary Polyphenols. Nutrients 2010, 2, 1231–1246. [CrossRef] [PubMed] 7. Cory, H.; Passarelli, S.; Szeto, J.; Tamez, M.; Mattei, J. The Role of Polyphenols in Human Health and Food Systems: A Mini-Review. Front. Nutr. 2018, 5, 87. [CrossRef] 8. Heydenreich, J.; Kayser, B.; Schutz, Y.; Melzer, K. Total Energy Expenditure, Energy Intake, and Body Composition in Endurance Athletes Across the Training Season: A Systematic Review. Sports Med. -Open 2017, 3, 8. [CrossRef] 9. Abbiss, C.R.; Laursen, P.B. Models to Explain Fatigue during Prolonged Endurance Cycling. Sports Med. 2005, 35, 865–898. [CrossRef] 10. Hausswirth, C.; Le Meur, Y. Physiological and nutritional aspects of post-exercise recovery: Specific recommendations for female athletes. Sports Med. 2011, 41, 861–882. [CrossRef] 11. D’Angelo, S. Polyphenols and Athletic Performance: A Review on Human Data. In Plant Physiological Aspects of Phenolic Compounds; Books on Demand: London, UK, 2019. [CrossRef] 12. Finkel, T. Signal transduction by reactive oxygen species. J. Cell Biol. 2011, 194, 7–15. [CrossRef] [PubMed] 13. Gomez-Cabrera, M.-C.; Domenech, E.; Viña, J. Moderate exercise is an antioxidant: Upregulation of antioxidant genes by training. Free. Radic. Biol. Med. 2008, 44, 126–131. [CrossRef] [PubMed] 14. Margaritelis, N.V.; Theodorou, A.A.; Paschalis, V.; Veskoukis, A.S.; Dipla, K.; Zafeiridis, A.; Panayiotou, G.; Vrabas, I.S.; Kyparos, A.; Nikolaidis, M.G. Adaptations to endurance training depend on exercise-induced oxidative stress: Exploiting redox interindividual variability. Acta Physiol. 2017, 222, e12898. [CrossRef] [PubMed] Nutrients 2023, 15, 421 17 of 20 15. Tryfidou, D.V.; McClean, C.; Nikolaidis, M.G.; Davison, G.W. DNA Damage Following Acute Aerobic Exercise: A Systematic Review and Meta-analysis. Sports Med. 2020, 50, 103–127. [CrossRef] 16. Donato, A.J.; Uberoi, A.; Bailey, D.M.; Wray, D.W.; Richardson, R.S. Exercise-induced brachial artery vasodilation: Effects of antioxidants and exercise training in elderly men. Am. J. Physiol. Circ. Physiol. 2010, 298, H671–H678. [CrossRef] 17. Powers, S.K.; Talbert, E.E.; Adhihetty, P.J. Reactive oxygen and nitrogen species as intracellular signals in skeletal muscle. J. Physiol. 2011, 589, 2129–2138. [CrossRef] 18. Perkins, I.C.; Vine, S.A.; Blacker, S.D.; Willems, M. New Zealand Blackcurrant Extract Improves High-Intensity Intermittent Running. Int. J. Sport Nutr. Exerc. Metab. 2015, 25, 487–493. [CrossRef] 19. Deley, G.; Guillemet, D.; Allaert, F.; Babault, N. An Acute Dose of Specific Grape and Apple Polyphenols Improves Endurance Performance: A Randomized, Crossover, Double-Blind versus Placebo Controlled Study. Nutrients 2017, 9, 917. [CrossRef] 20. Fisher, N.D.L.; Hughes, M.; Gerhard-Herman, M.; Hollenberg, N.K. Flavanol-rich cocoa induces nitric-oxide-dependent vasodila- tion in healthy humans. J. Hypertens. 2003, 21, 2281–2286. [CrossRef] 21. Davis, J.M.; Carlstedt, C.J.; Chen, S.; Carmichael, M.D.; Murphy, E.A. The Dietary Flavonoid Quercetin Increases VO2max and Endurance Capacity. Int. J. Sport Nutr. Exerc. Metab. 2010, 20, 56–62. [CrossRef] 22. Davis, J.M.; Murphy, E.A.; Carmichael, M.D.; Davis, B. Quercetin increases brain and muscle mitochondrial biogenesis and exercise tolerance. Am. J. Physiol. Integr. Comp. Physiol. 2009, 296, R1071–R1077. [CrossRef] 23. Westphal, C.; Dipp, M.; Guarente, L. A therapeutic role for sirtuins in diseases of aging? Trends Biochem. Sci. 2007, 32, 555–560. [CrossRef] 24. Chung, S.; Yao, H.; Caito, S.; Hwang, J.-W.; Arunachalam, G.; Rahman, I. Regulation of SIRT1 in cellular functions: Role of polyphenols. Arch. Biochem. Biophys. 2010, 501, 79–90. [CrossRef] 25. Kamei, Y.; Miura, S.; Suzuki, M.; Kai, Y.; Mizukami, J.; Taniguchi, T.; Mochida, K.; Hata, T.; Matsuda, J.; Aburatani, H.; et al. Skeletal Muscle FOXO1 (FKHR) Transgenic Mice Have Less Skeletal Muscle Mass, Down-regulated Type I (Slow Twitch/Red Muscle) Fiber Genes, and Impaired Glycemic Control. J. Biol. Chem. 2004, 279, 41114–41123. [CrossRef] 26. Lu, H.; Huang, H. FOXO1: A potential target for human diseases. Curr. Drug Targets 2011, 12, 1235–1244. [CrossRef] 27. Lagouge, M.; Argmann, C.; Gerhart-Hines, Z.; Meziane, H.; Lerin, C.; Daussin, F.; Messadeq, N.; Milne, J.; Lambert, P.; Elliott, P.; et al. Resveratrol improves mitochondrial function and protects against metabolic disease by activating SIRT1 and PGC-1α. Cell 2006, 127, 1109–1122. [CrossRef] 28. Wood dos Santos, T.; Cristina Pereira, Q.; Teixeira, L.; Gambero, A.; Villena, J.A.; Lima Ribeiro, M. Effects of polyphenols on thermogenesis and mitochondrial biogenesis. Int. J. Mol. Sci. 2018, 19, 2757. [CrossRef] 29. Bowtell, J.; Kelly, V. Fruit-Derived Polyphenol Supplementation for Athlete Recovery and Performance. Sports Med. 2019, 49, 3–23. [CrossRef] 30. Myburgh, K.H. Polyphenol supplementation: Benefits for exercise performance or oxidative stress? Sports Med. 2014, 44, 57–70. [CrossRef] 31. Bowtell, J.L.; Sumners, D.P.; Dyer, A.; Fox, P.; Mileva, K.N. Montmorency Cherry Juice Reduces Muscle Damage Caused by Intensive Strength Exercise. Med. Sci. Sports Exerc. 2011, 43, 1544–1551. [CrossRef] 32. Connolly, D.; McHugh, M.; Padilla-Zakour, O. Efficacy of a tart cherry juice blend in preventing the symptoms of muscle damage. Brit. J. Sports Med. 2006, 40, 679–683. [CrossRef] 33. Howatson, G.; McHugh, M.P.; Hill, J.A.; Brouner, J.; Jewell, A.P.; Van Someren, K.A.; Shave, R.E.; Howatson, S.A. Influence of tart cherry juice on indices of recovery following marathon running. Scand. J. Med. Sci. Sports 2010, 20, 843–852. [CrossRef] 34. Rietjens, S.J.; Bast, A.; de Vente, J.; Haenen, G.R.M.M. The olive oil antioxidant hydroxytyrosol efficiently protects against the oxidative stress-induced impairment of the NO• response of isolated rat aorta. Am. J. Physiol. Circ. Physiol. 2007, 292, H1931–H1936. [CrossRef] 35. Rietjens, S.J.; Bast, A.; Haenen, G.R.M.M. New Insights into Controversies on the Antioxidant Potential of the Olive Oil Antioxidant Hydroxytyrosol. J. Agric. Food Chem. 2007, 55, 7609–7614. [CrossRef] 36. Covas, M.-I.; Nyyssönen, K.; Poulsen, H.E.; Kaikkonen, J.; Zunft, H.-J.F.; Kiesewetter, H.; Gaddi, A.; de la Torre, R.; Mursu, J.; Bäumler, H. The effect of polyphenols in olive oil on heart disease risk factors: A randomized trial. Ann. Intern. Med. 2006, 145, 333–341. [CrossRef] 37. Marrugat, J.; the members of the SOLOS Investigators; Covas, M.-I.; Fitó, M.; Schroder, H.; Miró-Casas, E.; Gimeno, E.; López- Sabater, M.C.; de la Torre, R.; Farré, M. Effects of differing phenolic content in dietary olive oils on lipids and LDL oxidation. Eur. J. Nutr. 2004, 43, 140–147. [CrossRef] 38. Weinbrenner, T.; Fitó, M.; de la Torre, R.; Saez, G.T.; Rijken, P.; Tormos, C.; Coolen, S.; Albaladejo, M.F.; Abanades, S.; Schroder, H.; et al. Olive Oils High in Phenolic Compounds Modulate Oxidative/Antioxidative Status in Men. J. Nutr. 2004, 134, 2314–2321. [CrossRef] 39. Manna, C.; Galletti, P.; Cucciolla, V.; Moltedo, O.; Leone, A.; Zappia, V. The protective effect of the olive oil polyphenol (3,4- Dihydroxyphenyl)-ethanol counteracts reactive oxygen metabolite–induced cytotoxicity in Caco-2 cells. J. Nutr. 1997, 127, 286–292. [CrossRef] 40. Goya, L.; Mateos, R.; Bravo, L. Effect of the olive oil phenol hydroxytyrosol on human hepatoma HepG2 cells. Eur. J. Nutr. 2007, 46, 70–78. [CrossRef] 41. Faul, F.; Erdfelder, E.; Lang, A.-G.; Buchner, A. G* Power 3: A flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav. Res. Methods 2007, 39, 175–191. [CrossRef] Nutrients 2023, 15, 421 18 of 20 42. Riva, A.; Vitale, J.A.; Belcaro, G.; Hu, S.; Feragalli, B.; Vinciguerra, G.; Cacchio, M.; Bonanni, E.; Giacomelli, L.; Eggenhöffner, R.; et al. Quercetin phytosome® in triathlon athletes: A pilot registry study. Minerva Med. 2018, 109, 285–289. [CrossRef] 43. Hopkins, W.G.; Schabort, E.J.; Hawley, J.A. Reliability of Power in Physical Performance Tests. Sports Med. 2001, 31, 211–234. [CrossRef] 44. Ward, S.A. Open-circuit respirometry: Real-time, laboratory-based systems. Eur. J. Appl. Physiol. 2018, 118, 875–898. [CrossRef] 45. Roberts, J.D.; Lillis, J.; Pinto, J.M.; Willmott, A.G.B.; Gautam, L.; Davies, C.; López-Samanes, Á.; Del Coso, J.; Chichger, H. The Impact of a Natural Olive-Derived Phytocomplex (OliPhenolia®) on Exercise-Induced Oxidative Stress in Healthy Adults. Nutrients 2022, 14, 5156. [CrossRef] 46. James, C.; Richardson, A.; Watt, P.W.; Willmott, A.; Gibson, O.R.; Maxwell, N.S. Short-term heat acclimation improves the determinants of endurance performance and 5-km running performance in the heat. Appl. Physiol. Nutr. Metab. 2017, 42, 285–294. [CrossRef] 47. Longman, D.P.; Merzbach, V.; Pinto, J.M.; Atkinson, L.H.; Wells, J.C.K.; Gordon, D.; Stock, J.T. Alternative Metabolic Strategies are Employed by Endurance Runners of Different Body Sizes; Implications for Human Evolution. Adapt. Hum. Behav. Physiol. 2022, 8, 79–97. [CrossRef] 48. Jones, A.M.; Doust, J.H. A 1% treadmill grade most accurately reflects the energetic cost of outdoor running. J. Sports Sci. 1996, 14, 321–327. [CrossRef] 49. Bentley, D.J.; Newell, J.; Bishop, D. Incremental Exercise Test Design and Analysis. Sports Med. 2007, 37, 575–586. [CrossRef] 50. Faude, O.; Kindermann, W.; Meyer, T. Lactate Threshold Concepts. Sports Med. 2009, 39, 469–490. [CrossRef] 51. Winter, E.M.; Jones, A.M.; Davison, R.R.; Bromley, P.D.; Mercer, T.H. Sport and Exercise Physiology Testing Guidelines: Volume I–Sport Testing: The British Association of Sport and Exercise Sciences Guide; Routledge: London, UK, 2006. 52. Zacca, R.; Azevedo, R.; Figueiredo, P.; Vilas-Boas, J.P.; Castro, F.A.D.S.; Pyne, D.B.; Fernandes, R.J. VO2FITTING: A Free and Open-Source Software for Modelling Oxygen Uptake Kinetics in Swimming and other Exercise Modalities. Sports 2019, 7, 31. [CrossRef] 53. Jones, A.M.; Poole, D.C. Oxygen Uptake Dynamics: From Muscle to Mouth—An Introduction to the Symposium. Med. Sci. Sports Exerc. 2005, 37, 1542–1550. [CrossRef] 54. Robergs, R.A.; Dwyer, D.; Astorino, T. Recommendations for Improved Data Processing from Expired Gas Analysis Indirect Calorimetry. Sports Med. 2010, 40, 95–111. [CrossRef] 55. Roberts, J.; Willmott, A.; Beasley, L.; Boal, M.; Davies, R.; Martin, L.; Chichger, H.; Gautam, L.; Del Coso, J. The Impact of Decaffeinated Green Tea Extract on Fat Oxidation, Body Composition and Cardio-Metabolic Health in Overweight, Recreationally Active Individuals. Nutrients 2021, 13, 764. [CrossRef] 56. Shapiro, S.S.; Wilk, M.B.; Chen, H.J. A comparative study of various tests for normality. J. Am. Stat. Assoc. 1968, 63, 1343–1372. [CrossRef] 57. Bloomer, R.J.; Goldfarb, A.H.; Wideman, L.; Mckenzie, M.J.; Consitt, L.A. Effects of acute aerobic and anaerobic exercise on blood markers of oxidative stress. J. Strength Cond. Res. 2005, 19, 276–285. [CrossRef] 58. Bloomer, R.J.; Goldfarb, A.H.; McKenzie, M.J. Oxidative stress response to aerobic exercise: Comparison of antioxidant supple- ments. Med. Sci. Sport. Exerc. 2006, 38, 1098–1105. [CrossRef] 59. Feng, Z.; Bai, L.; Yan, J.; Li, Y.; Shen, W.; Wang, Y.; Wertz, K.; Weber, P.; Zhang, Y.; Chen, Y.; et al. Mitochondrial dynamic remodeling in strenuous exercise-induced muscle and mitochondrial dysfunction: Regulatory effects of hydroxytyrosol. Free Radic. Biol. Med. 2011, 50, 1437–1446. [CrossRef] 60. Friedel, A.; Raederstorff, D.; Roos, F.; Toepfer, C.; Wertz, K. Hydroxytyrosol Benefits Muscle Differentiation and Muscle Contraction and Relaxation. U.S. Patent Application No 13,550,972, 7 March 2013. 61. Hao, J.; Shen, W.; Yu, G.; Jia, H.; Liu, J.; Feng, Z.; Wang, Y.; Weber, P.; Wertz, K.; Sharman, E. Hydroxytyrosol promotes mitochondrial biogenesis and mitochondrial function in 3T3-L1 adipocytes. J. Nutr. Biochem. 2010, 21, 634–644. [CrossRef] 62. Signorile, A.; Micelli, L.; De Rasmo, D.; Santeramo, A.; Papa, F.; Ficarella, R.; Gattoni, G.; Scacco, S.; Papa, S. Regulation of the biogenesis of OXPHOS complexes in cell transition from replicating to quiescent state: Involvement of PKA and effect of hydroxytyrosol. BBA-Mol. Cell Res. 2014, 1843, 675–684. 63. Leri, M.; Scuto, M.; Ontario, M.L.; Calabrese, V.; Calabrese, E.J.; Bucciantini, M.; Stefani, M. Healthy Effects of Plant Polyphenols: Molecular Mechanisms. Int. J. Mol. Sci. 2020, 21, 1250. [CrossRef] 64. Al Fazazi, S.; Casuso, R.A.; Aragón-Vela, J.; Casals, C.; Huertas, J.R. Effects of hydroxytyrosol dose on the redox status of exercised rats: The role of hydroxytyrosol in exercise performance. J. Int. Soc. Sports Nutr. 2018, 15, 1–7. [CrossRef] 65. Casuso, R.A.; Al-Fazazi, S.; Hidalgo-Gutierrez, A.; López, L.C.; Plaza-Díaz, J.; Rueda-Robles, A.; Huertas, J.R. Hydroxytyrosol influences exercise-induced mitochondrial respiratory complex assembly into supercomplexes in rats. Free Radical Bio. Med. 2019, 134, 304–310. [CrossRef] 66. Simioni, C.; Zauli, G.; Martelli, A.M.; Vitale, M.; Sacchetti, G.; Gonelli, A.; Neri, L.M. Oxidative stress: Role of physical exercise and antioxidant nutraceuticals in adulthood and aging. Oncotarget 2018, 9, 17181–17198. [CrossRef] 67. Bailey, S.J.; Wilkerson, D.P.; DiMenna, F.J.; Jones, A.M. Influence of repeated sprint training on pulmonary O2 uptake and muscle deoxygenation kinetics in humans. J. Appl. Physiol. 2009, 106, 1875–1887. [CrossRef] Nutrients 2023, 15, 421 19 of 20 68. Breese, B.C.; McNarry, M.A.; Marwood, S.; Blackwell, J.R.; Bailey, S.J.; Jones, A.M. Beetroot juice supplementation speeds O2uptake kinetics and improves exercise tolerance during severe-intensity exercise initiated from an elevated metabolic rate. Am. J. Physiol. Integr. Comp. Physiol. 2013, 305, R1441–R1450. [CrossRef] 69. Christensen, P.M.; Jacobs, R.A.; Bonne, T.C.; Flück, D.; Bangsbo, J.; Lundby, C. A short period of high-intensity interval training improves skeletal muscle mitochondrial function and pulmonary oxygen uptake kinetics. J. Appl. Physiol. 2016, 120, 1319–1327. [CrossRef] 70. Navarro, A.; Boveris, A. The mitochondrial energy transduction system and the aging process. Am. J. Physiol. Physiol. 2007, 292, C670–C686. [CrossRef] 71. Zheng, A.; Li, H.; Xu, J.; Cao, K.; Li, H.; Pu, W.; Yang, Z.; Peng, Y.; Long, J.; Liu, J.; et al. Hydroxytyrosol improves mitochondrial function and reduces oxidative stress in the brain of db/db mice: Role of AMP-activated protein kinase activation. Br. J. Nutr. 2015, 113, 1667–1676. [CrossRef] 72. Cogliati, S.; Enriquez, J.A.; Scorrano, L. Mitochondrial Cristae: Where Beauty Meets Functionality. Trends Biochem. Sci. 2016, 41, 261–273. [CrossRef] 73. Whitfield, J.; Ludzki, A.; Heigenhauser, G.J.F.; Senden, J.M.G.; Verdijk, L.B.; van Loon, L.J.C.; Spriet, L.L.; Holloway, G.P. Beetroot juice supplementation reduces whole body oxygen consumption but does not improve indices of mitochondrial efficiency in human skeletal muscle. J. Physiol. 2015, 594, 421–435. [CrossRef] 74. Bailey, S.J.; Winyard, P.; Vanhatalo, A.; Blackwell, J.R.; DiMenna, F.J.; Wilkerson, D.P.; Tarr, J.; Benjamin, N.; Jones, A.M. Dietary nitrate supplementation reduces the O2 cost of low-intensity exercise and enhances tolerance to high-intensity exercise in humans. J. Appl. Physiol. 2009, 107, 1144–1155. [CrossRef] 75. Liu, F.; Wanigatunga, A.A.; Zampino, M.; Knuth, N.D.; Simonsick, E.M.; Schrack, J.A.; Ferrucci, L. Association of Mitochondrial Function, Substrate Utilization, and Anaerobic Metabolism With Age-Related Perceived Fatigability. J. Gerontol. Ser. A 2020, 76, 426–433. [CrossRef] 76. Zhu, L.; Liu, Z.; Feng, Z.; Hao, J.; Shen, W.; Li, X.; Sun, L.; Sharman, E.; Wang, Y.; Wertz, K.; et al. Hydroxytyrosol protects against oxidative damage by simultaneous activation of mitochondrial biogenesis and phase II detoxifying enzyme systems in retinal pigment epithelial cells. J. Nutr. Biochem. 2010, 21, 1089–1098. [CrossRef] 77. Ditano-Vázquez, P.; Torres-Peña, J.D.; Galeano-Valle, F.; Pérez-Caballero, A.I.; Demelo-Rodríguez, P.; Lopez-Miranda, J.; Katsiki, N.; Delgado-Lista, J.; Alvarez-Sala-Walther, L.A. The Fluid Aspect of the Mediterranean Diet in the Prevention and Management of Cardiovascular Disease and Diabetes: The Role of Polyphenol Content in Moderate Consumption of Wine and Olive Oil. Nutrients 2019, 11, 2833. [CrossRef] 78. vila-Román, J.; Soliz-Rueda, J.R.; Bravo, F.I.; Aragonès, G.; Suárez, M.; Arola-Arnal, A.; Mulero, M.; Salvadó, M.-J.; Arola, L.; Torres-Fuentes, C. Phenolic compounds and biological rhythms: Who takes the lead? Trends Food Sci. Tech. 2021, 113, 77–85. [CrossRef] 79. Corona, G.; Tzounis, X.; Dessì, M.A.; Deiana, M.; Debnam, E.S.; Visioli, F.; Spencer, J.P.E. The fate of olive oil polyphenols in the gastrointestinal tract: Implications of gastric and colonic microflora-dependent biotransformation. Free Radic. Res. 2006, 40, 647–658. [CrossRef] 80. Caruso, D.; Visioli, F.; Patelli, R.; Galli, C.; Galli, G. Urinary excretion of olive oil phenols and their metabolites in humans. Metabolism 2001, 50, 1426–1428. [CrossRef] 81. Castillo-Luna, A.; Ledesma-Escobar, C.; Gómez-Díaz, R.; Priego-Capote, F. The secoiridoid profile of virgin olive oil conditions phenolic metabolism. Food Chem. 2022, 395, 133585. [CrossRef] 82. Ricelli, A.; Gionfra, F.; Percario, Z.; De Angelis, M.; Primitivo, L.; Bonfantini, V.; Antonioletti, R.; Bullitta, S.M.; Saso, L.; Incerpi, S.; et al. Antioxidant and Biological Activities of Hydroxytyrosol and Homovanillic Alcohol Obtained from Olive Mill Wastewaters of Extra-Virgin Olive Oil Production. J. Agric. Food Chem. 2020, 68, 15428–15439. [CrossRef] 83. Lee, H.C.; Jenner, A.M.; Low, C.S.; Lee, Y.K. Effect of tea phenolics and their aromatic fecal bacterial metabolites on intestinal microbiota. Res. Microbiol. 2006, 157, 876–884. [CrossRef] 84. Morgan, D.W.; Martin, P.E.; Krahenbuhl, G.S. Factors Affecting Running Economy. Sports Med. 1989, 7, 310–330. [CrossRef] 85. Saunders, P.U.; Pyne, D.B.; Telford, R.D.; Hawley, J.A. Factors Affecting Running Economy in Trained Distance Runners. Sports Med. 2004, 34, 465–485. [CrossRef] 86. Jacobs, R.A.; Lundby, C. Mitochondria express enhanced quality as well as quantity in association with aerobic fitness across recreationally active individuals up to elite athletes. J. Appl. Physiol. 2013, 114, 344–350. [CrossRef] 87. Menendez, J.A.; Joven, J.; Aragonès, G.; Barrajón-Catalán, E.; Beltrán-Debón, R.; Borrás-Linares, I.; Camps, J.; Corominas-Faja, B.; Cufí, S.; Fernández-Arroyo, S. Xenohormetic and anti-aging activity of secoiridoid polyphenols present in extra virgin olive oil: A new family of gerosuppressant agents. Cell Cycle 2013, 12, 555–578. [CrossRef] 88. Enríquez, J.A. Supramolecular Organization of Respiratory Complexes. Annu. Rev. Physiol. 2016, 78, 533–561. [CrossRef] 89. Greggio, C.; Jha, P.; Kulkarni, S.S.; Lagarrigue, S.; Broskey, N.T.; Boutant, M.; Wang, X.; Alonso, S.C.; Ofori, E.; Auwerx, J.; et al. Enhanced Respiratory Chain Supercomplex Formation in Response to Exercise in Human Skeletal Muscle. Cell Metab. 2016, 25, 301–311. [CrossRef] 90. Rasmussen, P.; Nielsen, J.; Overgaard, M.; Krogh-Madsen, R.; Gjedde, A.; Secher, N.H.; Petersen, N.C. Reduced muscle activation during exercise related to brain oxygenation and metabolism in humans. J. Physiol. 2010, 588, 1985–1995. [CrossRef] Nutrients 2023, 15, 421 20 of 20 91. Maughan, R.J.; Burke, L.M.; Dvorak, J.; Larson-Meyer, D.E.; Peeling, P.; Phillips, S.M.; Rawson, E.S.; Walsh, N.P.; Garthe, I.; Geyer, H.; et al. IOC Consensus Statement: Dietary Supplements and the High-Performance Athlete. Int. J. Sport Nutr. Exerc. Metab. 2018, 28, 104–125. [CrossRef] 92. Rickards, L.; Lynn, A.; Barker, M.E.; Russell, M.; Ranchordas, M.K. Comparison of the polyphenol content and in vitro antioxidant capacity of fruit-based nutritional supplements commonly consumed by athletic and recreationally active populations. J. Int. Soc. Sports Nutr. 2022, 19, 336–348. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

The Effect of a Hydroxytyrosol-Rich, Olive-Derived Phytocomplex on Aerobic Exercise and Acute Recovery.

01-13-2023

Roberts, Justin D,Lillis, Joseph B,Pinto, Jorge Marques,Chichger, Havovi,López-Samanes, Álvaro,Coso, Juan Del,Zacca, Rodrigo,Willmott, Ashley G B

eng

PMC10030110

Figure 3—source data 1. Footstep counts for each subject on all terrain. subject flat uneven I uneven II 1 363 321 448 2 228 78 473 3 473 131 436 4 373 471 519 5 366 557 109 6 224 160 327 7 218 398 442 8 503 489 479 9 477 390 392

How human runners regulate footsteps on uneven terrain.

02-22-2023

Dhawale, Nihav,Venkadesan, Madhusudhan

eng

PMC7871487

Vol. 37 No.2. May-August 2019 • ISSNe: 2216-0280 Original article Metabolic fatigue in resuscitators using personal protection equipment against biological hazard Francisco Martín-Rodríguez1,2 1 M.Sc, Ph.D. Professor, Centro de Simulación Clínica Avanzada. Facultad de Medicina. Universidad de Valladolid. 2 Unidad Móvil de Emergencias Valladolid I, Gerencia de Emergencias Sanitarias de Castilla y León (SACYL). España. Email: [email protected] Conflicts of interest: none. Received: October 8th, 2018. Approvedptado: June 4th, 2019. How to cite this article: Martín-Rodríguez F. Metabolic fatigue in resuscitators using personal protection equip- ment against biological hazard. Invest. Educ. Enferm. 2019; 37(2):e04. DOI: 10.17533/udea.iee.v37n2e04 Metabolic fatigue in resuscitators using personal protection equipment against biological hazard Abstract Objective. To describe the effects of wearing individual protection equipment against biological hazard when performing a simulated resuscitation. Methods. Uncontrolled quasi-experimental study involving 47 volunteers chosen by random sampling stratified by sex and professional category. We determined vital signs, anthropometric parameters and baseline lactate levels; subsequently, the volunteers put on level D individual protection equipment against biological hazard and performed a simulated resuscitation for 20 minutes. After undressing and 10 minutes of rest, blood was extracted again to determine lactate levels. Metabolic fatigue was defined as a level of lactic acid above 4 mmol/L at the end of the intervention. Results. 25.5% of the participants finished the simulation with an unfavorable metabolic tolerance pattern. The variables that predict metabolic Metabolic fatigue in resuscitators using personal protection equipment against biological hazard Invest Educ Enferm. 2019; 37(2): e04 fatigue were the level of physical activity and bone mass -in a protective form- and muscle mass. People with a low level of physical activity had ten times the probability of metabolic fatigue compared to those with higher levels of activity (44% versus 4.5%, respectively). Conclusion. Professionals who present a medium or high level of physical activity tolerate resuscitation tasks better with a level D individual biological protection suit in a simulated resuscitation. Descriptors: cardiopulmonary resuscitation; personal protective equipment; anaerobic threshold; containment of biohazards; stress, physiological. Fatiga metabólica en reanimadores usando equipos de protección personal frente a riesgos biológicos Resumen Objetivo. Describir cómo afecta llevar puesto un equipo de protección individual frente a riesgos biológicos durante la realización de una reanimación simulada. Métodos. Estudio cuasi-experimental no controlado en el que participaron 47 voluntarios elegidos mediante un muestreo aleatorio estratificado por sexo y categoría profesional. Se realizó una toma de contantes vitales y parámetros antropométricos, así como una determinación basal de lactato; posteriormente, los voluntarios se pusieron un equipo de protección individual nivel D frente a riesgos biológicos y realizaron una reanimación simulada durante 20 minutos; después del desvestido y de 10 minutos de reposo se realizó otra extracción de sangre para conocer los niveles de lactato. Se definió fatiga metabólica si el nivel de ácido láctico al final de la intervención estaba por encima de 4 mmol/L. Resultados El 25.5% de los participantes terminó la simulación con un mal patrón de tolerancia metabólica. Las variables que predicen la fatiga metabólica son el nivel de actividad física y la masa ósea –en forma protectora- y la masa muscular. Las personas con un nivel bajo de actividad física tuvieron diez veces la probabilidad de fatiga metabólica comparadas con las de niveles más altos de actividad (44% versus 4.5%, respectivamente). Conclusión. Los profesionales que presentan un nivel de actividad física media o alta toleran mejor las labores de reanimación con un traje de protección biológica individual nivel D, en el caso de reanimación simulada. Francisco Martín-Rodríguez Invest Educ Enferm. 2019; 37(2): e04 Descriptores: reanimación cardiopulmonar; equipo de protección personal; umbral anaerobio; contención de riesgos biológicos; estrés fisiológico. Fatiga metabólica em reanimadores usando equipamentos de proteção pessoal frente a riscos biológicos Resumo Objetivo. Descrever como afeta vestir um equipamento de proteção individual frente a riscos biológicos durante a realização de uma reanimação simulada. Métodos. Estudo quase-experimental não controlado no qual participaram 47 voluntários elegidos mediante uma amostragem aleatória estratificado por sexo e categoria profissional. Se realizou uma toma de concreta e de parâmetros antropométricos, assim como uma determinação basal de lactato; posteriormente, os voluntários vestiram um equipamento de proteção individual nível D frente a riscos biológicos e realizaram uma reanimação simulada durante 20 minutos; depois do desvestido e de 10 minutos de repouso se realizou outra extração de sangue para conhecer os níveis de lactato. Se definiu fatiga metabólica se o nível de ácido láctico ao final da intervenção estava por encima de 4 mmol/L. Resultados 25.5% dos participantes terminou a simulação com um mal padrão de tolerância metabólica. As variáveis que predizem a fatiga metabólica são o nível de atividade física e a massa óssea –em forma protetora- e a massa muscular. As pessoas com um nível baixo de atividade física tiveram dez vezes a probabilidade de fatiga metabólica comparadas com as de níveis mais altos de atividade (44% versus 4.5%, respectivamente). Conclusão. Os profissionais que apresentam um nível de atividade física média ou alta toleram melhor os trabalhos de reanimação com um equipamento de proteção biológica individual nível D, no caso de reanimação simulada. Descritores: reanimação cardiopulmonar; equipamento de proteção individual; limiar anaeróbio; contenção de riscos biológicos; estresse fisiológico. Invest Educ Enferm. 2019; 37(2): e04 Metabolic fatigue in resuscitators using personal protection equipment against biological hazard Introduction I n “conventional” resuscitation, health personnel must perform cardiopulmonary resuscitation techniques according to protocol.(1) However, the consequences of a cardiorespiratory arrest occurring in highly complex situations, such as an incident with biological hazard, are not known. The risk to health personnel necessarily implies the use of personal protective equipment (PPE). The recent Ebola virus epidemic in West Africa(2) has confronted health systems around the world with an alarming reality of biological hazard situations. It is increasingly common to attend numerous incidents - either provoked or unanticipated - that generate situations of collective emergency in which certain substances with biological hazard are implicated. These are situations that require a highly specialized response; in short: situations that must be handled comprehensively by the Emergency Services. The usual work of emergency teams is per se difficult, changing and often unfolding before a complex background. Professionals placed in such scenarios with diverse requirements must have received special attitudes and skills from education and training and in providing materials and resources. The risk of a situation with biological hazard occurring is percentually low compared with other types of disasters,(3) but, due to its multiple and varied repercussions, the system must be specially prepared and trained. The various PPEs must represent the backbone of protective systems, the prevention of contagion and, by definition, the control of the situation. This type of incidents, despite currently being isolated cases, occur with certain frequency,(4) so we must be prepared to intervene in these scenarios. Most are caused either by accidental situations, or because appropriate safety measures for the handling or transport of certain substances have not been taken.(4) To these accidents we must add the possibility of terrorist acts; a situation that, unfortunately, is happening more frequently, as has been demonstrated in several attempts frustrated by the police worldwide in recent years.(5) The use of PPE has improved both the assistance to victims and the survival of those involved in chemical or biological incidents, but this type of protection could otherwise reduce a person’s operational capacity. When selecting protective equipment for biological and chemical preparation, a balance must be struck between the degree of protection necessary for the potential hazard in question and the resulting difficulty in carrying out user functions.(6) Performing their work in situations of biological hazard with the necessary level of protection directly affects the physiology of health workers, since it generates significant metabolic fatigue. This metabolic fatigue can increase the risk of accidents with PPE, increase cross-contamination, and lead to hasty termination of the procedures due to physiological stress, among others. Emergency services should contemplate these situations when planning interventions.(7) Therefore, the question arises: Are all workers in the emergency services able to tolerate physiologically performing a resuscitation with PPE in the face of biological hazard? To answer this question, we selected a level of Invest Educ Enferm. 2019; 37(2): e04 Francisco Martín-Rodríguez lactic acid above 4 mmol/L at the end of the intervention as the parameter for the appearance of metabolic fatigue.(8) The objective of this study was to describe the effects of wearing individual protection equipment against biohazard when performing a simulated resuscitation. Methods Type of study and sample. An uncontrolled quasi- experimental study was performed in 2016 including 47 volunteers chosen through random sampling stratified by sex and professional category (doctors and nurses of the Hospital Emergency Services and Prehospital Emergencies) of an opportunity sample of 104 volunteers. We included professionals from the hospital emergency services of the University Clinic and Río Hortega University Hospital in Valladolid and professionals of the prehospital emergencies system of Castilla y León (mobile emergency units of: Palencia, Salamanca and Valladolid, both urban and rural) in Spain. We included voluntary participants between 22 and 65 years. Any volunteer who presented at least one of the following exclusion criteria was rejected: severe motor, visual or hearing impairment, acute phase skin disease, body mass index greater than 40 kg/ m2, systolic blood pressure below 80 mmHg and baseline heart rate above 150 bpm. Environmental conditions and personal protective equipment used. All participants performed the same simulated clinical case, in the same diaphanous laboratory room of 20 m2, with an average controlled temperature of 33.6±4.3ºC and an average controlled humidity of 51.1 ± 1.5%. All PPE elements used in the performance of the practice case conformed with European Community standards and had an instruction manual. Elements are listed in the order of placement: boot covers, protective overall, inner nitrile gloves, hood, FFP3 mask, panoramic and self-ventilated protective goggles and outer nitrile gloves. Photo 1. Enactment of the emergency services with biological protection suits (Photo by Francisco Martín-Rodríguez) Invest Educ Enferm. 2019; 37(2): e04 Metabolic fatigue in resuscitators using personal protection equipment against biological hazard Variables studied and measurement equipment. An anthropometric study was conducted to assess the following parameters: height, weight, body fat, muscle mass, bone mass, body mass index and total water content. For measuring the volunteers, we used a SECA® model 206 mechanical metric tape and for measuring weight and bio-impedance, we used a Tanita® precision scale model BC-601. The volunteers were asked to sit in a chair, roll up their sleeves and wait 5 minutes calmly, to have their vital signs taken: heart rate, systolic and diastolic blood pressure, respiratory rate, tympanic temperature, total hemoglobin, perfusion index, oxygen saturation and basal glycaemia. For determining systolic blood pressure, diastolic blood pressure and heart rate, we used a SCHILLER brand BP-200 plus meter. The temperature was measured with a tympanic brand BRAUN model ThermoScan PRO 6000 thermometer with ExacTemp technology. The values of total hemoglobin, oxygen saturation and perfusion index were obtained with a MASIMO model Pronto 7 multiparameter monitor, with software version b99e80000004ef796 (2.2.15), and revision version of sensor a83f90f0000c53f2, and glucose levels in blood with an Accu-Chek Mobile meter from Roche®. For determining lactic acid levels, we used an Accutrend® Plus meter from Roche® with a measuring range of 0.8- 21.7 mmol/L, with three measurements: baseline determination, just after the volunteer perform the cardiac massage and once the case was concluded after 10 minutes of rest. In addition, each volunteer completed the IPAQ physical activity questionnaire.(9) The test has seven items with high reliability (α = 0.80), suitable for people aged 15 and above. The full version of the questionnaire can be found on the website: www. ipaq.ki.se. The unit of measurement is called METs (unit of measurement of the metabolic index), and corresponds to the sum of the following activities: walking, moderate physical activity and vigorous physical activity. Once the physical activity questionnaire is completed, the volunteers are classified into three levels based on the exercise performed in the last seven days, as follows:(10) high level: vigorous physical activity at least 3 days per week achieving a total of at least 1500 METs, or 7 days of any combination of walking, with moderate physical activity and/or vigorous physical activity, achieving a total of at least 3000 METs; moderate level: 3 or more days of vigorous physical activity for at least 20 minutes per day, or 5 or more days of moderate physical activity and/or walking at least 30 minutes per day, or 5 or more days of any combination of walking, moderate or vigorous physical activity achieving a total of at least 600 METs; low or inactive level: not meeting any of the above criteria. Development of the clinical simulation scenario. All participants in the study had the same information and the same materials and medical devices to solve the same case. The volunteers, guided by a biohazard specialist, had ten minutes to equip themselves completely, following by checking their suits. Once equipped with the PPE, they entered a room with controlled temperature and humidity and had to attend to a convulsing patient with possible biological hazard. After 10 minutes of simulation, the patient suffered a cardiac arrest and the volunteers had to perform a regulated resuscitation during 20 minutes. The total duration of the case inside the laboratory was 30 minutes. Once the practical case was completed, the PPE was taken off under supervision, and 10 minutes after the removal of the PPE, vital signs were taken again. Statistical analysis. The qualitative variables are summarized with their frequency distribution, and the quantitative variables in their mean and standard deviation (SD). In all cases, the distribution of the variable was checked against the theoretical models; and, in the case of asymmetry, we calculated the median and its interquartile range (IQR). The association between qualitative variables was evaluated with the c2 test or Fisher’s exact test if more than 25% of the expected were less than 5. The behavior of the quantitative variables was analyzed for each of the independent variables categorized by the Student t test. We calculated mean absolute effects and their 95% confidence Invest Educ Enferm. 2019; 37(2): e04 Francisco Martín-Rodríguez intervals (95% CI). A logistic regression model was adjusted, in order to evaluate the association of those variables that predicted poor tolerance. This model allowed to identify the relationship between a set of explanatory variables and the probability of control of the variables studied. The calibration capacity of the model was evaluated with the Hosmer and Lemeshow test (p near 1 denoting high calibration). In all hypothesis contrasts, the null hypothesis was rejected with a type I error or alpha error of less than 0.05. The software package used for the analysis was SPSS version 20.0. Ethical aspects. The study was approved on April 6, 2016 by the Clinical Research Ethics Committee of the Río Hortega University Hospital of Valladolid (Spain) with registration code #412016. All volunteers had to read and sign the informed consent document. Results Of 47 participants, 22 were men (46.8%) and 25 women (53.1%), with an average age of 40.2±8.7 years. By profession, 25 were nurses (53.1%) and 22 medical doctors (46.8%); 26 worked in hospital emergency services (55.3%) and 21 in prehospital emergency services (44.6%). On the IPAQ physical activity questionnaire, 25 participants presented a low level of physical activity (53.2%), 14 scored a moderate level of physical activity (29.8%) and 8 presented a level of high physical activity (17%). Table 1 shows the mean values and standard deviation of the parameters at baseline and according to the final lactic acid values. Table 2 shows that one in four participants concluded the simulation with an unfavourable metabolic tolerance pattern. No statistically significant differences were found in terms of poor metabolic tolerance due to sex or profession variables. In contrast, statistically significant differences could be observed in the variables of life support training level in environments with biological hazard, where the proportion of subjects with fatigue was greater in the category with basic training (37.5%). By physical activity category performed in the last 7 days, participants with a low level had ten times the probability of metabolic fatigue compared to those with higher levels of activity (44% versus 4.5%, respectively). We adjusted a multivariate logistic regression model in which the variables of professional group, workplace (hospital emergencies or prehospital emergencies), age, physical activity level, body mass index, muscle mass and bone mass were included. The prediction capacity of the model was very good, with an AUC of 0.901 (95% CI 0.81-0.99) and p<0.001. The variables that predicted metabolic fatigue were the level of physical activity, muscle mass and bone mass. With decreasing physical activity and increasing muscle mass, tolerance worsened, whereas higher bone mass correlated with better tolerance (Table 3). Invest Educ Enferm. 2019; 37(2): e04 Metabolic fatigue in resuscitators using personal protection equipment against biological hazard Table 1. Distribution of vital signs and anthropometric parameters at baseline and according to final lactic acid values Variables Baseline parameters Final lactate <4 mmol/L ≥4 mmol/L p-value Mean SD* Mean SD Mean SD Age (years) 40.2 8.7 39.3 9.3 42.9 5.8 0.210 Height (cm) 168.9 8.4 168.0 8.4 171.4 8.2 0.225 Weight (kg) 73.5 16.4 70.3 14.0 82.4 19.8 0.026 Body fat (%) 24.0 8.0 22.6 7.2 28.2 9.0 0.034 Muscle mass (%) 52.7 11.5 51.6 10.8 55.9 13.4 0.270 Bone mass (kg) 2.8 0.6 2.7 0.5 2.9 0.7 0.320 Body mass index (kg/m2) 25.5 4.2 24.7 3.6 27.9 5.2 0.024 Total water (%) 55.7 5.5 56.8 4.9 52.5 6.3 0.018 Pulse (bpm) 79.4 12.6 78.8 13.0 81.3 11.4 0.553 Systolic arterial pressure (mmHg) 129.7 13.4 127.4 13.5 136.7 10.5 0.036 Diastolic arterial pressure (mmHg) 83.5 9.4 82.5 9.7 86.3 8.0 0.225 Respiratory rate (rpm) 16.2 1.7 16.1 1.8 16.6 1.5 0.365 Temperature (ºC) 36.5 0.5 36.5 0.6 36.4 0.5 0.945 Saturation (%) 98 1.5 97.9 1.6 98.4 1.2 0.280 Hemoglobin (mg/dl) 13.7 1.4 13.7 1.4 13.7 1.4 0.992 Perfusion (%) 3.6 2.9 3.5 2.7 3.8 3.6 0.755 Glycemia (mg/dl) 114.3 21.8 112.5 18.9 119.8 29.0 0.424 Baseline lactate (mmol/L) 2.3 1.4 2.2 1.2 2.5 2.1 0.587 Lactate during CPR (mmol/L) 9.4 5.2 8.2 5.1 12.7 4.1 0.006 Final lactate (mmol/L) 3.2 1.8 2.3 0.9 5.6 1.7 <0.001 Variation between final and baseline lactate (mmol/L) 1.0 0.40 0.17 1.35 3.03 2.11 <0.001 * Standard deviation Invest Educ Enferm. 2019; 37(2): e04 Francisco Martín-Rodríguez Table 2. Metabolic fatigue according to study variable categories Variable n (%) p-value Total (n=47) 12 (25.5) - Sex Male (n=22) 6 (27.3) 0.797 Female (n=25) 6 (24.0) Profession Nurse (n=25) 4 (16.0) 0.110 Doctor (n=22) 8 (36.4) Workplace Hospital emergency dept. (n=26) 8 (30.8) 0.360 Emergency services (n=21) 4 (19.0) Training level in life support in biological hazard conditions Without training (n=6) 1 (16.6) 0.001 Basic training (n=8) 3 (37.5) Advanced training (n=33) 8 (24.2) Level of physical activity Low (n=25) 11 (44.0) 0.008 Moderate (n=14) 1 (7.1) High (n=8) 0 (0.0) Level of physical activity low (n=25) 11 (44.0) 0.002 Moderate to high (n=22) 1 (4.5) Table 3. Variables for the logistic regression model to predict metabolic fatigue Variables Odds ratio 95% CI OR p-value Minimum Maximum Physical activity (intense or moderate compared to low) 0.02 0.00 0.45 0.013 Muscle mass (units) 6.59 1.29 33.75 0.024 Bone mass (units) 0.00 0.00 0.02 0.027 Invest Educ Enferm. 2019; 37(2): e04 Metabolic fatigue in resuscitators using personal protection equipment against biological hazard Discussion With the generated predictive model, we know a priori with excellent reliability which professionals are going to conclude a cardiopulmonary resuscitation with more than 4 mmol/L of lactic acid in blood, an analytical value that characterizes the presence of metabolic fatigue, and value that insinuates the appearance of accidental errors of the workers and decrease in the quality or intensity of the maneuvers necessary for such a critical situation. The results of this study are especially relevant, since they allow establishing the profile of people who would inadequately tolerate the performance of a job with PPE against biological hazard. Knowing the possible behavior of workers at the physiological level, a more efficient selection can be made, and avoiding as much as possible situations of unnecessary risk in the interventions. The anthropometric parameters behaved in the expected way in the face of physiological stress that requires increased physical activity to generate more bioavailable energy, with increases in muscle mass and water and decrease in body fat.(11) During any moderately intense or highly intense physical exercise (such as a resuscitation with a PPE against biological hazard), the blood pressure is increased to compensate the higher demand for energy. At the end of the exercise, a generalized vasodilatation results, and, as a consequence, a redistribution of blood, lowering blood pressure. After 5-6 minutes of concluding the exercise, the blood pressure decreases to the previous levels at rest, and the blood pressure decreases more to a level below baseline, maintaining this decrease for the following 5-6 hours.(12) The physiological model explains these variations, as a response to intense exercise and the release of catecholamines, leading to peripheral vasoconstriction and to blood redistribution.(13) We found no significant differences by sex, group or study subgroup among subjects with metabolic fatigue.(14) Regarding the variation of lactic acid, during exercise of high intensity and short duration, the organism does not have enough oxygen immediately available, and must get energy through less efficient routes that generate more metabolic waste (glycolytic metabolism).(15) Consequently, high levels of lactic acid form that decrease muscle capacity and the ability to generate energy, causing early fatigue.(16) In this study, the lactate threshold direct correlated with the physical form of each subject, and revealed substantial differences between people with a high level of physical activity and people with a sedentary lifestyle.(17) In healthy people, we can observe an increase in the levels of lactic acid during exercise of high intensity and short duration (more so in less trained persons). Lactic acid is generated as a metabolic byproduct, becoming recycled as it originates, to a point where the body is unable to recycle lactic acid and it accumulates above 4 mmol/L,(15) exceeding the anaerobic threshold. Consequently, in high levels of lactic acid, the ability to generate energy decreases and muscle capacity decreases, appearing early fatigue.(18) In trained people, this threshold may be higher, and even more important, the capacity to recycle lactic acid is higher, so large quantities cannot accumulate.(19) Many authors have evaluated the realization of techniques with protection equipment. Szarpak et al.(20) studied advanced airway management by paramedical personnel wearing protective suits; the same authors similarly compared the use of intraosseous puncture equipment with suit and without suit,(21) and the performance of conventional vascular access techniques with and without suit.(22) Another study by Szarpak et al.(23) evaluated the correct performance of external cardiac massage techniques on a mannequin by professionals in protective suits, evaluating the correct position, depth or quality, among other aspects, but none of these studies evaluated how this physical exertion affected the resuscitators. The study by Stein et al.,(24) which analyzes the reaction time of workers carrying PPE and their physiological response, should be highlighted. The authors describe and compare changes in heart rate, venous pH, pCO2, bicarbonate, lactate level, oxygen saturation and temperature. They analyze the variations of these parameters in 19 healthy subjects, in two cases of 20 minutes Invest Educ Enferm. 2019; 37(2): e04 Francisco Martín-Rodríguez of exercise without protective equipment, and then the variation during 20 minutes of exercise wearing protective suits. The heart rate and temperature of volunteers in protective equipment were substantially more elevated than in control condition; however, due to the size of the sample, the results were not statistically significant. If we combine the physiological overload that is caused by the use of specific protection equipment, together with the effort involved in resuscitation tasks, working with biological hazard protection equipment generates discomfort and decreases in the level of attention and response capacity. (24) These circumstances increase the probability of suffering occupational accidents and the risk of exacerbating pre-existing diseases, decreasing effective work time or generating work situations where it is impossible to perform the assigned tasks safely, for the patient, the healthcare worker themselves or the rest of the staff.(25) Our research is limited to the study of the physiological and anthropometric parameters cited in the methodology, but the usefulness of other parameters such as cortisol, pH or insulin levels, among others, is not discussed. They were discarded from the study due to the complexity involved in measuring them; this limitation has to be taken into account in the study. Broader prospective studies are necessary in order to generalize the results and expand the parameters studied. Hospital and pre-hospital emergency services must contemplate within their curricular design of competences the handling of incidents with biological hazard, be it as acts of chance or stemming from intentional terrorist acts.(26) Generally, we can affirm that the use of PPE against biological hazard is especially hard and arduous for workers, imposing a burden of additional physiological stress for the intervention. (27) It is easy to demonstrate that work with protective equipment against biological hazard complicates technical procedures; however, so far, no extensive studies have shown that the use of PPE requires highly intense physical effort that prohibits working for large time intervals.(28) We can conclude that the parameters studied reveal a metabolic pattern of poor physiological tolerance after the use of individual protection equipment level D in the observed sample. Consequently, future studies could derive a predictive rule that allows us to assess which professionals may tolerate and adapt better to work in a biological incident. Invest Educ Enferm. 2019; 37(2): e04 Metabolic fatigue in resuscitators using personal protection equipment against biological hazard References 1. Barsuk JH, Cohen ER, Wayne DB, Siddall VJ, McGaghie WC. Developing a Simulation-Based Mastery Learning Curriculum: Lessons from 11 Years of Advanced Cardiac Life Support. Simul. Healthc. 2016; 11(1):52–9. 2. Kwon JH, Burnham CAD, Reske KA, Liang SY, Hink T, Wallace MA, et al. Assessment of Healthcare Worker Protocol Deviations and Self-Contamination During Personal Protective Equipment Donning and Doffing. Infect. Control Hosp. Epidemiol. 2017; 38(9):1077-83. 3. Schoch-Spana M, Cicero A, Adalja A, Gronvall G, Kirk Sell T, Meyer D, et al. Global Catastrophic Biological Risks: Toward a Working Definition. Health Secur. 2017; 15(4):323-8. 4. Millett P, Snyder-Beattie A. Existential Risk and Cost-Effective Biosecurity. Health Secur. 2017; 15(4):373-83. 5. Fogel I, David O, Balik CH, Eisenkraft A, Poles L, Shental O, et al. The association between self-perceived proficiency of personal protective equipment and objective performance: An observational study during a bioterrorism simulation drill. Am. J. Infect. Control. 2017; 45(11): 1238-42. 6. Calfee MW, Tufts J, Meyer K, McConkey K, Mickelsen L, Rose L, et al. Evaluation of standardized sample collection, packaging, and decontamination procedures to assess cross-contamination potential during Bacillus anthracis incident response operations. J. Occup. Environ. Hyg. 2016; 13(12): p. 980-92. 7. Narayanan N, Lacy CR, Cruz JE, Nahass M, Karp J, Barone JA, et al. Disaster Preparedness: Biological Threats and Treatment Options. Pharmacotherapy. 2018; 38(2):217-34. 8. Hunt L, Gupta-Wright A, Simms V, al. e. Clinical presentation, biochemical, and haematological parameters and their association with outcome in patients with Ebola virus disease: an observational cohort study. Lancet Infect. Dis. 2015; 15(11):1292–9. 9. Nicaise V. The Sensitivity And Specificity Of The IPAQ For Detecting Intervention Related Changes In Physical Activity. Med. Sci. Sports Exerc. 2011; 43(Sup. 1):607. 10. van Poppel MNM, Chinapaw MJM, Mokkink LB, van Mechelen W, Terwee CB. Physical activity questionnaires for adults: A systematic review of measurement properties. Sports Med. 2010; 40(7):565-600. 11. Baur DA, Bach CW, Hyder WJ, Ormsbee MJ. Fluid retention, muscle damage, and altered body composition at the Ultraman triathlon. Eur. J. Appl. Physiol. 2016; 116(3):447-58. 12. Spartano LN, Lyass GA, Larson DM, Lewis SG, Vasan SR. Abstract 19256: Predicting Exercise Systolic Blood Pressure and Heart Rate at 20 Years of Follow-up: Correlates in the Framingham Heart Study. Circulation. 2015; 132(3):A19256-A19256. 13. Jayasinghe S, Lambert G, Torres S, Fraser S, Eikelis N, Turner A. Hypothalamo-pituitary adrenal axis and sympatho- adrenal medullary system responses to psychological stress were not attenuated in women with elevated physical fitness levels. Endocrine. 2016; 51(2):369-79. Invest Educ Enferm. 2019; 37(2): e04 Francisco Martín-Rodríguez 14. Pattani R, Marquez C, Dinyarian C, Sharma M, Bain J, Moore JE, et al. The perceived organizational impact of the gender gap across a Canadian department of medicine and proposed strategies to combat it: a qualitative study. BMC Medicine. 2018; 16(1): p. 48. 15. ¿Morales-Alamo D, Losa-Reyna J, Torres-Peralta R, Martin-Rincon M, Perez-Valera M, Curtelin D, ¿et al. What limits performance during whole-body incremental exercise to exhaustion in humans? J. Physiol. 2015; 593(20):4631– 48. 16. Hall MM, Rajasekaran S, Thomsen TW, Peterson AR. Lactate: Friend or Foe. PM R. 2016; 8(3):S8-S15. 17. Ekkekakis P, Hall EE, Petruzzello SJ. Practical markers of the transition from aerobic to anaerobic metabolism during exercise: rationale and a case for affect-based exercise prescription. Prev. Med. 2014; 38(2):149-59. 18. Vikmoen O, Raastad T, Seynnes O, Bergstrøm K, Ellefsen S, Rønnestad BR. Effects of Heavy Strength Training on Running Performance and Determinants of Running Performance in Female Endurance Athletes. PLoS One. 2016; 11(3):e0150. 19. Devlin J, Paton B, Poole L, Sun W, Ferguson C, Wilson J, et al. d lactate clearance after maximal exercise depends on active recovery intensity. J. Sports Med. Phys. Fitness. 2014; 54(3):271-8. 20. Szarpak L, Madziała M, Smereka J. Comparison of endotracheal intubation performed with 3 devices by paramedics wearing chemical, biological, radiological, and nuclear personal protective equipment. Am. J. Emerg Med. 2016; 34(9):1902-3. 21. Szarpak L, Ramirez JG, Buljan D, Drozd A, Madziała M, Czyzewski L. Comparison of Bone Injection Gun and Jamshidi intraosseous access devices by paramedics with and without chemical-biological-radiological-nuclear personal protective equipment: a randomized, crossover, manikin trial. Am. J. Emerg. Med. 2016; 34(7):1307-8. 22. Szarpak L, Truszewski Z, Smereka J, Madziała M, Czyzewski L. Comparison of two intravascular access techniques when using CBRN-PPE: A randomized crossover manikin trial. Am. J. Emerg. Med. 2016; 34(6):1170-2. 23. Szarpak L, Truszewski Z, Gałązkowski R, Czyzewski L. Comparison of two chest compression techniques when using CBRN-PPE: a randomized crossover manikin trial. Am. J. Emerg. Med. 2016; 34(5): 913-5. 24. Stein C, Makkink A, Vincent-Lambert C. The effect of physical exertion in chemical and biological personal protective equipment on physiological function and reaction time. Prehosp Emerg Care. 2010; 14(1):36-44. 25. Ji T, Qian X, Yuan M, Jiang J. Experimental study of thermal comfort on stab resistant body armor. Springerplus. 2016; 5(1):1168. 26. Carter H, Amlôt R. Mass Casualty Decontamination Guidance and Psychosocial Aspects of CBRN Incident Management: A Review and Synthesis. PLoS Curr. 2016; September 27; 8. 27. Verbeek JH. Personal protective equipment for preventing highly infectious diseases due to exposure to contaminated body fluids in healthcare staff. Cochrane Database Syst. Rev. 2016; 4:CD011621. 28. Coca A. Physiological Evaluation of Personal Protective Ensembles Recommended for Use in West Africa. Disaster Med. Public Health Prep. 2017; 11(5): 580-6. Invest Educ Enferm. 2019; 37(2): e04 Metabolic fatigue in resuscitators using personal protection equipment against biological hazard

Metabolic fatigue in resuscitators using personal protection equipment against biological hazard.

[]

Martín-Rodríguez, Francisco

eng

PMC9690603

Citation: Neves, L.N.S.; Gasparini Neto, V.H.; Araujo, I.Z.; Barbieri, R.A.; Leite, R.D.; Carletti, L. Is There Agreement and Precision between Heart Rate Variability, Ventilatory, and Lactate Thresholds in Healthy Adults? Int. J. Environ. Res. Public Health 2022, 19, 14676. https:// doi.org/10.3390/ijerph192214676 Academic Editor: Paulina Hebisz Received: 9 August 2022 Accepted: 3 September 2022 Published: 9 November 2022 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affil- iations. Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). International Journal of Environmental Research and Public Health Article Is There Agreement and Precision between Heart Rate Variability, Ventilatory, and Lactate Thresholds in Healthy Adults? Letícia Nascimento Santos Neves 1,* , Victor Hugo Gasparini Neto 1 , Igor Ziviani Araujo 1, Ricardo Augusto Barbieri 2, Richard Diego Leite 1 and Luciana Carletti 1 1 Laboratory of Exercise Physiology (LAFEX), Physical Education and Sports Center, Federal University of Espírito Santo (CEFD-UFES), Vitória 29075-910, Brazil 2 Postgraduate Program in Physical Education and Sport, School of Physical Education and Sport of Ribeirão Preto, University of São Paulo (EEFERP-USP), São Paulo 05360-160, Brazil * Correspondence: [email protected] Abstract: This study aims to analyze the agreement and precision between heart rate variability thresholds (HRVT1/2) with ventilatory and lactate thresholds 1 and 2 (VT1/2 and LT1/2) on a treadmill. Thirty-four male students were recruited. Day 1 consisted of conducting a health sur- vey, anthropometrics, and Cardiopulmonary Exercise Test (CPx). On Day 2, after 48 h, a second incremental test was performed, the Cardiopulmonary Stepwise Exercise Test consisting of 3 min stages (CPxS), to determine VT1/2, LT1/2, and HRVT1/2. One-way repeated-measures ANOVA and effect size (ηp2) were used, followed by Sidak’s post hoc. The Coefficient of Variation (CV) and Typical Error (TE) were applied to verify the precision. Bland Altman and the Intraclass Correlation Coefficient (ICC) were applied to confirm the agreement. HRVT1 showed different values compared to LT1 (lactate, RER, and R-R interval) and VT1 ( ˙VE, RER, ˙VCO2, and HR). No differences were found in threshold 2 (T2) between LT2, VT2, and HRVT2. No difference was found in speed and ˙VO2 for T1 and T2. The precision was low to T1 (CV > 12% and TE > 10%) and good to T2 (CV < 12% and TE < 10%). The agreement was good to fair in threshold 1 (VT1, LT1, HRVT1) and excellent to good in T2 (VT1, LT1, HRVT1). HRVT1 is not a valid method (low precision) when using this protocol to estimate LT1 and VT1. However, HRVT2 is a valid and noninvasive method that can estimate LT2 and VT2, showing good agreement and precision in healthy adults. Keywords: anaerobic threshold; lactate threshold; ventilatory threshold; exercise test; athletic performance; prescription 1. Introduction Thresholds in physiology are essential concepts that establish the boundaries between different domains of exercise intensity (moderate/heavy/severe) [1,2]. In addition, thresh- olds can predict athletic performance, monitor training progress, and clinically assess a patient’s physiological function or dysfunction (e.g., in heart failure) [1]. Furthermore, some methods for threshold determination are costly, require qualified professionals, and need specific software. Thus, using more economical and less complex tools to determine thresholds is interesting. However, in this article, threshold 1 (T1) (e.g., first lactate, ven- tilatory or gas exchange, and heart rate variability (HRV) thresholds) lies between the moderate and heavy domains [2–4] and threshold 2 (T2) (e.g., second lactate threshold, second ventilatory threshold, or respiratory compensation point) lies between the heavy and severe domains [1–5]. Thresholds can be identified invasively by blood analysis (e.g., lactate, glucose, catecholamines) and noninvasively (e.g., gas exchange, electromyogra- phy, heart rate deflection point of heart rate, and infrared spectroscopy) [6–8]. Although different thresholds can be estimated, some still generate confusion about the moment of Int. J. Environ. Res. Public Health 2022, 19, 14676. https://doi.org/10.3390/ijerph192214676 https://www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2022, 19, 14676 2 of 14 demarcation by indicating different points on the curve, so they still need to be studied and analyzed [1]. A potentially cheaper and less invasive method than LT and VT to identify thresholds can be performed through autonomic balance, determined by heart rate variability (HRV), which is defined by the oscillation in the interval between consecutive heartbeats as well as the oscillations between consecutive instantaneous heart rates [9]. The literature suggests using HRV to estimate physiological thresholds, such as the ventilatory or lactate threshold, in an incremental test. During the test, there was a pro- gressive increase in exercise intensity over time with a concomitant reduction in HRV and parasympathetic activity, followed by vagal withdrawal [10,11]. The vagal withdrawal marks the first heart rate variability threshold (HRVT1), which occurs at approximately 50–60% of ˙VO2max [12] and may coincide with the first lactate and ventilatory thresholds (LT1 and VT1) [10]. In contrast, the second lactate threshold (LT2) and second ventila- tory threshold (VT2) are also investigated to determine whether HRV can estimate them. However, little evidence has shown comparisons between VT2 and LT2 with the heart rate variability threshold (HRVT2), and even rarer are treadmill studies [13,14]. Ventilatory thresholds 1 and 2 [3,15] and lactate thresholds 1 and 2 [2,10] are validated and widely used methods for threshold identification. However, although they agree, their occurrence may not occur at identical work rates, i.e., with insufficient precision to be determined simultaneously [1,7]. Therefore, comparing both to a potentially cheaper and recent proposal (HRV) is essential. As far as we know, no literature has shown the comparison between these six thresh- olds simultaneously (VT1/VT2, LT1/LT2, and HRVT1/HRVT2) on a treadmill. However, some authors compared three thresholds but only HRV1 with LV1 and LT1 on a cycle ergometer [10]. Other authors who used the first and second thresholds compared HRV only with the ventilatory thresholds (HRVT 1 vs. VT1 and HRVT2 vs. VT2) [13,14,16]. Therefore, it is interesting to present and compare the six thresholds in the same individual, allowing a more significant contribution to studies that demonstrate the thresholds in isolation, such as the comparison between only the first thresholds or only the second ones, and even only the comparison of HRV with the ventilatory threshold or with lactate threshold alone. Furthermore, when LT, VT, and HRVT are involved, the studies used a cycle ergometer, which makes it difficult to transpose HRV behavior into a practical application such as running. Since ˙VO2, ventilation, fiber recruitment patterns, and muscle coordination according to the specificity of each test or exercise can influence threshold identification [17], studies are needed to verify whether both thresholds (VT1/VT2 and LT1/LT2) can be estimated using HRVT1 and HRVT2 on a treadmill. Therefore, we aimed to test if there is agreement and precision between HRVT1/2 with LT1/2 and VT1/2 on a treadmill. 2. Materials and Methods 2.1. Participants The study consisted of 34 healthy male university students who were physically independent (≥150 min·week−1 of physical exercise), 22 ± 2 years, height: 176.2 ± 6.6 cm, body mass: 72.89 ± 8.84 kg; body fat (%): 8.77 ± 3.69; BMI: 23.5 ± 2.66; and R-R intervals (ms): 935.7 ± 151.0). Informed consent was obtained from all subjects involved in the study. The study followed the ethical guidelines outlined in the Declaration of Helsinki and was approved by an ethical committee (CAAE: 76607717.5.0000.5542–18/09/2017). For this sample, the statistical power found was 80% (effect size f: 0.25-equivalent to the moderate ηp2 found in the present study for ( ˙VO2); alpha: 5%), calculated by G*Power 3.1. University students aged 18 to 30 years were included, and none of the participants had cardiovascular, metabolic, respiratory, or osteoarticular diseases, and their maximal oxygen consumption ( ˙VO2max) was considered below 34 mL·kg−1·min−1 according to the American Heart Association [18]. Alcohol was forbidden for 48 h, and caffeine-containing products for 24 h prior to the beginning of the study. Int. J. Environ. Res. Public Health 2022, 19, 14676 3 of 14 2.2. Study Design Data collection occurred over two days with a 48-h interval in the morning. On the first day, questionnaires to identify health status were applied, followed by anthropometry, an electrocardiogram, and a cardiopulmonary exercise test (CPx) applied by a cardiologist to familiarize and determine the initial intensity of the second day. A Cardiopulmonary Stepwise Exercise Test (CPxS) was performed on the second day to compare the six thresh- olds (LT1/2, VT1/2, HRVT1/2). Two evaluators analyzed the criteria of VTs, LTs, and HRVTs blindly and independently. When necessary, a third evaluator was requested. The intraclass correlation coefficient between evaluators was used with values varying between LT1: 0.889, LT2: 0.948; VT1: 0.965, VT2: 0.954; HRVT1: 1.00, and HRVT2: 1.00. HRV, ˙VO2, and Lactate were collected during the CPxS every 3 min and after the test (10 min recovery). All participants were instructed to have a light meal at least 2 h before the test and avoid caffeinated and strenuous exercise 24 h before the tests. 2.3. Anthropometry Body mass and height were measured using a digital anthropometric scale (Marte Scientific, L200, SP), and the Body Mass Index (BMI) was calculated. In addition, seven skinfolds were measured (subscapular, triceps, mid-axillary, pectoral, abdominal, suprailiac, mid-thigh) (Mitutoyo Cescorf, Porto Alegre, Brazil) to calculate the body fat percentage (%BF). To calculate %BF, it is necessary to calculate body density (BD), and Jackson and Pollock’s seven skinfold equation was used: BD = 1.11200000 − 0.00043499 (X1) + 0.00000055 (X1)2 – 0.00028826 (X4), where X1 = sum of seven skinfolds and X4 = age in Years [19]. The %BF for the anthropometric technique was estimated using the SIRI (1961) equation %BF = (495/BD) – 450 [20]. 2.4. Cardiopulmonary Exercise Test (CPx) Before the CPx, the individuals remained in the supine position on a stretcher for 5 min, where a 12-lead resting electrocardiogram was performed (USB Micromed electro- cardiograph, Brasília, Brazil). In addition, electrocardiographic records were performed during pre-exertion, standing on the treadmill, and during exertion. Then, the CPx was performed on a treadmill (Inbra Sports Super ATL, Porto Alegre, Brazil), with a slope of 1% in the ramp protocol, with a duration of 8 to 12 min. The initial speed was 5 km·h−1 with gradual increments of 1 km·h−1 each minute, supervised by a cardiologist (room temperature: 23 ◦C–25 ◦C). At least four criteria were considered to define the CPx as the maximum test: (a) Voluntary exhaustion; (b) ≥ 90% of the age-adjusted estimate of HRmax; (c) z respiratory exchange ratio (RER) equal to or greater than 1.05; (d) ˙VO2max plateau or peak with increased exercise intensity; (e) peak blood [lactate] of ≥ 8 mM. During the CPx, a Metabolic Gas Analyzer (Cortex Metalyzer 3B, Leipzig, Germany) was used, with breath-by-breath measurement. Before each test, a gas mixture was used to calibrate (11.97% O2 and 4.95% CO2) the ambient gas. Then, the volume was calibrated with a 3 L syringe. Ventilation ( ˙VE), oxygen consumption ( ˙VO2), carbon dioxide production ( ˙VCO2), respiratory exchange ratio (RER), and speed were analyzed for an average of 15 s. 2.5. Cardiopulmonary Stepwise Exercise Test (CPxS) The CPxS was used to identify the six thresholds (VT1/2, LT1/2, and HRVT1/2) for comparing and validating the HRVTs. CPxS was applied 48 h after CPx, following the same precautions and procedures described in CPx. The CPx determined the initial intensity of this protocol on the first day. The CPxS incremental test starts at 4 km·h−1 below the speed in VT1 determined by CPx with an increment of 1 km·h−1 every 3 min (considering the first two stages as a warm-up, with a slope of 1% until maximum effort), measuring lactate, gas analysis, and HRV. Before the test, the participants remained at rest for 10 min in the supine position to measure the HRV. Int. J. Environ. Res. Public Health 2022, 19, 14676 4 of 14 2.6. Blood Lactate Concentrations Samples of 25 µL of capillary blood from the ear lobe were collected, at rest, during the CPxS test (every 3 min, participants jump off the treadmill and come back quickly in approximately 15 to 20 s, with no time wasted at each stage), and in passive recovery after 1, 3, 5, 7, and 10 min. The lactate peak in recovery was considered the highest value measured in this interval. The samples were stored under refrigeration (−4 ◦C) for further analysis on the equipment YSI 2300 STAT plus (Ohio, USA) [21]. 2.7. Heart Rate Variability (HRV) A Polar H7 Bluetooth heart rate monitor (Polar Electro Oy, Kempele, Finland) was connected to a smartphone to collect HRV, beat by beat (R-R intervals). The Polar H7 was validated to measure HRV data [22]. Each step described in CPxS was recorded in the Elite HRV app (HRV Elite, Asheville-North Carolina) [23] and edited in the Kubios HRV 3.0 Program. Data were collected at rest and during and after CPxS. 2.8. Determination of the Ventilatory Threshold (VT) VT1 was identified by the increase in the ventilatory equivalent of O2 ( ˙VE/ ˙VO2) without increasing the ventilatory equivalent of CO2 ( ˙VE/ ˙VCO2) (Figure 1). When necessary, V- slope and CO2 excess were observed to help identification (MetasoftTM software) [3,15]. VT2 was identified as the moment of the lowest point of ˙VE/ ˙VCO2 with subsequent elevation, in addition to the moment of the gradual decrease in PetCO2 [3]. The same two evaluators always determined the VT1 and VT2, and a third opinion was requested in case of disagreement. The participants were excluded when it was not possible to identify the VT (a success rate of 85.3% for T1 and 100% for T2). All intervals between steps, such as when participants jumped off the treadmill to take blood samples, were excluded from the analysis to avoid misinterpretation. The average for the last 30 s of each stage was analyzed. Int. J. Environ. Res. Public Health 2022, 19, x FOR PEER REVIEW 5 of 16 Figure 1. Represents an example of threshold identifications 1 and 2 on the CPxS (single participant). A represents the lactate threshold 1 and 2 (black line) and heart rate variability thresholds 1 and 2 (gray line). B represents the ventilatory threshold 1 and 2 (dark gray and black lines) and heart rate variability thresholds 1 and 2 (light gray line). Lactate thresholds 1 and 2 (LT1 and LT2), ventilatory thresholds 1 and 2 (VT1 and VT2), and heart rate variability thresholds 1 and 2 (HRVT1 and HRVT2). 2.11. Statistical Analysis All data were tabulated and double-verified by independent researchers. The analy- sis was performed using SPSS 20.0 software, and the figures were created by Excel soft- ware and GraphPad Prism 6. The normality was tested using the Shapiro–Wilk test and submitted to evaluate the histogram, kurtosis, and skewness. The results were presented as mean ± standard deviation (SD). Student’s t-test was used to compare the maximum values of CPx with CPxS. A repeated-measures ANOVA was used to compare the meth- ods of identifying VT with LT and HRVT, followed by Sidak’s post-hoc (T1, n = 29 and T2, n = 34). The Greenhouse-Geisser correction was considered when a lack of sphericity was Figure 1. Represents an example of threshold identifications 1 and 2 on the CPxS (single participant). (A) represents the lactate threshold 1 and 2 (black line) and heart rate variability thresholds 1 and 2 (gray line). (B) represents the ventilatory threshold 1 and 2 (dark gray and black lines) and heart rate variability thresholds 1 and 2 (light gray line). Lactate thresholds 1 and 2 (LT1 and LT2), ventilatory thresholds 1 and 2 (VT1 and VT2), and heart rate variability thresholds 1 and 2 (HRVT1 and HRVT2). Int. J. Environ. Res. Public Health 2022, 19, 14676 5 of 14 2.9. Determination of the Lactate Threshold (LT) The LT was identified by the visual method. LT1 was determined by the first increase in the blood lactate concentration above resting values [2,10] and LT2 by the second linearity breakpoint and exponential lactate accumulation [2,5,24] (Figure 1). 2.10. Determination of the Heart Rate Variability Threshold (HRVT) The R-R intervals were grouped into three-minute sequences for HRV analysis. Data filtering was performed using the Kubios HRV Standard 3.0 Program (Gronau, Germany), and it was automatically filtered to remove missing or premature intervals and artifacts not exceeding 5% [10]. The first 90 s of physical effort at each stage were excluded from the analysis due to the HR and HRV kinetic adjustment. This study chose the Root Mean Square of the Successive Differences (RMSSD) and Poincaré plot indexes (SD1 and SD2) as the HRV indexes (time-domain and non-linear HRV parameters). HRVT1 was identified by the first linearity breakpoint determined by visual inspection for the RMSSD variables, and when necessary, SD1 was used to confirm the identification [25,26]. HRVT2 was determined by the linearity breakpoint after the lowest value with a subsequent increase in the RMSSD and confirmed by the linearity break of the SD1/SD2 variables only when necessary (determined by visual inspection) [26–28]. They were identified in Excel software (Microsoft Excel® 2022) (Figure 1). Only two evaluators were necessary for HRVT identification because there was no disagreement. The same evaluators were used to determine VTs, LTs, and HRVTs, and they did not have experience with HRVTs. All thresholds (LT1/2, VT1/2, HRVT1/2) were represented in Figure 1 (individual example) in CPxS. 2.11. Statistical Analysis All data were tabulated and double-verified by independent researchers. The analysis was performed using SPSS 20.0 software, and the figures were created by Excel software and GraphPad Prism 6. The normality was tested using the Shapiro–Wilk test and submitted to evaluate the histogram, kurtosis, and skewness. The results were presented as mean ± standard deviation (SD). Student’s t-test was used to compare the maximum values of CPx with CPxS. A repeated-measures ANOVA was used to compare the methods of identifying VT with LT and HRVT, followed by Sidak’s post-hoc (T1, n = 29 and T2, n = 34). The Greenhouse-Geisser correction was considered when a lack of sphericity was noted (F statistics, degrees of freedom, degrees of freedom error). Partial eta-squared (ηp2) was utilized as a measure of effect size in ANOVA, using small (ηp2 = 0.01), medium (ηp2 = 0.06), and large (ηp2 = 0.14) effects [29]. The typical error (TE = SDdiff/ √ 2; where SDdiff is the standard deviation of difference) was expressed as an absolute value and as a percentage of the mean value, and the coefficient of variation (CV) was expressed in a percentage [CV = (SD/X)·100; where SD is the standard deviation of data and X is the mean], both used to test precision [30,31]. Bland Altman plots were used to identify agreement between the different methods (one-way ANOVA was used to compare the bias between the methods). The intraclass correlation coefficient (ICC) was used to scale the agreement (reliability). The reference values for the ICC were < 0.5 (poor), 0.5–0.75 (moderate), 0.75–0.90 (good), and ≥0.90 (excellent) [32]. Statistical significance was set at p < 0.05 for all analyses. 3. Results The maximal values at CPxS are demonstrated in Table 1. The average initial velocity of participants in the CPxS was 5.3 ± 1.2 km·h−1. HRVT1 showed statistically higher values when compared to LT1 for the variables lactate and RER (p < 0.05) and lower for the R-R interval in LT1 (Table 2). HRVT, ˙VE, RER, ˙VCO2, and HR were significantly higher than VT1 (p > 0.05). The R-R intervals were lower than VT1 and LT1 (p < 0.05) (Table 2). For the other variables, no statistical differences were found between LT1 and VT1 (p > 0.05) (Table 2). The percentage demonstrated in Tables 2 and 3 (speed, ˙VO2, RR, and HR) is relative to the maximum value reached during the CPxS. Int. J. Environ. Res. Public Health 2022, 19, 14676 6 of 14 Table 1. Maximal physiological characteristics of participants at CPxS. Variables Mean ± SD HRmax (bpm) 194 ± 8 HRmax predicted (%) 97.75 ± 4.07 ˙VEmax (L·min−1) 125.42 ± 15.91 ˙VO2max (ml·kg−1·min−1) 47.86 ± 4.96 ˙VO2max (L·min−1) 3.47 ± 0.42 RERmax 1.02 ± 0.08 [La]peak (mM) 10.47 ± 2.01 Mean ± SD; Stepwise Progressive Test (CPxS), Ventilation ( ˙VE), Respiratory Exchange Ratio (RER), Lactate ([La]). Table 2. Comparison between the means of the methods for identifying thresholds 1 in CPxS. LT1 VT1 HRVT1 Within-Participants Effects Mean ± SD Mean ± SD Mean ± SD F (df, df) p ηp2 Speed (km·h−1) 7.2 ± 1.5 6.8 ± 1.4 7.3 ± 1.0 Speed (%) 50.63 ± 8.68 49.26 ± 9.08 52.52 ± 5.01 1.93 (2, 56) 0.155 0.064 Lactate (mM) 1.57 ± 0.89 1.58 ± 0.85 2.01 ± 0.82 * 4.83 (2, 56) 0.012 § 0.147 ˙VE (L) 44.44 ± 15.51 39.80 ± 12.57 47.96 ± 9.68 † 3.95 (2, 56) 0.025 § 0.124 ˙VO2 (ml·kg−1·min−1) 25.02 ± 6.48 23.75 ± 6.58 27.01 ± 5.22 ˙VO2 (%) 51.62 ± 13.43 50.28 ± 13.71 55.68 ± 8.43 2.94 (2, 56) 0.061 0.095 RER 0.84 ± 0.08 0.82 ± 0.07 0.88 ± 0.07 *† 18.06 (2, 56) 0.000 § 0.392 ˙VCO2 (L·min−1) 1.56 ± 0.53 1.43 ± 0.48 1.71 ± 0.34 † 4.15 (2, 56) 0.021 § 0.129 ˙VO2 (L·min−1) 1.84 ± 0.55 1.72 ± 0.48 1.95 ± 0.32 2.36 (2, 56) 0.103 0.078 RR (ms) 473.8 ± 85.9 498.0 ± 93.0 439.2 ± 52.9 *† RR (%) 52.04 ± 10.37 53.30 ± 9.33 47.63 ± 7.16 *† 6.56 (2, 56) 0.003 § 0.190 HR (bpm) 131 ± 23 125 ± 25 139 ± 16 † 5.12 (2, 56) 0.009 § 0.155 HR (%) 66.20 ± 7.59 64.53 ± 12.29 71.12 ± 11.14 † Lactate threshold 1 (LT1), ventilatory threshold 1 (VT1), and heart rate variability threshold 1 (HRVT1). Ventilation ( ˙VE), relative oxygen consumption ( ˙VO2), respiratory exchange ratio (RER), carbon dioxide consumption ( ˙VCO2), R-R interval (RR), heart rate (HR), variable of heart rate variability (RMSSD), F statistics and degrees of freedom (F(df, df)), partial eta-squared (ηp2). * (p < 0.05) different from LT1; † (p < 0.05) different from VT1; § (p < 0.05) ANOVA effect. Percentage values relative to the maximum (for RR it is relative to baseline). Table 3. Comparison between the means of the methods for identifying thresholds 2 in CPxS. LT2 VT2 HRVT2 Within-Participants Effects Mean ± SD Mean ± SD Mean ± SD F (df, df) p ηp2 Speed (km·h−1) 10.7 ± 1.1 10.6 ± 1.4 10.9 ± 1.2 Speed (%) 76.59 ± 5.42 76.07 ± 7.22 78.49 ± 6.23 1.45 (2, 66) 0.242 0.042 Lactate (mM) 3.96 ± 1.44 3.95 ± 1.19 4.30 ± 1.36 1.77 (2, 66) 0.178 0.051 ˙VE (L) 79.25 ± 13.57 75.83 ± 12.86 80.18 ± 11.08 1.92 (2, 66) 0.155 0.055 ˙VO2 (ml·kg−1·min−1) 39.39 ± 3.60 39.19 ± 5.16 39.98 ± 3.86 ˙VO2 (%) 82.63 ± 6.32 81.97 ± 7.55 83.85 ± 6.79 0.96 (2, 66) 0.389 0.028 RER 0.92 ± 0.07 0.91 ± 0.06 0.92 ± 0.06 0.79 (2, 66) 0.460 0.023 ˙VCO2 (L·min−1) 2.62 ± 0.39 2.59 ± 0.39 2.66 ± 0.31 0.86 (2, 66) 0.429 0.025 ˙VO2 (L·min−1) 2.86 ± 0.38 2.84 ± 0.39 2.90 ± 0.31 0.89 (2, 66) 0.417 0.026 RR (ms) 347.2 ± 27.1 348.0 ± 27.0 342.8 ± 21.6 RR (%) 37.85 ± 5.49 37.91 ± 5.28 37.46 ± 5.81 1.26 (2, 66) 0.291 0.037 HR (bpm) 174 ± 13 173 ± 13 176 ± 11 1.18 (2, 66) 0.313 0.035 HR (%) 89.68 ± 4.14 89.49 ± 4.83 90.72 ± 3.71 Lactate threshold 2 (LT2), ventilatory threshold 2 (VT2), and heart rate variability threshold 2 (HRVT2). Ventilation ( ˙VE), relative oxygen consumption ( ˙VO2), respiratory exchange ratio (RER), carbon dioxide consumption ( ˙VCO2), R-R interval (RR), heart rate (HR), variable of heart rate variability (RMSSD), F statistics and degrees of freedom (F(df, df), partial eta-squared (ηp2). There were no differences between methods (p > 0.05). Percentage values relative to the maximum (for RR it is relative to baseline). No statistical differences were found in the methods’ comparison: HRVT2 vs. LT2, HRVT2 vs. VT2, and LT2 vs. VT2 (p > 0.05) (Table 3). Furthermore, repeated-measures ANOVA revealed a significant effect mainly for the variables Lactate (F2, 56 = 4.83, p = 0.012, ηp2 = 0.147, large), RER (F2, 56 = 18.06, p = 0.000, ηp2 = 0.392, large), RR (F2, 56 = 6.56, p = 0.003, ηp2 = 0.190, large), and HR (F2, 56 = 5,12, Int. J. Environ. Res. Public Health 2022, 19, 14676 7 of 14 p = 0.009, ηp2 = 0.155, large) between methods in T1 (Table 2), but no significant effects were found in T2 (Table 3). At T1 (LT1, VT1, and HRVT1), speed, ˙VO2, and HR demonstrated TE values greater than 10% (Table 4). At T2 (LT2, VT2, and HRVT2), all variables showed TE lower than 10% (Table 4). In T1, velocity, HR, and ˙VO2 present CV > 12%, while T2 CV < 12% for the same variables, showing that T2 had less variation than T1. The ICC at T1 was moderate (p < 0.05), and for T2, the ICC was moderate to good (p < 0.05). Table 4. Typical Error, Coefficient of Variation, and Intraclass Correlation Coefficient between the methods for identifying thresholds 1 and 2. LT1 vs. VT1 LT1 vs. HRVT1 VT1 vs. HRVT1 LT2 vs. VT2 LT2 vs. HRVT2 VT2 vs. HRVT2 Typical Error absolute (%) Speed (km·h−1) 0.95 (13.8) 0.86 (11.9) 0.81 (12.3) 0.87 (8.1) 0.78 (7.4) 0.85 (8.0) ˙VO2 (ml·kg−1·min−1) 4.51 (18.5) 4.13 (16.4) 4.32 (18.5) 2.61 (6.6) 2.38 (6.0) 2.37 (6.1) HR (bpm) 15 (12.2) 16 (12.2) 16 (14.1) 7 (3.9) 6 (3.6) 7 (4.1) Coefficient of Variation Speed (km·h−1) 19.31 16.82 16.27 11.56 10.28 11.14 ˙VO2 (ml·kg−1·min−1) 26.13 22.44 23.97 9.40 8.50 8.47 HR (bpm) 17.10 16.29 17.56 5.52 5.03 5.66 Intraclass Correlation Coefficient Speed (km·h−1) 0.693 * 0.671 * 0.637 * 0.684 * 0.663 * 0.709 * ˙VO2 (ml·kg−1·min−1) 0.684 * 0.616 * 0.616 * 0.797 * 0.744 * 0.840 * HR (bpm) 0.748 * 0.559 * 0.504 * 0.842 * 0.840 * 0.783 * Lactate threshold 1 (LT1), ventilatory threshold 1 (VT1), heart rate variability threshold 1 (HRVT1), lactate threshold 2 (LT2), ventilatory threshold 2 (VT2), heart rate variability threshold 2 (HRVT2). Relative oxygen consumption ( ˙VO2) and heart rate (HR). * p < 0.05. There were no statistical differences between the means of differences for all methods (p > 0.05) (Figures 2–4). Int. J. Environ. Res. Public Health 2022, 19, x FOR PEER REVIEW 8 of 16 Figure 2. Bland Altman plots for speed at thresholds 1 (A–C) and 2 (D–F). The dashed lines at the ends represent the limits of agreement (1.96 SD) and the dashed central lines (bias). Lactate thresh- old 1 (LT1), ventilatory threshold 1 (VT1), heart rate variability threshold 1 (HRVT1), lactate thresh- old 2 (LT2), ventilatory threshold 2 (VT2), heart rate variability threshold 2 (HRVT2). Figure 2. Bland Altman plots for speed at thresholds 1 (A–C) and 2 (D–F). The dashed lines at the ends represent the limits of agreement (1.96 SD) and the dashed central lines (bias). Lactate threshold 1 (LT1), ventilatory threshold 1 (VT1), heart rate variability threshold 1 (HRVT1), lactate threshold 2 (LT2), ventilatory threshold 2 (VT2), heart rate variability threshold 2 (HRVT2). Int. J. Environ. Res. Public Health 2022, 19, 14676 8 of 14 Int. J. Environ. Res. Public Health 2022, 19, x FOR PEER REVIEW 9 of 16 Figure 3. Bland Altman plots for V̇ O2 at thresholds 1 (A–C) and 2 (D–F). The dashed lines at the ends represent the limits of agreement (1.96 SD) and the dashed central lines (bias). Lactate thresh- old 1 (LT1), ventilatory threshold 1 (VT1), heart rate variability threshold 1 (HRVT1), lactate thresh- old 2 (LT2), ventilatory threshold 2 (VT2), heart rate variability threshold 2 (HRVT2). Figure 3. Bland Altman plots for ˙VO2 at thresholds 1 (A–C) and 2 (D–F). The dashed lines at the ends represent the limits of agreement (1.96 SD) and the dashed central lines (bias). Lactate threshold 1 (LT1), ventilatory threshold 1 (VT1), heart rate variability threshold 1 (HRVT1), lactate threshold 2 (LT2), ventilatory threshold 2 (VT2), heart rate variability threshold 2 (HRVT2). Int. J. Environ. Res. Public Health 2022, 19, x FOR PEER REVIEW 10 of 16 Figure 4. Bland Altman plots for HR at thresholds 1 (A–C) and 2 (D–F). The dashed lines at the ends represent the limits of agreement (1.96 SD) and the dashed central lines (bias). Lactate threshold 1 (LT1), ventilatory threshold 1 (VT1), heart rate variability threshold 1 (HRVT1), lactate threshold 2 (LT2), ventilatory threshold 2 (VT2), heart rate variability threshold 2 (HRVT2). 4. Discussion As far as we know, this is the first study to compare these six thresholds on a tread- mill (LT1/LT2, VT1/VT2, and HRVT1/HRVT2). The main findings of our study suggest that HRVT2 has good agreement and precision to estimate LT2 and VT2 in healthy adults, increasing the ecological validity of these noninvasive methods and allowing a real-world HRVT1/2 determination. In contrast, HRVT1 did not show enough precision to estimate LT1 and VT1. 4.1. LTs vs. VTs L d l h h ld d d d l l Figure 4. Bland Altman plots for HR at thresholds 1 (A–C) and 2 (D–F). The dashed lines at the ends represent the limits of agreement (1.96 SD) and the dashed central lines (bias). Lactate threshold 1 (LT1), ventilatory threshold 1 (VT1), heart rate variability threshold 1 (HRVT1), lactate threshold 2 (LT2), ventilatory threshold 2 (VT2), heart rate variability threshold 2 (HRVT2). Int. J. Environ. Res. Public Health 2022, 19, 14676 9 of 14 4. Discussion As far as we know, this is the first study to compare these six thresholds on a treadmill (LT1/LT2, VT1/VT2, and HRVT1/HRVT2). The main findings of our study suggest that HRVT2 has good agreement and precision to estimate LT2 and VT2 in healthy adults, increasing the ecological validity of these noninvasive methods and allowing a real-world HRVT1/2 determination. In contrast, HRVT1 did not show enough precision to estimate LT1 and VT1. 4.1. LTs vs. VTs Lactate and ventilatory thresholds presented a moderate to good intraclass correla- tion and no statistical difference. However, it is essential to highlight that LT1 vs. VT1 demonstrated less precision (CV > 12% in all variables) and ICC (<0.750) than LT2 vs. VT2. Some studies also found different results when comparing these methods (LT1 vs. VT1 or LT2 vs. VT2) [8,10,14,16,25,33], which made our findings important to confirm the use of both LTs and VTs to estimate the exercise intensity. VT1 is challenging to identify due to several factors that can influence the agreement and precision identification, such as the cardiorespiratory fitness level [34], walking-running transition [35], control of food intake [1], and ergometer used [17]. Still, the present study showed an 85.3% success rate in identifying VT1, similar to research on a cycle ergometer [10], and a 100% success rate in second threshold identification. The CPx performed before CPxS likely affects the small loss percentage because it helps define the initial load of CPxS since the intensity of the protocol start can influence the identification of the VT1. 4.2. LT1 and VT1 vs. HRVT1 At HRVT1, Lactate, ˙VE, RER, CO2, HR, and RR were statistically different compared to LT1 and VT1. However, no difference was found in speed, which could contribute to practical application in training (Table 2). However, more differences were found between HRVT1 and VT1 than HRVT1 and LT1 when observing the mean values on average (Table 2). HRVT1 seems closer to LT1 than VT1, especially for the HR parameter, which differed between HRVT1 and VT1, and it is among the three primary parameters for training prescription (speed, ˙VO2, and HR). Still, we can infer that HRTV1 tends to overestimate the values of LT1 and VT1 for some parameters. Most studies that compared HRVT1 with LT1 or VT1 demonstrated good agreement in ˙VO2, HR, lactate, and speed [8,10,25,33]. However, one study showed difficulty in identifying VT1 using HRV analyses [16], which is close to our findings, reinforcing that T1 (HRVT1, VT1, LT1) is less precise to determine using different variables. On the other hand, some researchers who found good agreement at T1 only presented ˙VO2 and did not present the parameter of speed or HR (km·h−1) [8,10], not allowing a comparative analysis with parameters such as speed (km·h−1). Our findings demonstrate a good mean difference (Bland Altman) with moderate ICC in ˙VO2, velocity, and HR. However, we found low precision for all these variables (TE > 10% and CV > 12%) in T1 (LT1, VT1, HRVT1), which makes it difficult to show good acceptance of the HRVT1 to estimate LT1 and VT1. Ramos-Campos et al. (2017) did not identify reproducibility between HRVT1 and VT1 for speed [14]. These results differed from the present study because no difference was found in speed in the present study. However, the precision was low. Besides that, both studies showed difficulties in determining VT1 using HRV analysis. Different methods to identify heart rate variability thresholds, such as the fixed value of 1 ms between stages, the visual method, and mathematical analysis, can make it difficult to compare different thresholds [8]. Despite that, the difficulty in determining VT1 is also essential because the occurrence or not of thresholds can result in a sample loss [10] and may harm the prescription of training intensity [36,37]. In addition, our study was careful to use an incremental tread- mill test 48h before the Cardiopulmonary Stepwise Exercise Test to determine the initial intensity and minimize the loss of the first threshold. Different ergometers, treadmills, or Int. J. Environ. Res. Public Health 2022, 19, 14676 10 of 14 cycle ergometers can be another problematic factor because they demonstrated different results. For example, in maximum progressive protocols on the cycle ergometer, many studies have accurately shown a relationship between vagal withdrawal and thresholds 1 (T1), determined by HRV, concentrations of blood lactate, and gas exchange [10,11,25,33]. Nevertheless, motor differences in exercise specificity in the different protocols are essential factors for determining thresholds and must be considered [17]. Consequently, protocols performed on a treadmill can present different behavior than the cycle ergometer, preferably at T1 [17]. Furthermore, the literature demonstrated that T1 could occur in the moment of transition from walking to running between 6 km·h−1 and 8 km·h−1 [35], which does not happen in the cycle ergometer. These findings are similar to our averages, in which they presented means of 7.15, 7.34, and 6.78 km·h−1 for LT1, HRVT1, and VT1, respectively. Furthermore, the walk-to-run transition can modify some physiological variables, such as HR, ˙VE, and ˙VO2, causing interference in autonomic control, HRV responses, and gas exchange, creating confusion in threshold identification [35]. However, further research is needed to compare different ergometers using HRVT to determine T1 and confirm whether these transitions influenced the threshold identifications. Therefore, the use of HRVT1 to estimate LT1 and VT1 is not suggested under these conditions because, although they agree, their occurrence happened with insufficient precision. Another possibility to determine VT1 using HRV would be through the short-term scaling exponent alpha1 (DFA a1) [38]. However, it is still a recent proposal, in which the authors themselves suggest that the use of this index still needs to be better studied [38] since it appears to be influenced by spontaneous breathing during exercise [39], whereas this is less influenced when using RMSSD or the Poincaré plot [40]. 4.3. LT2 and VT2 vs. HRVT2 HRVT2 demonstrated excellent agreement for all parameters compared to LT2 and VT2. Our study used the time domain and the Poincaré plot only when necessary to identify HRTV2 [26], which is simple to analyze and interpret since a third evaluator was not required. Furthermore, HRV measurement was performed with lactate and gas analysis, showing agreement and precision in using HRVT2 to estimate LT2 and VT2, which is helpful for monitoring and prescribing the intensity zones in exercise. Thus, research that evaluated VT2 and LT2 with HRVT2 demonstrated good agreement between the methods even in different sports modalities [14,41,42]. Some authors used variables in the frequency domain to identify T2 [42]. However, using these variables requires a more complex and careful analysis because the frequency domain is easily influenced by breathing, which may have less influence when applied to respiratory sinus arrhythmia (RSA) but is used only at rest and needs to have a fixed breathing rate [40]. However, changes in the breathing rate do not markedly influence the RMSSD or Poincaré plot dimension [40], which does not affect our results. Therefore, it is essential to use simple methods to identify HRVT2, such as the time domain [9]. These tests to estimate LT1/LT2 and VT1/VT2 by HRVT1/HRVT2 provide new in- sights into the relationship between the first and second thresholds. Some studies empha- size only the excellent agreement of the first thresholds, leaving the second thresholds out due to their wide range of identifications, and often do not show the exact timing [1,2]. However, this is not what our study demonstrated. Instead, we demonstrated excellent agreement on T2 (LT2 and VT2) estimated by HRV. 4.4. Importance to Use Visual Methods to Identify HRVTs The methodological choices were constrained primarily by visual analysis to deter- mine thresholds. It is easier to analyze (agreement in 100% of identifications) and identifies thresholds in the daily routine of health professionals because the use of the visual method allows identification and does not need expensive software (just need the free Kubios software or a Microsoft Office or Libre Office to analyze). Researchers strongly recommend Int. J. Environ. Res. Public Health 2022, 19, 14676 11 of 14 determining T1 for health participants but are not highly trained, employing visual, mathe- matical, or a combination of both methods, as was performed in the present study [1,10]. Furthermore, some authors found that the visual method to determine HRV was vali- dated in the literature and had better reliability than a mathematical method (Dmax) [27]. Therefore, all the methods used in this study are valid and recommended for all these reasons. 4.5. Importance of T1 and T2 in Training Thresholds 1 and 2 are significant indices to help establish rigorous exercise inten- sity domains to individualize training and rehabilitation. They can help predict athletic performance, monitor training progress, and clinically assess the patient’s physiological function or dysfunction [1]. Exercise intensity is an important training element for effective cardiovascular and metabolic adaptations [43], mainly in T2. However, some researchers have also raised the importance of training volume even at low intensity (below or at T1), suggesting that duration is crucial to inducing training effects [44]. In this sense, researchers have demonstrated a polarized model of intensity distribution for training, in which participants who spend approximately 75% of the training below T1, 5–10% between T1 and T2, and 15–20% above T2 may have less autonomic stress, better motivation, and performance [36,37]. Consequently, the present study demonstrates the importance of determining the intensity in T1 and T2 in favor of better agreement and precision (accuracy) in practical and prescribing training. Future studies are needed to find better indices or methods to estimate T1. T1 is an important index when it comes to specific populations, for example, in patients with cardiovascular disease (ischemia or heart failure), who re- quire adequate workload control of low-intensity exercises (below T1) [45]. In addition, the position statement of the European Association for Cardiovascular Prevention and Rehabilitation suggests the use of threshold-based exercise intensity prescriptions in heart disease patients [46,47]. 4.6. Practical Significance These findings may be helpful to coaches, conditioning instructors, and laboratories’ daily routines to monitor and prescribe exercise intensity. The HRV presented a simple analysis and interpretation method, visually using the time domain and Poincaré plot. Furthermore, HRV measurement using the heart rate sensor via a smartphone reduces the cost and facilitates its analysis, allowing less complexity to assess and help identify different intensity domains, which is essential for monitoring exercise intensity and prescriptions. 4.7. Limitations Although the present study provides meaningful information, some limitations should be acknowledged. This study is cross-sectional, so future research should be considered to assess the training effect on HRVTs. In addition, studies are needed to measure HRV during exercise in different individuals with a wider range of BMI, ages, and fitness levels. The present study did not strictly control food intake or fluid intake in the pre-test, but all participants were instructed to eat 2 h before the test, and the same hydration was offered on the day of the test. Besides, all parameters were measured simultaneously and with the same conditions, impacting all thresholds equally. Furthermore, to obtain stable RR intervals, the 3-min stage protocol was used in this study; however, some researchers recommend the 1-min stage for better VT detection [10]. 5. Conclusions HRVT1 presents low precision using this protocol to estimate LT1 and VT1. However, HRVT2 is a valid and noninvasive method that can estimate LT2 and VT2, showing good agreement and precision in healthy adults. Therefore, studying HRVTs using this simple and visual method on a treadmill must be encouraged to show consistency and increase ecological validity. Int. J. Environ. Res. Public Health 2022, 19, 14676 12 of 14 Author Contributions: Conceptualization, L.N.S.N. and L.C.; methodology, L.N.S.N. and I.Z.A.; software, L.N.S.N.; validation, L.N.S.N., V.H.G.N., and L.C.; formal analysis, L.N.S.N., V.H.G.N., and R.A.B.; investigation, L.N.S.N.; resources, L.N.S.N.; data curation, L.N.S.N.; writing—original draft preparation, L.N.S.N.; writing—review and editing, L.N.S.N., V.H.G.N., I.Z.A., R.D.L., R.A.B., and L.C.; visualization, L.N.S.N.; supervision, L.C.; project administration, L.N.S.N. and L.C. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of Federal University of Espírito Santo (CAAE: 76607717.5.0000.5542–18/09/2017). Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: Not applicable. Acknowledgments: We thank David C. Poole for suggesting important revisions to the publication of this article, and all the people who helped to obtain this final version. We would like to thank the Coordenação de Aperfeiçoamento de Pessoal de Nível Superior-Brasil (CAPES), Finance Code 001–for their financial support. Conflicts of Interest: The authors declare no conflict of interest. References 1. Poole, D.C.; Rossiter, H.B.; Brooks, G.A.; Gladden, L.B. The Anaerobic Threshold: 50+ Years of Controversy. J. Physiol. 2020, 599, 737–767. [CrossRef] 2. Faude, O.; Kindermann, W.; Meyer, T. Lactate Threshold Concepts: How Valid Are They? Sports Med. 2009, 39, 469–490. [CrossRef] [PubMed] 3. Beaver, W.L.; Wasserman, K.; Whipp, B.J. A New Method for Detecting Anaerobic Threshold by Gas Exchange. Am. Physiol. Soc. 1986, 60, 2020–2027. [CrossRef] [PubMed] 4. Pettitt, R.W.; Clark, I.E.; Ebner, S.M.; Sedgeman, D.T.; Murray, S.R. Gas Exchange Threshold and VO2max Testing for Athletes: An Update. J. Strength Cond. Res. 2013, 27, 549–555. [CrossRef] [PubMed] 5. Kindermann, W.; Simon, G.; Keul, J. Physiology The Significance of the Aerobic-Anaerobic Transition. Eur. J. Appl. Physiol. 1979, 34, 25–34. [CrossRef] [PubMed] 6. Azevedo, P.H.S.M.; Garcia, A.; Duarte, J.M.P.; Rissato, G.M.; Carrara, V.K.P.; Marson, R.A. Limiar Anaeróbio e Bioenergética: Uma Abordagem Didática e Integrada. Rev. Da Educ. Fís./UEM 2009, 20, 453–464. [CrossRef] 7. Caen, K.; Vermeire, K.; Bourgois, J.G.; Boone, J. Exercise Thresholds on Trial: Are They Really Equivalent? Med. Sci. Sports Exerc. 2018, 50, 1277–1284. [CrossRef] 8. Shiraishi, Y.; Katsumata, Y.; Sadahiro, T.; Azuma, K.; Akita, K.; Isobe, S.; Yashima, F.; Miyamoto, K.; Nishiyama, T.; Tamura, Y.; et al. Real-Time Analysis of the Heart Rate Variability during Incremental Exercise for the Detection of the Ventilatory Threshold. J. Am. Heart Assoc. 2018, 7, e006612. [CrossRef] 9. Electrophysiology, T.F. of the E.S. of C. the N.A. Heart Rate Variability. Circulation 1996, 93, 1043–1065. [CrossRef] 10. Karapetian, G.K.; Engels, H.J.; Gretebeck, R.J. Use of Heart Rate Variability to Estimate LT and VT. Int. J. Sports Med. 2008, 29, 652–657. [CrossRef] 11. Tulppo, M.P.; Makikallio, T.H. Quantitative Beat-to-Beat Analysis of Heart Rate Dynamics during Exercise. Am. Physiol. Soc. 1996, 271, 244–252. [CrossRef] 12. Tulppo, M.P.; Mäkikallio, T.H.; Seppänen, T.; Laukkanen, R.T.; Huikuri, H.v; Coote, J.H.; Fisher, J.P.; Ogoh, S.; Ahmed, A.; Aro, M.R.; et al. Vagal Modulation of Heart Rate during Exercise: Effects of Age and Physical Fitness. Am. Physiol. Soc. 1998, 274, 424–429. [CrossRef] 13. Mourot, L.; Fabre, N.; Savoldelli, A.; Schena, F. Second Ventilatory Threshold from Heart-Rate Variability: Valid When the Upper Body Is Involved? Int. J. Sports Physiol. Perform. 2014, 9, 695–701. [CrossRef] 14. Ramos-Campo, D.J.; Rubio-Arias, J.A.; Ávila-Gandía, V.; Marín-Pagán, C.; Luque, A.; Alcaraz, P.E. Heart Rate Variability to Assess Ventilatory Thresholds in Professional Basketball Players. J. Sport Health Sci. 2017, 6, 468–473. [CrossRef] 15. Gaskill, S.E.; Ruby, B.C.; Walker, A.V.A.J.; Sanchez, O.A.; Serfass, R.C.; Leon, A.S. Validity and Reliability of Combining Three Methods to Determine Ventilatory Threshold. Med. Sci. Sports Exerc. 2001, 33, 1841–1848. [CrossRef] 16. Cassirame, J.; Tordi, N.; Fabre, N.; Duc, S.; Durand, F.; Mourot, L. Heart Rate Variability to Assess Ventilatory Threshold in Ski-Mountaineering. Eur. J. Sport Sci. 2015, 15, 615–622. [CrossRef] 17. Millet, G.P.; Vleck, V.E.; Bentley, J.D. Physiological Differences Between Cycling and Running. Sports Med. 2009, 39, 179–206. [CrossRef] Int. J. Environ. Res. Public Health 2022, 19, 14676 13 of 14 18. Herdy, A.H.; Caixeta, A. Classificação Nacional Da Aptidão Cardiorrespiratória Pelo Consumo Máximo de Oxigênio. Arq. Bras. Cardiol. 2016, 106, 389–395. [CrossRef] 19. Jackson, A.S.; Pollock, M.L. Practical Assessment of Body Composition. Phys. Sportsmed. 1985, 13, 76–90. [CrossRef] 20. Siri, W.E. Body Composition from Fluid Spaces and Density: Analysis of Methods. Nutrition 1961, 9, 480–491, discussion 480, 492. 21. Simões, H.G.; Campbell, C.S.G.; Denadai, B.S.; Kokubun, E. Determinação Do Limiar Anaeróbio Por Meio de Dosagens Glicêmicas e Lactacidêmicas Em Teste de Pista Para Corredores. Rev. Paul. Educ. Fis. São Paulo. 1998, 12, 17–30. 22. Plews, D.J.; Scott, B.; Altini, M.; Wood, M.; Kilding, A.E.; Laursen, P.B. Comparison of Heart-Rate-Variability Recording with Smartphone Photoplethysmography, Polar H7 Chest Strap, and Electrocardiography. Int. J. Sports Physiol. Perform. 2017, 12, 1324–1328. [CrossRef] [PubMed] 23. Perrotta, A.S.; Jeklin, A.T.; Hives, B.A.; Meanwell, L.E.; Warburton, D.E.R. Validity of the Elite HRV Smartphone Application for Examining Heart Rate Variability in a Field-Based Setting. J. Strength Cond. Res. 2017, 31, 2296–2302. [CrossRef] [PubMed] 24. Yoshida, T.; Chida, M.; Ichioka, M.; Suda, Y. Blood Lactate Parameters Related to Aerobic Capacity and Endurance Performance. Eur. J. Appl. Physiol. Occup. Physiol. 1987, 56, 7–11. [CrossRef] [PubMed] 25. Leprêtre, P.-M.; Bulvestre, M.; Ghannem, M.; Ahmaidi, S.; Weissland, T.; Lopes, P. Determination of Ventilatory Threshold Using Heart Rate Variability in Patients with Heart Failure. Surg. Curr. Res. 2013, 1, 1–6. [CrossRef] 26. Nascimento, E.M.F.; Kiss, M.A.P.D.M.; Santos, T.M.; Lambert, M.; Pires, F.O. Determination of Lactate Thresholds in Maximal Running Test by Heart Rate Variability Data Set. Asian J. Sports Med. 2017, 8, e58480. [CrossRef] 27. Candido, N.; Okuno, N.; da Silva, C.; Machado, F.; Nakamura, F. Reliability of the Heart Rate Variability Threshold Using Visual Inspection and Dmax Methods. Int. J. Sports Med. 2015, 36, 1076–1080. [CrossRef] 28. Mankowski, R.T.; Michael, S.; Rozenberg, R.; Stokla, S.; Stam, H.J.; Praet, S.F.E. Heart-Rate Variability Threshold as an Alternative for Spiro-Ergometry Testing: A Validation Study. J. Strength Cond. Res. 2016, 32, 474–479. [CrossRef] 29. Lakens, D. Calculating and Reporting Effect Sizes to Facilitate Cumulative Science: A Practical Primer for t-Tests and ANOVAs. Front. Psychol 2013, 4, e00863. [CrossRef] 30. Hopkins, W.G. Measures of Reliability in Sports Medicine and Science. Sports Med. 2000, 30, 1–15. [CrossRef] 31. Reed, G.F.; Lynn, F.; Meade, B.D. Use of Coefficient of Variation in Assessing Variability of Quantitative Assays. Clin. Vaccine Immunol. 2002, 9, 1235–1239. [CrossRef] 32. Koo, T.K.; Li, M.Y. A Guideline of Selecting and Reporting Intraclass Correlation Coefficients for Reliability Research. J. Chiropr. Med. 2016, 15, 155–163. [CrossRef] 33. Sales, M.M.; Campbell, C.; Morais, P.K.; Ernesto, C.; Soares-Caldeira, L.F.; Russo, P.; Motta, D.F.; Moreira, S.R.; Nakamura, F.Y.; Simões, H.G. Noninvasive Method to Estimate Anaerobic Threshold in Individuals with Type 2 Diabetes. Diabetol. Metab. Syndr. 2011, 3, 1. [CrossRef] 34. Neves, L.N.S.; Neto, V.H.G.; alves, S.P.; Leite, R.D.; Barbieri, R.A.; Carletti, L. Cardiorespiratory Fitness Level Influences the Ventilatory Threshold Identification. J. Phys. Educ. 2021, 32, e3279. [CrossRef] 35. Monteiro, W.D.; Araújo, C.G.S. De Transição Caminhada-Corrida: Considerações Fisiológicas e Perspectivas Para Estudos Futuros. Rev. Bras. De Med. Do Esporte 2001, 7, 207–222. [CrossRef] 36. Seiler, K.S.; Kjerland, G.Ø. Quantifying Training Intensity Distribution in Elite Endurance Athletes: Is There Evidence for an “Optimal” Distribution? Scand. J. Med. Sci. Sports 2006, 16, 49–56. [CrossRef] 37. Muñoz, I.; Seiler, S.; Bautista, J.; España, J.; Larumbe, E.; Esteve-Lanao, J. Does Polarized Training Improve Performance in Recreational Runners? Int. J. Sports Physiol. Perform. 2014, 9, 265–272. [CrossRef] 38. Rogers, B.; Giles, D.; Draper, N.; Hoos, O.; Gronwald, T. A New Detection Method Defining the Aerobic Threshold for Endurance Exercise and Training Prescription Based on Fractal Correlation Properties of Heart Rate Variability. Front. Physiol. 2021, 11, 1806. [CrossRef] 39. Gronwald, T.; Rogers, B.; Hoos, O. Fractal Correlation Properties of Heart Rate Variability: A New Biomarker for Intensity Distribution in Endurance Exercise and Training Prescription? Front. Physiol. 2020, 11, 550572. [CrossRef] 40. Penttilä, J.; Helminen, A.; Jartti, T.; Kuusela, T.; Huikuri, H.v.; Tulppo, M.P.; Coffeng, R.; Scheinin, H. Time Domain, Geometrical and Frequency Domain Analysis of Cardiac Vagal Outflow: Effects of Various Respiratory Patterns. Clin. Physiol. 2001, 21, 365–376. [CrossRef] 41. Cottin, F.; Médigue, C.; Lopes, P.; Leprêtre, P.M.; Heubert, R.; Billat, V. Ventilatory Thresholds Assessment from Heart Rate Variability during an Incremental Exhaustive Running Test. Int. J. Sports Med. 2007, 28, 287–294. [CrossRef] 42. di Michele, R.; Gatta, G.; di Leo, A.; Cortesi, M.; Andina, F.; Tam, E.; da Boit, M.; Merni, F. Estimation of the Anaerobic Threshold from Heart Rate Variability in an Incremental Swimming Test. J. Strength Cond. Res. 2012, 26, 3059–3066. [CrossRef] 43. Stergiopoulos, D.C.; Kounalakis, S.N.; Miliotis, P.G.; Geladas, N.D. Second Ventilatory Threshold Assessed by Heart Rate Variability in a Multiple Shuttle Run Test. Int. J. Sports Med. 2020, 42, 48–55. [CrossRef] 44. Hofmann, P.; Tschakert, G. Intensity- and Duration-Based Options to Regulate Endurance Training. Front. Physiol. 2017, 8, 337. [CrossRef] 45. Rogers, B.; Mourot, L.; Gronwald, T. Aerobic Threshold Identification in a Cardiac Disease Population Based on Correlation Properties of Heart Rate Variability. J. Clin. Med. 2021, 10, 4075. [CrossRef] Int. J. Environ. Res. Public Health 2022, 19, 14676 14 of 14 46. Mezzani, A.; Hamm, L.F.; Jones, A.M.; McBride, P.E.; Moholdt, T.; Stone, J.A.; Urhausen, A.; Williams, M.A. Aerobic Exercise Intensity Assessment and Prescription in Cardiac Rehabilitation: A Joint Position Statement of the European Association for Cardiovascular Prevention and Rehabilitation, the American Association of Cardiovascular and Pulmonary Rehabilitation and the Canadian Association of Cardiac Rehabilitation. Eur. J. Prev. Cardiol. 2013, 20, 442–467. [CrossRef] 47. Marcin, T.; Eser, P.; Prescott, E.; Prins, L.F.; Kolkman, E.; Bruins, W.; van der Velde, A.E.; Peña Gil, C.; Iliou, M.-C.; Ardissino, D.; et al. Training Intensity and Improvements in Exercise Capacity in Elderly Patients Undergoing European Cardiac Rehabilitation— The EU-CaRE Multicenter Cohort Study. PLoS ONE 2020, 15, e0242503. [CrossRef]

Is There Agreement and Precision between Heart Rate Variability, Ventilatory, and Lactate Thresholds in Healthy Adults?

11-09-2022

Neves, Letícia Nascimento Santos,Gasparini Neto, Victor Hugo,Araujo, Igor Ziviani,Barbieri, Ricardo Augusto,Leite, Richard Diego,Carletti, Luciana

eng

PMC8199297

International Journal of Environmental Research and Public Health Article Exercise Intensity and Technical Involvement in U9 Team Handball: Effect of Game Format Georgios Ermidis 1 , Rasmus C. Ellegard 2 , Vincenzo Rago 3 , Morten B. Randers 2,4 , Peter Krustrup 2 and Malte N. Larsen 2,*



 Citation: Ermidis, G.; Ellegard, R.C.; Rago, V.; Randers, M.B.; Krustrup, P.; Larsen, M.N. Exercise Intensity and Technical Involvement in U9 Team Handball: Effect of Game Format. Int. J. Environ. Res. Public Health 2021, 18, 5663. https://doi.org/10.3390/ ijerph18115663 Academic Editors: Filipe Manuel Clemente and Hugo Sarmento Received: 31 March 2021 Accepted: 21 May 2021 Published: 25 May 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affil- iations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). 1 Department of Movement and Wellness Sciences, University of Naples “Parthenope”, 80133 Naples, Italy; [email protected] 2 Department of Sports Science and Clinical Biomechanics, SDU Sport and Health Sciences Cluster (SHSC), University of Southern Denmark, 5230 Odense, Denmark; [email protected] (R.C.E.); [email protected] (M.B.R.); [email protected] (P.K.) 3 Faculty of Health and Sport Sciences, Universidade Europeia, 1500-210 Lisbon, Portugal; [email protected] 4 School of Sports Sciences, UiT The Arctic University of Norway, 9037 Tromsø, Norway * Correspondence: [email protected] Abstract: The purpose of this study was to quantify the exercise intensity and technical involvement of U9 boys’ and girls’ team handball during different game formats, and the differences between genders. Locomotor activity (total distance, distance in speed zones, accelerations, and decelerations), heart rate (HR), and technical involvement (shots, goals, and duels) metrics were collected during various 15 min game formats from a total of 57 Danish U9 players (37 boys and 20 girls). Game formats were a small size pitch (20 × 13 m) with 3 vs. 3 players and offensive goalkeepers (S3 + 1) and 4 vs. 4 players (S4), a medium size pitch (25.8 × 20 m) with 4 vs. 4 (M4) and 5 vs. 5 (M5) players, and a large size pitch (40 × 20 m) with 5 vs. 5 (L5) players. Boys and girls covered a higher total distance (TD) of high-speed running (HSR) and sprinting during L5 games compared to all other game formats (p < 0.05; ES = (−0.9 to −2.1), (−1.4 to −2.8), and (−0.9 to −1.3) respectively). Players covered the highest amount of sprinting distance in L5 games compared to all other game formats (p < 0.01; ES = 0.8 to 1.4). In all the game formats, players spent from 3.04 to 5.96 min in 180–200 bpm and 0.03 min to 0.85 min in >200 bpm of the total 15 min. In addition, both genders had more shots in S3 + 1 than M5 (p < 0.01; ES = 1.0 (0.4; 1.7)) and L5 (p < 0.01; ES = 1.1 (0.6; 2.2)). Team handball matches have high heart rates, total distances covered, and high-intensity running distances for U9 boys and girls irrespective of the game format. Locomotor demands appeared to be even higher when playing on larger pitches, whereas the smaller pitch size and fewer players led to elevated technical involvement. Keywords: physiology; youth; heart rate; time motion; notational analysis 1. Introduction Team handball is an intermittent high-intensity body contact team sport, characterized by sprinting, jumping, throwing, blocking, and pushing [1]. Various studies have described the locomotor demands of team handball in different age and sex groups; on average, inter- national male players cover 4370 ± 702 m [2], elite female players cover 4002 ± 551 m [3], and elite male adolescent players (15 years old) cover 1777 ± 264 m [4]. Overall, the physical and physiological demands, and therefore, the potential as a health-promoting activity of team handball have been predominantly investigated in adults and adolescent players [5]. Only one study investigated U13 boys and girls across different formats [6]. Youth team handball games in Denmark are played on different pitch dimensions and with a different number of players compared to adult games, similar to other team sports such as football [7]. For instance, in youth team handball games, the pitch and goal di- mensions are smaller and the number of players is reduced. Extensive research in other Int. J. Environ. Res. Public Health 2021, 18, 5663. https://doi.org/10.3390/ijerph18115663 https://www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2021, 18, 5663 2 of 14 team sports showed that manipulating the player numbers and the pitch size can alter the exercise intensity (i.e., locomotor activity, physiological responses) during a game in different sports [8]. Indeed, higher exercise intensity (e.g., heart rate (HR)) is reached when decreasing the number of players and increasing the pitch area [9]. On the other hand, reducing the number of players and pitch dimensions appears to induce higher technical involvement [7,10]. Moreover, the sex-specific timing of maturation [11,12] and the gender differences in morphological and neuromuscular characteristics are still early at this stage of age, and gender-related differences in explosive actions are therefore unlikely. Investigating differences in exercise intensity between gender may provide practitioners with a greater understanding of sex-specific training prescription. Overall, several external factors can influence the physiological and technical demands of training drills and thus, the desired conditioning stimulus [13]. Thus, information regarding the exercise intensity in children of both genders across different game formats could be of interest for practitioners involved with youth handball. Based on previous findings in a study of the game format of U13 handball [6], we hypothesized that a larger court will increase the total distance and that fewer players on the court will increase the involvement of the players in terms of more shots and duels per player. The purpose of this study was, therefore, to quantify the exercise intensity, the technical involvement, and a gender comparison of U9 boys’ and girls’ team handball during different game formats. The study provides useful knowledge that might change the game format used in tournaments for U9 players for relevant development and health promotion. 2. Materials and Methods 2.1. Design U9 players from ten Danish teams (local handball clubs around the region of Funen) participated in a 1-day tournament. Up to 5 games per player were used. Game formats were classified according to the pitch size and number of players that represent possible official games: • (S3 + 1): 3 vs. 3 + offensive goalkeepers on a small size pitch size of 20 × 13 m (37 m2 per player); • (S4): 4 vs. 4 on a small size (33 m2 per player); • (M4): 4 vs. 4 on a medium size pitch size of 25.8 × 20 m (65 m2 per player); • (M5): 5 vs. 5 on a medium size (52 m2 per player); • (L5): 5 vs. 5 on a large size pitch size of 40 × 20 m (80 m2 per player). In all of the above formats, goalkeepers participated, but they only were tracked in S3 + 1. To remove the effect of exercise volume and fatigue, the game duration was maintained at 15 min, and the games were played in a randomized order on the same day. All games were played on indoor team handball pitches. The sizes of the goals were 1.6 × 2.4 m on the small pitch, 1.78 × 3 m on the medium pitch, and 2 × 3 m on the large pitch. The games were played with all the official rules of the team handball game. The study was carried out according to the Helsinki protocol. 2.2. Participants Six teams of U9 boys (n = 37) and four teams of U9 girls (n = 20) participated in the study. All participants were 8–9-year-old recreational handball players. 2.3. Activity Profile The activity patterns were recorded using a wearable device incorporating a 200 Hz accelerometer and gyroscope (Polar Team Pro system, Polar, Kempele, Finland), which was placed on the lower sternum using an elastic band. The following variables were adopted: total distance (TD) covered, peak speed (Vpeak) attained, and number of sprints (>18 km/h). Exercise intensity was also distributed in the following running zones: stand- Int. J. Environ. Res. Public Health 2021, 18, 5663 3 of 14 ing/walking (St/W; 0.00–2.99 km/h), jogging (3.00–7.99 km/h), moderate-speed running (MSR, 8.00–12.99 km/h), high-speed running (HSR, 13.00–17.99 km/h), and sprinting (>18 km/h), according to previous studies describing the locomotor demands of team sports’ children [14,15]. In addition, the number of accelerations and decelerations were measured with the fol- lowing zones: Acc < 1.49 m·s−2, Acc 1.50 to 2.30 m·s−2, Acc > 2.30 m·s−2, Dec < −1.49 m·s−2, Dec −1.50 to −2.30 m·s−2, and Dec < −2.30 m·s−2 [14,15]. The total number of acceler- ations and decelerations was also quantified. The activity profiles and HR data were stored in the device and downloaded using the manufacturer’s software (POLAR, software version 1.3.1, POLAR, Polar Electro Oy, Kempele, Finland) [16]. 2.4. Heart Rate and Subjective Perceptions HRs were recorded in 1 s intervals during each game. The HR data were downloaded and expressed as the mean and max HR for the full match. In addition, the HR data were expressed as the time spent in HR zones as follows: <120, 120 to 160, 160 to 180, 180 to 200, and >200 bpm, as previously described [7]. Furthermore, after each game, a Visual Analogue Scale was used to assess the rating of perceived exertion (RPE) and enjoyment/fun (RPF), as previously done in similar studies since it is a well-accepted method to describe subjective phenomena [17,18]. Immediately after the 15 min matches, every player had a paper and pencil to record their scores. All players underwent a brief familiarization session in which three researchers explained the procedure, underlining the importance of scoring their perception of exertion (not fatigue or tiredness). For physical exertion, the players placed a mark on a 17.4 cm line ranging from ‘maximally demanding’ to ‘not demanding at all’, while for perceived fun, a similar line was used, ranging from ‘maximal fun’ to ‘not fun at all’. The result was obtained by measuring with a ruler the length (in centimeters) from 0 to the mark made by the player. 2.5. Technical Analysis Notational analysis was performed by video analysis by five experienced handball coaches (an observer-to-player ratio of 1:1) engaged by the Danish Handball Federation (DHF). The operational definitions of these variables were the following: goal (an attempt with successful scoring), shot (an attempt to score a goal made with any (legal) part of the body, either on or off-target), successful shot (an attempt that successfully scores a goal, given by the ratio between goals and shots and expressed as a percentage), 1 vs. 1 duels (offensive breakthrough to an opponent with the ball) [19,20]. 2.6. Statistical Analyses Differences between game formats and between sexes were analyzed using a linear mixed model with unstructured covariance, considering the fact that participants differed regarding the number of game formats they participated in [21]. The game format was set as a fixed effect and the individual subjects and teams were set as random effects. Physical, physiological, and perceptual variables were dependent variables. If a significant effect was found, a pairwise comparison was tested using the Bonferroni post-hoc test. Magnitude- based inferences were adopted to interpret differences between game formats and sexes [22]. Effect sizes (ES) were calculated using mean differences and pooled standard deviation, and classified according to Hopkins and Marshall [22] as following: trivial (ES < 0.2), small (ES = 0.2–0.6), moderate (ES = 0.6–1.2), large (ES = 1.2–2.0), very large (ES = 2.0–4.0), and huge (ES > 4.0). When 90% confidence intervals overlapped positive and negative values, the effect was deemed as unclear. Otherwise, the effect was deemed as the observed magnitude [23]. Significance was set at p ≤ 0.05. Data analysis was performed using the Statistical Package for Social Science statistical software (version 23, IBM SPSS Statistics, Chicago, IL, USA) and an online-available Excel spreadsheet [24]. Int. J. Environ. Res. Public Health 2021, 18, 5663 4 of 14 3. Results 3.1. Activity Profile Boys covered more TD, HSR, and sprinting and performed more sprints in L5 com- pared to S3 + 1, S4, M4, and M5 (p < 0.05; ES = 0.9 to 1.9). Moreover, the TD was moderately higher in S3 + 1 compared to M5 (p = 0.026; ES = 0.9 [0.4; 1.3]). Higher peak speed were reached during L5 compared to S4 and M5 (p = 0.01; ES = 0.9 to 1.1) (Table 1) (Figure 1A). Table 1. Differences in peak and average values and total distance between game formats. Variables (U9) Sex S3 + 1 S4 M4 M5 L5 Activity profile TD (m) Boys 1133 ± 171 988 ± 141 1106 ± 157 977 ± 172 a 1320 ± 232 a,b,c,d Girls 965 ± 195 878 ± 159 999 ± 156 846 ± 124 c 1125 ± 134 a,b,d Vpeak (km·h−1) Boys 21.4 ± 3.2 20.1 ± 3.0 22.4 ± 3.1 20.7 ± 2.8 23.6 ± 3.0 b,d Girls 19.0 ± 2.3 19.5 ± 2.6 19.5 ± 2.8 19.9 ± 2.7 21.6 ± 2.4 a Sprints (counts) Boys 5.1 ± 5.4 2.9 ± 3.6 5.8 ± 5.2 4.6 ± 5.5 11.9 ± 7.5 a,b,c,d Girls 2.9 ± 3.5 1.5 ± 1.5 2.4 ± 2.6 2.7 ± 2.5 8.4 ± 7.6 a,b,c,d Acctotal (counts) Boys 212.0 ± 20.3 209.0 ± 21.2 200.7 ± 20.0 186.1 ± 21.6 a,b 172.2 ± 23.2 a,b,c,d Girls 198.0 ± 14.4 197.1 ± 25.3 200.3 ± 21.4 177.4 ± 21.7 a,b,c 160.7 ± 15.0 a,b,c,d Dectotal (counts) Boys 219.8 ± 17.3 209.8 ± 21.7 203.6 ± 17.8 191.5 ± 19.7 a,b,c 182.7 ± 24.0 a,b,d Girls 201.1 ± 18.1 197.1 ± 26.4 204.1 ± 17.8 184.7 ± 20.9 a,b,c 173.2 ± 13.6 a,b,d Heart rate HRavg (bpm) Boys 175.8 ± 10.3 165.9 ± 11.2 167.9 ± 11.5 165.8 ± 13.1 169.7 ± 14.8 Girls 174.7 ± 10.2 171.1 ± 9.7 168.8 ± 11.1 164.6 ± 13.5 172.8 ± 8.7 HRpeak (bpm) Boys 195.1 ± 10.3 185.6 ± 10.1 a 192.3 ± 10.5 189.2 ± 11.6 191.2 ± 11.1 Girls 196.3 ± 8.6 191.9 ± 11.4 189.4 ± 9.5 186.1 ± 12.8 192.3 ± 8.3 Subjective perceptions RPE (AU) Boys 8.4 ± 3.8 10.2 ± 4.2 8.5 ± 4.1 9.2 ± 5.2 6.9 ± 4.9 Girls 6.9 ± 3.3 10.1 ± 3.0 6.4 ± 3.3 6.8 ± 4.4 6.5 ± 4.8 RPF (AU) Boys 4.3 ± 4.6 5.2 ± 4.7 5.2 ± 4.3 4.5 ± 5.1 5.0 ± 5.4 Girls 3.8 ± 2.7 5.0 ± 3.7 5.1 ± 2.8 5.4 ± 3.6 4.1 ± 4.7 Data are mean ± SD. S3 + 1: small size, 3 vs. 3 + offensive goalkeeper; S4: small size, 4 vs. 4; M4: medium size, 4 vs. 4; M5: medium size, 5 vs. 5; L5: large size, 5 vs. 5. Acctotal: total accelerations; Dectotal: total decelerations; HRavg: average heart rate; HRpeak: peak heart rate; TD: total distance; Vpeak: peak speed attained; RPE: rating of perceived exertion; RPF: rating of perceived enjoyment/fun. a denotes significant differences compared to S3 + 1; b to S4; c to M4; d to M5 (p ≤ 0.05). Int. J. Environ. Res. Public Health 2021, 18, 5663 5 of 14 Figure 1. Distance covered in different speed zones in U9 (A) boy and (B) girl handball players by game formats. S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: medium size, 5 v 5; L5: large size, 5 v 5. St/W: standing/walking; MSR: moderate-speed running; HSR: high-speed running. a denotes significant differences compared to S3 + 1; b to S4; c to M4; d to M5 (p ≤ 0.05). Furthermore, number of Acctotal and Dectotal were lower in L5 compared to S3 + 1, S4, and M4 (p < 0.05; ES = 0.6 to 1.8). In addition, M5 exhibited lower number of Acctotal and Dectotal than S3 + 1 and S4 (p < 0.05; ES = 0.8 to 1.5). The numbers of Acc<1.5 and Acc1.5–2.3 were lower during L5 compared to S3 + 1, S4, and M4 (p < 0.05; ES = 0.7 to 1.6). Conversely, the number of Acc<1.5 was moderately higher in S4 compared to M5 (p = 0.033; ES = 0.8 (0.3; 1.3)), and the number of Acc1.5–2.3 was largely higher in S3 + 1 than M5 (p < 0.01; ES = 1.4 Figure 1. Distance covered in different speed zones in U9 (A) boy and (B) girl handball players by game formats. S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: medium size, 5 v 5; L5: large size, 5 v 5. St/W: standing/walking; MSR: moderate-speed running; HSR: high-speed running. a denotes significant differences compared to S3 + 1; b to S4; c to M4; d to M5 (p ≤ 0.05). Int. J. Environ. Res. Public Health 2021, 18, 5663 5 of 14 Girls covered more TD during L5 compared to S3 + 1, S4, and M5 (p < 0.05; ES = 0.9–2.1). In addition, the TD was moderately higher in M4 compared to M5 (p = 0.043; ES = 1.0 [0.4; 1.6]). Higher peak speed were reached during L5 compared to S3 + 1 (p = 0.044; ES = 1.0 [1.6; 0.4]). Girls covered more HSR and sprinting and performed more sprints in L5 compared to S3 + 1, S4, M4, and M5 (p < 0.05; ES = 0.8 to 2.9) (Table 1) (Figure 1B). Furthermore, number of Acctotal and Dectotal were lower in L5 compared to S3 + 1, S4, and M4 (p < 0.05; ES = 0.6 to 1.8). In addition, M5 exhibited lower number of Acctotal and Dectotal than S3 + 1 and S4 (p < 0.05; ES = 0.8 to 1.5). The numbers of Acc<1.5 and Acc1.5–2.3 were lower during L5 compared to S3 + 1, S4, and M4 (p < 0.05; ES = 0.7 to 1.6). Conversely, the number of Acc<1.5 was moderately higher in S4 compared to M5 (p = 0.033; ES = 0.8 (0.3; 1.3)), and the number of Acc1.5–2.3 was largely higher in S3 + 1 than M5 (p < 0.01; ES = 1.4 (0.8; 1.9)). Notably, lower number of decelerations were observed during L5 compared to S3 + 1, S4, and M4 (p < 0.05; ES = 1.0 to 1.3). Furthermore, higher number of Dec1.5–2.3, were observed in S3 + 1 compared to M5 and L5 (p < 0.05; ES = 0.9 to 1.3). Additionally, S4 showed higher number of Dec1.5–2.3 than M5 (p = 0.042; ES = 0.9 (0.4; 1.3)). S3 + 1 showed higher number of Dec>2.3 than M5 (p = 0.019; ES = 0.8 (0.3; 1.3)). For girls, number of Acctotal and Dectotal were lower in L5 compared to S3 + 1, S4, and M4 (p < 0.05; ES = 1.2 to 2.5). In addition, M5 exhibited lower number of Acctotal than S3 + 1, S4, and M4 (p < 0.05; ES = 0.8 to 1.0). M5 had moderately lower number of Dectotal than M4 (p = 0.047; ES = 0.9 (0.3; 1.5)). Girls had lower number of Acc<1.5 during L5 compared to S3 + 1, S4, M4, and M5 (p < 0.05; ES = 1.1 to 1.8). Similarly, during L5, girls had lower number of Acc1.5–2.3 than S3 + 1, S4, and M4 (p < 0.05; ES = 1.1 to 2.1). In addition, Acc1.5–2.3 had fewer efforts in M5 than in S3 + 1 and S4 (p < 0.05; ES = 0.8 to 1.4). In Dec1.5–2.3, L5 had lower number of efforts than S3 + 1, S4, and M4 (p < 0.05; ES = 1.0 to 1.5). In addition, Dec1.5-2.3 in S3 + 1 had higher number than M5 (p = 0.002; ES = 1.2 [0.6; 1.8]). Detailed representations of accelerations and decelerations are reported in Figures 2 and 3. 3.2. Heart Rate and Subjective Perceptions Boys attained higher HRpeak in S3 + 1 compared to S4 (p = 0.029; ES = 0.9 (0.4; 1.4)) (Table 1). In addition, boys spent more time within 180–200 bpm in S3 + 1 than in S4 (p = 0.045; ES = 0.8 (0.3; 1.3)) (Figure 4). No significant differences were found between game formats in the RPEs and RPFs of boys (p > 0.05). Girls had higher times below 120 bpm during S3 + 1 compared to M4 and L5 (p < 0.05; ES = 1.3 to 1.6) (Table 1). In addition, girls spent more time between 120–160 bpm in M5 than S3 + 1 (p = 0.028; ES = 0.9 (0.3; 1.5)) (Figure 4). No significant differences were found between game formats in the RPEs and RPFs for girls (p > 0.05) (Table 1). 3.3. Technical Analysis For the total number of shots, more shots occurred in S3 + 1 and S4 compared to M5 and L5 (p < 0.05; ES = 0.8 to 1.1). In contrast, no differences were observed for goals, successful shots, or duels in all the formats. For girls, the total amount of goals was higher in S3 + 1 than in M4, M5, and L5 (p < 0.05; ES = 0.9 to 1.3), as well as in S4 compared to M5 (p = 0.029; ES = 1.3 (0.3; 1.6)). In addition, situation S3 + 1 had more shots than M5 and L5 (p < 0.05; ES = 1.0 to 1.6). Furthermore, girls were less successful with shots in M5 than in S3 + 1 and S4 (p < 0.05; ES = 0.9 to 1.6) (Table 2). Int. J. Environ. Res. Public Health 2021, 18, 5663 6 of 14 blic Health 2021, 18, 5663 6 of 14 Figure 2. Number of accelerations in U9 (A) boy and (B) girl handball players by game formats. S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: me- dium size, 5 v 5; L5: large size, 5 v 5. a denotes significant differences compared to S3 + 1; b to S4; c to M4; d to M5 (p ≤ 0.05). Acc < 1.5 Acc 1.5 – 2.3 Acc > 2.3 0 60 120 180 Number of accelerations (n) S3+1 S4 M4 M5 L5 A) a,b,c b a a,b,c Acc < 1.5 Acc 1.5 – 2.3 Acc > 2.3 0 60 120 180 Number of accelerations (n) S3+1 S4 M4 M5 L5 B) a,b,c,d a,b a,b,c Figure 2. Number of accelerations in U9 (A) boy and (B) girl handball players by game formats. S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: medium size, 5 v 5; L5: large size, 5 v 5. a denotes significant differences compared to S3 + 1; b to S4; c to M4; d to M5 (p ≤ 0.05). Int. J. Environ. Res. Public Health 2021, 18, 5663 7 of 14 ublic Health 2021, 18, 5663 7 of 14 Figure 3. Number of decelerations in U9 (A) boy and (B) girl handball players by game formats. S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: me- dium size, 5 v 5; L5: large size, 5 v 5. a denotes significant differences compared to S3 + 1; b to S4; c to M4 (p ≤ 0.05). 3.2. Heart Rate and Subjective Perceptions Boys attained higher HRpeak in S3 + 1 compared to S4 (p = 0.029; ES = 0.9 (0.4; 1.4)) (Table 1). In addition, boys spent more time within 180–200 bpm in S3 + 1 than in S4 (p = 0.045; ES = 0.8 (0.3; 1.3)) (Figure 4). No significant differences were found between game f t i th RPE d RPF f b ( > 0 05) Gi l h d hi h ti b l 120 b Dec < -1.5 Dec -1.5 – -2.3 Dec < -2.3 0 60 120 180 Number of decelerations (n) A) S3+1 S4 M4 M5 L5 a,b,c a,b a a Dec < -1.5 Dec -1.5 – -2.3 Dec < -2.3 0 60 120 180 Number of decelerations (n) B) S3+1 S4 M4 M5 L5 a a,b,c Figure 3. Number of decelerations in U9 (A) boy and (B) girl handball players by game formats. S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: medium size, 5 v 5; L5: large size, 5 v 5. a denotes significant differences compared to S3 + 1; b to S4; c to M4 (p ≤ 0.05). Int. J. Environ. Res. Public Health 2021, 18, 5663 8 of 14 lic Health 2021, 18, 5663 8 of 14 Figure 4. Heart rate distribution during U9 (A) boy and (B) girl handball games. S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: medium size, 5 v 5; L5: large size 5 v 5. a denotes significant differences compared to S3 + 1 (p ≤ 0.05). 3.3. Technical Analysis For the total number of shots, more shots occurred in S3 + 1 and S4 compared to M5 and L5 (p < 0.05; ES = 0.8 to 1.1). In contrast, no differences were observed for goals, suc- cessful shots, or duels in all the formats. For girls, the total amount of goals was higher in S3 + 1 than in M4, M5, and L5 (p < 0.05; ES = 0.9 to 1.3), as well as in S4 compared to M5 (p Figure 4. Heart rate distribution during U9 (A) boy and (B) girl handball games. S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: medium size, 5 v 5; L5: large size 5 v 5. a denotes significant differences compared to S3 + 1 (p ≤ 0.05). Int. J. Environ. Res. Public Health 2021, 18, 5663 9 of 14 Table 2. Differences in technical demands between game formats. Data are mean ± SD. Variables Sex S3 + 1 S4 M4 M5 L5 Shots (counts) Boys 7.0 ± 4.1 6.4 ± 3.9 4.9 ± 3.3 3.5 ± 2.6 a,b 3.2 ± 2.1 a,b Girls 7.4 ± 3.5 6.0 ± 5.2 4.7 ± 2.8 3.8 ± 3.0 a 2.8 ± 2.0 a Goals (counts) Boys 2.5 ± 2.4 2.2 ± 2.1 2.0 ± 2.0 1.1 ± 1.3 1.1 ± 1.0 Girls 3.9 ± 2.9 2.4 ± 1.5 1.6 ± 1.4 a 0.7 ± 0.9 b 1.0 ± 1.3 a Successful shots (counts) Boys 35.2 ± 29.2 34.7 ± 23.9 38.0 ± 30.1 29.2 ± 32.0 32.2 ± 28.9 Girls 53.7 ± 23.9 47.4 ± 41.3 35.3 ± 25.8 15.4 ± 22.5 a,b 31.0 ± 38.9 Duels (counts) Boys 1.1 ± 1.2 0.7 ± 1.3 0.7 ± 1.1 0.5 ± 0.9 0.4 ± 0.6 Girls 1.7 ± 2.8 1.6 ± 2.1 1.2 ± 1.7 0.6 ± 1.0 0.6 ± 1.2 Data are mean ± SD. S3 + 1: small size, 3 vs. 3 + offensive goalkeeper; S4: small size, 4 vs. 4; M4: medium size, 4 vs. 4; M5: medium size, 5 vs. 5; L5: large size, 5 vs. 5. a denotes significant differences compared to S3 + 1; b to S4. 3.4. Gender The boys covered more TD in S3 + 1, S4, M5, and L5 compared to the girls (p < 0.05; ES = 0.8 to 0.9). Furthermore, the boys reached higher Vpeak during S4 and M5 compared to the girls (p < 0.05; ES = 0.7 to 0.9). Moderate higher sprints were observed in the M4 format for boys compared to girls (p = 0.020; ES = 0.7 [0.2; 1.3]). Notably, the jog distance was higher for boys during S3 + 1, S4, M4, M5, and L5 compared to girls (p < 0.05; ES = 0.6 to 1.2). In addition, sprinting in M4 and L5 was higher for boys than girls (p < 0.05; ES = 0.8 to 0.5). Moreover, Acctotal and Dectotal were higher in S3 + 1 for boys compared to girls (p < 0.05; ES = 0.7 to 1.0). Acc<1.5 was moderately higher during S4 in boys compared to girls (p = 0.044; ES = 0.6 (0.1; 1.2)). Notably, boys had higher numbers of decelerations during L5 compared to girls in Dec>2.3 (p = 0.021; ES = 0.6 (0.2; 1.1)). Furthermore, S3 + 1 and S4 formats had more decelerations for boys compared to girls (p < 0.05; ES = 0.8 to 1.2). Conversely, in Dec1.5-2.3, girls had more decelerations in M4 than boys (p = 0.009; ES = 0.9 (1.5; 0.4)). The girls had higher Time<120 in S3 + 1 and S4 compared to the boys (p < 0.05; ES = 0.7 to 1.9) (Table 1). In addition, Time>200 in S4 was moderately higher for girls compared to boys (p = 0.034; ES = 0.7 (1.2; 0.1)). A detailed representation of the differences in activity profile, heart rate, subjective ratings, and technical involvement is reported in Figure 5. Res. Public Health 2021, 18, 5663 9 of 14 Table 2. Differences in technical demands between game formats. Data are mean ± SD. Variables Sex S3 + 1 S4 M4 M5 L5 ots (counts) Boys 7.0 ± 4.1 6.4 ± 3.9 4.9 ± 3.3 3.5 ± 2.6 a,b 3.2 ± 2.1 a,b Girls 7.4 ± 3.5 6.0 ± 5.2 4.7 ± 2.8 3.8 ± 3.0 a 2.8 ± 2.0 a als (counts) Boys 2.5 ± 2.4 2.2 ± 2.1 2.0 ± 2.0 1.1 ± 1.3 1.1 ± 1.0 Girls 3.9 ± 2.9 2.4 ± 1.5 1.6 ± 1.4 a 0.7 ± 0.9 b 1.0 ± 1.3 a ful shots (counts) Boys 35.2 ± 29.2 34.7 ± 23.9 38.0 ± 30.1 29.2 ± 32.0 32.2 ± 28.9 Girls 53.7 ± 23.9 47.4 ± 41.3 35.3 ± 25.8 15.4 ± 22.5 a,b 31.0 ± 38.9 uels (counts) Boys 1.1 ± 1.2 0.7 ± 1.3 0.7 ± 1.1 0.5 ± 0.9 0.4 ± 0.6 Girls 1.7 ± 2.8 1.6 ± 2.1 1.2 ± 1.7 0.6 ± 1.0 0.6 ± 1.2 e mean ± SD. S3 + 1: small size, 3 vs. 3 + offensive goalkeeper; S4: small size, 4 vs. 4; M4: medium size, 4 vs. 4; M5: m size, 5 vs. 5; L5: large size, 5 vs. 5. a denotes significant differences compared to S3 + 1; b to S4. 3.4. Gender The boys covered more TD in S3 + 1, S4, M5, and L5 compared to the girls (p < 0.05; ES = 0.8 to 0.9). Furthermore, the boys reached higher Vpeak during S4 and M5 compared to the girls (p < 0.05; ES = 0.7 to 0.9). Moderate higher sprints were observed in the M4 format for boys compared to girls (p = 0.020; ES = 0.7 [0.2; 1.3]). Notably, the jog distance was higher for boys during S3 + 1, S4, M4, M5, and L5 compared to girls (p < 0.05; ES = 0.6 to 1.2). In addition, sprinting in M4 and L5 was higher for boys than girls (p < 0.05; ES = 0.8 to 0.5). Moreover, Acctotal and Dectotal were higher in S3 + 1 for boys compared to girls (p < 0.05; ES = 0.7 to 1.0). Acc<1.5 was moderately higher during S4 in boys compared to girls (p = 0.044; ES = 0.6 (0.1; 1.2)). Notably, boys had higher numbers of decelerations during L5 compared to girls in Dec>2.3 (p = 0.021; ES = 0.6 (0.2; 1.1)). Furthermore, S3 + 1 and S4 formats had more decelerations for boys compared to girls (p < 0.05; ES = 0.8 to 1.2). Conversely, in Dec1.5-2.3, girls had more decelerations in M4 than boys (p = 0.009; ES = 0.9 (1.5; 0.4)). The girls had higher Time<120 in S3 + 1 and S4 compared to the boys (p < 0.05; ES = 0.7 to 1.9) (Table 1). In addition, Time>200 in S4 was moderately higher for girls compared to boys (p = 0.034; ES = 0.7 (1.2; 0.1)). A detailed representation of the differences in activity profile, heart rate, subjective ratings, and technical involvement is reported in Figure 5. -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Sprinting HSR MSR Jog St / W Effect sizes S3+1 S4 M4 M5 L5 Favours girls Favours boys B) Figure 5. Cont. Int. J. Environ. Res. Public Health 2021, 18, 5663 10 of 14 ublic Health 2021, 18, 5663 10 of 14 Figure 5. Differences in (A) overall physical and physiological demands, (B) activity profile, (C) accelerations and decelerations, (D) heart rate, and (E) technical demands between boys and girls during different U9 handball games. The forest plots are effect sizes (90% CI). S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: medium size, 5 v 5; L5: large size, 5 v 5. St/W: standing/walking; MSR: moderate-speed running; HSR: high-speed run- ning. 4. Discussion This study provides the first detailed analysis of movement patterns and heart rates in U9 team handball for boys and girls, showing that the exercise intensity, heart rates, and technical involvement are high during small, medium, and large-sized games in all investigated formats. When comparing game formats, we observed higher distances cov- ered and more sprints with L5 but a lower number of accelerations and decelerations com- pared to all the other formats. Notably, heart rates were similar between game formats. Irrespective of game format, boys covered 977–1320 m and girls covered 846–1124 m. For boys and girls, remarkably in the L5 format, TD, Vpeak, sprints, HSR, and sprinting were higher, whereas St/W, JOG, Acctotal, Dectotal, Acc<1.5, Acc1.5–2.3, Acc>1.5, Dec<1.5, Dec1.5–2.3, and Dec>1.5 were lower than other formats and, on many occasions, significantly different. This may be because there is more room for sprinting and high-intensity running on larger pitches, which is supported by the greater distance covered with high-intensity running and higher Vpeak during games on larger pitch sizes (40 × 20 m) compared with small pitches (20 × 13 m) in adult football players [25]. Interestingly, no differences were found -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Dec ≤ -2.3 m⋅s-2 Dec -2.2 – -1.5 m⋅s-2 Dec ≥ 1.4 m⋅s-2 Acc ≥ 2.3 m⋅s-2 Acc 1.5 – 2.9 m⋅s-2 Acc ≤ 1.4 m⋅s-2 Effect sizes S3+1 S4 M4 M5 L5 Favours girls Favours boys C) -3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Time >200 bpm Time 180 – 200 bpm Time160 – 180 bpm Time 120 – 160 bpm Time < 120 bpm Effect sizes S3+1 S4 M4 M5 L5 Favours girls Favours boys D) -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0 Duels Succesful shots Goals Total shots Effect sizes S3+1 S4 M4 M5 L5 Favours girls Favours boys E) Figure 5. Differences in (A) overall physical and physiological demands, (B) activity profile, (C) accelerations and decelera- tions, (D) heart rate, and (E) technical demands between boys and girls during different U9 handball games. The forest plots are effect sizes (90% CI). S3 + 1: small size, 3 v 3 + offensive goalkeeper; S4: small size, 4 v 4; M4: medium size, 4 v 4; M5: medium size, 5 v 5; L5: large size, 5 v 5. St/W: standing/walking; MSR: moderate-speed running; HSR: high-speed running. 4. Discussion This study provides the first detailed analysis of movement patterns and heart rates in U9 team handball for boys and girls, showing that the exercise intensity, heart rates, and technical involvement are high during small, medium, and large-sized games in all investigated formats. When comparing game formats, we observed higher distances covered and more sprints with L5 but a lower number of accelerations and decelerations compared to all the other formats. Notably, heart rates were similar between game formats. Irrespective of game format, boys covered 977–1320 m and girls covered 846–1124 m. For boys and girls, remarkably in the L5 format, TD, Vpeak, sprints, HSR, and sprinting were higher, whereas St/W, JOG, Acctotal, Dectotal, Acc<1.5, Acc1.5–2.3, Acc>1.5, Dec<1.5, Dec1.5–2.3, and Dec>1.5 were lower than other formats and, on many occasions, significantly different. This may be because there is more room for sprinting and high-intensity running on larger pitches, which is supported by the greater distance covered with high-intensity running and higher Vpeak during games on larger pitch sizes (40 × 20 m) compared with small pitches (20 × 13 m) in adult football players [25]. Interestingly, no differences were found Int. J. Environ. Res. Public Health 2021, 18, 5663 11 of 14 between S4 and M4 in any variables (physical, physiological, subjective perception, and technical). As we already reported, other team sports showed that manipulating the player numbers and the pitch size can alter the exercise intensity (i.e., distance covered, jogging and walking, heart rate, and tackling, dribbling, goal attempts, and passes) during a game in different sports [8]. The forces generated while rapidly changing direction, stopping, and landing, as well as during jumping and shooting, may confer excellent osteogenic properties to team handball [26]. It is well known from cross-sectional studies that participation in sports activities is associated with markedly higher muscle mass and bone mineralization, as well as better coordination and postural balance [27,28], and a longitudinal intervention study with 8–10-year-old children has shown that participation in school-based small-sized ball games enhances the same parameters [29]. The mean HR was high for boys and girls, at 166–176 bpm and 165–175 bpm, respectively, in all game formats. A high HR during sports and, specifically, team handball match-play, irrespective of game format and gender, is important for the health profile of children [30]. Aerobic high-intensity training (>90% maximum HR) has been shown to be superior to moderate continuous training in improving cardiorespiratory fitness [31,32], which has been identified as a strong independent predictor of the risk of cardiovascular diseases and mortality [33]. Sports participation is an effective way to improve aerobic and anaerobic fitness, especially participation in high-intensity ball games [34]. For the Time180–200 and Time>200 in S3 + 1 format, young girl and boy team handball players spent more time above 180 bpm, which is not significantly different but working at a high intensity for more time could improve cardiorespiratory fitness positively [35]. No differences occurred in subjective perception between different game formats, in contrast with other studies [6,36] that found that larger courts felt more physically demanding. In our study, we had more goals and more shots in the small size pitch (S3 + 1, S4), as was also observed in a study by Randers and colleagues [7], where smaller pitches created more technical actions and may seem logical, as ball contacts are higher during a game with fewer players [37]. Interestingly, no differences were found in 1 v 1 duels in all the formats, that the players may try to score or shot faster in games with small size pitches. Involvement with many relevant activities is important in terms of motivation for children [38], as it helps the players to continue as active handball players. Maturation at this stage is still early, whereas it seems that the physiological load of the game is higher for boys than for girls, with many differences between them, as is supported by the work of Michalsik and colleagues [3] in the different distance zones, except for the TD, which females covered more of. A possible explanation is that boys have more self-confidence and perceived self-competence, making the game more demanding [39]. Only one significant difference was observed in favor of the girls in Dec1.5–2.3, which had more decelerations in the M4 format. However, for physical loading between sexes, similar HR values were found, with only three comparisons, girls spent more time below 120 bpm in S3 + 1 and S4 compared to boys for Time>200 in S4. Additionally, no significant differences were found for subjective perceptions or the technical analysis. In conclusion, having both genders mixed in the same format and game would possibly be very demanding for girls in terms of activity patterns at this age. It is important to underline some limitations inherent to this study, Firstly, physical and physiological demands were compared across game formats of various pitch sizes and numbers of players, and thus, relative space per player was not constant. Secondly, maximum HR, maximal aerobic speed, and maximal sprinting speed were not assessed. The use of fixed HR and speed zones does not reflect the actual individual capacity, possibly resulting in under- or overestimating the real physical and physiological demands of the game. Although the technical analysis was carried out by experienced handball coaches, this analysis could be somewhat subjective. Thus, our technical analysis should be interpreted with caution. Finally, for logistical reasons, we were unable to describe the physical levels of the players. Future studies are warranted to use individualized HR and speed zones to accurately quantify the physical and physiological demands of youth Int. J. Environ. Res. Public Health 2021, 18, 5663 12 of 14 team handball as well as physical evaluations of the players. In this context, the fitness component of max speed can be adopted in future studies as suggested by [40]. 5. Conclusions In summary, the HR and high-intensity distances are high in U9 team handball matches irrespective of the game format. The present data provide insight into how different game formats influence the physiological and the physical loading and evidence that various types of match-plays can contribute significantly to the improvement in the musculoskeletal and cardiovascular fitness of U9 boys because of high HRs and high-intensity running distances, along with multiple accelerations and specific actions with considerable impact. In all the game formats, physical loading seems similar but, interestingly, on the large pitch, the physiological load was higher. Playing with fewer players on smaller pitches resulted in minor changes to the physiological loading but elevated the technical involvement of players, which favors the use of smaller formats to emphasize technical demands. Several differences between girls and boys were found in U9 team handball players that should be considered when planning games for boys and girls separately or for mixed-gender games. The various game types could provide valuable information to coaches in the selection of players or training guidance. We would recommend the use of games with fewer players on smaller courts for U9 boys and girls since we believe that technical development is the most important factor at this age. Author Contributions: Conceptualization, G.E. and M.N.L.; methodology, M.B.R.; validation, M.N.L., P.K., M.B.R. and G.E.; formal analysis, G.E.; investigation, G.E. and R.C.E.; resources, P.K., M.B.R. and M.N.L.; data curation, G.E. and V.R.; writing—original draft preparation, G.E.; writing— review and editing, M.N.L., P.K., M.B.R., V.R. and G.E.; visualization, G.E. and V.R.; supervision, M.N.L., P.K. and M.B.R.; project administration, G.E.; funding acquisition, M.N.L., P.K. and M.B.R. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Danish Handball Federation, grant number 10-154-15655. Institutional Review Board Statement: Ethical review and approval were waived for this study, due to these types of academic research projects, formal ethical approval is not required by law. Informed Consent Statement: Patient consent was waived, due to these types of academic research projects, patient consent is not required by law but only oral consent. Conflicts of Interest: The authors declare no conflict of interest. References 1. Gorostiaga, E.M.; Granados, C.; Ibanez, J.; Izquierdo, M. Differences in physical fitness and throwing velocity among elite and amateur male handball players. Int. J. Sports Med. 2005, 26, 225–232. [CrossRef] 2. Povoas, S.C.; Seabra, A.F.; Ascensao, A.A.; Magalhaes, J.; Soares, J.M.; Rebelo, A.N. Physical and physiological demands of elite team handball. J. Strength Cond. Res. 2012, 26, 3365–3375. [CrossRef] 3. Michalsik, L.B.; Madsen, K.; Aagaard, P. Match performance and physiological capacity of female elite team handball players. Int. J. Sports Med. 2014, 35, 595–607. [CrossRef] [PubMed] 4. Chelly, M.S.; Hermassi, S.; Aouadi, R.; Khalifa, R.; Van den Tillaar, R.; Chamari, K.; Shephard, R.J. Match analysis of elite adolescent team handball players. J. Strength Cond. Res. 2011, 25, 2410–2417. [CrossRef] [PubMed] 5. Souhail, H.; Castagna, C.; Mohamed, H.Y.; Younes, H.; Chamari, K. Direct validity of the yo-yo intermittent recovery test in young team handball players. J. Strength Cond. Res. 2010, 24, 465–470. [CrossRef] [PubMed] 6. Madsen, M.; Ermidis, G.; Rago, V.; Surrow, K.; Vigh-Larsen, J.F.; Randers, M.B.; Krustrup, P.; Larsen, M.N. Activity Profile, Heart Rate, Technical Involvement, and Perceived Intensity and Fun in U13 Male and Female Team Handball Players: Effect of Game Format. Sports 2019, 7, 90. [CrossRef] 7. Randers, M.B.; Andersen, T.B.; Rasmussen, L.S.; Larsen, M.N.; Krustrup, P. Effect of game format on heart rate, activity profile, and player involvement in elite and recreational youth players. Scand. J. Med. Sci. Sports 2014, 24 (Suppl. 1), 17–26. [CrossRef] [PubMed] 8. Halouani, J.; Chtourou, H.; Gabbett, T.; Chaouachi, A.; Chamari, K. Small-sided games in team sports training: A brief review. J. Strength Cond. Res. 2014, 28, 3594–3618. [CrossRef] 9. Aroso, J.; Rebelo, A.; Gomes-Pereira, J. Physiological impact of selected game-related exercises. J. Sports Sci. 2004, 22, 522. Int. J. Environ. Res. Public Health 2021, 18, 5663 13 of 14 10. Owen, A.L.; Wong del, P.; McKenna, M.; Dellal, A. Heart rate responses and technical comparison between small- vs. large-sided games in elite professional soccer. J. Strength Cond. Res. 2011, 25, 2104–2110. [CrossRef] 11. Malina, R.M.; Eisenmann, J.C.; Cumming, S.P.; Ribeiro, B.; Aroso, J. Maturity-associated variation in the growth and functional capacities of youth football (soccer) players 13–15 years. Eur. J. Appl. Physiol. 2004, 91, 555–562. [CrossRef] 12. Malina, R.M.C.; Bar-Or, O. Growth, Maturation, and Physical Activity, 2nd ed.; Human Kinetics: Champaign, IL, USA, 2004. 13. Hill-Haas, S.V.; Rowsell, G.J.; Dawson, B.T.; Coutts, A.J. Acute physiological responses and time-motion characteristics of two small-sided training regimes in youth soccer players. J. Strength Cond. Res. 2009, 23, 111–115. [CrossRef] [PubMed] 14. Barbero-Alvarez, J.C.; Gomez-Lopez, M.; Castagna, C.; Barbero-Alvarez, V.; Romero, D.V.; Blanchfield, A.W.; Nakamura, F.Y. Game Demands of Seven-A-Side Soccer in Young Players. J. Strength Cond. Res. 2017, 31, 1771–1779. [CrossRef] 15. Sanchez-Sanchez, J.; Sanchez, M.; Hernandez, D.; Ramirez-Campillo, R.; Martinez, C.; Nakamura, F.Y. Fatigue In U12 Soccer-7 Players During Repeated One-Day Tournament Games—A Pilot Study. J. Strength Cond. Res. 2017, 33, 3092–3097. [CrossRef] 16. Akyildiz, Z.; Yildiz, M.; Clemente, F.M. The reliability and accuracy of Polar Team Pro GPS units. Proc. Inst. Mech. Eng. Part P J. Sports Eng. Technol. 2020, 1754337120976660. [CrossRef] 17. Rebelo, A.; Brito, J.; Seabra, A.; Oliveira, J.; Drust, B.; Krustrup, P. A new tool to measure training load in soccer training and match play. Int. J. Sports Med. 2012, 33, 297–304. [CrossRef] [PubMed] 18. Wewers, M.E.; Lowe, N.K. A critical review of visual analogue scales in the measurement of clinical phenomena. Res. Nurs. Health 1990, 13, 227–236. [CrossRef] 19. Liu, H.; Gomez, M.-Á.; Lago-Peñas, C.; Sampaio, J. Match statistics related to winning in the group stage of 2014 Brazil FIFA World Cup. J. Sports Sci. 2015, 33, 1205–1213. [CrossRef] [PubMed] 20. Michalsik, L.B.; Madsen, K.; Aagaard, P. Technical match characteristics and influence of body anthropometry on playing performance in male elite team handball. J. Strength Cond. Res. 2015, 29, 416–428. [CrossRef] 21. Cnaan, A.; Laird, N.M.; Slasor, P. Using the general linear mixed model to analyse unbalanced repeated measures and longitudinal data. Stat. Med. 1997, 16, 2349–2380. [CrossRef] 22. Hopkins, W.G.; Marshall, S.W.; Batterham, A.M.; Hanin, J. Progressive statistics for studies in sports medicine and exercise science. Med. Sci. Sports Exerc. 2009, 41, 3–13. [CrossRef] 23. Batterham, A.M.; Hopkins, W.G. Making meaningful inferences about magnitudes. Int. J. Sports Physiol. Perform. 2006, 1, 50–57. [CrossRef] 24. Hopkins, W. A spreadsheet for deriving a confidence interval, mechanistic inference and clinical inference from a P value. Sportscience 2007, 11, 16–21. 25. Casamichana, D.; Castellano, J. Time-motion, heart rate, perceptual and motor behaviour demands in small-sides soccer games: Effects of pitch size. J. Sports Sci 2010, 28, 1615–1623. [CrossRef] [PubMed] 26. Missawi, K.; Zouch, M.; Chakroun, Y.; Chaari, H.; Tabka, Z.; Bouajina, E. Handball Practice Enhances Bone Mass in Specific Sites Among Prepubescent Boys. J. Clin. Densitom. 2016, 19, 389–395. [CrossRef] [PubMed] 27. Seabra, A.; Marques, E.; Brito, J.; Krustrup, P.; Abreu, S.; Oliveira, J.; Rego, C.; Mota, J.; Rebelo, A. Muscle strength and soccer practice as major determinants of bone mineral density in adolescents. Jt. Bone Spine 2012, 79, 403–408. [CrossRef] [PubMed] 28. Vicente-Rodriguez, G.; Ara, I.; Perez-Gomez, J.; Dorado, C.; Calbet, J.A. Muscular development and physical activity as major determinants of femoral bone mass acquisition during growth. Br. J. Sports Med. 2005, 39, 611–616. [CrossRef] 29. Larsen, M.N.; Nielsen, C.M.; Helge, E.W.; Madsen, M.; Manniche, V.; Hansen, L.; Hansen, P.R.; Bangsbo, J.; Krustrup, P. Positive effects on bone mineralisation and muscular fitness after 10 months of intense school-based physical training for children aged 8–10 years: The FIT FIRST randomised controlled trial. Br. J. Sports Med. 2018, 52, 254. [CrossRef] 30. Larsen, M.N.; Nielsen, C.M.; Madsen, M.; Manniche, V.; Hansen, L.; Bangsbo, J.; Krustrup, P.; Hansen, P.R. Cardiovascular adaptations after 10 months of intense school-based physical training for 8- to 10-year-old children. Scand. J. Med. Sci. Sports 2018, 28 (Suppl. 1), 33–41. [CrossRef] 31. Helgerud, J.; Hoydal, K.; Wang, E.; Karlsen, T.; Berg, P.; Bjerkaas, M.; Simonsen, T.; Helgesen, C.; Hjorth, N.; Bach, R.; et al. Aerobic high-intensity intervals improve VO2max more than moderate training. Med. Sci. Sports Exerc. 2007, 39, 665–671. [CrossRef] 32. Nybo, L.; Sundstrup, E.; Jakobsen, M.D.; Mohr, M.; Hornstrup, T.; Simonsen, L.; Bulow, J.; Randers, M.B.; Nielsen, J.J.; Aagaard, P.; et al. High-intensity training versus traditional exercise interventions for promoting health. Med. Sci. Sports Exerc. 2010, 42, 1951–1958. [CrossRef] [PubMed] 33. Archer, E.; Blair, S.N. Physical activity and the prevention of cardiovascular disease: From evolution to epidemiology. Prog. Cardiovasc. Dis. 2011, 53, 387–396. [CrossRef] [PubMed] 34. Bendiksen, M.; Williams, C.A.; Hornstrup, T.; Clausen, H.; Kloppenborg, J.; Shumikhin, D.; Brito, J.; Horton, J.; Barene, S.; Jackman, S.R.; et al. Heart rate response and fitness effects of various types of physical education for 8- to 9-year-old schoolchildren. Eur. J. Sport Sci. 2014, 14, 861–869. [CrossRef] [PubMed] 35. Sperlich, B.; Zinner, C.; Heilemann, I.; Kjendlie, P.L.; Holmberg, H.C.; Mester, J. High-intensity interval training improves VO(2peak), maximal lactate accumulation, time trial and competition performance in 9–11-year-old swimmers. Eur. J. Appl. Physiol. 2010, 110, 1029–1036. [CrossRef] [PubMed] 36. Corvino, M.; Tessitore, A.; Minganti, C.; Sibila, M. Effect of Court Dimensions on Players’ External and Internal Load during Small-Sided Handball Games. J. Sports Sci. Med. 2014, 13, 297–303. Int. J. Environ. Res. Public Health 2021, 18, 5663 14 of 14 37. Abrantes, C.I.; Nunes, M.I.; Maçãs, V.M.; Leite, N.M.; Sampaio, J.E. Effects of the number of players and game type constraints on heart rate, rating of perceived exertion, and technical actions of small-sided soccer games. J. Strength Cond. Res. 2012, 26, 976–981. [CrossRef] 38. Bangsbo, J.; Krustrup, P.; Duda, J.; Hillman, C.; Andersen, L.B.; Weiss, M.; Williams, C.A.; Lintunen, T.; Green, K.; Hansen, P.R.; et al. The Copenhagen Consensus Conference 2016: Children, youth, and physical activity in schools and during leisure time. Br. J. Sports Med. 2016, 50, 1177–1178. [CrossRef] 39. O’Connor, D.; Gardner, L.; Larkin, P.; Pope, A.; Williams, A.M. Positive youth development and gender differences in high performance sport. J. Sports Sci 2020, 38, 1399–1407. [CrossRef] 40. Mendez-Villanueva, A.; Buchheit, M.; Simpson, B.; Bourdon, P.C. Match play intensity distribution in youth soccer. Int. J. Sports Med. 2013, 34, 101–110. [CrossRef]

Exercise Intensity and Technical Involvement in U9 Team Handball: Effect of Game Format.

05-25-2021

Ermidis, Georgios,Ellegard, Rasmus C,Rago, Vincenzo,Randers, Morten B,Krustrup, Peter,Larsen, Malte N

eng

PMC7084740

International Journal of Environmental Research and Public Health Article Participation and Performance Analysis in Children and Adolescents Competing in Time-Limited Ultra-Endurance Running Events Volker Scheer 1,2 , Stefania Di Gangi 3, Elias Villiger 3 , Thomas Rosemann 3, Pantelis T. Nikolaidis 4 and Beat Knechtle 3,5,* 1 Ultra Sports Science Foundation, 69130 Pierre-Bénite, France; [email protected] 2 Health Science Department, Universidad a Distancia de Madrid (UDIMA), 28400 Collado Villaba, Madrid, Spain 3 Institute of Primary Care, University of Zurich, 8091 Zürich, Switzerland; [email protected] (S.D.G.); [email protected] (E.V.); [email protected] (T.R.) 4 Exercise Physiology Laboratory, 18450 Nikaia, Greece; [email protected] 5 Medbase St. Gallen Am Vadianplatz, 9001 St. Gallen, Switzerland * Correspondence: [email protected]; Tel.: +41-71-226-93-00 Received: 8 February 2020; Accepted: 29 February 2020; Published: 3 March 2020



 Abstract: Ultra-endurance running is of increasing popularity in the adult population, mainly due to master runners older than 35 years of age. However, youth runners younger than 19 years of age are also competing in ultra-endurance events, and an increase has been observed in distance-limited events, but no data is available on time-limited ultra-endurance events in this age group. This study investigated participation and performance trends in time-limited ultra-endurance races, including multi-day events, in runners younger than 19 years of age. Between the period 1990 and 2018, the most popular events recorded a total of 214 finishes (from 166 unique finishers (UF)) for 6-h events, 247 (212 UF) for 12-h events, and 805 (582 UF) for 24-h events, respectively. The majority of athletes originated from Europe and North America. Only a minority participated in multi-day events. Overall, speed increased with age, but the overall performance speed decreased across calendar years for 6- and 24-h events as participation numbers grew. In summary, in youth ultra-endurance runners, differences were observed regarding participation and performance across the different time-limited events, the age of the athletes and their country of origin Keywords: boy; girl; ultra-endurance; running; ultramarathon 1. Introduction Ultra-endurance running can be defined as running activities lasting longer than six hours [1–3]. These activities can be distance- or time-based, with typical time-based ultra-endurance events ranging from six hours to several days, the distance covered during this time period being recorded and ranked among competitors [4]. Popular time-based events in the adult population include races over 24 h and participation numbers have increased over the years, particularly among master athletes and women [3,5]. Men are generally faster than women, however, women have closed the gap in the last decade [6]. Most adult ultra-endurance athletes in 24-h events originate from Europe, mostly France and Germany [6]. Multi-day ultra-endurance events are also quite popular, with an exponential increase between the 1990s and 2010 [7]; however, competitor numbers are generally lower compared to other ultra-endurance races. Again, most finishers come from Europe, mainly France, the United Kingdom and Germany, followed by runners from the USA, Asia, Africa, Australia and South America [7]. Ultra-endurance races can be held in challenging and remote environments, and can include races in Int. J. Environ. Res. Public Health 2020, 17, 1628; doi:10.3390/ijerph17051628 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2020, 17, 1628 2 of 12 the heat, desert, cold or jungle environments [3,8–11]. Peak running performance has increased with increasing event duration in races lasting from hours to ten days in adults [4]. As outlined, the participation and performance trends of ultra-endurance running are well described in the adult population [12,13], as well as the performance differences among sexes [14,15]; however, very little is known about childhood participation in ultra-endurance running events and if, indeed, they should be participating in them at all [16]. For ultra-marathons, mainly the aspects of age [4,17] and nationality [18,19] have been investigated. One concern is that running and training for ultra-endurance distances at a young age can have acute negative effects on an immature or developing muscular skeletal system or another organ system. Other concerns may relate to the long-term negative health effects this could have; however, this is currently not known and has not been investigated [15]. Until recently there was only anecdotal evidence that children and adolescents participate in ultra-endurance events; however, one study was able to demonstrate that the participation of children and adolescents younger than 19 years old is a reality [3]. This study looked at participation trends among youth ultramarathon runners and described an exponential increase in participation in the last 20 years [16]. The most popular race distances were those of 100 km, followed by 50 km and 50 miles, with the majority of finishers being older boys between 16–18 years of age [20]. However, ultramarathon running is only distance based, defined as races over marathon distance (42.195 km), and is quite distinct to time-based or multi-day events. Youth ultra-endurance athletes first participated in multi-day events in the year 2000, with events ranging from two to eight days, with distances covering 81 to 293 km, respectively; however, less than 50 runners participated in these events in the last two decades [20]. To date, there are no data available on the country of origin of the ultra-endurance youth participants and similarly no data are available on their performance times. This is the first study examining participation and performances in time-limited ultra-endurance and multi-day events in youth runners. This is of practical interest to scientists, coaches and health care professionals, looking after youth ultra-endurance runners, to get a better understanding of participation trends and performance times. Our aim was, therefore, to examine participation numbers and trends, including countries of origin, race performance times and speeds and sex differences, in children and adolescents younger than 19 years of age in time-based ultra-endurance and multi-day running events. Our hypothesis was that participation numbers would increase over calendar years and in time, more boys than girls would participate in these events, and that the performance times from boys would be faster. 2. Materials and Methods 2.1. Ethical Approval The study was approved by the medical council (Ärztekammer Westfalen Lippe, Germany) and the University of Münster, Germany (Chairperson Prof Berdel, protocol number 2018-304-f-S). 2.2. Data Sampling and Data Analysis All data were obtained from the Deutsche Ultramarathon Vereinigung (DUV) website where all the race results of ultramarathons are recorded (https://statistik.d-u-v.org/index.php). The DUV is the largest ultra-running database worldwide, containing more than 5.8 million performances of more than 1.4 million runners in approximately 60,000 ultramarathon events and is widely used to gain insights in participation and performance trends in ultra-running (www.ultra-marathon.org/) [3,16,17]. A computer script was written to retrieve a list for every event recorded on the website. Each event’s web page was then read by the script to extract the complete data table available. The script compiled all that data into one large Excel file, which was our starting point for further manual filtering of relevant information. We extracted data of time-limited races (i.e., 6, 8, 12, 24, 48, 72 h, 6 and 8 days) from 1990 to 2018. Int. J. Environ. Res. Public Health 2020, 17, 1628 3 of 12 The following variables were extracted: year of race, race distance/duration, name of the race, race performance (km or miles), name of athlete, year of birth, nationality of athlete, and sex of athlete. Running speed (km/h) was calculated from the performance and duration variables. Age was obtained by subtracting the year of birth from the year when the race was held. Continent variable was defined from the nationality of the athletes. 2.3. Statistical Analysis The outcome was the running speed (km/h). Information for all races: number of observations, mean (SD) and minimum and maximum of speed (km/h) is provided in Table 1. For the main analyses, time-limited races of 8, 48, 72 h, 6 and 8 days were excluded due to insufficient data (< 100 observations). Descriptive statistics were presented as means (SD = standard deviations) by sex, age groups, continents and time groups. The age groups were 10–13, 14–15, 16–17 and 18 years. The continent groups, with reference to the nationality of the athletes, were: Africa, Asia, Central-South America, Europe, North-America, Oceania. When the number of observations of each continent group, within each race, was not greater than ten, continents were grouped together into other continents. To show a performance by a period of time, the calendar year of the race was grouped into time periods of 10 years. Age and calendar year were considered as continuous variables, in 1-year intervals, when defined as predictor variables for ultra-running speed. In fact, non-linear regression mixed models, with basis splines (BS), were performed to examine the time trend together with the effects of sex, age and continent on the speed time of each duration race. The mixed models were used to correct repeated measurements within runners (clusters) through the random effects of intercepts. The statistical models were specified as follows: Ultra-running speed (Y) ~ [Fixed effects (X) = BS(Year, df = 3) + BS(Age, df = 3) *sex + continent + [random effects of intercept=runners] where BS(Year, df=3) and BS(Age, df = 3) are three degrees of freedom (df) basis splines changing with calendar year and age, respectively; BS(Age, df = 3) * sex denote the age–sex interaction term. Different analyses were performed, one for each duration (6, 12, and 24 h). The interaction term age–sex was significant and considered only in the 24-h events. In the 6- and 12-h race analyses, a linear term on year, instead of a spline term, was considered. Results of the regression models were presented as estimates and standard errors. Statistical significance was defined as p < 0.05. All statistical analyses were carried out with R, R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL https://www.R-project.org/. The R packages ggplot2, lme4, and lmerTest were used, respectively, for data visualization and for the mixed models. Table 1. Ultra-endurance performance—speed km/h by duration. Mean (SD) and minimum (Min), maximum (Max) were reported. Duration Running Speed (km/h) N Mean (SD) Min Max 6 h 214 8.84 (1.06) 7.50 12.83 8 h 68 6.97 (0.92) 5.68 10.24 12 h 247 5.39 (1.36) 3.75 10.42 24 h 805 2.83 (1.00) 1.87 8.00 48 h 46 2.40 (1.06) 0.94 4.37 72 h 50 1.54 (0.72) 0.66 3.24 6 days 13 1.57 (0.77) 0.45 3.37 8 days 7 1.72 (0.35) 1.15 2.30 Int. J. Environ. Res. Public Health 2020, 17, 1628 4 of 12 3. Results Between 1990 and 2018, the total number of observations, over 6-, 12- and 24-h races was n = 214, n = 247, and n = 805 records, respectively. Instead, the number of individual finishers was, respectively, n = 166, n = 212 and n = 582. We observed that the percentage of children aged 10–13 years was relatively high. In fact, in 24-h races, the majority were 10–13 years, 376 (46.7%). Instead, in 6-h races, the majority of finishers, 81 (37.9%), were 18 years and in the 12-h race, the majority were 16-17 years, 83 (33.6%). In the 6- and 24-h races, 146 (68.2%) and 441 (54.8%), respectively, came from Europe but in the 12-h races, the majority of 130 (52.6%) came from North America. In time-limited races, however, no participation was recorded before 1990 and the vast majority participated within the last 8 years. The number of observations and the average performance by sex, age groups, continent and time groups are reported for distance races, respectively, in Table 2. In Figures 1–3, the participation (%) and average performances (km/h) by nationality for each time-limited race are reported. Table 3 describes the results of the statistical models, as described in the methods section. Boys were significantly faster than girls only in the 12-h race. There were no significant differences between North America, Europe and other continents. Table 2. Mean ultra-endurance running speed in km/h and (SD): duration of races (6, 12 and 24 h) by sex, age, country (continent) and calendar year groups. Africa, Asia, Central-South America, Oceania, due to small sample size, were combined into “Other” group. Duration 6 h, N = 214 12 h, N = 247 24 h, N = 805 Age Sex N Mean (SD) N Mean (SD) N Mean (SD) 10–13 F 5 8.23 (0.34) 24 4.94 (0.93) 136 2.45 (0.56) M 18 8.16 (0.51) 22 5.05 (0.98) 240 2.53 (0.77) 14–15 F 7 8.68 (0.66) 18 4.91 (1.10) 84 2.84 (0.89) M 32 8.47 (0.93) 36 5.01 (1.10) 101 2.93 (0.86) 16–17 F 17 8.52 (0.91) 23 5.21 (1.08) 55 2.83 (0.94) M 54 8.82 (0.91) 60 5.34 (1.32) 104 2.97 (0.93) 18 F 20 9.09 (0.78) 17 5.35 (0.80) 26 3.19 (1.01) M 61 9.33 (1.31) 47 6.42 (1.79) 86 3.85 (1.56) Continent Sex N Mean (SD) N Mean (SD) N Mean (SD) North America F 11 8.65 (0.96) 53 5.11 (1.02) 106 2.90 (0.94) M 50 8.71 (0.91) 77 5.51 (1.34) 246 3.03 (1.07) Europe F 35 8.77 (0.81) 25 5.07 (0.98) 162 2.50 (0.63) M 111 8.96 (1.20) 68 5.54 (1.70) 279 2.77 (1.08) Other F M 3 4 8.87 (0.75) 8.22 (0.53) 4 20 5.02 (0.79) 5.62 (1.43) 6 6 3.60 (0.95) 4.21 (0.71) Year Sex N Mean (SD) N Mean (SD) N Mean (SD) 1990–1999 F 1 8.77 1 5.00 2 3.10 (0.57) M 8 9.64 (1.26) 6 6.92 (2.10) 17 4.03 (1.79) 2000–2009 F 19 8.87 (0.77) 16 4.76 (0.97) 23 3.26 (1.05) M 52 8.99 (1.26) 43 5.35 (1.69) 69 3.50 (1.52) 2010–2018 F 29 8.67 (0.88) 65 5.18 (0.99) 248 2.61 (0.75) M 105 8.75 (1.00) 116 5.53 (1.36) 439 2.73 (0.82) Int. J. Environ. Res. Public Health 2020, 17, 1628 5 of 12 Figure 1. Participation (%) and average performance (km/h) by nationality in 6-h races. Int. J. Environ. Res. Public Health 2020, 17, 1628 6 of 12 Figure 2. Participation (%) and average performance (km/h) by nationality in 12-h races. Int. J. Environ. Res. Public Health 2020, 17, 1628 7 of 12 Figure 3. Participation (%) and average performance (km/h) by nationality in 24-h races. Int. J. Environ. Res. Public Health 2020, 17, 1628 8 of 12 Table 3. Regression analysis (mixed model) of ultra-endurance events (6, 12 and 24 h). Estimates and standard errors (SE) of fixed effects are reported. p-values ranges are marked with asterisks (see note). Smoothing terms, basis splines (BS), are denoted with BS(x) t, where x = year, age; t = 1,2,3. Predictor Time-Limited Races 6 h 12 h 24 h Age BS(Age)1 0.842 (0.940) 1.423 (1.173) 0.227 (0.382) BS(Age)2 −0.226 (0.538) −0.584 (0.628) 0.708 *(0.331) BS(Age)3 1.323 **(0.482) 1.708 **(0.590) 0.696 ***(0.203) Sex = M (ref=F) 0.236 (0.185) 0.377 *(0.189) −0.134 (0.167) Age:Sex interaction terms BS(Age)1:SexM 0.551 (0.475) BS(Age)2:SexM −0.402 (0.402) BS(Age)3:SexM 0.703 **(0.243) Year −0.024 (0.014) 0.003 (0.017) BS(Year)1 −1.349 (1.210) BS(Year)2 −3.417 ***(0.612) BS(Year)3 −3.215 ***(0.677) Continent (ref. North America) Europe 0.142 (0.173) −0.140 (0.196) −0.115 (0.069) Other −0.238 (0.328) 0.395 (0.296) Constant 56.062 *(27.523) −1.086 (33.910) 5.495 ***(0.684) Observations Runners 207 159 247 212 805 582 Notes: * p < 0.05; ** p < 0.01; *** p < 0.001. For time-limited races (Figure 4), a time effect was not significant in 6- and 12-h events but it was significant in 24-h races, where running speed decreased over time. In all time-limited races, running speed increased across age groups (Figure 5) and this trend was different between boys and girls. Figure 4. Running speed across years for time-limited ultra-endurance events for 6, 12 and 24 h. Fitted values=line, points=observed mean values. Int. J. Environ. Res. Public Health 2020, 17, 1628 9 of 12 Figure 5. Running speed across age groups for time-limited ultra-endurance events for 6, 12 and 24 h. Fitted values=line, points=observed mean values. 4. Discussion This is the first study that examined the participation and performances in time-limited ultra-endurance events and multi-day events in youth runners. The aim of the present study was to investigate the age-related participation and performance trends of children and adolescent ultra-endurance runners, younger than 19 years of age, in time-limited events. Our hypothesis was that participation numbers would increase over calendar years and time, more boys than girls would participate in these events, and that performance times from boys would be faster. The main findings were (i) an increase in the number of ultra-endurance participation over time and across races and sexes, (ii) performance differences between boys and girls, with boys being significantly faster than girls only in the 12-h race, (iii) differences between running speed across age groups and continents and (iv) variations in running speed over the years and different age–sex trend in 24-h events. 4.1. Participation Trends An exponential increase in participation numbers among youth ultra-runners was observed in the last 30 years. The most popular time-limited race competitions were 24, 12 and 6 h long; however, participation numbers were considerably smaller. The majority of finishers belonged to the older age groups (16–18 years of age) and were mostly male. Our findings confirm previously observed participation trends [16]. In time-limited races, most of the finishers in the 6- and 24-h races came from Europe but, for 12-h races, most originated from North America. A study investigating the sex difference in 24-h adult ultramarathoners showed that most of the starters originated from Europe, mainly France and Germany [6]. Only a small minority participated in multi-day events. 4.2. Differences in Age Regarding Duration of Races The percentage of runners aged 10-13 years was rather high in time-limited races. The existing literature for adult ultramarathoners investigated the age trends only for distance-limited races [2,21] but it made no comparison between the different kinds of ultra-endurance events. In the 6-h races, the majority of the finishers were 18 years-old, whereas in the 12-h races, the majority of the competitors were between 16–17 years of age. For adult ultramarathoners, it has been reported that the age of peak running performance increased with race duration with time-limited events ranging from 6 h to 10 days [4]. Obviously, there is a difference between youth and adult ultra-endurance runners, where Int. J. Environ. Res. Public Health 2020, 17, 1628 10 of 12 youth athletes seem to be older in longer time-limited races, in contrast to adult runners, where the opposite was found. 4.3. Analysis among Different Races and Sexes Boys were faster than girls in the 12-h races, but not in the 6- and 24-h races. This is hardly surprising, taking the physiological differences during and after maturation into account. For adult ultramarathoners, men were generally faster than women in time-limited ultramarathons (e.g., 24-h races) [6]. Women, however, have closed the gap with men in the last decade [6,17,22,23]. The mean ultra-endurance running speed was generally faster in boys than girls across all time limited ultra-endurance races; however, this was only significant for the 12-h races. Younger girls, under the age of 15 years, were faster than boys in the 6-h races. One possible explanation for the latter finding may be that girls mature physiologically earlier than boys and this may have given them an advantage at this race distance. Another explanation may be that there were far fewer girls participating at this particular race distance and that they may have been better prepared for this event. 4.4. Analysis in Running Speed Across Age Groups In all time-limited races, running speed increased across age groups and was different between boys and girls in 24-h races, with boys being faster than girls. In other terms, boys and girls improved their running performance with increasing age in time-limited races. However, they were still far away from the age of peak ultramarathon performance which is generally achieved at ages beyond 35 years [2,4,21] and increases with increasing race duration in time-limited races [4]. It is well-known that the age of peak performance in endurance sports increases with increasing length or duration of the endurance performance [24]. 4.5. Analysis in Running Speed over Years The last important finding was that the running speed showed differences in the trend across calendar years for the different ultra-endurance events. A time effect was not significant in 6- and 12-h events but it was significant in 24-h races, where running speed decreased over time. Taken together, these youth ultra-runners were not able to improve their running performance in recent years, although in some races in earlier years their running speed was higher than in recent years. Obviously, there is a general trend in long-distance races that running speed decreased in recent years, and this may be related to the general increase in participation numbers, but not necessarily an increase in faster or more elite runners, that would increase the overall performance times. 4.6. Limitations The analyzed data originate from one database (DUV-www.ultra-marathon.org/). We recognize that there are several other national databases; however, the DUV is the largest ultra-running database worldwide and has been used widely to address similar research questions in the adult population. Inaccuracies in reporting or missing datasets are possible in such a large database. For the analysis of particular national races, data from the race websites could be analyzed for future studies. Several races have been grouped together, without taking into account the specific ambient, environmental or terrain particularities, which can have an impact on the average performance time. To address this, we recommend analyzing specific races in future. 4.7. Practical Applications In last decade, a large increase in the number of finishers and annual races of ultra-marathons has been observed. Following this trend, the number of children and adolescents competing in these races has increased, too. Consequently, strength and conditioning trainers could face new challenges when working with children and adolescent ultramarathoners, since the existing knowledge is based mostly Int. J. Environ. Res. Public Health 2020, 17, 1628 11 of 12 in studies on adult athletes. The present study added valuable practical information in the existing literature with regards to trends in the participation and performance of children and adolescent ultramarathoners. For instance, these trends varied by country; thus, the findings would especially be of practical interest for strength and conditioning trainers working in countries with increased participation of these age groups in ultramarathons. Future studies may investigate the same trends in distance-limited ultra-marathons such as 100 km and 100 miles. 5. Conclusions Comparing time-limited races from 6 h to 8 days, it was concluded that ultramarathoners younger than 19 years of age participated mostly in 6-, 12- and 24-h races, and the majority of these athletes originated from Europe and North America. Only a minority participated in multi-day events. Overall, speed was faster in the older rather than in the younger athletes. Finally, the overall speed mostly decreased across calendar years as participation numbers grew. Author Contributions: Conceptualization, V.S. and B.K.; methodology, S.D.G.; software, S.D.G.; formal analysis, S.D.G.; resources, E.V.; data curation, E.V.; writing—original draft preparation, V.S., S.D.G., E.V., T.R., P.T.N. and B.K.; writing—review and editing, V.S., S.D.G., E.V., T.R., P.T.N. and B.K. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflict of interest. References 1. Zaryski, C.; Smith, D.J. Training principles and issues for ultra-endurance athletes. Curr. Sports Med. Rep. 2005, 4, 165–170. [CrossRef] [PubMed] 2. Nikolaidis, P.T.; Knechtle, B. Performance in 100-km ultra-marathoners—At which age it reaches its peak? J. Strength Cond. Res. 2018. In print. [CrossRef] [PubMed] 3. Scheer, V. Participation trends of ultra endurance events. Sports Med. Arthrosc. Rev. 2019, 27, 3–7. [CrossRef] [PubMed] 4. Knechtle, B.; Valeri, F.; Zingg, M.A.; Rosemann, T.; Rüst, C.A. What is the age for the fastest ultra-marathon performance in time-limited races from 6 h to 10 days? Age (Dordr) 2014, 36, 9715. [CrossRef] [PubMed] 5. Zingg, M.; Rüst, C.A.; Lepers, R.; Rosemann, T.; Knechtle, B. Master runners dominate 24-h ultramarathons worldwide-a retrospective data analysis from 1998 to 2011. Extrem. Physiol. Med. 2013, 2, 21. [CrossRef] [PubMed] 6. Peter, L.; Rüst, C.A.; Knechtle, B.; Rosemann, T.; Lepers, R. Sex differences in 24-h ultra-marathon performance—A retrospective data analysis from 1977 to 2012. Clinical 2014, 69, 38–46. [CrossRef] 7. Shoak, M.A.; Knechtle, B.; Rust, C.A.; Lepers, R.; Rosemann, T. European dominance in multistage ultramarathons: An analysis of finisher rate and performance trends from 1992 to 2010. Open Access J. Sports Med. 2013, 4, 9–18. 8. Scheer, B.V.; Murray, A. Al Andalus Ultra Trail: An observation of medical interventions during a 219-km, 5-day ultramarathon stage race. Clin. J. Sport Med. 2011, 21, 444–446. [CrossRef] 9. Knoth, C.; Knechtle, B.; Rust, C.A.; Rosemann, T.; Lepers, R. Participation and performance trends in multistage ultramarathons-the ’Marathon des Sables’ 2003-2012. Extr. Physiol. Med. 2012, 1, 13. [CrossRef] 10. Costa, R.J.; Teixeira, A.; Rama, L.; Swancott, A.J.; Hardy, L.D.; Lee, B.; Camoes-Costa, V.; Gill, S.; Waterman, J.P.; Freeth, E.C.; et al. Water and sodium intake habits and status of ultra-endurance runners during a multi-stage ultra-marathon conducted in a hot ambient environment: An observational field based study. Nutr. J. 2013, 12, 13. [CrossRef] 11. Alcock, R.; McCubbin, A.; Camoes-Costa, V.; Costa, R.J.S. Case Study: Providing Nutritional Support to an Ultraendurance Runner in Preparation for a Self-Sufficient Multistage Ultramarathon: Rationed Versus Full Energy Provisions. Wilderness Environ. Med. 2018, 29, 508–520. [CrossRef] [PubMed] 12. da Fonseca-Engelhardt, K.; Knechtle, B.; Rüst, C.A.; Knechtle, P.; Lepers, R.; Rosemann, T. Participation and performance trends in ultra-endurance running races under extreme conditions—’Spartathlon’ versus Badwater. Extr. Physiol. Med. 2013, 2. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2020, 17, 1628 12 of 12 13. Knechtle, B.; Rüst, C.A.; Rosemann, T.; Lepers, R. Age-related changes in 100-km ultra-marathon running performance. Age 2012, 34, 1033–1045. [CrossRef] 14. Eichenberger, E.; Knechtle, B.; Rust, C.A.; Rosemann, T.; Lepers, R. Age and sex interactions in mountain ultramarathon running—the Swiss Alpine Marathon. Open Access. J. Sports Med. 2012, 3, 73–80. 15. Knechtle, B.; Nikolaidis, P.T. Physiology and pathophysiology in ultra-marathon running. Front. Physiol. 2018, 9. [CrossRef] 16. Scheer, V.; Hoffman, M.D. Should children be running ultramarathons? Curr. Sports Med. Rep. 2018, 17, 282–283. [CrossRef] [PubMed] 17. Waldvogel, K.J.; Nikolaidis, P.T.; Di Gangi, S.; Rosemann, T.; Knechtle, B. Women reduce the performance difference to men with increasing age in ultra-marathon running. Int. J. Environ. Res. Pub. Health 2019, 16. [CrossRef] 18. Cejka, N.; Rust, C.A.; Lepers, R.; Onywera, V.; Rosemann, T.; Knechtle, B. Participation and performance trends in 100-km ultra-marathons worldwide. J. Sports Sci. 2014, 32, 354–366. [CrossRef] 19. Knechtle, B.; Nikolaidis, P.T.; Valeri, F. Russians are the fastest 100-km ultra-marathoners in the world. PLoS ONE 2018, 13, e0199701. [CrossRef] 20. Scheer, V.; Hoffman, M.D. Too much too early? An analysis of worldwide childhood ultramarathon participation and attrition in adulthood. J. Sports Med. Phys. Fit. 2019, 59, 1363–1368. [CrossRef] 21. Nikolaidis, P.T.; Knechtle, B. Age of peak performance in 50-km ultramarathoners—is it older than in marathoners? Open Access J. Sports Med. 2018, 9, 37–45. [CrossRef] [PubMed] 22. Zingg, M.A.; Karner-Rezek, K.; Rosemann, T.; Knechtle, B.; Lepers, R.; Rüst, C.A. Will women outrun men in ultra-marathon road races from 50 km to 1,000 km? Springerplus 2014, 3, 97. [CrossRef] [PubMed] 23. Zingg, M.A.; Knechtle, B.; Rosemann, T.; Rust, C.A. Performance differences between sexes in 50-mile to 3,100-mile ultramarathons. Open Access J. Sports Med. 2015, 6, 7–21. [PubMed] 24. Allen, S.V.; Hopkins, W.G. Age of Peak Competitive Performance of Elite Athletes: A Systematic Review. Sports Med. 2015, 45, 1431–1441. [CrossRef] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Participation and Performance Analysis in Children and Adolescents Competing in Time-Limited Ultra-Endurance Running Events.

03-03-2020

Scheer, Volker,Di Gangi, Stefania,Villiger, Elias,Rosemann, Thomas,Nikolaidis, Pantelis T,Knechtle, Beat

eng

PMC9821460

Experiment 1 Experiment 2 #Participant Running speed (km/h) #Participant Running speed (km/h) 1 13.3 1 12.0 2 12.9 2 12.9 3 12.9 3 12.0 4 13.5 4 13.0 5 12.7 11 13.3 6 12.6 12 12.0 7 11.8 13 10.5 8 12.5 14 12.5 9 13.5 15 13.0 10 11.8 16 12.9 17 13.0 18 12.0 19 12.0 20 11.0 21 12.0 Mean 12.75 Mean 12.27 Note. A total of four participants (#1-4) participated in Experiments 1 and 2.

Auditory interaction between runners: Does footstep sound affect step frequency of neighboring runners?

01-06-2023

Furukawa, Hiroaki,Kudo, Kazutoshi,Kubo, Kota,Ding, Jingwei,Saito, Atsushi

eng

PMC10747732

Citation: Rasmussen, J.; Skejø, S.; Waagepetersen, R.P. Predicting Tissue Loads in Running from Inertial Measurement Units. Sensors 2023, 23, 9836. https://doi.org/10.3390/ s23249836 Academic Editor: Georg Fischer Received: 19 October 2023 Revised: 28 November 2023 Accepted: 13 December 2023 Published: 15 December 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). sensors Article Predicting Tissue Loads in Running from Inertial Measurement Units John Rasmussen 1,* , Sebastian Skejø 2,3 and Rasmus Plenge Waagepetersen 4 1 Department of Materials and Production, Aalborg University, Fibigerstraede 16, 9220 Aalborg East, Denmark 2 Department of Public Health, Aarhus University, Bartholins Allé 2, 8000 Aarhus, Denmark; [email protected] 3 Research Unit for General Practice, Aarhus University, Bartholins Allé 2, 8000 Aarhus, Denmark 4 Department of Mathematical Sciences, Aalborg University, Skjernvej 4A, 9220 Aalborg East, Denmark; [email protected] * Correspondence: [email protected] Abstract: Background: Runners have high incidence of repetitive load injuries, and habitual runners often use smartwatches with embedded IMU sensors to track their performance and training. If accelerometer information from such IMUs can provide information about individual tissue loads, then running watches may be used to prevent injuries. Methods: We investigate a combined physics- based simulation and data-based method. A total of 285 running trials from 76 real runners are subjected to physics-based simulation to recover forces in the Achilles tendon and patella ligament, and the collected data are used to train and test a data-based model using elastic net and gradient boosting methods. Results: Correlations of up to 0.95 and 0.71 for the patella ligament and Achilles tendon forces, respectively, are obtained, but no single best predictive algorithm can be identified. Conclusions: Prediction of tissues loads based on body-mounted IMUs appears promising but requires further investigation before deployment as a general option for users of running watches to reduce running-related injuries. Keywords: running; injuries; Achilles tendon; patella ligament; IMU; data science; biomechanics; public health 1. Introduction Running is a popular physical activity not only for recreational purposes, but also among elite athletes. For instance, a survey covering the years 2015 through 2022 in England [1] showed that about 10% of the population, corresponding to roughly 6 million people, regularly engage in running. Running thus engages enough participants to have an impact on the overall activity level of the population, and the health benefits of active lifestyles are well documented [2,3]. Unfortunately, running is also associated with a high risk of sustaining a running- related injury, and injury incidence between 8 and 18 injuries per 1000 h of running has been reported [4]. Especially the knee and the ankle are susceptible to injuries with Achilles tendinopathy being the most incident injury and patellofemoral pain syndrome being the most prevalent injury [5]. Apart from the immediate discomfort and long recovery time associated with running-related injuries [6,7], injuries are also the most common reason for stopping running [8,9]. Therefore, it is imperative to mitigate the risk of running-related injuries, which requires a profound understanding of why these injuries occur [10]. Running-related injuries are commonly assumed to be caused by the repetitive loading of the affected tissues, resulting in inflammation and tissue failure over time [11]. Loads on muscles, tendons, and bones are, however, difficult—bordering on impossible—to measure in vivo, even in advanced laboratories [12]. Advances in model-based simulation of musculoskeletal forces over the past two decades have provided algorithms [13,14] and models [15] that verifiably predict internal musculoskeletal forces, but these methods rely Sensors 2023, 23, 9836. https://doi.org/10.3390/s23249836 https://www.mdpi.com/journal/sensors Sensors 2023, 23, 9836 2 of 12 on complete kinematic input, typically obtained from optical motion capture systems [16] or motion capture suits [17]. Optical systems are impractical in field settings, and both types are too comprehensive for research with large cohorts as well as to inform the individual runner. Therefore, there is a need for light-weight technologies that can accurately estimate tissue loads and injury risk factors in field settings. Such technology could be running watches, which many recreational and elite runners rely on to track their training efforts and performance [18,19], and whose popularity has also been attributed to independence from organized coaching [20]. Running watches typically record biometrics, such as pulse, and kinematic data, such as accelerations and GPS positions, from which running speed, step frequency and step length can be derived. Recent developments in data science have inspired data-based models linking inertial measurement unit (IMU) kinematics, complete kinematics [21] and musculoskeletal kinet- ics [22–28]. Such models can be trained on musculoskeletal simulations based on optical motion capture data and have shown promising results for modeling of a variety of tasks, but on relatively small and uniform cohorts of test subjects. The data space of all possible movements performed by all possible people is very large, and the connection between kinematics and kinetics expressed by the laws of dynam- ics is highly nonlinear, so it is unlikely that machine learning will outcompete physics-based models in general. However, running is a small subset of human movements, and given runners’ need for fast and lightweight feedback from IMUs, it is worth investigating whether reliable kinetic estimations, i.e., tissue loads, are possible from the combination of IMU data, a database of verified running kinematics for a medium-size heterogeneous population, and simple anthropometrics for the individual runner. When choosing predictors, there are a few considerations to keep in mind. First and foremost, the predictors should be feasible and inexpensive to collect in field-based settings and, therefore, require as little additional equipment as possible. Secondly, to be relevant for performance and injury mechanisms, the predictors should carry as much information as possible about running biomechanics. Thirdly, the total of number of predictors should be curbed to avoid overfitting. IMUs, such as those embedded in running watches and smartphones, are inexpensive and feasible to use in field-based settings. They provide kinematic data only for their attachment points, but they typically sample at frequencies of 400–500 Hz [29] and therefore accumulate a large amount of data over a short time. In the interest of data reduction, some studies extract a few features from the raw signal’s time domain, such as the maximum angle of a segment or peak accelerations, which capture some, but not all, aspects of running kinematics. An alternative approach is to transform the entire signal to the frequency domain using discrete Fourier transformation. Given the periodic nature of the data, this provides an accurate, yet compressed, description of the entire running kinematics [30]. We investigate the prediction of Achilles tendon and patella ligament loads based on data from IMUs positioned on easily accessible anatomical locations, i.e., wrists, ankles, and the sternum, and we compare IMU positions and sets of predictors to optimize the results. 2. Materials and Methods Figure 1 illustrates the data flow and computational investigations. 2.1. Experimental Data The experimental and biomechanical simulation procedures were described previously in detail [30], and are briefly summarized here for completeness: a nine-camera Qualisys Miqus system (Qualisys AB, Gothenburg, Sweden) was used to collect full-body optical marker data for treadmill running for 78 runners (30 female, 48 male) in 285 trials (180 male trials and 105 female trials) with speeds ranging between 6 and 20 km/h at 300 frames per second and a resolution of 2 megapixels. The subjects were asymptomatic and ranged in skill level from beginner to elite. The optical marker data were transferred to a physics- based simulation in the AnyBody Modeling System version 7.2 (AnyBody Technology Sensors 2023, 23, 9836 3 of 12 A/S, Aalborg, Denmark) [13], which converted the marker data to anthropometrical data, i.e., individual segment dimensions, and anatomical joint angle time series [31]. Using the Twente Lower Extremity Model version 2.0 [32], this process also simulates internal biomechanical kinetics. For the purposes of this paper, simulated maximum values of the patella ligament force and the Achilles tendon force over the running cycle were stored for each recorded trial and normalized by body mass. Sensors 2023, 23, x FOR PEER REVIEW 3 of 13 Figure 1. Data flow and computational methods. Figures in parentheses designate numbers of var- iables. The collected data set contains a total of 1229 variables of which 43 are virtual accelerometer data and 132 are anthropometric parameters. The prediction methods are, respectively, Elastic Net with two tuning parameters and XGB with three tuning parameters. The experimental data are ran- domly split into training and validation set. 2.1. Experimental Data The experimental and biomechanical simulation procedures were described previ- ously in detail [30], and are briefly summarized here for completeness: a nine-camera Qualisys Miqus system (Qualisys AB, Gothenburg, Sweden) was used to collect full-body optical marker data for treadmill running for 78 runners (30 female, 48 male) in 285 trials (180 male trials and 105 female trials) with speeds ranging between 6 and 20 km/h at 300 frames per second and a resolution of 2 megapixels. The subjects were asymptomatic and ranged in skill level from beginner to elite. The optical marker data were transferred to a physics-based simulation in the AnyBody Modeling System version 7.2 (AnyBody Tech- nology A/S, Aalborg, Denmark) [13], which converted the marker data to anthropomet- rical data, i.e., individual segment dimensions, and anatomical joint angle time series [31]. Using the Twente Lower Extremity Model version 2.0 [32], this process also simulates in- ternal biomechanical kinetics. For the purposes of this paper, simulated maximum values of the patella ligament force and the Achilles tendon force over the running cycle were stored for each recorded trial and normalized by body mass. The time series of anatomical joint angles were segmented into strides, which were transferred to the frequency domain by Fast Fourier Transform (FFT), retaining five sine and five cosine terms and the DC (constant) component, i.e., 11 coefficients to describe the motion of each of the 88 independent anatomical degrees-of-freedom. This procedure was described in detail previously [30]. Using the model, we emulated five virtual, three-axis accelerometers located on the two wrists, the two ankles and the sternum. The accelerometers were constructed in the model as body segment-fixed local reference frames, whose acceleration vectors could be extracted. The placements and local coordinate systems are illustrated on Figure 2. Gyro- scopic information was not included The virtual accelerations were combined after the Methods 10% 90% PREDICTION Anthropometrics (43) Methods Elastic net XGB Parameters: α and λ Parameters: m, d, η Random split Validation data Training data Experimental data (1229) Accelerations (132) Max. Achilles tendon force (1) Max. patella ligament force (1) Max. Achilles tendon force (1) Max. patella ligament force (1) Max. Achilles tendon force Max. patella ligament force Validation Figure 1. Data flow and computational methods. Figures in parentheses designate numbers of variables. The collected data set contains a total of 1229 variables of which 43 are virtual accelerometer data and 132 are anthropometric parameters. The prediction methods are, respectively, Elastic Net with two tuning parameters and XGB with three tuning parameters. The experimental data are randomly split into training and validation set. The time series of anatomical joint angles were segmented into strides, which were transferred to the frequency domain by Fast Fourier Transform (FFT), retaining five sine and five cosine terms and the DC (constant) component, i.e., 11 coefficients to describe the motion of each of the 88 independent anatomical degrees-of-freedom. This procedure was described in detail previously [30]. Using the model, we emulated five virtual, three-axis accelerometers located on the two wrists, the two ankles and the sternum. The accelerometers were constructed in the model as body segment-fixed local reference frames, whose acceleration vectors could be extracted. The placements and local coordinate systems are illustrated on Figure 2. Gyroscopic information was not included. The virtual accelerations were combined, after the appropriate coordinate transform, with the gravity component to mimic the accel- erations including gravity that would have been measured by body-fixed IMUs. Time series for the resulting virtual accelerations in segment-fixed x, y, and z coordinates along with the magnitude of the acceleration vectors were also subjected to FFT transformation, retaining 11 Fourier coefficients for each direction and the magnitude, leading to a total of 44 coefficients for each accelerometer. Sensors 2023, 23, 9836 4 of 12 including gravity that would have been measured by body-fixed IMUs. Time series for the resulting virtual accelerations in segment-fixed x, y, and z coordinates along with the magnitude of the acceleration vectors were also subjected to FFT transformation, retaining 11 Fourier coefficients for each direction and the magnitude, leading to a total of 44 coef- ficients for each accelerometer. Figure 2. Placement of virtual IMUs on the left wrist, sternum and left ankle, and their respective local coordinate systems. The IMU on the right wrist is placed similarly to the IMU on the left wrist. The purple dots are floor contact points, and the red dots are spinal joint centers. This study used a previously anonymized version of the data resulting from the data processing described above. The process results in a matrix with 285 rows corresponding to the recorded trials. The number of columns is 1229, of which 968 contain Fourier coef- ficients for the model’s kinematic degrees-of-freedom, and 220 = 5 × 44 columns contain Fourier coefficients for the virtual accelerations of the five selected placements. The max- imal forces over the cycle for the right and left patella ligaments and Achilles tendons were extracted from the kinetic analysis, yielding four additional columns. Furthermore, 35 col- umns contain anthropometric measurements such as segment lengths, body weight, stat- ure, and gender. Finally, the table contains running speed and angular step frequency. Initial investigations of virtual accelerometer data revealed a high degree of sym- metry between the left and right ankles. It was therefore decided to discard the right ankle and include only accelerations of the two wrists, the left ankle, and the sternum, in total 176 columns of Fourier coefficients of acceleration data. It was furthermore decided that combinations of more than two IMUs would be impractical for most runners and, conse- quently, the following combinations were selected for further investigation: 1. left wrist 2. left wrist and right wrist 3. left wrist and sternum 4. left wrist and left ankle 5. sternum and left ankle 6. left ankle 7. sternum. Single-IMU options 1, 6 and 7 lead to inclusion of 1 × 4 × 11 = 44 columns of Fourier coefficients, and double-IMU options 2–5 each lead to inclusion of 88 columns. In the in- terest of simplicity of the procedure for the runner and to minimize the risk of overfitting, only the step frequency, sex, age, thigh length, shank length, foot length, body mass, Figure 2. Placement of virtual IMUs on the left wrist, sternum and left ankle, and their respective local coordinate systems. The IMU on the right wrist is placed similarly to the IMU on the left wrist. The purple dots are floor contact points, and the red dots are spinal joint centers. This study used a previously anonymized version of the data resulting from the data processing described above. The process results in a matrix with 285 rows corresponding to the recorded trials. The number of columns is 1229, of which 968 contain Fourier coefficients for the model’s kinematic degrees-of-freedom, and 220 = 5 × 44 columns contain Fourier coefficients for the virtual accelerations of the five selected placements. The maximal forces over the cycle for the right and left patella ligaments and Achilles tendons were extracted from the kinetic analysis, yielding four additional columns. Furthermore, 35 columns contain anthropometric measurements such as segment lengths, body weight, stature, and gender. Finally, the table contains running speed and angular step frequency. Initial investigations of virtual accelerometer data revealed a high degree of symmetry between the left and right ankles. It was therefore decided to discard the right ankle and include only accelerations of the two wrists, the left ankle, and the sternum, in total 176 columns of Fourier coefficients of acceleration data. It was furthermore decided that combinations of more than two IMUs would be impractical for most runners and, consequently, the following combinations were selected for further investigation: 1. left wrist 2. left wrist and right wrist 3. left wrist and sternum 4. left wrist and left ankle 5. sternum and left ankle 6. left ankle 7. sternum. Single-IMU options 1, 6 and 7 lead to inclusion of 1 × 4 × 11 = 44 columns of Fourier coefficients, and double-IMU options 2–5 each lead to inclusion of 88 columns. In the interest of simplicity of the procedure for the runner and to minimize the risk of overfitting, only the step frequency, sex, age, thigh length, shank length, foot length, body mass, stature, BMI and running speed were included as additional variables, leading to 54 or 98 predictor variables for single and double-IMU configurations, respectively. Sensors 2023, 23, 9836 5 of 12 2.2. Prediction Algorithms We are in a setting with a relatively low number of observations (285 running trials) compared to the number of predictor variables (54 or 98, depending on whether a single- IMU or double-IMU configuration is used). A standard multivariate linear regression model is therefore not suitable since it will be prone to overfitting. Thus, we consider regularized versions of multivariate linear regression using the elastic net approach [33], which encompasses ridge regression, LASSO (Least Absolute Shrinkage and Selection Operator), and combinations of ridge regression and LASSO. The elastic net further enables variable selection that discards weak predictors. This has the potential to reduce the number of variables needed for prediction, which can make the method more feasible for future users. A regularized linear regression is ”easy” to interpret, but clearly has shortcomings in case of nonlinear relationships or interaction effects (unless the latter are explicitly included in the model). Therefore, we also consider a flexible tree-based machine learning method based on gradient boosting using specifically the computationally efficient XGB algorithm [34]. In this case, the prediction is obtained from a sequence of relatively shallow trees. These trees are fitted sequentially, where each new tree is fitted to residuals arising from prediction by the current sequence of trees. The splits employed when constructing prediction trees are very useful for handling nonlinearities and interactions. 2.2.1. Tuning Parameters The elastic net approach relies on two tuning parameters, α and λ. The first of them, α, ranges between 0 and 1, where 0 gives ridge regression, 1 gives LASSO, and intermediate values provide combinations of ridge and LASSO. For α, we consider the values 0, 0.5 and 1. The other parameter, λ, determines the strength of regularization and is determined by minimizing a cross validation score. Following the implementation in the glmnet package [35], the data is partitioned into 10 subsets. For each subset, the model is trained on the remaining data and is then used to predict the observations in the subset. The resulting mean square prediction errors over each subset are averaged to get the cross-validation score. For XGB, there is a wide range of tuning parameters. Here, we restrict attention to the important number of sequentially fitted trees m, tree depth d and learning rate η and leave the remaining tuning parameters at their default values. The number of trees m is chosen by cross-validation as implemented in the XGB package (with 5 subsets, referring to the explanation above of cross-validation). Following recommendations in the machine learning literature, we further consider rather shallow trees, d = 3, 6 and η = 0.1, 0.3. There is a trade-off between the latter two parameters, so that deeper trees should in general be combined with a lower learning rate and vice versa. 2.2.2. Model Evaluation and Selection The models with varying values of tuning parameters are trained on a random subset of the data containing 90% of the observations (the training set) and the prediction perfor- mance is next evaluated on the remaining 10% of the observations (the test set). To avoid sensitivity to the random splitting into training and test data sets, we consider 2500 in- dependent random splits (with replacement), carry out the training and test procedure on each resulting training/test data set, and average test results over the 2500 splits. We measure prediction performance in terms of average correlation and normalized root mean square error (nRMSE). For each test set, the correlation is between the physics-based values of the dependent variables and the predicted values obtained from the model trained on the training data set. The root mean square error (RMSE) is the square root of the average of squared differences between physics-based values in the test data set and corresponding predicted values. The RMSE is normalized by dividing by the average of the dependent variable over the full data set. In a few cases, the elastic net method produced a constant Sensors 2023, 23, 9836 6 of 12 predictor. Then, the correlation is not well-defined due to division by zero. We hence ultimately use nRMSE for identifying the best models. 3. Results 3.1. Comparison of Prediction Methods and Configurations of Accelerometers Figures 3 and 4 show performance measures (correlation and normalized RMSE) when patella ligament forces and Achilles tendon forces, respectively, are predicted us- ing elastic net and XGB with different settings of tuning parameters and configurations of accelerometers. average of squared differences between physics-based values in the test data set and cor- responding predicted values. The RMSE is normalized by dividing by the average of the dependent variable over the full data set. In a few cases, the elastic net method produced a constant predictor. Then, the correlation is not well-defined due to division by zero. We hence ultimately use nRMSE for identifying the best models. 3. Results 3.1. Comparison of Prediction Methods and Configurations of Accelerometers Figures 3 and 4 show performance measures (correlation and normalized RMSE) when patella ligament forces and Achilles tendon forces , respectively, are predicted using elastic net and XGB with different settings of tuning parameters and configurations of accelerometers. Figure 3. Performance measures for prediction of patella ligament forces by elastic net and XGB, respectively. The left frame shows the normalized root mean square error, and the right frame shows the correlation. Conditions: lw = left wrist, rw = right wrist, ste = sternum, ank = ankle. Figure 3. Performance measures for prediction of patella ligament forces by elastic net and XGB, respectively. The left frame shows the normalized root mean square error, and the right frame shows the correlation. Conditions: lw = left wrist, rw = right wrist, ste = sternum, ank = ankle. Sensors 2023, 23, x FOR PEER REVIEW 7 of 13 Figure 4. Performance measures for prediction of Achilles tendon forces by elastic net and XGB, respectively. The left frame shows the normalized root mean square error, and the right frame shows the correlation. Conditions: lw = left wrist, rw = right wrist, ste = sternum, ank = ankle. Considering patella ligament force, best prediction results were obtained when at least one measurement unit was placed at the ankle. In that case, elastic net is superior to XGB and correlations up to 0.96 are achieved. For the Achilles tendon force, XGB generally outperforms elastic net, and the best results were obtained with at least one measurement Figure 4. Performance measures for prediction of Achilles tendon forces by elastic net and XGB, respectively. The left frame shows the normalized root mean square error, and the right frame shows the correlation. Conditions: lw = left wrist, rw = right wrist, ste = sternum, ank = ankle. Sensors 2023, 23, 9836 7 of 12 Considering patella ligament force, best prediction results were obtained when at least one measurement unit was placed at the ankle. In that case, elastic net is superior to XGB and correlations up to 0.96 are achieved. For the Achilles tendon force, XGB generally outperforms elastic net, and the best results were obtained with at least one measurement unit placed at the sternum. For the Achilles tendon force, correlations up to 0.72 were achieved. Generally, both for the patella ligament force and the Achilles tendon force, and over the seven configurations of accelerometers, alpha equal to 0.5 or 1 (elastic net) and shallow trees of depth 3 (XGB) gave the best results in terms of nRMSE. For the patella ligament force and Achilles tendon force alike, the best results were obtained when just one measurement unit was used (at the ankle for the patella ligament force and at the sternum for Achilles tendon force). 3.2. Detailed Inspection of Best Prediction Methods and Measurement Unit Configurations To closer inspect the prediction methodology, we considered the best configurations of prediction algorithms and accelerometer configurations, i.e., elastic net with one mea- surement unit at the ankle and α = 1 for the patella ligament force, and XGB with one unit at the sternum, d = 3 and η = 0.3 for the Achilles tendon force. Figures 5 and 6 show scatter plots (so-called calibration plots) of physics-based values and predictions from the test data sets. To avoid too dense scatterplots, we only included 250 randomly sampled points from the test data sets. The solid and dashed lines are the identity and the least squares lines, respectively. The normalized RMSE and correlation are 0.12 and 0.95, respectively for the patella ligament scatter plot and 0.18 and 0.71 for the Achilles tendon scatter plot. A moderate bias in the predictions for the Achilles tendon force is revealed by the discrepancy between the identity and least squares lines, whereas bias is essentially absent for patella ligament force predictions. Sensors 2023, 23, x FOR PEER REVIEW 8 of 13 Figure 5. Correlation between predicted and physics-based simulated patella ligament forces. The legend shows intercepts and slopes of identity line (solid) and least squares line (dashed). Figure 6. Correlation between predicted and physics-based simulated Achilles tendon forces. The legend shows intercepts and slopes of identity line (solid) and least squares line (dashed). 3.3. Variable Importance The importance of the various predictor values was assessed for the patella ligament force with elastic net and one measurement unit at the ankle. With α = 1, elastic net may Figure 5. Correlation between predicted and physics-based simulated patella ligament forces. The legend shows intercepts and slopes of identity line (solid) and least squares line (dashed). Sensors 2023, 23, x FOR PEER REVIEW 8 of 13 Figure 5. Correlation between predicted and physics-based simulated patella ligament forces. The legend shows intercepts and slopes of identity line (solid) and least squares line (dashed). Figure 6. Correlation between predicted and physics-based simulated Achilles tendon forces. The legend shows intercepts and slopes of identity line (solid) and least squares line (dashed). 3.3. Variable Importance The importance of the various predictor values was assessed for the patella ligament force with elastic net and one measurement unit at the ankle. With α = 1, elastic net may estimate some predictor coefficients to be exactly zero, so that the corresponding predictor Figure 6. Correlation between predicted and physics-based simulated Achilles tendon forces. The legend shows intercepts and slopes of identity line (solid) and least squares line (dashed). Sensors 2023, 23, 9836 8 of 12 3.3. Variable Importance The importance of the various predictor values was assessed for the patella ligament force with elastic net and one measurement unit at the ankle. With α = 1, elastic net may estimate some predictor coefficients to be exactly zero, so that the corresponding predictor has no effect on the prediction. Thirteen variables were included in less than 10% of the test/training data sets. These are age, body mass, stature and ten acceleration measurement coefficients. Nineteen variables were included in at least 90% of the cases. These are the angular frequency, gender, shank length and 16 acceleration coefficients. The predictors can also be ranked according to their estimated effect sizes. The 10 variables with highest average rank over test/training data sets were (in rank order): 1. the average acceleration magnitude of the left ankle 2. the second sine coefficient of the acceleration magnitude of the left ankle 3. the first cosine coefficient of the acceleration magnitude of the left ankle 4. shank length 5. the first sine coefficient of the acceleration magnitude of the left ankle 6. the third cosine coefficient of the anterior/posterior acceleration of the left ankle 7. the first sine coefficient of the anterior/posterior acceleration of the left ankle 8. the third sine coefficient of the acceleration magnitude of the left ankle 9. the fourth cosine coefficient of the acceleration magnitude of the left ankle 10. the second sine coefficient of the anterior/posterior acceleration of the left ankle The angular frequency of the Fourier series and gender have ranks 15 and 22, respec- tively. Overall, the data do not support reducing the number of predictors for the patella ligament force significantly. In case of the XGB predictions for the Achilles tendon force using sternum acceleration measurements, the importance of variables may be assessed by their so-called gain, which measures the contribution of a variable to the prediction. We assessed the average gains and average ranks of variables according to their gains over the 2500 test/training data sets. The ten variables with the highest average gains are: 1. The fourth sine coefficient of the vertical acceleration of the sternum; 2. The third sine coefficient of the lateral acceleration of the sternum; 3. BMI; 4. The third cosine coefficient of the lateral acceleration of the sternum; 5. The second sine coefficient of the anterior/posterior sternum acceleration; 6. Shank length; 7. The second sine coefficient of the magnitude of the sternum acceleration; 8. Foot length; 9. The first cosine coefficient of the anterior/posterior sternum acceleration; 10. The second sine coefficient of the vertical acceleration of the sternum. It is worth noticing that the gains taper off rapidly, with variables from rank 8 and upwards sharing similar, small gains with variables outside the list. The variables speed and gender have lowest and third lowest average ranks, respectively. The remaining additional (non-acceleration) variables are among the 50% variables with the highest rank; except perhaps for speed and gender, there does not seem to be an obvious opportunity to eliminate additional variables. 4. Discussion Running can be performed at a wide range of speeds, and runners exhibit different styles depending on their anthropometry and other physiological preconditions. The participants in the present study are all able-bodied runners, but the data represent speeds between 6 and 20 km/h and skill levels between beginner and elite, and the resulting biomechanical loads vary considerably. Nonetheless, despite the complexities of the data, we obtain satisfactory correlations and prediction errors, and the presented method is a promising approach to predicting running biomechanics in field settings. Sensors 2023, 23, 9836 9 of 12 The present study differs from most previously reported results by targeting so-called structure-specific loads, i.e., the force on the Achilles tendon and patella ligament. To our knowledge, only one other study has attempted to predict these structure-specific loads from wearable devices [36]. In this study, Brund et al. found mean absolute percentage errors ranging from 13 to 30%, which appear similar to the normalized root mean squares error of 15–20% we report for the best performing models. However, the predictions by Brund et al. show markedly worse calibration upon visual inspection, which may be a result of the much lower number of predictors (five) and less sophisticated prediction models (simple multiple linear regression) compared to the present study. A couple of studies have predicted tibial bone forces from wearable devices and found accurate predictions with mean absolute percentage errors ranging from 2.6 to 17.9% [28,37]. However, these studies are dependent on measurements from pressure-sensing insoles, which are not common equipment for runners currently. Most other prediction studies using wearable devices have not predicted structure-specific loads but rather net joint moments, ground reaction forces (and derivatives thereof), or kinematics such as joint angles or stride length [38]. While such measures might be interesting in other contexts, an injury prevention context begs for structure-specific loads as these are closest to the actual injury mechanism [39]. In terms of correlations, the patella ligament loads are clearly better predicted than Achilles tendon loads for all IMU configurations. Our predictions are further inferior to those reported by Long et al. [40], who, however, considered a much smaller and more homogeneous cohort of four male basketball players. The Patella ligament force is strongly related to quadriceps activity, which again is related to the exerted knee extension moment. The latter is given by the product of the vertical acceleration of the masses above the knee and the moment arm, which is given by the knee flexion. IMU data reflecting these properties would convey the necessary information to accurately assess patella ligament loads, as shown by the results. Prediction of the Achilles tendon force, on the other hand, appears to be more challeng- ing. The explanation might be found in the binary nature of the foot contact mechanism, where small motion differences determine contact or non-contact of specific parts of the foot. Indeed, forefoot versus heel strike running styles cause different ground reaction force patterns and different load patterns of the Achilles tendon, even though the kinematic differences in terms of heel position can be quite small. Possibly for this reason, correlations for the Achilles tendon force peak around 0.70. It is possible that inclusion of a categorical variable for heel versus forefoot strike would improve the predictions. The normalized root mean square errors for both output variables, i.e., patella ligament force and Achilles tendon force, appear to be more similar than the correlations, with best results for both cases in the range of 15–20%. The non-negligible root mean square errors mean that it is commendable to present, e.g., 95% prediction intervals to users rather than just one number. Correlations, on the other hand, do not contain information about the precision of the absolute numbers but rather about the ability of the model to capture changes correctly. This can be useful for an individual runner who considers a change in running style and wants to know whether that would influence the load on a given tissue positively or negatively. The ten first predictors in terms of rank contain several sine and cosine terms of higher order, which might be surprising given that the higher-order Fourier coefficients are supposed to diminish. However, a running cycle comprises the time from right heel strike to right heel strike, which causes the leg and arm movements to have their dominant frequency at the step cycle frequency, while torso motions tend to repeat with double frequency, where terms from second order and upwards dominate. The results do not support identification of a single best type of algorithm. Elastic net performs better than XGB for patella ligament force and vice versa for Achilles tendon force. However, XGB may have some merit over elastic net, in the sense that prediction results for XGB are more stable over IMU placement while, especially for Achilles tendon Sensors 2023, 23, 9836 10 of 12 force, elastic net results are quite sensitive to IMU placement. For both algorithms and forces to be predicted, using a single IMU gave better results than using two. The reason may be that the resulting lower number of Fourier acceleration coefficients (44 rather than 88) guards against overfitting. The high correlations for especially the patella ligament force suggest that the method has the potential to evolve into a reliable information source for injury prevention ef- forts. Speculatively, our method could be used to estimate whether a change in running biomechanics would lead to changes in injury risk for a given runner. However, running biomechanics is complex, and an attempt to offload a given anatomical structure might increase the load of other structures. Running habits, such as speed and weekly distance, might also be affected by a person’s running biomechanics and affect injury risk indirectly. In this study, the two structure-specific loads were included as their respective max- imum values over the running cycle. These values may occur at different times in the running cycle depending on the running style, for instance heel or forefoot strike. Future studies may consider predicting the entire cycle of the forces as a more informative measure of the load exposure of the tissues. It should be noted that the accelerations, from which the loads are predicted, are perfect virtual accelerations from treadmill running that are not affected by measurement noise, soft tissue artefacts, uncertain positioning, and variable terrain circumstances, which will be confounding factors for real measurements. A similar approach was previously employed [28,37] by first developing a tibial bone force prediction model using virtual instruments in [28] and subsequently validating the prediction model using physical instruments in [37]. 5. Conclusions Data-based methods for predicting structure loads in running from a small number of accelerometers appear to have merit in the case of running. However, further inves- tigations of data-based methods against physics-based models and in vivo experimental data, including injury registration, and the inclusion of accelerations measured in field conditions, are commendable. Author Contributions: Conceptualization and physics-based models, J.R.; development of statistical models and machine learning protocols, R.P.W.; running injury state-of-the-art, statistics and etiology, S.S. All authors contributed equally to the writing and reviewing process. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: Ethical review and approval were waived for this study because the underlying data were not collected as a part of the study but was insourced in an anonymized form from previous studies. Informed Consent Statement: Not applicable. Data Availability Statement: Anonymized data can be made available by personal contact to the authors. Acknowledgments: The authors thank Kaiser Sport og Ortopædi A/S for their contributions to data collection. Conflicts of Interest: The authors declare no conflict of interest. References 1. Pontefract, N. Active Lives Adult Survey November 2021–22 Report; Sportengland.org: Loughborough, UK, 2023. 2. Church, T.S. Invited Commentary: Little Steps Lead to Huge Steps—It’s Time to Make Physical Inactivity Our Number 1 Public Health Enemy. Am. J. Epidemiol. 2016, 184, 633–635. [CrossRef] [PubMed] 3. Hespanhol Junior, L.C.; Pillay, J.D.; van Mechelen, W.; Verhagen, E. Meta-Analyses of the Effects of Habitual Running on Indices of Health in Physically Inactive Adults. Sports Med. 2015, 45, 1455–1468. [CrossRef] [PubMed] Sensors 2023, 23, 9836 11 of 12 4. Videbæk, S.; Bueno, A.M.; Nielsen, R.O.; Rasmussen, S. Incidence of Running-Related Injuries Per 1000 h of Running in Different Types of Runners: A Systematic Review and Meta-Analysis. Sports Med. 2015, 45, 1017–1026. [CrossRef] [PubMed] 5. Kakouris, N.; Yener, N.; Fong, D.T.P. A Systematic Review of Running-Related Musculoskeletal Injuries in Runners. J. Sport Health Sci. 2021, 10, 513–522. [CrossRef] [PubMed] 6. Mulvad, B.; Nielsen, R.O.; Lind, M.; Ramskov, D. Diagnoses and Time to Recovery among Injured Recreational Runners in the RUN CLEVER Trial. PLoS ONE 2018, 13, e0204742. [CrossRef] [PubMed] 7. Nielsen, R.O.; Rønnow, L.; Rasmussen, S.; Lind, M. A Prospective Study on Time to Recovery in 254 Injured Novice Runners. PLoS ONE 2014, 9, e99877. [CrossRef] [PubMed] 8. Koplan, J.P.; Rothenberg, R.B.; Jones, E.L. The Natural History of Exercise: A 10-Yr Follow-up of a Cohort of Runners. Med. Sci. Sports Exerc. 1995, 27, 1180–1184. [CrossRef] 9. Fokkema, T.; Hartgens, F.; Kluitenberg, B.; Verhagen, E.; Backx, F.J.G.; van der Worp, H.; Bierma-Zeinstra, S.M.A.; Koes, B.W.; van Middelkoop, M. Reasons and Predictors of Discontinuation of Running after a Running Program for Novice Runners. J. Sci. Med. Sport 2019, 22, 106–111. [CrossRef] 10. van Mechelen, W.; Hlobil, H.; Kemper, H.C.G. Incidence, Severity, Aetiology and Prevention of Sports Injuries. Sports Med. 1992, 14, 82–99. [CrossRef] 11. Edwards, W.B. Modeling Overuse Injuries in Sport as a Mechanical Fatigue Phenomenon. Exerc. Sport Sci. Rev. 2018, 46, 224–231. [CrossRef] 12. Erdemir, A.; McLean, S.; Herzog, W.; van den Bogert, A.J. Model-Based Estimation of Muscle Forces Exerted during Movements. Clin. Biomech. 2007, 22, 131–154. [CrossRef] [PubMed] 13. Damsgaard, M.; Rasmussen, J.; Christensen, S.T.; Surma, E.; de Zee, M. Analysis of Musculoskeletal Systems in the AnyBody Modeling System. Simul. Model Pract. Theory 2006, 14, 1100–1111. [CrossRef] 14. Delp, S.L.; Anderson, F.C.; Arnold, A.S.; Loan, P.; Habib, A.; John, C.T.; Guendelman, E.; Thelen, D.G. OpenSim: Open-Source Software to Create and Analyze Dynamic Simulations of Movement. IEEE Trans. Biomed. Eng. 2007, 54, 1940–1950. [CrossRef] [PubMed] 15. Marra, M.A.; Vanheule, V.; Fluit, R.; Koopman, B.H.F.J.M.; Rasmussen, J.; Verdonschot, N.; Andersen, M.S. A Subject-Specific Musculoskeletal Modeling Framework to Predict in Vivo Mechanics of Total Knee Arthroplasty. J. Biomech. Eng. 2015, 137, 020904. [CrossRef] 16. Topley, M.; Richards, J.G. A Comparison of Currently Available Optoelectronic Motion Capture Systems. J. Biomech. 2020, 106, 109820. [CrossRef] 17. Nijmeijer, E.M.; Heuvelmans, P.; Bolt, R.; Gokeler, A.; Otten, E.; Benjaminse, A. Concurrent Validation of the Xsens IMU System of Lower-Body Kinematics in Jump-Landing and Change-of-Direction Tasks. J. Biomech. 2023, 154, 111637. [CrossRef] [PubMed] 18. Benson, L.C.; Räisänen, A.M.; Clermont, C.A.; Ferber, R. Is This the Real Life, or Is This Just Laboratory? A Scoping Review of IMU-Based Running Gait Analysis. Sensors 2022, 22, 1722. [CrossRef] 19. Mason, R.; Pearson, L.T.; Barry, G.; Young, F.; Lennon, O.; Godfrey, A.; Stuart, S. Wearables for Running Gait Analysis: A Systematic Review. Sports Med. 2023, 53, 241–268. [CrossRef] 20. Janssen, M.; Scheerder, J.; Thibaut, E.; Brombacher, A.; Vos, S. Who Uses Running Apps and Sports Watches? Determinants and Consumer Profiles of Event Runners’ Usage of Running-Related Smartphone Applications and Sports Watches. PLoS ONE 2017, 12, e0181167. [CrossRef] 21. Arac, A. Machine Learning for 3D Kinematic Analysis of Movements in Neurorehabilitation. Curr. Neurol. Neurosci. Rep. 2020, 20, 29. [CrossRef] 22. Dasgupta, A.; Sharma, R.; Mishra, C.; Nagaraja, V.H. Machine Learning for Optical Motion Capture-Driven Musculoskeletal Modelling from Inertial Motion Capture Data. Bioengineering 2023, 10, 510. [CrossRef] [PubMed] 23. Saxby, D.J.; Killen, B.A.; Pizzolato, C.; Carty, C.P.; Diamond, L.E.; Modenese, L.; Fernandez, J.; Davico, G.; Barzan, M.; Lenton, G.; et al. Machine Learning Methods to Support Personalized Neuromusculoskeletal Modelling. Biomech. Model Mechanobiol. 2020, 19, 1169–1185. [CrossRef] [PubMed] 24. Sharma, R.; Dasgupta, A.; Cheng, R.; Mishra, C.; Nagaraja, V.H. Machine Learning for Musculoskeletal Modeling of Upper Extremity. IEEE Sens. J. 2022, 22, 18684–18697. [CrossRef] 25. Sohane, A.; Agarwal, R. Knee Muscle Force Estimating Model Using Machine Learning Approach. Comput. J. 2022, 65, 1167–1177. [CrossRef] 26. Lim, H.; Kim, B.; Park, S. Prediction of Lower Limb Kinetics and Kinematics during Walking by a Single IMU on the Lower Back Using Machine Learning. Sensors 2019, 20, 130. [CrossRef] 27. Stetter, B.J.; Ringhof, S.; Krafft, F.C.; Sell, S.; Stein, T. Estimation of Knee Joint Forces in Sport Movements Using Wearable Sensors and Machine Learning. Sensors 2019, 19, 3690. [CrossRef] 28. Matijevich, E.S.; Branscombe, L.M.; Scott, L.R.; Zelik, K.E. Ground Reaction Force Metrics Are Not Strongly Correlated with Tibial Bone Load When Running across Speeds and Slopes: Implications for Science, Sport and Wearable Tech. PLoS ONE 2019, 14, e0210000. [CrossRef] 29. Liu, Y.; Boshoff, D.; Hancke, G.P. Feasibility of using Gyroscope to Derive Keys for Mobile Phone and Smart Wearable. In Proceedings of the 2022 IEEE 20th International Conference on Industrial Informatics (INDIN), Perth, Australia, 25–28 July 2022; pp. 151–156. [CrossRef] Sensors 2023, 23, 9836 12 of 12 30. Skejø, S.D.; Lund, M.E.; Stensvig, M.; Kaae, N.M.; Rasmussen, J. Running in Circles: Describing Running Kinematics Using Fourier Series. J. Biomech. 2021, 115, 110187. [CrossRef] 31. Andersen, M.S.; Damsgaard, M.; MacWilliams, B.; Rasmussen, J. A Computationally Efficient Optimisation-Based Method for Parameter Identification of Kinematically Determinate and over-Determinate Biomechanical Systems. Comput. Methods Biomech. Biomed. Eng. 2010, 13, 171–183. [CrossRef] 32. Carbone, V.; Fluit, R.; Pellikaan, P.; van der Krogt, M.M.; Janssen, D.; Damsgaard, M.; Vigneron, L.; Feilkas, T.; Koopman, H.F.J.M.; Verdonschot, N. TLEM 2.0—A Comprehensive Musculoskeletal Geometry Dataset for Subject-Specific Modeling of Lower Extremity. J. Biomech. 2015, 48, 734–741. [CrossRef] 33. Zou, H.; Hastie, T. Regularization and Variable Selection via the Elastic Net. J. R. Stat. Soc. Series B Stat. Methodol. 2005, 67, 301–320. [CrossRef] 34. Chen, T.; Guestrin, C. XGBoost. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August 2016; ACM: New York, NY, USA, 2016; pp. 785–794. 35. Friedman, J.; Hastie, T.; Tibshirani, R. Regularization Paths for Generalized Linear Models via Coordinate Descent. J. Stat. Softw. 2010, 33, 1–22. [CrossRef] [PubMed] 36. Brund, R.B.K.; Waagepetersen, R.O.; Nielsen, R.; Rasmussen, J.; Nielsen, M.S.; Andersen, C.H.; de Zee, M. How Precisely Can Easily Accessible Variables Predict Achilles and Patellar Tendon Forces during Running? Sensors 2021, 21, 7418. [CrossRef] [PubMed] 37. Elstub, L.J.; Nurse, C.A.; Grohowski, L.M.; Volgyesi, P.; Wolf, D.N.; Zelik, K.E. Tibial Bone Forces Can Be Monitored Using Shoe-Worn Wearable Sensors during Running. J. Sports Sci. 2022, 40, 1741–1749. [CrossRef] 38. Xiang, L.; Wang, A.; Gu, Y.; Zhao, L.; Shim, V.; Fernandez, J. Recent Machine Learning Progress in Lower Limb Running Biomechanics With Wearable Technology: A Systematic Review. Front. Neurorobot. 2022, 16, 913052. [CrossRef] 39. Bertelsen, M.L.; Hulme, A.; Petersen, J.; Brund, R.K.; Sørensen, H.; Finch, C.F.; Parner, E.T.; Nielsen, R.O. A Framework for the Etiology of Running-related Injuries. Scand. J. Med. Sci. Sports 2017, 27, 1170–1180. [CrossRef] 40. Long, T.; Outerleys, J.; Yeung, T.; Fernandez, J.; Bouxsein, M.L.; Davis, I.S.; Bredella, M.A.; Besier, T.F. Predicting Ankle and Knee Sagittal Kinematics and Kinetics Using an Ankle-Mounted Inertial Sensor. Comput. Methods Biomech. Biomed. Eng. 2023, 1–14. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Predicting Tissue Loads in Running from Inertial Measurement Units.

12-15-2023

Rasmussen, John,Skejø, Sebastian,Waagepetersen, Rasmus Plenge

eng

PMC7379642

Supplement Table 6. Change in VO2max (L·min-1 and ml·min-1·kg-1) from 1995-1997 to 2016-2017 in relation to sex and age. L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 Year n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change 95-97 646 2.82 (0.05) Ref 43.8 (1.09) Ref 1 160 2.47 (0.03) Ref 37.4 (0.84) Ref 589 2.13 (0.05) Ref 31.5 (1.03) Ref 98-99 767 2.82 (0.05) -0,1% 43.1 (1.14) -1,6% 1 302 2.50 (0.02) 1,2% 37.5 (0.53) 0,3% 895 2.16 (0.04) 1,3% 31.9 (0.81) 1,4% 00-01 1 626 2.80 (0.06) -0,5% 42.6 (1.26) -2,7% 2 592 2.51 (0.05) 1,7% 37.6 (1.11) 0,6% 1 988 2.13 (0.02) -0,2% 31.5 (0.71) -0,1% 02-03 2 897 2.72 (0.05) -3,7% 41.7 (1.01) -4,7% 5 026 2.45 (0.04) -1,0% 36.5 (1.03) -2,4% 3 935 2.06 (0.04) -3,3% 30.5 (0.79) -3,3% 04-05 4 397 2.73 (0.04) -3,2% 41.9 (0.95) -4,3% 8 724 2.45 (0.04) -0,6% 36.4 (1.00) -2,6% 6 379 2.07 (0.03) -2,7% 30.5 (0.62) -3,2% 06-07 4 257 2.74 (0.03) -2,7% 41.8 (0.84) -4,6% 8 185 2.47 (0.04) 0,2% 36.4 (0.94) -2,7% 6 272 2.09 (0.05) -1,8% 30.6 (0.93) -2,7% 08-09 4 765 2.74 (0.04) -3,0% 41.6 (1.16) -5,1% 8 660 2.50 (0.07) 1,3% 36.6 (1.32) -2,0% 6 643 2.13 (0.05) -0,1% 30.9 (0.94) -2,0% 10-11 4 087 2.74 (0.07) -2,9% 41.4 (1.46) -5,4% 8 000 2.51 (0.05) 1,7% 36.5 (1.26) -2,3% 5 214 2.13 (0.05) 0,0% 30.7 (0.81) -2,5% 12-13 5 727 2.73 (0.06) -3,3% 41.1 (1.44) -6,2% 10 823 2.50 (0.05) 1,1% 36.5 (1.33) -2,4% 6 786 2.13 (0.04) 0,1% 30.6 (0.91) -2,8% 14-15 5 671 2.68 (0.04) -5,1% 40.4 (1.00) -7,8% 9 457 2.46 (0.06) -0,2% 35.8 (1.42) -4,2% 5 766 2.12 (0.06) -0,5% 30.3 (1.09) -3,8% 16-17 3 743 2.67 (0.07) -5,3% 40.2 (1.38) -8,2% 5 904 2.45 (0.05) -0,9% 35.7 (1.23) -4,6% 3 817 2.12 (0.07) -0,6% 30.4 (1.14) -3,6% L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 Year n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change 95-97 708 3.54 (0.08) Ref 43.9 (1.19) Ref 1 035 3.16 (0.03) Ref 38.5 (0.59) Ref 436 2.74 (0.03) Ref 33.5 (0.42) Ref 98-99 1 073 3.54 (0.05) -0,1% 44.3 (0.65) 0,8% 1 547 3.10 (0.04) -1,9% 37.4 (0.69) -2,8% 959 2.69 (0.04) -1,8% 32.6 (0.72) -2,7% 00-01 1 843 3.49 (0.09) -1,3% 43.4 (1.09) -1,2% 2 656 3.12 (0.06) -1,4% 37.4 (1.10) -2,9% 1 840 2.67 (0.07) -2,5% 32.2 (0.97) -4,0% 02-03 3 666 3.44 (0.08) -2,9% 42.7 (1.03) -2,6% 4 403 3.11 (0.07) -1,7% 37.2 (1.14) -3,5% 2 702 2.64 (0.06) -3,6% 32.0 (0.79) -4,6% 04-05 5 220 3.44 (0.05) -2,8% 42.4 (0.82) -3,4% 7 570 3.10 (0.07) -1,8% 36.9 (1.15) -4,1% 5 130 2.63 (0.04) -4,0% 31.7 (0.73) -5,5% 06-07 5 486 3.40 (0.06) -3,9% 41.7 (1.06) -5,1% 8 682 3.11 (0.07) -1,7% 36.6 (1.21) -4,9% 5 637 2.65 (0.05) -3,3% 31.7 (0.65) -5,3% 08-09 6 503 3.40 (0.07) -4,1% 41.7 (1.07) -5,1% 9 992 3.13 (0.08) -1,0% 36.6 (1.42) -5,0% 6 916 2.66 (0.06) -3,0% 31.5 (0.89) -6,0% 10-11 6 253 3.37 (0.07) -4,8% 41.2 (0.96) -6,0% 9 618 3.13 (0.08) -0,8% 36.6 (1.46) -5,0% 6 005 2.69 (0.06) -1,8% 31.6 (0.92) -5,8% 12-13 10 010 3.34 (0.07) -5,6% 40.8 (1.15) -7,0% 14 828 3.07 (0.09) -2,7% 36.0 (1.59) -6,6% 9 072 2.64 (0.06) -3,8% 31.0 (0.84) -7,5% 14-15 10 757 3.28 (0.07) -7,3% 39.9 (1.10) -9,0% 14 268 3.03 (0.06) -4,2% 35.3 (1.39) -8,4% 9 665 2.63 (0.07) -4,1% 30.7 (0.96) -8,5% 16-17 8 201 3.28 (0.06) -7,3% 39.6 (0.99) -9,9% 8 789 3.01 (0.07) -4,9% 34.9 (1.38) -9,2% 6 107 2.65 (0.08) -3,2% 30.6 (1.10) -8,5% 18-34 years 35-49 years 50-74 years Women Men 18-34 years 35-49 years 50-74 years

Decline in cardiorespiratory fitness in the Swedish working force between 1995 and 2017.

11-15-2018

Ekblom-Bak, Elin,Ekblom, Örjan,Andersson, Gunnar,Wallin, Peter,Söderling, Jonas,Hemmingsson, Erik,Ekblom, Björn

eng

PMC7143174

International Journal of Environmental Research and Public Health Article Tower Running—Participation, Performance Trends, and Sex Difference Daniel Stark 1, Stefania Di Gangi 2, Caio Victor Sousa 3 , Pantelis Nikolaidis 4 and Beat Knechtle 2,5,* 1 Department of Orthopedic Surgery and Traumatology, Kantonsspital Baden, 5404 Baden, Switzerland; [email protected] 2 Institute of Primary Care, University Hospital Zurich, 8091 Zurich, Switzerland; [email protected] 3 Bouve College of Health Sciences, Northeastern University, 360 Huntington Ave., Boston, MA 02115, USA; [email protected] 4 Exercise Physiology Laboratory, 18450 Nikaia, Greece; [email protected] 5 Medbase St. Gallen Am Vadianplatz, 9000 St. Gallen, Switzerland * Correspondence: [email protected]; Tel.: +41-(0)-71-226-93-00 Received: 14 February 2020; Accepted: 10 March 2020; Published: 14 March 2020



 Abstract: Though there are exhaustive data about participation, performance trends, and sex differences in performance in different running disciplines and races, no study has analyzed these trends in stair climbing and tower running. The aim of the present study was therefore to investigate these trends in tower running. The data, consisting of 28,203 observations from 24,007 climbers between 2014 and 2019, were analyzed. The effects of sex and age, together with the tower characteristics (i.e., stairs and floors), were examined through a multivariable statistical model with random effects on intercept, at climber’s level, accounting for repeated measurements. Men were faster than women in each age group (p < 0.001 for ages ≤69 years, p = 0.003 for ages > 69 years), and the difference in performance stayed around 0.20 km/h, with a minimum of 0.17 at the oldest age. However, women were able to outperform men in specific situations: (i) in smaller buildings (<600 stairs), for ages between 30 and 59 years and >69 years; (ii) in higher buildings (>2200 stairs), for age groups <20 years and 60–69 years; and (iii) in buildings with 1600–2200 stairs, for ages >69 years. In summary, men were faster than women in this specific running discipline; however, women were able to outperform men in very specific situations (i.e., specific age groups and specific numbers of stairs). Keywords: tower running; sex differences; age; running speed; vertical run 1. Introduction Distance running is of high popularity and includes different distances, from 5 to 10 km [1], half-marathon [2,3], marathon [2,4], and up to ultra-marathon of different distances [5,6]. It is well-known that men are faster than women from 5 km to marathon [7], and in ultra-marathon running [8]. However, women were able to reduce the gap with men in ultra-marathon running, with increasing age and at longer race distances [9]. Stair climbing or tower running is a very specific running discipline, in which stair climbing has developed into the organized sport of tower running. Nowadays, tower running is a sport discipline that involves running up tall buildings, such as internal staircases of skyscrapers. However, tower running can cover any running race that involves a course that ascends a building. To date, we have knowledge about the health benefits of stair climbing [10–13]. However, no data exist about participation and performance trends in tower running, and especially about the sex difference in this specific running discipline. Such information is valuable for athletes and coaches, Int. J. Environ. Res. Public Health 2020, 17, 1902; doi:10.3390/ijerph17061902 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2020, 17, 1902 2 of 9 to better understand and plan a race strategy, and also for race organizers, for insights regarding future events. Therefore, the aim of the present study was to investigate participation and performance trends in tower running, with the hypothesis that men would also be faster than women in this discipline. Regarding age groups, we expected that women might close the performance gap in the older groups as already shown in long distance races [9]. 2. Materials and Methods 2.1. Ethics Approval This study was approved by the Institutional Review Board of Kanton St. Gallen, Switzerland, with a waiver of the requirement for informed consent of the participants, as the study involved the analysis of publicly available data. 2.2. Methodology There exists a tower running world association that presents all the results of the known races around the world on their homepage (www.towerrunning.com). In an older version of this homepage, there were only the results of the current year, and sometimes, of the preceding year. We contacted the person in charge at the association to find out whether he could provide us with older data as well. For some races, however, it was not possible to find the older results. For example, for the race at the Willis tower in Chicago, the results before 2018 were not available. For other races, such as the hustle up race in Chicago, the direct link did not work, but the results could be found by searching for the link to the race, which is also provided on the homepage of the tower running world association. Table 1 summarizes all considered events listed by the number of steps of the buildings. Table 1. Data included in the present study. Building City Steps Data Available (Years) Included (Years) Millennium Tower Wien 2529 2014–2016 2014–2016 Willis Tower (Sears Tower until 2009) Chicago 2109 2014–2019 2018 Taipei 101 Taipeh 2046 2014–2019 2017–2018 CN Tower Toronto 1776 2014–2019 2017–2018 Reunion Tower Dallas 1674 2018–2019 2018 Eiffelturm Paris 1665 2015–2020 2015–2018 AON Center Chicago 1643 none on towerrunning.com 2018 John Hancock Center (875 North Michigan Avenue) Chicago 1632 2014–2019 2017–2018 Empire State Building New York 1576 2014–2019 2017–2014 Bank of America Plaza Dallas 1540 none on towerrunning.com 2018 US Bank Tower Los Angeles 1500 2014–2019 2018 thyssenkrupp Testturm Rottweil 1390 2018–2019 2018 Swissôtel The Stamford Singapur 1336 2014–2018 2017 Rockefeller Center New York City 1214 2014–2016, 2018, 2019 2019 MesseTurm Frankfurt am Main 1202 2014–2019 2014–2017 Three Logan Square Philadelphia 1088 2014–2019 2014, 2018, 2019 Valliance Bank Oklahoma City 837 2014–2019 2019 Holmenkollbakken Oslo 800 2015–2018 2015–2017 Run Up Berlin (Park Inn Hotel) Berlin 770 2015–2019 2015–2018 KölnTurm Köln 714 none on towerrunning.com 2016–2019 Oakbrook Terrace Tower Oakbrook 680 2014–2020 2019 Münsterturm Ulm 560 none on towerrunning.com 2014–2018 Towerrun Berlin 465 2014–2020 2018 St.George’s Tower Leicester 351 none on towerrunning.com 2018 Matzleinsdorfer Hochhaus Wien 342 2017 2017 Windradlauf Lichtenegg 300 2014 2014 Haus des Meeres Wien 271 2015–2019 2016–2018 Oluempia Hotel Tallinn N/A 2015–2019 2017 Int. J. Environ. Res. Public Health 2020, 17, 1902 3 of 9 From the race results, the year of the event, the completed time, the sex, and the name of both the athletes and the building were available. We further looked for the height of the building and the number of stairs and floors. Race time in m:sec was converted to running speed in km/h, using the height of the building. We removed observations from unknown climbers (where the name of the climber was not reported or not known) in order to correctly account for repeated measurements. We also considered multi-climbing. 2.3. Statistical Analysis The outcome was the tower climbing speed (km/h). Descriptive statistics are presented as means (SD = standard deviations) by sex and age groups. T-tests were performed to assess the outcome difference between sex, overall and for each age groups. Two-way ANOVA tests were also performed to evaluate the multivariable effect of sex and age on the outcome. Then, to control also for repeated measurements and the other covariates, the effects of sex and age, together with the tower characteristics (i.e., stairs and floors) were examined more rigorously through a multivariable mixed effects model, with random effects (intercept) for climbers. The model was specified as follows: Tower climbing speed (Y) ~ [Fixed effects (X) = Sex*Age*BS (Stairs, df = 5) + BS (Floors, df = 5) + [random effects of intercept = runners] where BS (Stairs, df = 5) and BS (Floors, df = 5) are 5 degrees of freedom (df) basis splines changing with the number of stairs and floors, respectively; Sex*Age*BS (Stairs, df = 5) denoted the three-way interaction term Sex–Age–number of stairs. Calendar year was not considered in the above model because it was not significant. Results of the regression model are presented as estimates and standard errors. Statistical significance was defined as p < 0.05. All statistical analyses were carried out with R, R Core Team (2016). R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria (www.r-project.org/foundation/). The R packages ggplot2, lme4, and lmerTest were used, respectively, for data visualization and for the mixed model. The R code to reproduce the analysis is provided as supplementary information (Supplement 1 R-code). 3. Results Between 2014 and 2019, the total number of observations was 28,203 (24,007 climbers). However, the total number of observations, with non-missing sex, was 28,156 (23,960 climbers). The participation and men-to-women ratio is shown in Figure 1. We observed that we had a low number of participants and a high men-to-women ratio before 2017 (i.e., the number of men was three times the number of women in 2015). The highest number of participants was recorded in 2018. In fact, the number of women in 2018 was eight times the number of women in 2014, and the number of men in 2018 was four times the number of men in 2014. In 2019, the number of available observations decreased again. The men-to-women ratio reached a minimum in 2019 with 0.89, meaning that the number of women was higher than the number of men. In Table 2, the mean performance by sex and age group is reported. Men were faster than women in each age group (p < 0.001 for all ages until 69 years, p = 0.003 for ages >69 years), and the difference in performance stayed around 0.20 km/h, with a minimum of 0.17 at the oldest age. In Table 3, summary statistics of performance, together with tower characteristics: height, number of floors and stairs are reported by sex. Overall, the sex difference in performance was significant (p < 0.001); sex differences were also significant (p < 0.001) in average floors and stairs climbed. The results of the multivariable statistical analysis are displayed in Figure 2, to allow an easier interpretation and understanding. Moreover, we had no significant difference between men and women alone, but in the interaction with age groups and stairs climbed (Supplemental 2 Table). The variability, in terms of performance, was greater in very young and very old age groups (<20 years, 60–69 years, and >69 years). This also Int. J. Environ. Res. Public Health 2020, 17, 1902 4 of 9 had an effect on sex differences. Women performed better than men in the following situations: (i) smaller buildings (<600 stairs), for ages between 30 and 59 years and >69 years; (ii) higher buildings (>2200 stairs), for age <20 years and ages between 60 and 69 years; and (iii) buildings with 1600 to 2200 stairs, for age >69 years. In all other cases, men performed better than women, with the sex difference reducing when the number of stairs increased. In Figure 3, the effect of the number of floors on performance, by sex, is shown. When the number of floors increased, the average speed of tower climbing decreased, but then increased around 90 floors, and decreased again in climbing the highest buildings. Int. J. Environ. Res. Public Health 2020, 17, x FOR PEER REVIEW 4 of 9 the number of women in 2015). The highest number of participants was recorded in 2018. In fact, the number of women in 2018 was eight times the number of women in 2014, and the number of men in 2018 was four times the number of men in 2014. In 2019, the number of available observations decreased again. The men-to-women ratio reached a minimum in 2019 with 0.89, meaning that the number of women was higher than the number of men. Figure 1. Participation and men-to-women ratio. In Table 2, the mean performance by sex and age group is reported. Men were faster than women in each age group (p<0.001 for all ages until 69 years, p=0.003 for ages >69 years), and the difference in performance stayed around 0.20 km/h, with a minimum of 0.17 at the oldest age. In Table 3, summary statistics of performance, together with tower characteristics: height, number of floors and stairs are reported by sex. Overall, the sex difference in performance was significant (p<0.001); sex differences were also significant (p<0.001) in average floors and stairs climbed. The results of the multivariable statistical analysis are displayed in Figure 2, to allow an easier interpretation and understanding. Moreover, we had no significant difference between men and women alone, but in the interaction with age groups and stairs climbed (Supplemental 2 Table). The variability, in terms of performance, was greater in very young and very old age groups (<20 years, 60–69 years, and >69 years). This also had an effect on sex differences. Women performed better than men in the following situations: (i) smaller buildings (<600 stairs), for ages between 30 and 59 years and >69 years; (ii) higher buildings (>2200 stairs), for age <20 years and ages between 60 and 69 years; and (iii) buildings with 1600 to 2200 stairs, for age >69 years. In all other cases, men performed better than women, with the sex difference reducing when the number of stairs increased. In Figure 3, the effect of the number of floors on performance, by sex, is shown. When the number of floors increased, the average speed of tower climbing decreased, but then increased around 90 floors, and decreased again in climbing the highest buildings. Figure 1. Participation and men-to-women ratio. Table 2. Summary statistics of tower climbing performance, running speed (km/h), by sex and age groups. p-values from t-tests for each subgroup are reported. p-values from ANOVA were both p < 0.001 for sex and age. Men-to-women ratio, computed with the number of participants, is reported. Age Group Sex N Mean (SD) p Men-to-Women Ratio <20 F 501 0.73 (0.27) <0.001 1.30 M 652 0.91 (0.38) 20–29 F 1887 0.81 (0.24) <0.001 1.39 M 2615 0.99 (0.35) 30–39 F 2552 0.80 (0.30) <0.001 1.34 M 3415 1.03 (0.39) 40–49 F 1941 0.78 (0.32) <0.001 1.33 M 2583 1.00 (0.39) 50–59 F 1220 0.76 (0.30) <0.001 1.60 M 1951 0.97 (0.38) 60–69 F 239 0.72 (0.25) <0.001 2.62 M 626 0.90 (0.27) >69 F 44 0.66 (0.33) 0.003 4.57 M 201 0.83 (0.33) Int. J. Environ. Res. Public Health 2020, 17, 1902 5 of 9 Int. J. Environ. Res. Public Health 2020, 17, x FOR PEER REVIEW 6 of 9 Figure 2. Speed (km/h) by stairs, age, and sex. Lines represent the predicted values from the mixed model and points represent the average of the observed values. Figure 3. Speed (km/h) by floors and sex. Lines represent the predicted values from the mixed model and points represent the average of the observed values. Figure 2. Speed (km/h) by stairs, age, and sex. Lines represent the predicted values from the mixed model and points represent the average of the observed values. Int. J. Environ. Res. Public Health 2020, 17, x FOR PEER REVIEW 6 of 9 Figure 2. Speed (km/h) by stairs, age, and sex. Lines represent the predicted values from the mixed model and points represent the average of the observed values. Figure 3. Speed (km/h) by floors and sex. Lines represent the predicted values from the mixed model and points represent the average of the observed values. Figure 3. Speed (km/h) by floors and sex. Lines represent the predicted values from the mixed model and points represent the average of the observed values. Int. J. Environ. Res. Public Health 2020, 17, 1902 6 of 9 Table 3. Summary statistics of running speed (km/h) and race time (min), tower height (m), floors, and stairs by sex. Data expressed as mean (± SD). Females (n = 11,886) Males (n = 16,270) p-Value Speed km/h 0.85 (0.37) 1.06 (0.46) <0.001 Time (min) 24.26 (14.16) 18.43 (11.69) <0.001 Tower height (m) 296.25 (111.37) 276.27 (108.72) <0.001 Floors 85.44 (36.37) 76.00 (35.97) <0.001 Stairs 1466.43 (420.36) 1401.18 (429.59) <0.001 4. Discussion The aim of the present study was to investigate participation trends, performance trends, and trends in sex difference in tower running, with the hypothesis that men would be faster than women in this discipline. The main findings were: (1) more men than women competed before 2017, (2) men were faster than women in each age group and the difference in performance stayed around 0.20 km/h, with a minimum of 0.17 km/h at the oldest age, and (3) women aged between 30 and 59 years and >69 years performed better than men in smaller buildings (<600 stairs). 4.1. Change in the Men-to-Women Ratio Across Years Before 2017, we observed a low number of participants and a high men-to-women ratio. The highest number of participants was recorded in 2018. In 2019, the number of participants decreased again and the men-to-women ratio reached the minimum of 0.89, which means that the number of women was higher than the number of men. This could also be due to a selection bias. At the time of the data collection (2017–2019), there were more results available from the earlier races and since the aim of the selection was to represent the sport and include the most important races all over the world, we did not pay attention to compare for every year the exact same number of races. This fact should encourage race directors to join the ‘Towerrunning World Association’ (www.towerrunning.com), in order to build up a firm data base for future analyses. Generally, in races of long traditions, the men-to-women ratio is > 1.0, indicating that more men than women competed [14], but the men-to-women ratio can decrease over the years, indicating that the number of women increased over time [15]. Future studies with larger data sets are needed to investigate this trend. 4.2. Sex Difference in Performance Looking at the anatomical aspect of sex difference, studies have shown that there are differences in the anatomy and physiology of the heart [16], and in the oxygen uptake in repetitive muscle activity [17] between men and women. This fact suggests that there must also be differences in performance between genders in the sport of tower running. Men were faster than women in each age group and the difference in performance stayed around 0.20 km/h, with a minimum of 0.17 at the oldest age. However, women outperformed men in the following situations: (i) smaller buildings (<600 stairs) and ages between 30 and 59 years and >69 years; (ii) higher buildings (>2200 stairs) and ages <20 years and between 60 and 69 years; and (iii) buildings with 1600–2200 stairs and ages >69 years. When the number of floors increased, the average running speed of tower climbing decreased, but then increased around 90 floors, and decreased again in climbing of highest buildings. A possible explanation for this fact could be the diversity of the runners. One could think that recreational runners take part in races until a certain height, because of their estimated stamina. Therefore, their running speed decreases until they reach their maximum of the height of the building. More professional runners again might only start in the races in which they have to climb the higher buildings, starting around 90 floors. Again, these professional runners will have to decrease their average running speed, Int. J. Environ. Res. Public Health 2020, 17, 1902 7 of 9 to be able to climb even the highest building. This, on the other hand, is only a hypothesis that we did not investigate, and would need further studies to be verified. Another explanation could be the men-to-women ratio by age group. When female and male age group ultra-marathoners were investigated, women could close the gap to men in older age groups (>60 years) and longer race distances (i.e., 100 miles compared with 50 miles) [9]. This relative improvement in female performance at higher ages is most likely due to the change in the men-to-women ratio in older age groups. It has been shown for female and male age group freestyle swimmers, from 25–29 to 85–89 years, competing in the FINA World Masters Championships between 1986 and 2014, that women were faster than men for age groups 80–84 and 85–89 years. When the trend for the men-to-women ratio for age groups 25–29 to 75–79 years (i.e., men were faster than women) and age groups 80–84 to 85–89 years (i.e., women were faster than men) was analyzed, the men-to-women ratio remained unchanged in 50 m, 100 m, and 400 m in age groups 25–29 to 75–79 years, but increased in 200 m and 800 m. For age groups 80–84 to 85–89 years, the men-to-women ratio remained unchanged in 50 m and 100 m, but decreased in 200 to 800 m [18]. However, in the present tower runners, the men-to-women ratio increased with increasing age, but was lowest in the youngest age group (Table 2). Other variables could explain that women outperformed men in some specific situations (e.g., specific age groups and building heights) of this running discipline. Generally, women are lighter than men [19–21], which might help in running upwards. Body mass was, however, not predictive in female mountain ultra-marathoners [21]. Unfortunately, body mass was not available in these runners. Another explanation could be the motivation of female athletes [22]. For example, motivation differs between female and male marathon runners [22]. It has been shown that female marathon finishers exceeded men on the motivational scales for body weight concern, affiliation, psychological coping, life meaning, and self-esteem, and they scored lower on competitive motivation [23]. Future studies might investigate the motivation of female and male tower runners by age group and performance level. Regarding the health aspect, it has already been investigated that stair climbing brings certain benefits. It could be shown that it helps decrease blood glucose levels [12] and that it brings a cardiac benefit in senior citizens [13]. Therefore, there is a certain interest in investigating this subject regarding public health. 5. Conclusions Men are generally faster than women in tower running, but women are closing the gap with men, with increasing stairs and increasing age. The reason for the better performance in women with increasing stairs remains unclear and might be a subject for further research. Supplementary Materials: The following are available online at http://www.mdpi.com/1660-4601/17/6/1902/s1, Table S1: R-Code, Table S2: Regression analysis (mixed model) of speed (km/h) in tower climbing. Author Contributions: Conceptualization, D.S. and B.K.; methodology, S.D.G.; software, S.D.G.; validation, D.S., S.D.G., and B.K.; formal analysis, S.D.G.; investigation, D.S.; resources, D.S.; data curation, D.S.; writing—original draft preparation, D.S., S.D.G., C.V.S., P.N., and B.K.; writing—review and editing, D.S, S.D.G., C.V.S., P.N., and B.K.; visualization, S.D.G.; supervision, B.K.; project administration, B.K. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflicts of interest. References 1. Deaner, R.O.; Addona, V.; Carter, R.E.; Joyner, M.J.; Hunter, S.K. Fast men slow more than fast women in a 10 km road race. PeerJ 2016, 2016. [CrossRef] 2. Cuk, I.; Nikolaidis, P.T.; Markovic, S.; Knechtle, B. Age differences in pacing in endurance running: Comparison between marathon and half-marathonMen and Women. Medicina 2019, 55, 479. [CrossRef] Int. J. Environ. Res. Public Health 2020, 17, 1902 8 of 9 3. Knechtle, B.; Nikolaidis, P.T. Sex- and age-related differences in half-marathon performance and competitiveness in the world’s largest half-marathon-the GöteborgsVarvet. Res. Sports Med. 2018, 26, 75–85. [CrossRef] 4. Vitti, A.; Nikolaidis, P.T.; Villiger, E.; Onywera, V.; Knechtle, B. The “New York City Marathon”: Participation and performance trends of 1.2M runners during half-century. Res. Sports Med. 2019. [CrossRef] 5. Nikolaidis, P.T.; Knechtle, B. Russians are the fastest and the youngest in the “Comrades Marathon”. J. Sports Sci. 2019, 37, 1387–1392. [CrossRef] 6. Knechtle, B.; Valeri, F.; Zingg, M.A.; Rosemann, T.; Rüst, C.A. What is the age for the fastest ultra-marathon performance in time-limited races from 6 h to 10 days? Age 2014, 36. [CrossRef] [PubMed] 7. Knechtle, B.; Nikolaidis, P.T.; Di Gangi, S. World single age records in running from 5 km to marathon. Front. Psychol. 2018, 9. [CrossRef] [PubMed] 8. Senefeld, J.; Smith, C.; Hunter, S.K. Sex differences in participation, performance, and age of ultramarathon runners. Int. J. Sports Physiol. Perform. 2016, 11, 635–642. [CrossRef] [PubMed] 9. Waldvogel, K.J.; Nikolaidis, P.T.; Di Gangi, S.; Rosemann, T.; Knechtle, B. Women reduce the performance difference to men with increasing age in ultra-marathon running. Int. J. Environ. Res. Public Health 2019, 16, 2377. [CrossRef] 10. Jenkins, E.M.; Nairn, L.N.; Skelly, L.E.; Little, J.P.; Gibala, M.J. Do stair climbing exercise “snacks” improve cardiorespiratory fitness? Appl. Physiol. Nutr. Metab. 2019, 44, 681–684. [CrossRef] 11. Hongu, N.; Shimada, M.; Miyake, R.; Nakajima, Y.; Nakajima, I.; Yoshitake, Y. Promoting Stair Climbing as an Exercise Routine among Healthy Older Adults Attending a Community-Based Physical Activity Program. Sports 2019, 7, 23. [CrossRef] [PubMed] 12. Honda, H.; Igaki, M.; Hatanaka, Y.; Komatsu, M.; Tanaka, S.; Miki, T.; Suzuki, T.; Takaishi, T.; Hayashi, T. Stair climbing/descending exercise for a short time decreases blood glucose levels after a meal in people with type 2 diabetes. BMJ Open Diabetes Res. Care 2016, 4, e000232. [CrossRef] [PubMed] 13. Donath, L.; Faude, O.; Roth, R.; Zahner, L. Effects of stair-climbing on balance, gait, strength, resting heart rate, and submaximal endurance in healthy seniors. Scand. J. Med. Sci. Sports 2014, 24, e93–e101. [CrossRef] [PubMed] 14. Romancuk, N.; Nikolaidis, P.T.; Villiger, E.; Chtourou, H.; Rosemann, T.; Knechtle, B. Performance and Participation in the ‘Vasaloppet’ Cross-Country Skiing Race during a Century. Sports 2019, 7, 86. [CrossRef] [PubMed] 15. Nikolaidis, P.T.; Villiger, E.; Knechtle, B. Participation and Performance Trends in the ITU Duathlon World Championship From 2003 to 2017. J. Strength Cond. Res. 2018. [CrossRef] 16. Legato, M.J. Gender and the heart: Sex-specific differences in normal anatomy and physiology. J. Gend. Specif. Med. JGSM Off. J. Partnersh. Women’s Health Columbia 2000, 3, 15–18. 17. Nindl, B.C.; Sharp, M.A.; Mello, R.P.; Rice, V.J.; Murphy, M.M.; Patton, J.F. Gender comparison of peak oxygen uptake: Repetitive box lifting versus treadmill running. Eur. J. Appl. Physiol. Occup. Physiol. 1998, 77, 112–117. [CrossRef] 18. Knechtle, B.; Nikolaidis, P.T.; König, S.; Rosemann, T.; Rüst, C.A. Performance trends in master freestyle swimmers aged 25–89 years at the FINA World Championships from 1986 to 2014. Age 2016, 38, 1–8. [CrossRef] 19. Wright, A.; Marino, F.E.; Kay, D.; Micalos, P.; Fanning, C.; Cannon, J.; Noakes, T.D. Influence of lean body mass on performance differences of male and female distance runners in warm, humid environments. Am. J. Phys. Anthropol. 2002, 118, 285–291. [CrossRef] 20. Lewis, D.A.; Kamon, E.; Hodgson, J.L. Physiological Differences Between Genders Implications for Sports Conditioning. Sports Med. 1986, 3, 357–369. [CrossRef] 21. O’Loughlin, E.; Nikolaidis, P.T.; Rosemann, T.; Knechtle, B. Different predictor variables for women and men in ultra-marathon running—the wellington urban ultramarathon 2018. Int. J. Environ. Res. Public Health 2019, 16, 1844. [CrossRef] Int. J. Environ. Res. Public Health 2020, 17, 1902 9 of 9 22. Nikolaidis, P.T.; Chalabaev, A.; Rosemann, T.; Knechtle, B. Motivation in the athens classic marathon: The role of sex, age, and performance level in Greek recreational marathon runners. Int. J. Environ. Res. Public Health 2019, 16, 2549. [CrossRef] [PubMed] 23. Wa´skiewicz, Z.; Nikolaidis, P.T.; Gerasimuk, D.; Borysiuk, Z.; Rosemann, T.; Knechtle, B. What motivates successful marathon runners? The role of sex, age, education and training experience in Polish runners. Front. Psychol. 2019, 10. [CrossRef] [PubMed] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Tower Running-Participation, Performance Trends, and Sex Difference.

03-14-2020

Stark, Daniel,Di Gangi, Stefania,Sousa, Caio Victor,Nikolaidis, Pantelis,Knechtle, Beat

eng

PMC9305115

798  |       Scand J Med Sci Sports. 2022;32:798–806. wileyonlinelibrary.com/journal/sms Received: 15 October 2021  |  Revised: 15 December 2021  |  Accepted: 9 January 2022 DOI: 10.1111/sms.14129   O R I G I N A L A R T I C L E Disturbance of desire- goal motivational dynamics during different exercise intensity domains Ian M. Taylor | Summer Whiteley | Richard A. Ferguson This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided  the original work is properly cited. © 2022 The Authors. Scandinavian Journal of Medicine & Science In Sports published by John Wiley & Sons Ltd. School of Sport, Exercise & Health  Sciences, Loughborough University,  Leicestershire, United Kingdom Correspondence Ian M. Taylor, School of Sport, Exercise  & Health Sciences, Loughborough  University, Leicestershire, United  Kingdom. Email: [email protected] Abstract Purpose: The  desire- goal  motivational  conflict  helps  explain  endurance  per- formance; however, the physiological concomitants are unknown. The present  study  examined  disturbances  in  desire  to  reduce  effort  and  performance  goal  value across moderate, heavy, and severe exercise intensity domains, demarcated  by the first (LT1) and second (LT2) lactate thresholds. In addition, the within-  person relationships among blood lactate concentration, heart rate, and desire-  goal conflict were examined. Methods: Thirty participants (53% female, Mage = 21.03 years; SD = 2.06 years)  completed an incremental cycling exercise test, in which work rate was increased  by 25 watts every four minutes, until voluntary exhaustion or sufficient data from  the severe intensity domain had been collected. Desire to reduce effort, perfor- mance  goal  value,  blood  lactate  concentration  (for  determination  of  LT1  and  LT2), and heart rate were measured at the end of each stage and analyzed using  multilevel models. Results: The desire to reduce effort increased over the exercise test with addi- tional shifts and accelerations after each lactate threshold. The performance goal  did not show general declines, nor did it shift at LT1. However, the performance  goal value shifted at LT2, and the rate of change increased at both thresholds.  Within- person variation in blood lactate concentration positively correlated with  the desire to reduce effort and negatively correlated with the performance goal.  Within- person variation in heart rate correlated with desire to reduce effort but  not the performance goal. Conclusion: Transitioning through both lactate thresholds is important phases  for motivation during progressive exercise, particularly for the desire to reduce  effort. Within- person variation in blood lactate concentration is more influential  for motivation, compared with heart rate. K E Y W O R D S desire- goal conflict, exercise domains, lactate threshold, motivation, self- control  | 799 TAYLOR et al. 1  |  INTRODUCTION The ability to endure discomfort is critical in sport and  many  other  areas  of  human  performance,  such  as  the  military and survival in extreme circumstances. It is un- surprising,  therefore,  that  endurance  is  a  popular  topic  of  investigation  in  the  sport,  exercise,  and  performance  sciences.  Having  been  dominated  by  physiological  per- spectives,  our  understanding  of  human  endurance  has  benefited recently from researchers integrating psycholog- ical and physiological perspectives to offer more compre- hensive models of human endurance.1,2 Building on this  trend, the present study investigates the potential physi- ological underpinnings of motivational processes during  an  endurance  task.  Rather  than  defining  endurance  in  terms of time or distance, as is done in applied contexts,  we refer to an endurance act as requiring persistence in  the  face  of  psychological  and  physiological  difficulties.  This definition, therefore, can be applied to many scenar- ios beyond “endurance” classified events (e.g., 10 000 m  athletics event). Endurance  in  performance  contexts  requires  consid- erable  tolerance  and  management  of  psychophysiolog- ical  discomfort.  Afferent  signals  (i.e.,  sensory  impulses  transferred from parts of the body to the central nervous  system) associated with physiological responses (includ- ing those associated with exercise) are affectively labeled3  and  integrated  into  a  collective  representation  of  the  current  physiological  condition  compared  with  homeo- stasis.4 This  depiction  subsequently  emerges  as  a  single  conscious,  motivational  state,5  which  becomes  increas- ingly aversive as a function of exercise intensity.6 Hence,  a desire to reduce effort will evolve because humans have  a proclivity to avoid discomfort7 and maintain homeosta- sis.8 This proximal desire vies with the distal goal of suc- cessful performance. In endurance settings, the content of  the goal may vary, for example, winning a race, achieving  a pre- specified time, or exerting a specific amount of effort  during the task. This content is less important, compared  with the motivational strength of the goal and its conflict  with the desire to reduce effort. Such desire- goal conflicts  represent  a  central  aspect  of  all  self- control  dilemmas9  and  provide  an  empirically  supported  framework  to  in- vestigate endurance performance.10,11 For example, lower  desire to reduce effort at the beginning of a cycling time  trial and slower reductions in goal importance across the  trial have been found to be characteristic of better perfor- mance12 Measures of the desire and goal value at the mid- point of a cycling trial also predicted cycling performance  at a high intensity.12 Within  this  desire- goal  conflict  model  of  endurance,  it is assumed that the underlying basis for the desire to  reduce  effort  is  hedonic  and  stems  from  basic  drives  to  maintain  homeostasis.10  This  implies  that  the  desire  to  reduce  effort  may  have  significant  physiological  under- pinnings. In contrast, the value of the performance goal  is primarily underpinned by internal (e.g., improvement  and personal value) and external (squad selection, prize  money) incentive structures, and physiological responses  to exercise are less influential. Nonetheless, this assump- tion remains to be empirically tested until now. Broadly  speaking,  the  physiological  foundation  of  motivation  during  endurance  acts  is  not  a  new  area  of  research.  Endeavors  have  typically  focused  on  the  rela- tionship between physiological responses to exercise and  perceived exertion or effort.2 Responses include heart rate,  oxygen  uptake,  respiratory  rate,  and  blood  lactate  con- centration, yet no single parameter consistently explains  feelings of exertion.13 The psychobiological model of en- durance proposes that afferent signals from physiological  perturbances have little motivational value because they  are independent of perceived effort.1 This body of work,  therefore, implies that physiological responses to exercise  may have little influence on motivational processes during  endurance. By focusing on perceived effort; however, an  incomplete portrayal of motivation and its physiological  concomitants is presented. Physiological responses to ex- ercise and their generalized core affective labels (i.e., states  that vary simply on pleasantness and activation) are mo- tivationally salient because they form the basis of desires  that  are  often  contrary  to  valued  goals.14,15  Indeed,  the  central purpose of affect associated with afferent bodily  signals is to motivate action.4,16 The desire- goal conflict is  a motivational framework that can assimilate this psycho- physiological knowledge. The  physiological  response  to  exercise  varies  as  a  function  of  the  intensity  at  which  it  is  performed.17  These responses have been characterized into exercise  intensity domains,17,18 which are delineated by specific  physiological  thresholds.  The  moderate  intensity  do- main refers to intensities below the so- called first lactate  threshold, defined as the intensity after which there is a  sustained increase in blood lactate concentration above  resting values (LT1).19,20 This domain is characterized by  a steady state cardiopulmonary response and little or no  sustained  increase  in  blood  lactate  concentration. The  heavy intensity domain refers to intensities above LT1,  but below critical power, which is analogous to the so-  called second lactate threshold when there is a second  rise  in  blood  lactate  concentration  above  resting  lev- els (LT2).21 This domain is characterized by a delayed  steady state and emergence of a V̇O2 slow- component,  which eventually stabilizes after 20– 30 minutes, as well  as a sustained but gradual increase in blood lactate con- centration. During severe domain exercise (at intensities  above LT2), no steady state is achieved, V̇O2 progresses  800 | TAYLOR et al. to  reach  V̇O2max,  and  blood  lactate  concentration  in- creases  progressively.  Evaluation  of  these  domains  is  possible via a progressive exercise test in which blood  lactate  concentration  is  regularly  measured.  This  al- lows estimation of the two primary physiological (lac- tate) thresholds that delineate the moderate- heavy, and  heavy- severe  boundaries,  respectively,  as  well  as  the  subsequent  progression  through  each  of  the  domains.  As  such,  these  two  thresholds  represent  an  opportu- nity to analyze disruptions to the desire- goal conflict at  points when broad metabolic and cardiopulmonary sys- tem responses to exercise become unstable. Complementary  to  inspecting  specific  thresholds  and  domains,  it  is  worth  establishing  whether  the  desire- goal  conflict  is  sensitive  to  continuous  changes  in  metabolic  and  cardiopulmonary  responses  over  the  course  of  an  endurance  act.  Comparing findings  from  the  two  approaches  will  establish  whether  the  desire-  goal  conflict  is  sensitive  to  general  micro- fluctuations  (i.e.,  within- person  variation)  in  physiological  motiva- tional inputs or stronger macro- fluctuations (i.e., lactate  thresholds) are required to disrupt the desire- goal con- flict. In addition to repeated measurements of blood lac- tate concentration as a metabolic response to exercise,  cardiopulmonary responses are also likely to influence  motivational states.22 Heart rate, for example, is known  to produce interoceptive information that informs emo- tional  states.23,24  During  treadmill  exercise,  heart  rate,  unlike other respiratory factors (e.g., V̇O2 and respira- tory exchange ratio), remains stable when perceptions  of effort are fixed,25 albeit the relationship between HR  and  effort  is  likely  correlational  rather  than  causal.13  Collectively,  this  research  implies  that  heart  rate  may  be  an  underlying  input  to  motivational  states  during  exercise. In  sum,  recent  work  has  established  the  desire- goal  conflict as an important framework to study human en- durance.10- 12 The present study aims to build on this work  by  investigating  the  physiological  concomitants  of  the  desire- goal conflict during progressive exercise. It was hy- pothesized that the trajectories of the desire to reduce ef- fort and performance goal value would be disrupted when  participants  transition  through  relevant  physiological  thresholds from the moderate to heavy intensity domain  (hypothesis 1a and 1b), and from heavy to severe intensity  domain (hypothesis 2a and 2b). Specifically, the desire to  reduce effort was expected to shift and/or accelerate pos- itively, whereas the performance goal value was expected  to shift and/or accelerate negatively. Moreover, it was hy- pothesized that within- person variation in blood lactate  and heart rate would positively predict the desire to re- duce effort (hypothesis 3a and 3b) and negatively predict  the performance goal value (hypothesis 4a and 4b). 2  |  MATERIALS & METHODS 2.1  |  Participants Thirty  participants  (14  males,  16  females,  Mage  =  21.03  years;  SD  =  2.06  years)  were  recruited  through a university scheme in which students can partic- ipate in studies for course credit, as well as adverts placed  with university triathlon and cycling teams. Participants  were  required  to  be  18– 35  years  old,  physically  active  (i.e., a minimum of 30 minutes moderate intensity activ- ity three days a week for three months) and free of pre-  existing medical conditions or family history that made  high  intensity  exercise  potentially  unsafe.  Sample  size  targets were based on a minimum of 30 level- 2 units (i.e.,  participants  in  the  present  study)  required  for  minimal  bias in statistical parameters, when combined with at least  five level- 1 units (i.e., measurement points in the present  study) for multilevel modeling.26 2.2  |  Procedure All experimental procedures were approved by a univer- sity ethics approvals committee and conformed with the  Declaration of Helsinki. Participants were fully informed  of study details and the risks and discomforts associated  with all experimental trials. It was clarified that participa- tion was voluntary, data would be stored anonymously,  and they had a right to withdraw at any point during the  study without consequences. Participants provided writ- ten informed consent and completed questionnaires to es- tablish that they met the inclusion criteria (i.e., they were  generally healthy and physically active). Participants were  initially instructed that the goal of the session was to place  as high as possible on a leaderboard containing all partici- pants’ performance scores and, to enhance the meaning- fulness of the goal, they were to set a target position on  the leaderboard. Leaderboards are commonly used to set  meaningful goals in psychology research, especially when  the goal is difficult.27 Participants  performed  an  incremental  cycling  test  on  an  electronically  braked  cycle  ergometer  (Lode  Excalibur Sport, Lode B.V. Gronigen, The Netherlands).  Ergometer saddle and handlebar dimensions were setup  to  suit  individual  specifications.  Participants  cycled  at  a freely chosen pedal rate at an initial work rate of ei- ther  50  or  100  watts  (W)  depending  on  personal  pref- erence  and  experience  (e.g.,  cyclist/triathlete).  This  flexibility was permitted because we were not interested  in performance data (e.g., time to exhaustion), only un- derlying psychological and physiological data during ex- ercise. Work rate was increased by 25 W at four- minute   | 801 TAYLOR et al. intervals. Visual  information  regarding  pedal  rate  and  workload was obscured to avoid participants using this  information to regulate their performance. During the  third  minute  of  each  four- minute  stage,  participants  were presented with measures of their desire to reduce  effort and performance goal value. During the last thirty  seconds of each stage, capillary blood samples were ob- tained for the immediate measurement of blood lactate  concentration.  Heart  rate  was  recorded  at  the  end  of  each stage. The test continued until voluntary exhaus- tion or when two stages had been completed in which  a  clear  increase  in  blood  lactate  concentration  above  4 mmol.L−1 had occurred (to obtain data in the severe  intensity domain), whichever came first. We did not ask  participants to continue beyond two stages in the severe  domain  because  this  was  not  necessary  to  answer  the  research questions. 2.3  |  Measures Desire and goal value. The desire to reduce effort was  measured by verbal responses to the instruction “Please  rate to what extent do you want to reduce your effort” on  a 20- point scale, ranging from 1 (not wanting to reduce  effort at all) to 20 (definitely want to reduce effort imme- diately). The value of the performance goal was measured  by responding to the instruction “Please rate how impor- tant is it to achieve your goal” on a 20- point scale, ranging  from 1 (not important at all) to 20 (extremely important).  Similar  scales  have  demonstrated  predictive  and  nomo- logical validity in previous work.12 Blood lactate concentration, lactate thresholds, and establishment of exercise domains. Capillary blood  samples  were  taken  from  the  earlobe  and  immediately  analyzed  for  blood  lactate  concentration  (Lactate  Pro  2,  Arkray,  Japan).  A  blood  lactate/work  rate  curve  was  modelled  for  each  participant  using  publicly  available  software.28 The work rate corresponding to an initial in- crease of 1 mmol.L−1 above baseline concentration during  the initial stage of the exercise test, and fixed blood lac- tate  concentration  of  4  mmol.L−1  were  defined  as  LT1  and  LT2,  respectively,  and  were  used  to  demarcate  the  moderate, heavy, and severe domains of exercise. Blood  lactate  concentration  is  a  reliable  method  to  determine  exercise intensity,29 generally superior to heart rate30 and  circumvents the need to assess expired gases for V̇O2. Heart rate. Heart rate was continually monitored (T31  transmitter  and  FT1  watch,  Polar  Electro  Oy,  Kempele,  Finland). 2.4  |  Data analysis MLwiN  software  (version  3.0531)  was  used  to  construct  multilevel models to test study hypotheses. This method  was used because of the hierarchical structure of the data  with each measurement of desire, goal value, blood lactate  concentration, and heart rate (Level- 1 time- varying units)  nested within each participant (Level- 2 units32). First, un- conditional  means  models  (i.e.,  no  predictor  variables)  were formed to describe the variance of study variables  associated  with  level- 1  (i.e.,  within- person)  and  level- 2  errors (i.e., between- person). To test hypothesis 1 and 2,  two multilevel growth models (for desire and goal value,  respectively)  were  constructed  by  simultaneously  add- ing a linear time predictor variable (each time point was  coded as 1, 2, 3, etc.), two dichotomous “threshold” pre- dictor variables indexing pre- (coded as 0) to post- (coded  as 1) LT1 and LT2 threshold measures, respectively. In ad- dition, two higher order interaction terms between each  threshold variable and linear time were included. These  models estimate a) the degree of linear change in desire  and goal value over the course of the trial, b) the change  in mean levels as a function of the respective thresholds,  and c) the alterations in rate of change as a function of the  thresholds. To examine hypotheses 3 and 4, two multilevel models  without time and threshold variables included, but with  lactate concentration and heart rate as predictor variables  of desire and goal value. The predictor variables were cen- tered around each participant's unique mean (i.e., group  mean centered), therefore, the models estimated whether  within- person variation (as opposed to individual differ- ences) in blood lactate or heart rate predicted desire to re- duce effort and performance goal value. Variable Lactate Threshold 1 Lactate threshold 2 Mean (SD) Range Mean (SD) Range Desire to reduce effort 4.67 (3.60) 1– 15 7.04 (4.80) 1– 18 Performance goal value 14.67 (4.55) 3– 20 14.22 (4.76) 3– 20 Heart rate (beats.min−1) 151 (14) 119– 175 165 (14) 133– 186 Power (Watts) 155.56 (58.56) 75– 300 184.26 (56.82) 75– 325 TABLE 1  Descriptive statistics of  study variables 802 | TAYLOR et al. 3  |  RESULTS 3.1  |  Descriptive Statistics Descriptive  statistics  for  the  study  variables  at  LT1  and  LT2  can  be  found  in  Table  1.  On  average,  participants  completed 7.57 stages (SD = 1.74), which are equivalent  to approximately 30 minutes of work. Twenty- eight par- ticipants cycled until voluntary exhaustion, and two par- ticipants were stopped twice by an investigator because  two stages in the severe intensity domain had been com- pleted. Average heart rate at LT1 and LT2 was 83 percent  and 91 percent, respectively, of participants’ heart rate at  task cessation. Average workload at LT1 and LT2 was 68  percent and 81 percent, respectively, of participants’ peak  workload. Unconditional means models revealed that 20  percent of the variance in performance goal value was at- tributable to within- person variation, and 80 percent at- tributable to individual variation. In contrast, 85 percent  of the variance in desire to reduce effort was attributable  to within- person variation, and 15 percent was attribut- able to individual variation. 3.2  |  Trajectories of desire to reduce effort and performance goal value across intensity domains (hypotheses 1 and 2) In Table 2, multilevel growth model 1 describes the trajec- tory of the desire to reduce effort over the trial, therefore,  investigating hypotheses 1a and 2a. The model indicated  that the desire to reduce effort generally increased over  the course of the trial (linear time coefficient), shifted at  LT1  and  LT2  (threshold  coefficients),  and  accelerated  after each threshold (interaction terms). This pattern in  desire to reduce effort is illustrated in Figure 1, which de- scribes predicted trajectories in desire and goal value over  time and when participants transition through LT1 and  LT2. Variance terms indicated that the linear increase and  the shift at the LT2 varied in magnitude across individu- als, but the shift at the LT1 did not. When allowing the  interaction terms to vary across individuals the model did  not converge; hence, these parameters are not reported. Model 2 describes the trajectory of performance goal  value  over  the  trial,  therefore,  investigating  hypotheses  1b and 2b. The model revealed that the performance goal  value did not linearly decline over the course of the trail  (linear time coefficient), nor did it shift at LT1 (threshold  coefficient). However, the performance goal value shifted  at LT2 (threshold coefficient), and the rate of change was  disturbed  at  both  thresholds  (interaction  terms).  This  pattern in goal value across the course of the trial is illus- trated in Figure 1. Variance terms indicated that the linear  change, but not the shift at LT2, significantly varied across  individuals. When allowing the remaining effects to vary  across  individuals,  the  model  did  not  converge;  hence,  these parameters are not reported. 3.3  |  Within- person variation of blood lactate concentration and heart rate predicting desire and goal value (hypotheses 3 and 4) Multilevel  models  revealed  that  within- person  varia- tion  in  blood  lactate  concentration  positively  correlated  with the desire to reduce effort (b = 0.69, p < .001) and  negatively  correlated  with  the  performance  goal  value  (b = −0.15, p = .02). A small positive correlation between  within- person variation in heart rate and desire to reduce  effort (b = 0.05, p < .001) was observed, but no significant  relationship  between  heart  rate  and  performance  goal  value (b = 0.00, p = .37). 4  |  DISCUSSION The  desire- goal  conflict  has  been  proposed  as  a  valid  framework to study motivational dynamics during en- durance  performance,10  but  potential  underpinning  TABLE 2  Multilevel growth models describing desire to reduce  effort and performance goal value across LT1 and LT2 thresholds Outcome Desire to reduce effort (model 1) Goal value (model 2) Fixed Effects (SE in parentheses) Intercept −0.37 (0.37) 14.84 (0.99) Linear time 1.09 (0.21) 0.01 (0.14) LT1 threshold −3.50 (0.99) 1.21 (0.80) Time ×LT1 threshold 0.84 (0.20) −0.40 (0.16) LT2 threshold −5.26 (1.34) 2.08 (1.01) Time ×LT2 threshold 1.04 (0.22) −0.39 (0.18) Variance Terms Intercept 1.20 (0.84) 27.92 (7.49) Linear time 0.85 (0.30) 0.39 (0.13) LT1 threshold 1.21 (1.34) Time ×LT1 threshold LT2 threshold 5.31 (2.40) 1.76 (1.03) Time ×LT2 threshold Note: Bold figures indicate statistical significance (p ≤ .05). Exact values can  be calculated from the Z scores (b/SE for the fixed effects). Variance terms  are not reported in instances when the predictor variable was not permitted  to vary across individuals to aid convergence of models. LT = Lactate  Threshold.  | 803 TAYLOR et al. physiological  concomitants  were  unknown  until  now.  Two different blood lactate thresholds were inspected to  observe changes in the desire to reduce effort and value  of a performance goal at these periods, which are char- acterized by physiological instability. Moreover, within-  person  variation  in  blood  lactate  concentration  and  heart rate was examined as correlates of the desire and  performance goal value. Results indicated that progres- sion from the moderate through to the heavy and severe  domains of exercise by transitioning through LT1 and  LT2, respectively, are points in which motivational dy- namics are perturbed, particularly the desire to reduce  effort. Variation in blood lactate concentrations was a  stronger correlate of the desire to reduce effort and per- formance goal value, compared with variation in heart  rate. In  any  activity  requiring  self- control,  the  desire  to  stop will increase as a function of time.33 Unsurprisingly,  therefore, a general increase in the desire to reduce effort  was observed in the present study. As expected, however,  this desire deviated from its trajectory and began to ac- celerate when the LT1 occurred, and participants entered  the heavy exercise domain. The same picture is presented  regarding LT2 and entry into the severe exercise domain.  At this point, the desire to reduce effort shifts again and  further  accelerates.  Previous  research  has  not  modeled  changes in motivational factors as a function of these two  physiological thresholds. Nonetheless, core affect has been  shown to be stable below the ventilatory threshold (which  typically occurs at a similar exercise intensity as the LT113)  but  become  increasingly  negative  after  it.34  Our  results  align with the idea that hedonic motivational factors that  encourage avoidance of unpleasant states, such as nega- tive affect and desire to reduce effort, become increasingly  powerful when aversive physiological responses accumu- late. The causal pathway linking physiological responses  and generalized core affective labels,3 which subsequently  manifest into a motivational state,5 were not investigated  here but seems the most plausible explanation for this pat- tern of findings. Both  physiological  thresholds  also  have  implications  for the value of performance goal. The value of the goal  did not shift at LT1; however, after a period of stability it  began to significantly decline at this point. LT2 (and entry  into the severe intensity domain) coincided with a shift  in level and a change in rate of decline. The smaller re- gression coefficients compared with the desire to reduce  effort imply that underlying physiological responses con- tribute less to the performance goal value. Nonetheless, it  may be inaccurate to suggest that the performance goal  value  is  entirely  underpinned  by  more  stable  reflective  factors,  such  as  internal  and  external  incentives.10  The  lactate thresholds may signify that goal achievement is be- coming increasingly difficult, therefore, reductions in goal  value may occur to protect self- worth (i.e., defensive pes- simism35). The comparable (albeit in opposing directions)  disturbances of desire and goal value at the two lactate  thresholds imply that they are not isolated motivational  components  with  distinct  concomitants,  but  they  share  some physiological correlates during progressive exercise. The primary aim of some types of endurance training  is to delay reaching the lactate thresholds and associated  physiological consequences resulting in fatigue. The col- lective findings of the present study imply that achieving  this  aim  also  has  important  motivational  ramifications.  These  physiological  thresholds  are  associated  with  in- creases in the motivational potency of the desire to reduce  effort and simultaneous decreases in the value of the per- formance goal. As such, the relative weight of each moti- vational component shifts in favor of the proximal desire,  and diminished endurance performance occurs.12 Positive  motivational consequences can, therefore, be added to the  list of benefits resulting from delayed lactate thresholds. In  addition  to  the  examination  of  physiological  thresholds  and  intensity  domains,  the  results  demon- strated  that  within- person  variation  in  blood  lactate  FIGURE 1 Illustrative trajectories of  desire and performance goal value during  the cycling trial 804 | TAYLOR et al. concentration  was  positively  associated  with  the  de- sire to reduce effort and negatively associated with the  performance goal value. The relationship was stronger  with the desire to reduce effort, compared with the per- formance goal value, again implying that blood lactate  concentration is more salient for the hedonic desire to  reduce  effort  than  the  performance  goal  value.  A  rel- atively  smaller  positive  correlation  between  within-  person variation in heart rate and desire to reduce effort  was observed, and no significant relationship between  heart rate and performance goal value. Hence, the mag- nitude  of  the  regression  coefficients  suggests  that  the  relationship between lactate and motivation is stronger  than heart rate and motivation. Previous work offers no  definitive conclusion regarding the relative coupling of  heart rate and lactate with motivational factors. Some  evidence  suggest  perceived  exertion  is  correlated  to  a  greater  extent  with  heart  rate  compared  with  lactate  concentrations.36 However, the present analysis focuses  on within- person variation, rather than absolute levels,  which  makes  comparison  with  previous  research  dif- ficult. It is likely that deviations of physiological state  within  an  individual  are  more  likely  to  stimulate  mo- tivational  responses  (within- person  variation),  as  op- posed to individual differences leading to corresponding  differences in motivation (between- person differences).  The impact of lactate concentration on motivational fac- tors is likely not direct, of course, but indirect via pH  changes  associated  with  lactic  acid  production.37  The  affective sensation associated with this process may be  more potent, given that heart rate interoceptive signals  can be easily suppressed in favor of alternative stimuli.38 4.1  |  Limitations and Future Research Directions Future research should experimentally manipulate un- derpinning  physiological  states  to  establish  causal  ef- fects  on  the  desire- goal  conflict.  This  can  be  achieved  acutely through, for example, nutritional supplementa- tion  such  as  prior  ingestion  of  sodium  bicarbonate  or  ammonium  chloride  to  induce  metabolic  alkalosis  or  acidosis, respectively,39 or chronically through exercise  training. A familiarization trial would be necessary to  include in this type of experiment, which did not occur  in  the  present  study.  In  addition,  a  measure  of  affect  alongside  the  desire  to  reduce  effort  would  provide  a  more detailed examination of the assumed relationship  between  affective  responses  and  the  desire  to  reduce  effort. Third, our healthy and active sample may limit  the generalizability of some of the findings in the study.  For example, participants with experience of strenuous  activity  may  appraise  the  associated  sensations  of  lac- tate accumulation less negatively, compared with an un- healthy and inactive sample. The prevailing method of analyzing physiological re- sponses to exercise is to treat each response as a directly  observed  variable  working  independently.  Instead,  each  physiological  parameter  can  be  used  as  a  composite  la- tent variable describing overall homeostatic disturbance.  That is, each individual parameter contributes to a higher  order construct, but they do not cause any motivational  disturbances on their own. This latent factor approach is  common in psychology but would be physiologically inno- vative and align with the idea that system- wide responses  to exercise manifest into an overall gauge of homeostatic  integrity.5 Future research should also examine how to delay the  desire- goal  conflict,  particularly  when  the  desire  begins  to  overcome  the  performance  goal  value.  Delay  can  be  achieved by lowering the desire to reduce effort, increasing  the magnitude of the performance goal value, or a com- bination of these strategies. Existing work has suggested  that enhancing the congruence between the performance  goal and one's identity12 can disturb the desire- goal con- flict favorably. This approach fits with other motivational  frameworks, such as the identity- value model40 and self-  determination  theory.41  Other  strategies  have  potential,  such as enhancing the congruence between actively man- aging discomfort and the performance goal (i.e., means-  end fusion42). 4.2  |  Perspective The  desire- goal  conflict  is  a  recently  applied  motiva- tional framework that helps explain endurance perfor- mance.10,12 Unlike most other models of endurance that  considers  motivation  as  a  unidimensional  construct,1  motivation  is  viewed  as  a  network  of  constructs  and  related processes. Moreover, it has the potential to rec- oncile  different  physiological  and  psychological  ideas,  such as the role of physiological responses to exercise  and their affective labels in shaping motivational pro- cesses.3,4 Overall, the present study represents the first  analysis of how the desire- goal conflict is disturbed by  two  physiological  thresholds  (i.e.,  LT1  and  LT2,  and  how  these  delineate  fundamental  exercise  intensity  domains).  These  thresholds,  therefore,  represent  im- portant  points  when  motivational  factors  become  in- creasingly detrimental to performance and intervention  is required. In addition, analyzing within- person varia- tion in blood lactate and heart rate may portray a clearer  picture of their relationship with motivation constructs,  compared with absolute levels.  | 805 TAYLOR et al. DATA AVAILABILITY STATEMENT The data that support the findings of this study are avail- able from the corresponding author (IT), upon reasonable  request. ORCID Ian M. Taylor   https://orcid.org/0000-0001-6291-4025  Richard A. Ferguson   https://orcid. org/0000-0002-2508-8358  REFERENCES   1.  Marcora  S.  Counterpoint:  Afferent  Feedback  From  Fatigued  Locomotor  Muscles  Is  Not  An  Important  Determinant  Of  Endurance  Exercise  Performance.  J Appl Physiol.  (1985).  2010;108(2):454- 457. doi:10.1152/jappl physi ol.00976.2009a   2.  Pageaux  B.  The  psychobiological  model  of  endurance  per- formance:  an  effort- based  decision- making  theory  to  ex- plain  self- paced  endurance  performance.  Sports Med.  2014;44(9):1319- 1320. doi:10.1007/s4027 9- 014- 0198- 2   3.  Critchley  HD,  Harrison  NA.  Visceral  influences  on  brain  and  behavior.  Neuron.  2013;77(4):624- 638.  doi:10.1016/j. neuron.2013.02.008   4.  Craig  AD.  An  interoceptive  neuroanatomical  perspective  on  feelings, energy, and effort. Behav Brain Sci. 2013;36(6):685- 726.  doi:10.1017/S0140 525X1 3001489   5.  Ramirez JM, Cabanac M. Pleasure, the common currency of  emotions. Ann N Y Acad Sci. 2003;1000:293- 295. doi:10.1196/ annals.1280.028   6.  Ekkekakis P, Parfitt G, Petruzzello SJ. The pleasure and displea- sure people feel when they exercise at different intensities: de- cennial update and progress towards a tripartite rationale for  exercise intensity prescription. Sports Med. 2011;41(8):641- 671.  doi:10.2165/11590 680- 00000 0000- 00000   7.  Rozin P. Preadaptation and the puzzles and properties of plea- sure. In: Kahnemann D, Diener E, Schwartz N, eds. Well- being: The Foundations of Hedonic Psychology. Sage; 1999:109- 133.   8.  St Clair Gibson A, Swart J, Tucker R. The interaction of psycho- logical and physiological homeostatic drives and role of general  control  principles  in  the  regulation  of  physiological  systems,  exercise  and  the  fatigue  process  -  The  Integrative  Governor  theory.  Eur J Sport Sci.  2018;18(1):25- 36.  doi:10.1080/17461  391.2017.1321688   9.  Kotabe  HP,  Hofmann  W.  On  Integrating  the  Components  of  Self- Control.  Perspect Psychol Sci.  2015;10(5):618- 638.  doi:10.1177/17456 91615 593382  10.  Taylor IM. A motivational model of endurance and persistence.  In: Englert C, Taylor IM, eds. Motivation & Self- Regulation in Sport and Exercise. Routledge; 2021:87- 101.  11.  Taylor  IM,  Boat  R,  Murphy  SL.  Integrating  theories  of  self-  control and motivation to advance endurance performance. Int Rev of Sport Exerc Psychol. 2020;13(1):1– 20. doi:10.1080/17509  84X.2018.1480050  12.  Taylor  IM,  Smith  K,  Hunte  R.  Motivational  processes  during  physical  endurance  tasks.  Scand J Med Sci Sports.  2020;30(9):1769- 1776. doi:10.1111/sms.13739  13.  Hampson  DB,  St  Clair  Gibson  A,  Lambert  MI,  Noakes  TD.  The  influence  of  sensory  cues  on  the  perception  of  exertion  during exercise and central regulation of exercise performance.  Sports Med.  2001;31(13):935- 952.  doi:10.2165/00007 256- 20013  1130- 00004  14.  Ekkekakis P, Brand R. Affective responses to and automatic  affective valuations of physical activity: Fifty years of prog- ress  on  the  seminal  question  in  exercise  psychology.  Psych of Sport & Exercise.  2019;42:130- 137.  doi:10.1016/j.psych  sport.2018.12.018  15.  Loewenstein G. Out of control: Visceral influences on behav- ior.  Org Behavior Human Dec Processes.  1996;65(3):272- 292.  doi:10.1006/obhd.1996.0028  16.  Carver  CS,  Scheier  MF.  On  the  structure  of  behavioral  self-  regulation.  In:  Boekaerts  M,  Pintrich  PR,  Zeidner  M,  eds.  Handbook of Self- Regulation. Academic Press; 2000:41- 84.  17.  Burnley M, Jones AM. Power- duration relationship: Physiology,  fatigue, and the limits of human performance. Eur J Sport Sci.  2018;18(1):1- 12. doi:10.1080/17461 391.2016.1249524  18.  Whipp BJ. Domains of aerobic function and their limiting pa- rameters. In: Steinacker JM, Ward SA, eds. The Physiology and Pathophysiology of Exercise Tolerance. Plenum; 1996:83- 89.  19.  Coyle EF, Coggan AR, Hopper MK, Walters TJ. Determinants  of  endurance  in  well- trained  cyclists.  J Appl Physiol.  (1985).  1988;64(6):2622- 2630. doi:10.1152/jappl.1988.64.6.2622  20.  Faude  O,  Kindermann  W,  Meyer  T.  Lactate  threshold  con- cepts:  how  valid  are  they?  Sports Med.  2009;39(6):469- 490.  doi:10.2165/00007 256- 20093 9060- 00003  21.  Valenzuela PL, Alejo LB, Montalvo- Pérez A, et al. Relationship  Between  Critical  Power  and  Different  Lactate  Threshold  Markers  in  Recreational  Cyclists.  Front Physiol.  2021;12:  doi:10.3389/fphys.2021.676484  22.  Scherr  J,  Wolfarth  B,  Christle  JW,  Pressler  A,  Wagenpfeil  S,  Halle M. Associations between Borg's rating of perceived exer- tion and physiological measures of exercise intensity. Eur J Appl Physiol. 2013;113(1):147- 155. doi:10.1007/s0042 1- 012- 2421- x  23.  Critchley HD, Garfinkel SN. Interoception and emotion. Curr Opin Psychol. 2017;17:7- 14. doi:10.1016/j.copsyc.2017.04.020  24.  Hassanpour MS, Yan L, Wang DJ, et al. How the heart speaks  to  the  brain:  neural  activity  during  cardiorespiratory  in- teroceptive  stimulation.  Philos Trans R Soc Lond B Biol Sci.  2016;371(1708):20160017. doi:10.1098/rstb.2016.0017  25.  Cochrane- Snyman KC, Housh TJ, Smith CM, Hill EC, Jenkins  NDM. Treadmill running using an RPE- clamp model: media- tors  of  perception  and  implications  for  exercise  prescription.  Eur J Appl Physiol.  2019;119(9):2083- 2094.  doi:10.1007/s0042  1- 019- 04197 - 4  26.  Maas  CJM,  Hox  JJ.  Sufficient  Sample  Sizes  for  Multilevel  Modeling.  Methodol.  2005;1(3):86- 92.  doi:10.1027/161 4- 1881.1.3.86  27.  Landers RN, Bauer KN, Callan RC. Gamification of task perfor- mance with leaderboards: A goal setting experiment. Comput Human Behav. 2017;71:508- 515. doi:10.1016/j.chb.2015.08.008  28.  Newell J, Higgins D, Madden N, et al. Software for calculating  blood lactate endurance markers. J Sports Sci. 2007;25(12):1403-  1409. doi:10.1080/02640 41060 1128922  29.  Beneke R, Leithäuser RM, Ochentel O. Blood lactate diagnos- tics in exercise testing and training. Int J Sports Physiol Perform.  2011;6(1):8- 24. doi:10.1123/ijspp.6.1.8  30.  Goodwin  ML,  Harris  JE,  Hernández  A,  Gladden  LB.  Blood  lactate  measurements  and  analysis  during  exercise:  a  guide  for  clinicians.  J Diabetes Sci Technol.  2007;1(4):558- 569.  doi:10.1177/19322 96807 00100414 806 | TAYLOR et al.  31.  Rasbash J, Browne WJ, Healy M, Cameron B, Charlton C. A User’s Guide to MLwiN. University of Bristol; 2015:296.  32.  Snijders  TAB,  Bosker  RJ.  Multilevel analysis: An introduc- tion to basic and advanced multilevel modeling (2nd ed). Sage;  2012:368.  33.  Bieleke M, Wolff W. It's not a bug, it's boredom: Effortful will- power balances exploitation and exploration. Behav Brain Sci.  2021;44:e33. doi:10.1017/S0140 525X2 0001053. Published 2021  Apr 26.  34.  Hall EE, Ekkekakis P, Petruzzello SJ. The affective beneficence  of vigorous exercise revisited. Br J Health Psychol. 2002;7(1):47-  66. doi:10.1348/13591 07021 69358  35.  Norem JK, Cantor N. Defensive pessimism: harnessing anxiety  as motivation. J Pers Soc Psychol. 1986;51(6):1208- 1217. doi:10. 1037//0022- 3514.51.6.1208  36.  Green  JM,  McLester  JR,  Crews  TR,  Wickwire  PJ,  Pritchett  RC,  Lomax  RG.  RPE  association  with  lactate  and  heart  rate  during  high- intensity  interval  cycling.  Med Sci Sports Exerc.  2006;38(1):167- 172. doi:10.1249/01.mss.00001 80359.98241.a2  37.  Swank A, Robertson RJ. Effect of induced alkalosis on percep- tion  of  exertion  during  intermittent  exercise.  J Appl Physiol.  (1985). 1989;67(5):1862- 1867. doi:10.1152/jappl.1989.67.5.1862  38.  Salomon R, Ronchi R, Dönz J, et al. The Insula Mediates Access  to  Awareness  of  Visual  Stimuli  Presented  Synchronously  to  the Heartbeat. J Neurosci. 2016;36(18):5115- 5127. doi:10.1523/ JNEUR OSCI.4262- 15.2016  39.  Carr  AJ,  Hopkins  WG,  Gore  CJ.  Effects  of  acute  alkalosis  and  acidosis  on  performance:  a  meta- analysis.  Sports Med.  2011;41(10):801- 814. doi:10.2165/11591 440- 00000 0000- 00000  40.  Berkman  ET,  Livingston  JL,  Kahn  LE.  Finding  the  “self”  in  self- regulation:  The  identity- value  model.  Psychol Inq.  2017;28(2– 3):77- 98. doi:10.1080/10478 40X.2017.1323463  41.  Ryan RM, Deci EL. Brick by brick: The origins, development,  and  future  of  self- determination  theory.  In:  Elliot  AJ,  ed.  Advances in Motivation Science. Elsevier; 2019:111- 156.  42.  Kruglanski  AW,  Fishbach  A,  Woolley  K,  et  al.  A  structural  model  of  intrinsic  motivation:  On  the  psychology  of  means-  ends  fusion.  Psychol Rev.  2018;125(2):165- 182.  doi:10.1037/ rev00 00095 How to cite this article: Taylor IM, Whiteley S,  Ferguson RA. Disturbance of desire- goal  motivational dynamics during different exercise  intensity domains. Scand J Med Sci Sports.  2022;32:798– 806. doi:10.1111/sms.14129

Disturbance of desire-goal motivational dynamics during different exercise intensity domains.

01-27-2022

Taylor, Ian M,Whiteley, Summer,Ferguson, Richard A

eng

PMC9012817

1 Vol.:(0123456789) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports Non‑South East Asians have a better running economy and different anthropometrics and biomechanics than South East Asians Aurélien Patoz1,2*, Thibault Lussiana2,3,4, Bastiaan Breine2,5, Cyrille Gindre2,3, Laurent Mourot4 & Kim Hébert‑Losier6,7 Running biomechanics and ethnicity can influence running economy (RE), which is a critical factor of running performance. Our aim was to compare RE of South East Asian (SEA) and non‑South East Asian (non‑SEA) runners at several endurance running speeds (10–14 km/h) matched for on‑road racing performance and sex. Secondly, we explored anthropometric characteristics and relationships between RE and anthropometric and biomechanical variables. SEA were 6% less economical (p = 0.04) than non‑SEA. SEA were lighter and shorter than non‑SEA, and had lower body mass indexes and leg lengths (p ≤ 0.01). In terms of biomechanics, a higher prevalence of forefoot strikers in SEA than non‑SEA was seen at each speed tested (p ≤ 0.04). Furthermore, SEA had a significantly higher step frequency (p = 0.02), shorter contact time (p = 0.04), smaller footstrike angle (p < 0.001), and less knee extension at toe‑off (p = 0.03) than non‑SEA. Amongst these variables, only mass was positively correlated to RE for both SEA (12 km/h) and non‑SEA (all speeds); step frequency, negatively correlated to RE for both SEA (10 km/h) and non‑SEA (12 km/h); and contact time, positively correlated to RE for SEA (12 km/h). Despite the observed anthropometric and biomechanical differences between cohorts, these data were limited in underpinning the observed RE differences at a group level. This exploratory study provides preliminary indications of potential differences between SEA and non‑SEA runners warranting further consideration. Altogether, these findings suggest caution when generalizing from non‑SEA running studies to SEA runners. Running economy (RE), which refers to steady-state oxygen consumption at a given submaximal running speed, is a critical factor of running performance1. RE has been shown to differ between ethnic groups2–5. Indeed, Weston, et al.2 noted greater RE in African than Caucasian distance runners though not elucidating the origin of these differences. Similarly, elite Kenyans were found more economical than their Caucasian counterparts3–5. This difference was attributed to body dimensions, with longer legs (~ 5%), thinner and lighter calf musculature, as well as lower body mass and body mass index (BMI) in Kenyans than Caucasians, but not to differences in muscle fibre type3–6. These findings may partially explain the success of African runners at the elite level. Indeed, the longer, slenderer legs of Kenyans could be advantageous when running as RE is correlated with leg mass6. However, the precise mechanisms underpinning anthropometric and economy relationships are not clear7. OPEN 1Institute of Sport Sciences, University of Lausanne, 1015 Lausanne, Switzerland. 2Research and Development Department, Volodalen Swiss Sport Lab, 1860 Aigle, Switzerland. 3Research and Development Department, Volodalen, 39270 Chavéria, France. 4Research Unit EA3920 Prognostic Markers and Regulatory Factors of Cardiovascular Diseases and Exercise Performance, Health, Innovation Platform, University of Bourgogne Franche-Comté, 2500 Besançon, France. 5Department of Movement and Sports Sciences, Ghent University, 9000 Ghent, Belgium. 6Division of Health, Engineering, Computing and Science, Te Huataki Waiora School of Health, Adams Centre for High Performance, University of Waikato, Tauranga 3116, New Zealand. 7Department of Sports Science, National Sports Institute of Malaysia, 57000 Kuala Lumpur, Malaysia. *email: aurelien.patoz@ unil.ch 2 Vol:.(1234567890) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ Research into running and ethnic differences has mostly compared Caucasian and African runners2–5,8–13. These studies highlight differences in physiological2–5,12, anthropometrical9,14, neuromuscular15, and running gait patterns8,10,11 between ethnicities. Altogether, these results indicate caution in the generalization of results from one ethnic group to another. There exists only limited inclusion of Asian cohorts in running studies14–16 and, to the best of our knowledge, no study comparing their RE to another ethnic group. Nonetheless, road race participation continues to grow in Asia despite a decline in the number of participants since 2016 outside of Asia17. Therefore, the relative underrep- resentation of Asian runners in research is of concern, especially when considering their unique anthropometric features18,19, autonomic responses to exercise20, muscle–tendon unit properties15, walking gait characteristics21, and footstrike patterns16 compared to other ethnic groups. Although running biomechanics can influence RE1, the relationships between select biomechanical vari- ables and RE are unclear and even conflicting in the scientific literature. For instance, Gruber, et al.22 observed no difference in RE between rearfoot (RFS) and non-rearfoot (non-RFS) strike patterns, while both RFS23 and non-RFS24 patterns were suggested as more economical than the other. Similarly, superior RE has been linked with both long25 and short26 ground contact times (tc), while Williams and Cavanagh27 found no significant relation between RE and tc. These divergent findings might be due to differences between the cohorts examined, including ethnic differences. For these reasons, our primary aim was to explore whether South East Asian (SEA) and non-South East Asian (non-SEA) runners demonstrate similar RE at several endurance running speeds when matched for on-road running performance and sex. Secondly, we aimed to explore anthropometric differences between groups and potential relationships between RE and anthropometric and biomechanical variables in these groups. Materials and methods Participants. An existing database of 54 runners was explored to match SEA and non-SEA runners based on sex and on-road running performance on 21.1  km28. The matching led to the inclusion of 34 trained run- ners, 20 males (variable: mean ± standard deviation, age: 36 ± 6 years, mass: 68 ± 11 kg, height: 176 ± 7 cm, leg length: 92 ± 5 cm, BMI: 22 ± 2 kg/m2, running distance: 56 ± 20 km/week, running experience: 9 ± 7 y, and best half-marathon time: 93 ± 9 min) and 14 females (age: 36 ± 6 y, mass: 53 ± 6 kg, height: 162 ± 4 cm, leg length: 84 ± 3 cm, BMI: 20 ± 2 kg/m2, running distance: 58 ± 17 km/week, running experience: 7 ± 5 years, and best half- marathon time: 100 ± 9 min) in this study. For study inclusion, participants were required to be in good self- reported general health with no current or recent (< 3 months) musculoskeletal injuries and to meet a certain level of running performance. More specifically, runners were required to have competed in a road race in the last year with finishing times of ≤ 50 min for 10 km, ≤ 1 h 50 min for 21.1 km or ≤ 3 h 50 min for 42.2 km. The ethical committee of the National Sports Institute of Malaysia approved the study protocol prior to participant recruitment (ISNRP: 26/2015), which was conducted in accordance with international ethical standards29 and adhered to the latest Declaration of Helsinki of the World Medical Association. Runners were classified in two ethnic groups based on their nationality: SEA and non-SEA, which led to a total of 17 participants per group. SEA runners were from China (n = 12), Malaysia (n = 14), and Indonesia (n = 1); while non-SEA runners were from England (n = 7), Sweden (n = 2), Australia, Brazil, Canada, Denmark, France, Norway, Poland, and Scotland (n = 1 each). All non-SEA runners identified as “white”. Experimental procedure. Each participant completed one experimental laboratory session. After provid- ing written informed consent, the right leg length of participants was measured (from anterior superior iliac spine to medial malleolus in supine). Participants then ran 5 min at 9 km/h on a treadmill (h/p/cosmos mer- cury®, h/p/cosmos sports & medical gmbh, Nussdorf-Traunstein, Germany) as a warm-up. Participants then completed 3 × 4-min runs at 10, 12, and 14 km/h (with 2-min recovery periods between runs) on the tread- mill, during which time RE was assessed. Retro-reflective markers were subsequently positioned on individuals (described in Data Collection section) to assess running kinematics. For each participant, a 1-s static calibration trial was recorded, which was followed by 3 × 30-s runs at 10, 12, and 14 km/h (with 1-min recovery periods between each runs) to collect three-dimensional (3D) kinematic data in the last 10-s segment of these runs (30 ± 2 running steps), resulting in at least 25 steps being analysed30. RE and biomechanics were assessed sepa- rately given laboratory constraints and interference with data quality (e.g., presence of testing equipment that occluded markers). All participants were familiar with running on a treadmill as part of their usual training programs and wore their habitual running shoes during testing. Data collection. Gas exchange was measured using TrueOne 2400 (ParvoMedics, Sandy, UT, USA) during the 3 × 4-min runs. Prior to the experiment, the gas analyzer was calibrated using ambient air (O2: 20.93% and CO2: 0.03%) and a gas mixture of known concentration (O2: 16.00% and CO2: 4.001%). Volume calibration was performed at different flow rates with a 3 L calibration syringe (5530 series, Hans Rudolph, Shawnee, KS, USA). Oxygen consumption ( ˙VO2 ), carbon dioxide production ( ˙VCO2 ), and respiratory exchange ratio (RER) values were averaged over the last minute of each 4-min run. Steady state was confirmed through visual inspection of the ˙VO2 and ˙VCO2 curves for all running trials. RER had to remain below unity during the trials for data to be included in the analysis, otherwise the corresponding data were excluded as deemed to not represent a submaxi- mal effort. No trial was excluded on this basis. RE was expressed as the oxygen cost per mass to the power of 0.75 per kilometer (ml/kg0.75/km) to minimize the influence of body mass per se on ˙VO2 during running31. RE expressed in ml/kg/km was also computed for reference and is provided as supplementary materials. A higher RE value indicates a less economical runner. 3 Vol.:(0123456789) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ 3D kinematic data were collected at 200 Hz using seven infrared Oqus cameras (five Oqus 300+, one Oqus 310+, and one Oqus 311+) and Qualisys Track Manager software version 2.1.1 build 2902 together with the Pro- ject Automation Framework Running package version 4.4 (Qualisys AB, Göteborg, Sweden). A virtual laboratory coordinate system was generated such that the x–y–z axes denoted the medio-lateral (pointing towards the right side of the body), posterior-anterior, and inferior-superior directions, respectively. Thirty-five retro-reflective markers (Fig. 1) of 12 mm in diameter were used for static calibration and running trials, and were affixed to the skin and shoes of individuals over anatomical landmarks using double-sided tape following standard guide- lines from the Project Automation Framework Running package32. The 3D marker data were exported in .c3d format and processed in Visual3D Professional software version 5.02.25 (C-Motion Inc., Germantown, MD, USA). More explicitly, the 3D marker data were interpolated using a third-order polynomial least-square fit algorithm, allowing a maximum of 20 frames for gap filling, and subsequently low-pass filtered at 20 Hz using a fourth-order Butterworth filter. Biomechanical variables. From the marker set, a full-body biomechanical model with six degrees of free- dom at each joint and 15 rigid segments was constructed. The model included the head, upper arms, lower arms, hands, thorax, pelvis, thighs, shanks, and feet. Segments were assigned inertial properties and centre of mass (COM) locations based on their shape33 and attributed relative mass based on standard regression equations34. Kinematic variables were calculated using rigid-body analysis and whole-body COM location was calculated from the parameters of all 15 segments. Ankle ( θankle ) and knee ( θknee ) joint angles were defined as the orientation of the distal segment relative to the proximal one35. Angles were computed using an x–y–z Cardan sequence36,37 equivalent to the joint coordinate system36,38, leading to rotations with functional and ana- tomical meaning (flexion–extension, abduction–adduction, and internal–external rotation). Noteworthy, only the flexion–extension Cardan angle was considered for analysis due to possible errors linked with kinematic crosstalk39–41. Joint angles were calculated at footstrike and toe-off events. Footstrike angle (FSA) was calculated following the procedure described in Altman and Davis42. FSA was normalized by taking the angle of the foot at footstrike and subtracting the angle of the foot during standing trial. The mean FSA was used to categorise footstrike patterns of runners in two categories: RFS when the FSA was greater than 8°, and non-RFS when 8° or less42. Among all running trials, 5% and 7% were borderline (within 1°) RFS and non-RFS, respectively. These borderline footstrike patterns were only present in SEA runners. Visual inspection confirmed the footstrike pat- tern classifications were correct. Running events were derived from the trajectories of the 3D marker data using similar procedures to those previously reported 43,44. All events were verified to ensure correct identification and were manually adjusted when required. Swing time (ts) and tc were defined as the time from toe-off to footstrike and from footstrike to toe-off of the same foot, respectively. Flight time (tf) was defined as the time from toe-off to footstrike of the contralateral foot. Step frequency (SF) was calculated as SF = 1 tc+tf , and step length (SL) as SL = s/SF , where s represents running speed. In addition to raw units, SL was expressed as a percentage of participant’s leg length. The spring-mass characteristics of the lower limb were estimated using a sine-wave model following the procedure defined by Morin, et al.45. More explicitly, leg stiffness (kleg) was calculated as [Eq. (1)] Figure 1. Retro-reflective markers (N = 35) placed on anatomical landmarks of participants for biomechanical data collection. R and L at the start of the acronyms denote right and left, respectively. 4 Vol:.(1234567890) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ where Fz,max represents the maximal vertical force and was estimated using Fz,max = mg π 2  tf tc + 1  , L is the maximal leg length deformation, i.e., the leg spring compression and given by L =  z2 COM,FS + s2t2 b − zCOM,MS , where s defines running speed, tb denotes the braking time, i.e., the time from footstrike to mid-stance, and zCOM,FS and zCOM,MS are the COM heights at footstrike and mid-stance, respectively. For all biomechanical meas- ures, the values extracted from the 10-s data collection for each participant were averaged for subsequent statisti- cal analyses. Statistical analysis. Descriptive statistics are presented using mean ± standard deviation (SD). Data nor- mality and homogeneity of variances were verified using Kolmogorov–Smirnov and Levene’s test, respectively. Participant characteristics between SEA and non-SEA runners were compared using unpaired two-sided Welch’s t-tests when homogeneity of variance assumptions were violated and unpaired two-sided Student’s t-tests oth- erwise. The effect of group (SEA, non-SEA) and running speed on RE and biomechanical variables was evalu- ated using a linear mixed effects model fitted by restricted maximum likelihood. The within-subject nature was controlled for by including random effects for participants (individual differences in the intercept of the model). The fixed effects included group and running speed (both categorical variables). Cohen’s d effect size was calcu- lated when a significant group effect was observed46, and classified as small, moderate, and large when d values were larger than 0.2, 0.5, and 0.8, respectively46. Footstrike distribution between SEA and non-SEA runners were compared at all running speeds using Fisher exact tests given that some of the expected frequencies were less than five. A correlation matrix between anthropometric characteristics (mass and height, leg length, BMI, and ratio of leg length over height) was generated to identify unrelated anthropometric characteristics. Pearson correlation coefficients (r) between RE and the identified independent anthropometric variables were computed using RE values at the three running speeds separately, as well as with and without subgrouping of participants based on ethnicity. Similarly, Pearson correlation coefficients (r) between RE and biomechanical variables were computed at the three running speeds separately, as well as with and without subgrouping of participants based on ethnicity. Correlations were considered very high, high, moderate, low, and negligible when absolute r values were between 0.90–1.00, 0.70–0.89, 0.50–0.69, 0.30–0.49, and 0.00–0.29, respectively47. Given the number of correlations and exploratory nature of these analyses, only significant correlations reaching the moderate threshold were deemed meaningful. Statistical analyses were performed using Jamovi (version 1.2.17, Computer Software, retrieved from https:// www. jamovi. org) and R (version 3.5.0, The R Foundation for Statistical Computing, Vienna, Austria) with a level of significance set at p ≤ 0.05. Results Participant characteristics. Non-SEA runners were significantly heavier and taller, had a larger BMI and longer legs, had footwear with a larger heel-to-toe drop, and were more experienced than SEA runners (p ≤ 0.02; Table 1). Otherwise, demographic and footwear characteristics of non-SEA and SEA runners were similar (see Table 1). Running economy. SEA runners were significantly less economical (6%) than non-SEA runners (average across speeds: 522.6 ± 47.4 vs 492.4 ± 42.2 ml/kg0.75/km), with a moderate main effect of group on RE (p = 0.04, (1) kleg = Fz,max L Table 1. Participant and footwear characteristics for South East Asian (SEA) and non-South East Asian (non- SEA) runners. Significant differences (p ≤ 0.05) identified by Student’s or Welch’s t-tests are reported in bold. M male, F female, BMI body mass index, and NA not applicable. Characteristics SEA Non-SEA p Sex M = 10; F = 7 M = 10; F = 7 NA Age (y) 34 ± 4 38 ± 7 0.08 Mass (kg) 56 ± 9 68 ± 12 0.002 Height (cm) 167 ± 8 175 ± 9 0.01 Leg length (cm) 86 ± 4 91 ± 6 0.01 BMI (kg/m2) 20 ± 2 22 ± 2 0.004 Leg length over height (%) 52 ± 1 52 ± 1 0.54 Running distance (km/week) 60 ± 19 54 ± 18 0.32 Running experience (y) 6 ± 3 11 ± 7 0.02 Running performance on 21.1 km (min) 96 ± 9 96 ± 10 0.81 Shoe mass (g) 231 ± 32 215 ± 39 0.22 Shoe stack height (mm) 25 ± 3 25 ± 3 0.83 Shoe heel-to-toe drop (mm) 8 ± 3 6 ± 3 0.01 5 Vol.:(0123456789) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ d = 0.67; Fig. 2). There was no significant main effect of speed (p = 0.27) or group x speed interaction effect (p = 0.89) on RE. Larger differences were seen between SEA and non-SEA runners when expressing RE in ml/ kg/km instead of ml/kg0.75/km (see section S1 of supplementary materials). Biomechanical characteristics. There was a significant main effect of group on SF, SL, and tc (p ≤ 0.04; Table 2), with SEA having a higher SF (moderate effect; d = 0.75), smaller SL (small effect; d = 0.36), and shorter tc (moderate effect; d = 0.67) than non-SEA runners. There was no group effect on normalized SL, tf and kleg (p ≥ 0.23; Table 2). A significant speed effect was observed for all temporal variables (p ≤ 0.01; Table 2). SF, SL, and tf increased with increasing speed, whereas tc and kleg decreased with increasing speed. None of these variables demonstrated a group x speed interaction (p ≥ 0.32; Table 2). There was a significant group effect on θankle at footstrike and θknee at toe-off (p ≤ 0.03; Table 3), with SEA having less ankle dorsiflexion than non-SEA at footstrike (large effect; d = 1.20) and less knee extension at toe-off (moderate effect; d = 0.75). A significant speed effect was observed for θankle and θknee at toe-off (p ≤ 0.02; Table 3), with greater flexion at footstrike and extension at toe-off with increasing speed. None of these variables showed a group x speed interaction except θankle at footstrike (p = 0.007; Table 3), with SEA decreasing dorsiflexion with increasing speed while non-SEA increased dorsiflexion with increasing speed. Footstrike angle and pattern. SEA had a significantly lower FSA than non-SEA runners (large effect; d = 1.67), as depicted by the group effect on FSA (p < 0.001; Table  4). A speed effect was observed on FSA (p < 0.001; Table 4), indicating an increase of FSA with increasing running speed, while no significant group x speed interaction effect was noted (p = 0.13; Table 4). Footstrike distribution between SEA and non-SEA runners differed significantly at all speeds, with non-SEA being more commonly RFS (p ≤ 0.04; Table 4). Figure 2. Running Economy (RE) of South East Asian (SEA) and non-South East Asian (non-SEA) runners at several endurance running speeds. Linear mixed effects modelling identified a significant group effect (p ≤ 0.05). Table 2. Step frequency (SF), step length (SL), contact time (tc), flight time (tf), and spring-mass characteristics of the lower limb as given by leg stiffness (kleg) for South East Asian (SEA) and non-South East Asian (non-SEA) runners at endurance running speeds. Significant differences (p ≤ 0.05) identified by linear mixed effects modelling are indicated in bold. SL was expressed as a percentage of participant’s leg length in addition to raw units. a Step length normalized to leg length. Running speed (km/h) Group SF (steps/min) SL (cm) SL (%)a tc (ms) tf (ms) kleg (kN/m) 10 SEA 176 ± 9 95 ± 5 110 ± 7 268 ± 24 78 + 21 12.3 ± 2.5 Non-SEA 168 ± 9 100 ± 5 110 ± 6 287 ± 31 84 + 23 13.5 ± 2.8 12 SEA 181 ± 10 111 ± 6 128 ± 8 237 ± 22 96 ± 21 12.3 ± 2.4 Non-SEA 173 ± 10 116 ± 7 127 ± 6 253 ± 23 98 ± 25 13.5 ± 3.0 14 SEA 187 ± 11 125 ± 7 145 ± 9 215 ± 20 107 ± 19 12.0 ± 2.2 Non-SEA 179 ± 11 131 ± 8 144 ± 7 231 ± 21 107 ± 23 12.8 ± 2.7 Group effect 0.02 0.03 0.78 0.04 0.67 0.23 Running speed effect < 0.001 < 0.001 < 0.001 < 0.001 < 0.001 0.009 Interaction effect 0.93 0.48 0.68 0.81 0.44 0.32 6 Vol:.(1234567890) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ Relationship between RE and anthropometric characteristics. High positive correlations were identified between mass and height (r ≥ 0.83; p < 0.001), mass and leg length (r ≥ 0.74; p < 0.001), and mass and BMI (r ≥ 0.84; p < 0.001), while the correlation between mass and ratio of leg length over height was negligible and not significant (r ≤ 0.17; p ≥ 0.35). Hence, relationships between RE and mass and ratio of leg length over height were further explored (Table 5). For SEA runners, a high positive correlation was observed between RE and mass at 12 km/h (r = 0.69, p < 0.001; Table 5), while high positive correlations were observed between RE and mass for non-SEA runners at all speeds (r ≥ 0.65, p ≤ 0.005; Table 5). For runners combined, the strongest cor- relations were low. Table 6 presents all correlations, including the low and negligible ones. Relationships between RE expressed in ml/kg/km and anthropometric characteristics are provided in section S1 of supplementary materials. Table 3. Flexion–extension angle of the lower limb for South East Asian (SEA) and non-South East Asian (non-SEA) runners at endurance running speeds. Significant differences (p ≤ 0.05) identified by linear mixed effects modelling are indicated in bold. θankle : ankle joint angle, θknee : knee joint angle, FS: footstrike, and TO: toe-off. Running speed (km/h) Group θankle(°) θknee(°) FS TO FS TO 10 SEA 9 ± 5 − 12 ± 8 17 ± 2 27 ± 4 Non-SEA 14 ± 6 − 9 ± 3 18 ± 3 24 ± 7 12 SEA 8 ± 5 − 14 ± 8 17 ± 3 24 ± 4 Non-SEA 15 ± 6 − 11 ± 3 18 ± 4 21 ± 5 14 SEA 8 ± 6 − 14 ± 9 18 ± 3 24 ± 4 Non-SEA 15 ± 6 − 11 ± 4 18 ± 4 20 ± 4 Group effect 0.001 0.18 0.57 0.03 Running speed effect 0.31 0.02 0.65 < 0.001 Interaction effect 0.007 0.95 0.09 0.65 Table 4. Footstrike angle (FSA) and footstrike distribution [rearfoot strike (RFS) for FSA > 8° and non- rearfoot strike (non-RFS) otherwise42] for South East Asian (SEA) and non-South East Asian (non-SEA) runners at endurance running speeds. Significant differences (p ≤ 0.05) identified by linear mixed effects modelling and by Fisher exact tests are indicated in bold. Running speed (km/h) Group FSA (°) RFS—non-RFS p 10 SEA 6 ± 4 4–13 < 0.001 Non-SEA 13 ± 5 16–1 12 SEA 7 ± 4 6–11 < 0.001 Non-SEA 15 ± 5 16–1 14 SEA 9 ± 4 10–7 0.04 Non-SEA 17 ± 6 16–1 Group effect < 0.001 NA Running speed effect < 0.001 NA Interaction effect 0.13 NA Table 5. Pearson correlation coefficients between running economy and anthropometric characteristics (mass and ratio of leg length over height), together with their corresponding p-values underneath for South East Asian (SEA), non-South East Asian (non-SEA), as well as all runners pooled together (ALL). Note. Only the relationships between running economy and mass and ratio of leg length over height were considered because mass was highly and significantly correlated to height, leg length, and body mass index. Statistical significances (p ≤ 0.05) gray shaded boxes denote correlation coefficients above an absolute value of 0.5 (moderate). 7 Vol.:(0123456789) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ Relationships between RE and biomechanics. For SEA runners, a high positive correlation was seen between RE and tc at 12 km/h (|r|≥ 0.70, p ≤ 0.002; Table 6). SF and θankle at footstrike at 10 km/h were moderately and negatively correlated to RE, whereas SL (10 km/h) was moderately and positively correlated to RE (|r|≥ 0.50, p ≤ 0.04; Table 6). For non-SEA runners, a moderate and negative correlation was observed between RE and SF at 12 km/h (|r|≥ 0.51, p ≤ 0.04; Table 6). Besides, moderate positive correlations between RE and SL (12 km/h) and kleg (10 km/h) were identified (|r|≥ 0.51, p ≤ 0.04; Table 6). For runners combined, the strongest correlations were low. Table 6 presents all correlations, including the low and negligible ones. Relationships between RE expressed in ml/kg/km and biomechanics are given in section S1 of supplementary materials. Discussion Differences in RE were observed between SEA and non-SEA runners despite being matched for recent (< 1 year) road running performance and sex. SEA runners were less economical than non-SEA runners at endurance running speeds. Anthropometric differences were observed between groups, depicting that SEA were lighter and shorter than non-SEA runners, and had a lower BMI and shorter legs. Differences in running biomechan- ics between cohorts were also observed, but correlations between anthropometric and biomechanical variables and RE measures at a group-level were of small magnitudes at best, and provided limited explanations of the underlying differences in RE. Non-SEA were 6% more economical than SEA runners at endurance running speeds (Fig. 2). The lower RE in SEA than non-SEA runners could in part be due to anthropometric differences. We observed that SEA were lighter and shorter than non-SEA runners, and had a lower BMI and shorter legs (Table 1). Mass was significantly related to RE in both ethnic groups, with more economical runners having lower body mass. Mass was highly related to RE in SEA runners at 12 km/h and non-SEA at all speeds, but correlations became low or non-significant when pooling all runners together (Table 5). Previous studies showing that elite Caucasian runners were less economical than Kenyans attributed RE differences to longer legs (~ 5%), thinner and lighter calf musculature, and lower mass and BMI of Kenyan than Caucasian runners3–6. Indeed, RE being correlated with leg mass, Kenyan runners could benefit from their long, slender legs6. In our case, the ratio of leg length over height was not related to RE (Table 5) and was similar between SEA and non-SEA, indicating similar lower limb proportions in these two groups (Table 1). In fact, due to both smaller mass and shorter legs (Table 1), SEA might have had a proportionally similar leg mass than non-SEA runners. Table 6. Pearson correlation coefficients between running economy and biomechanical variables [step frequency (SF), step length (SL), contact time (tc), flight time (tf), spring-mass characteristics of the lower limb as given by leg stiffness (kleg), footstrike angle (FSA), and flexion–extension ankle ( θankle ) and knee ( θknee ) joint angle at footstrike (FS) and toe-off (TO)], together with their corresponding p-values underneath for South East Asian (SEA), non-South East Asian (non-SEA), as well as all runners pooled together (ALL). Statistical significances (p ≤ 0.05) are indicated in bold. Gray shaded boxes denote correlation coefficients above an absolute value of 0.5 (moderate). SL was expressed as a percentage of participant’s leg length in addition to raw units. a Step length normalized to leg length. 8 Vol:.(1234567890) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ Participants wore their own running shoes during testing similar to previous research exploring differences in running mechanics between ethnic groups10. Given that differences in footwear characteristics can underpin differences in running biomechanics48 and RE49, using a standardised shoe might have led to different study outcomes. Noteworthy, however, is that there were no significant difference in shoe mass or stack height between groups, with the 2 mm difference in heel-to-toe drop between groups likely having limited biomechanical or performance implications50. Recreational runners are more comfortable wearing their own shoes51, and show individual responses to novel footwear51,52 and cushioning properties53. A recent meta-analysis indicates rec- reational runners demonstrate improved RE when wearing more comfortable shoes54, supporting the appropri- ateness of participants wearing their own footwear for this investigation. Nevertheless, it is possible that other footwear characteristics not assessed as part of this study differed between groups, such as midsole cushioning and/or the longitudinal bending stiffness50, and contributed to the biomechanical and RE differences observed. Among all correlations between biomechanical variables and RE, only SF and SL were significantly related to RE in both ethnic groups. The SF and SL variables were moderately related to RE in SEA runners at 10 km/h and non-SEA at 12 km/h, but correlations became low and non-significant when all runners were pooled together (Table 6). Noteworthy, correlations between SL and RE were smaller and became non-significant when normal- ized to leg length. In addition, tc was highly and positively related to RE for SEA runners at 12 km/h. The identi- fied correlations between SF (and SL) and RE and between tc and RE suggest that individuals with higher SF (and shorter SL) and smaller tc (for SEA runners) are more economical. However, SEA had intrinsically higher SF (and shorter SL) and shorter tc, but worse RE than non-SEA runners (Table 2); therefore, contradicting the observed correlations. Based on the cost-of-generating-force hypothesis, one requires less metabolic energy with increased tc and longer leg lengths55–57, both observed in non-SEA (Table 1). The longer tc in non-SEA suggests that muscles had more time to shorten and produce the necessary forces to move the body than SEA runners. Based on the force–velocity relationship, if a muscle is shortening slower but only a given force is necessary (i.e., running on a treadmill), it could be speculated that the activation levels of the muscles were lower to reach the target force. These theories might partially explain the reduced metabolic cost in non-SEA than SEA runners, i.e., a longer tc, lower SF, and longer leg lengths are more economical. Nevertheless, studies indicate that increasing SF above self-selected ones in novice (156 ± 6 steps/min, 9.6 km/h) and trained (169 ± 11 steps/min, 12.6 km/h) runners acutely improves RE (+ 2%)58, as does undertak- ing a 10-day training programme to increase SF (from 166 ± 4 to 180 ± 1 steps/min, 12.3 km/h)59. At 12 km/h, mean SF values were 173 (range: 151 to 185) in non-SEA and 181 (range: 159 to 200) in SEA. Further increasing SF in runners with an intrinsically high SF might not be energetically optimal, but has yet to be examined. An extremely high SF might be suboptimal at endurance speeds given the greater mechanical power associated with increased frequency of reciprocal movements, which may require a greater reliance on less economical type II muscle fibers60. Indeed, Kaneko et al.60 suggested that SF and RE could be related through muscle fiber recruit- ment. Besides, given the shorter stature of our SEA vs non-SEA runners, their higher SF aligns with findings of moderate correlations between leg lengths and SF (r = −0.53, p < 0.001; 12 km/h), in agreement with previous literature (r = −0.45, p < 0.001) 61, whereby individuals with shorter legs tend to adopt higher SF. Alongside their higher SF and smaller SL, SEA had shorter tc, smaller FSA (more forefoot strike pattern), and smaller θankle at footstrike than non-SEA runners (Tables 2, 3, 4). Previous studies observed that running at a higher SF led to smaller tc 62 and FSA63, which is consistent with our findings. In addition, the prevalence of RFS was shown to be lower in Asian than North American recreational runners16, aligning with the findings of the present study. A smaller tc might be associated with smaller braking and propulsion phases. Although short braking phases are considered important for economical running64, SEA runners were less economical. Braking forces were not recorded herein due to unavailability of instrumented treadmills. Shorter braking times does not necessarily equate minimising braking forces, which is important in the context of RE65. Moreover, it could be that the orientation of the ground reaction forces in SEA runners was suboptimal. Indeed, Moore, et al.66 observed that a better alignment of the leg axis during propulsion and resultant ground reaction force improved RE, mainly via a more horizontal application of the ground reaction force. This idea is supported by our data, which show less extension of θknee at toe-off (Table 3), and thus potentially less horizontal propulsion for SEA than non-SEA runners. Nevertheless, θknee at toe-off was not correlated to RE (Table 6). Though SF, SL and tc significantly differed between groups, no difference in kleg was identified (Table 2), contradicting previous findings that kleg relates to the aforementioned variables67–69. These studies were all within-subject comparisons rather than between-subject ones; hence, at an individual level, the relationship might still hold within SEA and non-SEA participants. The lack of difference in kleg between groups despite differences in SF, SL, and tc potentially relates to the body mass difference between groups that is counterbalancing the spatiotemporal differences in the biomechanical variables [see Eq. (1)]. These biomechanical data were not clearly able to explain the variances in RE between groups, and support that RE improvements in various groups might need individualized training and considerations. A similar conclusion was made by Santos-Concejero, et al.10 when assessing RE differences between Eritrean and European runners. Moreover, these divergent findings overall suggest there is no unique or ideal running pattern that is the most economical amongst runners1. The running pattern of an individual results from a complex interaction between several biomechanical factors 70 that are interconnected and interact in a global and dynamic manner71 to optimize RE. A few limitations to the present study exist. Although the effect size was moderate (d = 0.67), the between- group difference in RE units was rather small (mean difference = 30.1 ml/kg0.75/km; p = 0.04). In addition, the within-group variability in RE and biomechanical variables at a given running speed were relatively small. There- fore, observed correlations between RE and biomechanical variables might have been greater in more heterogene- ous groups. Given the exploratory nature of this investigation, several variables were compared, leading to a high likelihood of finding a spurious difference or correlation. Nonetheless, our research provides preliminary indica- tions of potential differences between SEA and non-SEA runners warranting further consideration. Moreover, 9 Vol.:(0123456789) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ an underpinning factor to the differences in RE might be the running experience given that experienced runners self-optimize their running patterns better than novice runners1. Non-SEA runners were more experienced (years running) than SEA runners (Table 1), but all runners trained regularly and had a minimum of 2 years running experience, indicating they were all "experienced" and not "novice" runners. Nonetheless, a gradual improvement in RE (+ 15%) over an 11-year time span has been reported for a former world record holder in the women’s marathon72. Therefore, an effect due to running experience cannot be ruled out. Besides, several morphological factors which were not measured in this study might have partly explained differences in RE between SEA and non-SEA runners18,19,73–78 (more details are provided in section S2 of supplementary materials). Furthermore, although all SEA runners identified as “white”, the numerous nationalities of the non-SEA group potentially increased the heterogeneity of our cohort and influenced our results. Lastly, RE and biomechanics were collected within the same experimental session, but the two were not collected simultaneously (as common in running research79). Although possible that participants altered their runs, research indicates that metabolic equipment does not affect sagittal plane running kinematics and are comparable to running without metabolic testing80. Conclusion SEA and non-SEA runners were different in terms of RE, with SEA runners being less economical than non-SEA runners at endurance running speeds. Differences in anthropometric characteristics and running biomechanics between cohorts were also observed, but explained differences in RE to a limited extent. Other factors, which could be related to ethnicity, might be underpinning such differences. Unfortunately, these factors were not measured in this study. Nonetheless, caution must be taken when generalizing from non-SEA running studies to SEA runners. Data availability The dataset supporting this article is available on request to the corresponding author. Received: 8 August 2021; Accepted: 21 March 2022 References 1. Moore, I. S. Is there an economical tunning technique? A review of modifiable biomechanical factors affecting running economy. Sports Med. 46, 793–807. https:// doi. org/ 10. 1007/ s40279- 016- 0474-4 (2016). 2. Weston, A. R., Mbambo, Z. & Myburgh, K. H. Running economy of African and Caucasian distance runners. Med. Sci. Sports Exerc. 32, 1130–1134. https:// doi. org/ 10. 1097/ 00005 768- 20000 6000- 00015 (2000). 3. Larsen, H. B. & Sheel, A. W. The Kenyan runners. Scand. J. Med. Sci. Sports 25(Suppl 4), 110–118. https:// doi. org/ 10. 1111/ sms. 12573 (2015). 4. Larsen, H. B., Christensen, D. L., Nolan, T. & Søndergaard, H. Body dimensions, exercise capacity and physical activity level of adolescent Nandi boys in western Kenya. Ann. Hum. Biol. 31, 159–173. https:// doi. org/ 10. 1080/ 03014 46041 00016 63416 (2004). 5. Saltin, B. et al. Aerobic exercise capacity at sea level and at altitude in Kenyan boys, junior and senior runners compared with Scandinavian runners. Scand. J. Med. Sci. Sports 5, 209–221. https:// doi. org/ 10. 1111/j. 1600- 0838. 1995. tb000 37.x (1995). 6. Larsen, H. B. Kenyan dominance in distance running. Comp. Biochem. Physiol. A: Mol. Integr. Physiol. 136, 161–170, doi:https:// doi. org/ 10. 1016/ s1095- 6433(03) 00227-7 (2003). 7. Lucia, A. et al. Physiological characteristics of the best Eritrean runners-exceptional running economy. Appl. Physiol. Nutr. Metab. 31, 530–540. https:// doi. org/ 10. 1139/ h06- 029 (2006). 8. Lieberman, D. E. et al. Foot strike patterns and collision forces in habitually barefoot versus shod runners. Nature 463, 531–535. https:// doi. org/ 10. 1038/ natur e08723 (2010). 9. Marino, F. E., Lambert, M. I. & Noakes, T. D. Superior performance of African runners in warm humid but not in cool environ- mental conditions. J. Appl. Physiol. 96, 124–130. https:// doi. org/ 10. 1152/ jappl physi ol. 00582. 2003 (2004). 10. Santos-Concejero, J. et al. Gait-cycle characteristics and running economy in elite Eritrean and European runners. Int. J. Sports Physiol. Perform. 10, 381–387. https:// doi. org/ 10. 1123/ ijspp. 2014- 0179 (2015). 11. Santos-Concejero, J. et al. Differences in ground contact time explain the less efficient running economy in north African runners. Biol. Sport 30, 181–187. https:// doi. org/ 10. 5604/ 20831 862. 10591 70 (2013). 12. Tam, E. et al. Energetics of running in top-level marathon runners from Kenya. Eur. J. Appl. Physiol. 112, 3797–3806. https:// doi. org/ 10. 1007/ s00421- 012- 2357-1 (2012). 13. Wishnizer, R. R., Inbar, O., Klinman, E. & Fink, G. Physiological differences between Ethiopian and Caucasian distance runners and their effects on 10 km running performance. Adv Phys Educ 3, 9. https:// doi. org/ 10. 4236/ ape. 2013. 33023 (2013). 14. Shu, Y. et al. Foot morphological difference between habitually shod and unshod runners. PLoS ONE 10, e0131385. https:// doi. org/ 10. 1371/ journ al. pone. 01313 85 (2015). 15. Sano, K. et al. Can measures of muscle–tendon interaction improve our understanding of the superiority of Kenyan endurance runners?. Eur. J. Appl. Physiol. 115, 849–859. https:// doi. org/ 10. 1007/ s00421- 014- 3067-7 (2015). 16. Patoz, A., Lussiana, T., Gindre, C. & Hébert-Losier, K. Recognition of foot strike pattern in Asian recreational runners. Sports 7, 147. https:// doi. org/ 10. 3390/ sport s7060 147 (2019). 17. Andersen, J. J. & International Association of Athletics Federations. The State of Running 2019, (2020). 18. Hawes, M. R. et al. Ethnic differences in forefoot shape and the determination of shoe comfort. Ergonomics 37, 187–196. https:// doi. org/ 10. 1080/ 00140 13940 89636 37 (1994). 19. Zárate-Kalfópulos, B., Romero-Vargas, S., Otero-Cámara, E., Correa, V. C. & Reyes-Sánchez, A. Differences in pelvic parameters among Mexican, Caucasian, and Asian populations. J. Neurosurg. Spine 16, 516–519. https:// doi. org/ 10. 3171/ 2012.2. Spine 11755 (2012). 20. Sun, P. et al. Autonomic recovery is delayed in Chinese compared with Caucasian following treadmill exercise. PLoS ONE 11, e0147104. https:// doi. org/ 10. 1371/ journ al. pone. 01471 04 (2016). 21. Chen, W.-L., O’Connor, J. J. & Radin, E. L. A comparison of the gaits of Chinese and Caucasian women with particular reference to their heelstrike transients. Clin. Biomech. 18, 207–213. https:// doi. org/ 10. 1016/ S0268- 0033(02) 00187-0 (2003). 22. Gruber, A. H., Umberger, B. R., Braun, B. & Hamill, J. Economy and rate of carbohydrate oxidation during running with rearfoot and forefoot strike patterns. J. Appl. Physiol. 115, 194–201. https:// doi. org/ 10. 1152/ jappl physi ol. 01437. 2012 (2013). 23. Ogueta-Alday, A. N. A., RodrÍGuez-Marroyo, J. A. & GarcÍA-LÓPez, J. Rearfoot striking runners are more economical than midfoot strikers. Med. Sci. Sports Exerc. 46 (2014). 10 Vol:.(1234567890) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ 24. Di Michele, R. & Merni, F. The concurrent effects of strike pattern and ground-contact time on running economy. J. Sci. Med. Sport 17, 414–418. https:// doi. org/ 10. 1016/j. jsams. 2013. 05. 012 (2014). 25. Støren, Ø., Helgerud, J. & Hoff, J. Running stride peak forces inversely determine running economy in elite runners. J. Strength Cond. Res. 25 (2011). 26. Paavolainen, L. M., Nummela, A. T. & Rusko, H. K. Neuromuscular characteristics and muscle power as determinants of 5-km running performance. Med. Sci. Sports Exerc. 31, 124–130. https:// doi. org/ 10. 1097/ 00005 768- 19990 1000- 00020 (1999). 27. Williams, K. R. & Cavanagh, P. R. Relationship between distance running mechanics, running economy, and performance. J. Appl. Physiol. 63, 1236–1245. https:// doi. org/ 10. 1152/ jappl. 1987. 63.3. 1236 (1987). 28. Lussiana, T., Gindre, C., Mourot, L. & Hébert-Losier, K. Do subjective assessments of running patterns reflect objective parameters?. Eur. J. Sport Sci. 17, 847–857. https:// doi. org/ 10. 1080/ 17461 391. 2017. 13250 72 (2017). 29. Harriss, D. J., Macsween, A. & Atkinson, G. Standards for ethics in sport and exercise science research: 2018 update. Int. J. Sports Med. 38, 1126–1131. https:// doi. org/ 10. 1055/s- 0043- 124001 (2017). 30. Oliveira, A. S. & Pirscoveanu, C. I. Implications of sample size and acquired number of steps to investigate running biomechanics. Sci. Rep. 11, 3083. https:// doi. org/ 10. 1038/ s41598- 021- 82876-z (2021). 31. Svedenhag, J. & Sjödin, B. Body-mass-modified running economy and step length in elite male middle- and long-distance runners. Int. J. Sports Med. 15, 305–310. https:// doi. org/ 10. 1055/s- 2007- 10210 65 (1994). 32. Tranberg, R., Saari, T., Zügner, R. & Kärrholm, J. Simultaneous measurements of knee motion using an optical tracking system and radiostereometric analysis (RSA). Acta Orthop. 82, 171–176. https:// doi. org/ 10. 3109/ 17453 674. 2011. 570675 (2011). 33. Hanavan, E. A mathematical model of the human body. AMRL-TR. Aerospace Med Res Lab 1, 1–149 (1964). 34. Dempster, W. T. Space requirements of the seated operator: geometrical, kinematic, and mechanical aspects of the body with special reference to the limbs. (Wright Air Development Center, 1955). 35. Woltring, H. Representation and calculation of 3-D joint movement. Hum. Mov. Sci. 10, 603–616. https:// doi. org/ 10. 1016/ 0167- 9457(91) 90048-3 (1991). 36. Cole, G. K., Nigg, B. M., Ronsky, J. L. & Yeadon, M. R. Application of the joint coordinate system to three-dimensional joint atti- tude and movement representation: a standardization proposal. J. Biomech. Eng. 115, 344–349. https:// doi. org/ 10. 1115/1. 28954 96 (1993). 37. Davis, R. B., Õunpuu, S., Tyburski, D. & Gage, J. R. A gait analysis data collection and reduction technique. Hum. Mov. Sci. 10, 575–587. https:// doi. org/ 10. 1016/ 0167- 9457(91) 90046-Z (1991). 38. Grood, E. S. & Suntay, W. J. A joint coordinate system for the clinical description of three-dimensional motions: application to the knee. J. Biomech. Eng. 105, 136–144. https:// doi. org/ 10. 1115/1. 31383 97 (1983). 39. Blankevoort, L., Huiskes, R. & de Lange, A. The envelope of passive knee joint motion. J. Biomech. 21, 705–720. https:// doi. org/ 10. 1016/ 0021- 9290(88) 90280-1 (1988). 40. Kadaba, M. P., Ramakrishnan, H. K. & Wootten, M. E. Measurement of lower extremity kinematics during level walking. J. Orthop. Res. 8, 383–392. https:// doi. org/ 10. 1002/ jor. 11000 80310 (1990). 41. Piazza, S. J. & Cavanagh, P. R. Measurement of the screw-home motion of the knee is sensitive to errors in axis alignment. J. Biomech. 33, 1029–1034. https:// doi. org/ 10. 1016/ s0021- 9290(00) 00056-7 (2000). 42. Altman, A. R. & Davis, I. S. A kinematic method for footstrike pattern detection in barefoot and shod runners. Gait Posture 35, 298–300. https:// doi. org/ 10. 1016/j. gaitp ost. 2011. 09. 104 (2012). 43. Lussiana, T., Patoz, A., Gindre, C., Mourot, L. & Hébert-Losier, K. The implications of time on the ground on running economy: less is not always better. J. Exp. Biol. 222, jeb192047, doi:https:// doi. org/ 10. 1242/ jeb. 192047 (2019). 44. Maiwald, C., Sterzing, T., Mayer, T. A. & Milani, T. L. Detecting foot-to-ground contact from kinematic data in running. Footwear Sci. 1, 111–118. https:// doi. org/ 10. 1080/ 19424 28090 31339 38 (2009). 45. Morin, J.-B., Dalleau, G., Kyröläinen, H., Jeannin, T. & Belli, A. A simple method for measuring stiffness during running. J. Appl. Biomech. 21, 167–180. https:// doi. org/ 10. 1123/ jab. 21.2. 167 (2005). 46. Cohen, J. Statistical Power Analysis for the Behavioral Sciences. (Routledge, 1988). 47. Hinkle, D. E., Wiersma, W. & Jurs, S. G. Applied Statistics for the Behavioral Sciences. 768 (Houghton Mifflin (p. 109), 2002). 48. Sinclair, J., Fau-Goodwin, J., Richards, J. & Shore, H. The influence of minimalist and maximalist footwear on the kinetics and kinematics of running. Footwear Sci. 8, 33–39. https:// doi. org/ 10. 1080/ 19424 280. 2016. 11420 03 (2016). 49. Fuller, J. T., Bellenger, C. R., Thewlis, D., Tsiros, M. D. & Buckley, J. D. The effect of footwear on running performance and running economy in distance runners. Sports Med. 45, 411–422. https:// doi. org/ 10. 1007/ s40279- 014- 0283-6 (2015). 50. Sun, X., Lam, W.-K., Zhang, X., Wang, J. & Fu, W. Systematic review of the role of footwear constructions in running biomechanics: implications for running-related injury and performance. J. Sports Sci. Med. 19, 20–37 (2020). 51. Hébert-Losier, K. et al. Metabolic and performance responses of male runners wearing 3 types of footwear: Nike Vaporfly 4%, Saucony Endorphin racing flats, and their own shoes. J. Sport Health Sci. https:// doi. org/ 10. 1016/j. jshs. 2020. 11. 012 (2020). 52. Tam, N., Tucker, R. & Astephen Wilson, J. L. Individual Responses to a barefoot running program: insight into risk of injury. Am. J. Sports Med. 44, 777–784, doi:https:// doi. org/ 10. 1177/ 03635 46515 620584 (2016). 53. Tung, K. D., Franz, J. R. & Kram, R. A test of the metabolic cost of cushioning hypothesis during unshod and shod running. Med. Sci. Sports Exerc. 46, 324–329. https:// doi. org/ 10. 1249/ MSS. 0b013 e3182 a63b81 (2014). 54. Van Alsenoy, K., van der Linden, M. L., Girard, O. & Santos, D. Increased footwear comfort is associated with improved running economy - a systematic review and meta-analysis. Eur. J. Sport Sci., 1–13, doi:https:// doi. org/ 10. 1080/ 17461 391. 2021. 19986 42 (2021). 55. Roberts, T. J., Kram, R., Weyand, P. G. & Taylor, C. R. Energetics of bipedal running. I. Metabolic cost of generating force. J. Exp. Biol. 201, 2745–2751, doi:https:// doi. org/ 10. 1242/ jeb. 201. 19. 2745 (1998). 56. Roberts, T. J., Chen, M. S. & Taylor, C. R. Energetics of bipedal running. II. Limb design and running mechanics. J. Exp. Biol. 201, 2753–2762, doi:https:// doi. org/ 10. 1242/ jeb. 201. 19. 2753 (1998). 57. Kram, R. & Taylor, C. R. Energetics of running: a new perspective. Nature 346, 265–267. https:// doi. org/ 10. 1038/ 34626 5a0 (1990). 58. de Ruiter, C. J., Verdijk, P. W. L., Werker, W., Zuidema, M. J. & de Haan, A. Stride frequency in relation to oxygen consumption in experienced and novice runners. Eur. J. Sport Sci. 14, 251–258. https:// doi. org/ 10. 1080/ 17461 391. 2013. 783627 (2014). 59. Quinn, T. J., Dempsey, S. L., LaRoche, D. P., Mackenzie, A. M. & Cook, S. B. Step frequency training improves running economy in well-trained female runners. J. Strength Cond. Res. https:// doi. org/ 10. 1519/ jsc. 00000 00000 003206 (2019). 60. Kaneko, M., Matsumoto, M., Ito, A. & Fuchimoto, T. Optimum step frequency in constant speed running. (Human Kinetics, 1987). 61. Tenforde, A. S., Borgstrom, H. E., Outerleys, J. & Davis, I. S. Is cadence related to leg length and load rate?. J. Orthop. Sports Phys. Ther. 49, 280–283. https:// doi. org/ 10. 2519/ jospt. 2019. 8420 (2019). 62. Adams, D., Pozzi, F., Willy, R. W., Carrol, A. & Zeni, J. Altering cadence or vertical oscillation during running: effects on running related injury factors. Int. J. Sports Phys. Ther. 13, 633–642 (2018). 63. Allen, D. J., Heisler, H., Mooney, J. & Kring, R. The effect of step rate manipulation on foot strike pattern of long distance runners. Int. J. Sports Phys. Ther. 11, 54–63 (2016). 64. Nummela, A., Keränen, T. & Mikkelsson, L. Factors related to top running speed and economy. Int. J. Sports Med. 28, 655–661 (2007). 11 Vol.:(0123456789) Scientific Reports | (2022) 12:6291 | https://doi.org/10.1038/s41598-022-10030-4 www.nature.com/scientificreports/ 65. Lieberman, D. E., Warrener, A. G., Wang, J. & Castillo, E. R. Effects of stride frequency and foot position at landing on braking force, hip torque, impact peak force and the metabolic cost of running in humans. J. Exp. Biol. 218, 3406–3414. https:// doi. org/ 10. 1242/ jeb. 125500 (2015). 66. Moore, I. S., Jones, A. M. & Dixon, S. J. Reduced oxygen cost of running is related to alignment of the resultant GRF and leg axis vector: a pilot study. Scand. J. Med. Sci. Sports 26, 809–815. https:// doi. org/ 10. 1111/ sms. 12514 (2016). 67. Farley, C. T. & González, O. Leg stiffness and stride frequency in human running. J. Biomech. 29, 181–186. https:// doi. org/ 10. 1016/ 0021- 9290(95) 00029-1 (1996). 68. Morin, J. B., Samozino, P., Zameziati, K. & Belli, A. Effects of altered stride frequency and contact time on leg-spring behavior in human running. J. Biomech. 40, 3341–3348. https:// doi. org/ 10. 1016/j. jbiom ech. 2007. 05. 001 (2007). 69. Monte, A., Muollo, V., Nardello, F. & Zamparo, P. Sprint running: how changes in step frequency affect running mechanics and leg spring behaviour at maximal speed. J. Sports Sci. 35, 339–345. https:// doi. org/ 10. 1080/ 02640 414. 2016. 11643 36 (2017). 70. Saunders, P. U., Pyne, D. B., Telford, R. D. & Hawley, J. A. Factors affecting running economy in trained distance runners. Sports Med. 34, 465–485. https:// doi. org/ 10. 2165/ 00007 256- 20043 4070- 00005 (2004). 71. Dickinson, M. H. et al. How animals move: an integrative view. Science 288, 100. https:// doi. org/ 10. 1126/ scien ce. 288. 5463. 100 (2000). 72. Jones, A. M. The physiology of the world record holder for the women’s marathon. Int. J. Sports Sci. Coach. 1, 101–116. https:// doi. org/ 10. 1260/ 17479 54067 77641 258 (2006). 73. Yue, B. et al. Differences of knee anthropometry between Chinese and white men and women. J. Arthroplasty 26, 124–130. https:// doi. org/ 10. 1016/j. arth. 2009. 11. 020 (2011). 74. Scholz, M. N., Bobbert, M. F., van Soest, A. J., Clark, J. R. & van Heerden, J. Running biomechanics: shorter heels, better economy. J. Exp. Biol. 211, 3266. https:// doi. org/ 10. 1242/ jeb. 018812 (2008). 75. Hunter, G. R. et al. Tendon length and joint flexibility are related to running economy. Med. Sci. Sports Exerc. 43, 1492–1499. https:// doi. org/ 10. 1249/ MSS. 0b013 e3182 10464a (2011). 76. Ueno, H. et al. Relationship between Achilles tendon length and running performance in well-trained male endurance runners. Scand. J. Med. Sci. Sports 28, 446–451. https:// doi. org/ 10. 1111/ sms. 12940 (2018). 77. Kunimasa, Y. et al. Specific muscle-tendon architecture in elite Kenyan distance runners. Scand. J. Med. Sci. Sports 24, e269-274. https:// doi. org/ 10. 1111/ sms. 12161 (2014). 78. Mooses, M. et al. Dissociation between running economy and running performance in elite Kenyan distance runners. J. Sports Sci. 33, 136–144. https:// doi. org/ 10. 1080/ 02640 414. 2014. 926384 (2015). 79. Hunter, I. et al. Running economy, mechanics, and marathon racing shoes. J. Sports Sci. 37, 2367–2373. https:// doi. org/ 10. 1080/ 02640 414. 2019. 16338 37 (2019). 80. Sloan, R. S., Wight, J. T., Hooper, D. R., Garman, J. E. J. & Pujalte, G. G. A. Metabolic testing does not alter distance running lower body sagittal kinematics. Gait Posture 76, 403–408. https:// doi. org/ 10. 1016/j. gaitp ost. 2020. 01. 001 (2020). Acknowledgements The authors thank Mr. Chris Tee Chow Li for assistance during the data collection process. The authors also thank the participants for their time and participation. Author contributions Conceptualization: T.L., C.G., and K.H.-L.; Methodology: T.L., L.M., C.G., and K.H.-L.; Investigation: T.L. and K.H.-L.; Formal analysis: T.L., A.P., K.H.-L., and B.B.; Writing—original draft preparation: A.P. and B.B.; Writ- ing—review and editing: A.P., B.B., T.L., L.M., C.G., and K.H.-L.; Supervision: L.M. and K.H.-L; Project admin- istration: L.M.; Funding acquisition: T.L., L.M., and K.H.-L. Funding This study was financially supported by the University of Bourgogne Franche-Comté (France) and the National Sports Institute of Malaysia. Competing interests The authors declare no competing interests. Additional information Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1038/ s41598- 022- 10030-4. Correspondence and requests for materials should be addressed to A.P. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. © The Author(s) 2022

Non-South East Asians have a better running economy and different anthropometrics and biomechanics than South East Asians.

04-15-2022

Patoz, Aurélien,Lussiana, Thibault,Breine, Bastiaan,Gindre, Cyrille,Mourot, Laurent,Hébert-Losier, Kim

eng

PMC7143174

Supplement 1 R-code: #Packages used library(readxl) library(lubridate) library(plyr) library(dplyr) library(stringr) library(doBy) library(ggplot2) library(tableone) library(gamm4) library(splines) library(lme4) library(lmerTest) library(mgcv) library(itsadug) library(effects) library(stargazer) library(ggpubr) library(readr) #tower_an=database imported from Excel, recoded to be analysed #Outcome=speed_km_h (time_m=performance in minutes) #Effects= Gender, Age, Stairs, Floors #Year is the year of the Event #dataset for Figure 1 d3 <- subset(tower_an, !is.na(Gender)) %>% group_by(Year, Gender) %>% summarise(count=n()) %>% mutate(perc=count/sum(count), ratio=count[Gender=="M"]/count[Gender=="F"]) d3<-as.data.frame(d3) d3_s<-reshape(d3 , v.names=c("count", "perc") , idvar="Year" , timevar = "Gender" , direction="wide") d3_s$ratio<-d3_s$count.M/d3_s$count.F  #Figure 1 ggplot(d3) + geom_bar(stat="identity", aes(x=Year, y = count, fill = factor(Gender)), position = "dodge") + geom_line(aes(x = Year, y = ratio*1920, group=1, color = 'ratio'), linetype=3, size=1)+ geom_point(aes(x = Year, y = ratio*1920, group=1, color = 'ratio'), size = 1.5)+ labs(x = "Year", y = "Climbers (N)", fill = "Sex") + scale_y_continuous(expand = c(0,0), breaks=c(0, 500, 1000, 1500, 3000, 5000, 7000), limits=c(0, 7000), sec.axis = sec_axis(~./1920))+ scale_x_continuous(breaks=c(2014, 2015, 2016, 2017, 2018, 2019))+ scale_fill_manual(values = c("pink","cadetblue3")) + scale_color_manual('', labels = 'Men to Women ratio', values = 'black') + theme_bw(base_size = 18) #Table 2 #Age/Gender groups: mean(SD) mean_g<-subset(tower_an, !is.na(Gender) & !is.na(Age)) %>% group_by(Age, Gender) %>% summarise(count=n(), mean=mean(speed_km_h), sd=sd(speed_km_h)) %>% mutate(ratio=count[Gender=="M"]/count[Gender=="F"]) #t-tests sex groups for each age group by(tower_an, tower_an$Age, function(.tower_an) t.test(speed_km_h~Gender, data=.tower_an)) #Two-way ANOVA summary(aov(lm(speed_km_h~Gender+Age, data=tower_an))) #Table 3 - height in m (tower height in meters) CreateTableOne(vars=c("speed_km_h", "time_m", "height.in.m", "Floors", "Stairs"), strata = "Gender", data=tower_an) #Statistical model: Mixed models: bs=basis splines; Num_name is a number for identify climbers by their names - this model takes into #account repeated measurements fit_g2<-lmer(data=tower_an, speed_km_h ~ Gender*Age*bs(Stairs, 5)+bs(Floors,5)+(1|Num_name)) #Supplemental Table 1 class(fit_g2) <- "lmerMod" stargazer(fit_g2, type="text", out="rep3.html", star.cutoffs = c(0.05, 0.01, 0.001), keep.stat=c("n"), model.numbers=F, no.space=T, column.labels=c("Speed km/h"), dep.var.labels.include=F) #database for the fitted values ef <- Effect(c("Gender", "Age", "Stairs"), xlevels=list(Stairs=seq(300, 2500, 50)), fit_g2) x_22 <- as.data.frame(ef) ef <- Effect(c("Floors", "Gender"), xlevels=list(Floors=seq(5, 150, 5)), fit_g2) x_22b <- as.data.frame(ef) #Figure 2 fig2<-ggplot(subset(tower_an, !is.na(Age)), aes(x=Stairs, y=speed_km_h, color=Gender, linetype=Gender)) + #geom_errorbar(aes(ymin=mean-se, ymax=mean+se), width=.1,position=pd) + geom_line(data=subset(x_22, fit>0 & fit<7), aes(x=Stairs, y=fit, color=Gender, linetype=Gender), size=1.5) + stat_summary(geom = "point", fun.y = mean, size=1)+ labs(x="Stairs", y="Speed (km/h)", color="Sex")+ scale_color_manual(values = c("pink","cadetblue3"), labels=c("F", "M")) + scale_linetype_manual(name = "Sex" , values=c(1,6) , labels=c("F","M") ) + facet_wrap(~factor(as.vector(Age), levels = c("<20", "20-29", "30-39", "40-49", "50-59", "60-69", ">69")))+ # scale_y_continuous(breaks = c(0.08333333, 0.125, 0.1666667, 0.2083333),labels = c("02:00", "03:00", "04:00", "05:00"))+ theme_bw(base_size = 18) ggsave("Figure 2.tiff", ggarrange(fig2, legend="bottom"), height=25, width=30, units='cm', compression="lzw", dpi=300) #Figure 3 fig3<-ggplot(tower_an, aes(x=Floors, y=speed_km_h, color=Gender, linetype=Gender)) + #geom_errorbar(aes(ymin=mean-se, ymax=mean+se), width=.1,position=pd) + geom_line(data=x_22b, aes(x=Floors, y=fit, color=Gender, linetype=Gender), size=1.5) + stat_summary(geom = "point", fun.y = mean)+ labs(x="Floors", y="Speed (km/h)", color="Sex")+ # scale_y_continuous(breaks = c(0.08333333, 0.125, 0.1666667, 0.2083333),labels = c("02:00", "03:00", "04:00", "05:00"))+ scale_color_manual(values = c("pink","cadetblue3"), labels=c("F", "M")) + scale_linetype_manual(name = "Sex" , values=c(1,6) , labels=c("F","M") ) + theme_bw(base_size = 12) ggsave("Figure 3.tiff", ggarrange(fig3, legend="bottom"), height=12, width=15, units='cm', compression="lzw", dpi=300) Supplemental 2 Table: Predictor Speed (km/h) Sex (ref F) M 0.881 (0.586) Age (ref 20-29) <20 30-39 40-49 50-59 60-69 > 69 -4.579** (1.519) 1.765** (0.557) 2.367*** (0.579) 6.903*** (0.703) 1.278* (0.610) 10.368 (6.616) Stairs BS(Stairs, 5)1 BS(Stairs, 5)2 BS(Stairs, 5)3 BS(Stairs, 5)4 BS(Stairs, 5)5 1.059 (0.745) 0.835 (0.491) 0.859 (0.573) 0.211 (0.533) 1.584** (0.559) Floors BS(Floors, 5)1 BS(Floors, 5)2 BS(Floors, 5)3 BS(Floors, 5)4 BS(Floors, 5)5 0.033 (0.039) -0.053 (0.029) -0.285*** (0.028) 0.587*** (0.046) 0.258*** (0.051) SexM:Age SexM:Age< 20 SexM:Age30-39 SexM:Age40-49 SexM:Age50-59 SexM:Age60-69 SexM:Age> 69 -0.690 (1.893) -1.848** (0.607) -1.325* (0.624) -5.272*** (0.742) -0.706 (0.669) -14.030* (6.968) Sex M:Stairs SexM:BS(Floors, 5)1 SexM:BS(Floors, 5)2 SexM:BS(Floors, 5)3 SexM:BS(Floors, 5)4 SexM:BS(Floors, 5)5 -0.251 (0.797) -1.035 (0.528) -0.526 (0.616) -0.859 (0.575) -0.770 (0.600) Age:Stairs Age< 20:BS(Stairs, 5)1 Age30-39:BS(Stairs, 5)1 Age40-49:BS(Stairs, 5)1 Age50-59:BS(Stairs, 5)1 Age60-69:BS(Stairs, 5)1 Age> 69:BS(Stairs, 5)1 Age< 20:BS(Stairs, 5)2 Age30-39:BS(Stairs, 5)2 Age40-49:BS(Stairs, 5)2 Age50-59:BS(Stairs, 5)2 Age60-69:BS(Stairs, 5)2 Age> 69:BS(Stairs, 5)2 Age< 20:BS(Stairs, 5)3 6.122** (2.085) -2.287** (0.759) -3.074*** (0.789) -9.022*** (0.960) -1.691* (0.848) -14.657 (9.294) 3.869** (1.299) -1.625** (0.501) -2.259*** (0.521) -6.426*** (0.625) -1.587** (0.567) -9.068 (5.487) 5.273** (1.733)  Age30-39:BS(Stairs, 5)3 Age40-49:BS(Stairs, 5)3 Age50-59:BS(Stairs, 5)3 Age60-69:BS(Stairs, 5)3 Age> 69:BS(Stairs, 5)3 Age< 20:BS(Stairs, 5)4 Age30-39:BS(Stairs, 5)4 Age40-49:BS(Stairs, 5)4 Age50-59:BS(Stairs, 5)4 Age60-69:BS(Stairs, 5)4 Age> 69:BS(Stairs, 5)4 Age< 20:BS(Stairs, 5)5 Age30-39:BS(Stairs, 5)5 Age40-49:BS(Stairs, 5)5 Age50-59:BS(Stairs, 5)5 Age60-69:BS(Stairs, 5)5 Age> 69:BS(Stairs, 5)5 -1.895** (0.586) -2.514*** (0.610) -7.274*** (0.769) -1.081 (0.712) -12.012 (7.743) 1.982 (1.626) -1.652** (0.547) -2.319*** (0.570) -6.679*** (0.839) -2.404 (1.412) -5.324 (5.660) 12.506* (5.426) -1.986*** (0.574) -2.504*** (0.615) -7.358*** (2.217) 1.751 (4.352) -25.605 (22.977) Sex M:Age:Stairs SexM:Age< 20:BS(Stairs, 5)1 SexM:Age30-39:BS(Stairs, 5)1 SexM:Age40-49:BS(Stairs, 5)1 SexM:Age50-59:BS(Stairs, 5)1 SexM:Age60-69:BS(Stairs, 5)1 SexM:Age> 69:BS(Stairs, 5)1 SexM:Age< 20:BS(Stairs, 5)2 SexM:Age30-39:BS(Stairs, 5)2 SexM:Age40-49:BS(Stairs, 5)2 SexM:Age50-59:BS(Stairs, 5)2 SexM:Age60-69:BS(Stairs, 5)2 SexM:Age> 69:BS(Stairs, 5)2 SexM:Age< 20:BS(Stairs, 5)3 SexM:Age30-39:BS(Stairs, 5)3 SexM:Age40-49:BS(Stairs, 5)3 SexM:Age50-59:BS(Stairs, 5)3 SexM:Age60-69:BS(Stairs, 5)3 SexM:Age> 69:BS(Stairs, 5)3 SexM:Age< 20:BS(Stairs, 5)4 SexM:Age30-39:BS(Stairs, 5)4 SexM:Age40-49:BS(Stairs, 5)4 SexM:Age50-59:BS(Stairs, 5)4 SexM:Age60-69:BS(Stairs, 5)4 SexM:Age> 69:BS(Stairs, 5)4 SexM:Age< 20:BS(Stairs, 5)5 SexM:Age30-39:BS(Stairs, 5)5 SexM:Age40-49:BS(Stairs, 5)5 SexM:Age50-59:BS(Stairs, 5)5 SexM:Age60-69:BS(Stairs, 5)5 SexM:Age> 69:BS(Stairs, 5)5 0.286 (2.565) 2.241** (0.828) 1.384 (0.851) 6.509*** (1.013) 0.253 (0.930) 18.792 (9.748) 1.200 (1.643) 1.914*** (0.549) 1.644** (0.564) 5.166*** (0.663) 1.266* (0.624) 12.446* (5.808) 0.081 (2.104) 1.785** (0.639) 1.161 (0.657) 5.288*** (0.809) 0.152 (0.769) 15.608 (8.133) 3.225 (1.965) 1.884** (0.599) 1.380* (0.617) 5.208*** (0.873) 2.068 (1.440) 7.381 (6.050) -6.851 (5.547) 1.922** (0.630) 1.463* (0.664) 5.385* (2.231) -2.830 (4.364) 33.808 (24.159) Intercept 0.114 (0.543) Observations 19,851 Note *p<0.05;**p<0.01;***p<0.001 Table S1. Regression analysis (mixed model) of speed (km/h) in tower climbing. Estimates and standard errors (SE) of fixed effects are reported. P-values ranges are marked with asterisks (see note). Smoothing terms, basis splines (BS), are denoted with BS(x, 5) t, where x=stairs, floors; t==1,…,5.

Tower Running-Participation, Performance Trends, and Sex Difference.

03-14-2020

Stark, Daniel,Di Gangi, Stefania,Sousa, Caio Victor,Nikolaidis, Pantelis,Knechtle, Beat

eng

PMC6728027

RESEARCH ARTICLE Optimizing running a race on a curved track Amandine AftalionID1*, Pierre Martinon2 1 Ecole des Hautes Etudes en Sciences Sociales, PSL Research University, Centre d’Analyse et de Mathe´matique Sociales, Paris, France, 2 Inria Paris and LJLL Sorbonne Universite´, Paris, France * [email protected] Abstract In order to determine the optimal strategy to run a race on a curved track according to the lane number, we introduce a model based on differential equations for the velocity, the pro- pulsive force and the anaerobic energy which takes into account the centrifugal force. This allows us to analyze numerically the different strategies according to the types of track since different designs of tracks lead to straights of different lengths. In particular, we find that the tracks with shorter straights lead to better performances, while the double bend track with the longest straight leads to the worst performances and the biggest difference between lanes. Then for a race with two runners, we introduce a psychological interaction: there is an attraction to follow someone just ahead, but after being overtaken, there is a delay before any benefit from this interaction occurs. We provide numerical simulations in different cases. Overall, the results agree with the IAAF rules for lane draws in competition, where the highest ranked athletes get the center lanes, the next ones the outside lanes, while the lowest ranked athletes get the inside lanes. Introduction In athletics, inside lanes are considered a disadvantage due to curvature, while in outside lanes, there is no one to chase. The aim of this paper is to understand from a physical and mathemat- ical point of view the effect of the curved part of a track and of the lane number on the running performance both for a single runner and for a two-runner race. To our knowledge, no optimal control problem including these effects has been studied. There is a huge literature on the way of running on a curved track, see for instance [1–8]. Nev- ertheless, it is never coupled with the psychological effect to have a neighbor on the next lane, which is mentioned as important. Furthermore, though the IAAF regulations [9] do not impose a fixed shape of track, but allow the straights to vary between 80m and 100m, we are not aware of any study discussing the effect of the the lane and the track coupled with the psy- chological effect. In this paper, we will build on a model introduced by Keller [10] and extended by [11, 12], to investigate how the shape of the track and the centrifugal force change the optimal strategy in a race: this leads to longer race times for higher curvatures, and therefore favors the outer lanes. Estimating the performance of champions based on the modeling of Keller [10] has PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 1 / 23 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Aftalion A, Martinon P (2019) Optimizing running a race on a curved track. PLoS ONE 14(9): e0221572. https://doi.org/10.1371/journal. pone.0221572 Editor: Gordon Fisher, University of Alabama at Birmingham, UNITED STATES Received: December 13, 2018 Accepted: August 10, 2019 Published: September 5, 2019 Copyright: © 2019 Aftalion, Martinon. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper. Funding: The authors received no specific funding for this work. Competing interests: The authors have declared that no competing interests exist. been developed by various authors [6, 13–16], but never taking into account so many parame- ters as in this paper. We will also introduce a model taking into account the psychological effect between two runners. This is made up of two effects: on the one hand, the attraction by a runner close ahead, and on the other hand, the delay before any benefit from the interaction occurs again after being overtaken. This delay model is inspired by a paper on walking [17]. Let us point out that the mathematical problem encompassing delay in the equations is quite involved. We model the attraction by a runner close ahead as a decreased friction, since the focus on chasing someone ahead improves the runner’s economy. Due to the staggered start positions in the curved part of the track, this “rabbit effect” is less favorable on the outer lanes. After introducing the model, we perform simulations using the optimal control toolbox BOCOP [18]. Since the IAAF regulations do not impose a single shape of track, we ana- lyze the effect of the shape of the track on the optimal velocity profile, as well as the influence of the various parameters of the runner for a single runner. Then we perform simulations for two runners and our results show that the combination of the centrifugal force and the two runners interaction brings a numerical justification to the fact that the central lanes are the most favorable to win a race. Race model Model for a single runner race Single runner on a straight track. When a runner is running on a straight, as used by Keller [10], according to Newton’s second law, the acceleration is equal to the sum of forces. We can list two forces, the propulsive force f(t) in the direction of motion, and the friction force, that we assume to be linear in velocity. This leads to the first equation of motion for the velocity v(t) written by unit of mass: _vðtÞ ¼ f ðtÞ considered in this paper, we assume a linear function σ and note σf the final value, thus sðeÞ ¼ sf e0 centrifugal force, which, by unit of mass, is fc = v2/R where v is the velocity of the runner and R the curvature radius. In order to produce a mathematical model for the dynamics in the curved part, we have to take into account the centrifugal force in Newton’s law of motion and project this equation on the 3 directions of motion. Even on straights, there is an equation to be written in the z direction: the reaction of the ground, N, is equal to the weight. By the principle of action/reaction, the reaction of the ground is equal to the runner’s propulsive force in the z direction. Note that the runner does not have his feet on the ground all the time in the stride: he rather pushes (propulsive force) only for some time in a stride [19]. Some remarks in [26] can be found related to this issue. We point out that there is an interesting explanation of the effect of arms to counterbalance the torque, and that since there are two legs, the reaction on each leg is not exactly the same [1]. In this paper, we do not include these effects as we believe them to be of minor importance. The specificity of our work is that although we consider a mean force and mean velocity in a stride, our model allows us to compute an instantaneous force and speed along the race. On a curve, the runner makes an angle α with the vertical axis to balance the centrifugal force. The runner is subject to gravity g, to the reaction of the ground N along the angle α, and to the centrifugal force fc = v2/R (see Fig 1). One has to consider the equations of motion in the centrifugal direction and the z direction, which lead to v2 R ¼ N sin a; g ¼ N cosa ð9Þ which provides the angle according to the velocity and the value of N: tana ¼ v2 Rg ; N2 ¼ g2 þ v4 R2 : ð10Þ By the principle of action/reaction, the propulsive force in the transverse direction is the opposite of the reaction of the ground in the horizontal direction, hence is equal to N sin α. Moreover, the propulsive force in the vertical direction is N cos α. The total propulsive force F is therefore such that F2 = f2 + N2 where we recall that f is in the direction of movement. From (10), we find F2 ¼ f 2 þ N2 ¼ f 2 þ g2 þ v4 R2 : ð11Þ Since F has to be bounded and g is constant, this leads to the new constraint f 2 þ v4 R2  f 2 M: ð12Þ We point out that eventually the effect of the centrifugal force is taken into account in the force constraint. It cannot have an energy effect directly since the centrifugal force does not produce any work. Study of the track shape. It is important to know the exact shape of the track since it influences the runner’s optimal pacing strategy and performance. However, there is no fixed regulation to build an athletic track. Actually, as indicated in the IAAF manual [9], the length of the straight part can vary between 80 and 100m, while the curved part can be a half circle (‘standard’ track) or two different circular sections (‘double bend’ tracks). We choose to study a standard track with an 84.39m straight part, and then two double bend tracks with straight parts of 79.996m and 98.52m respectively. The shapes and dimensions of theses tracks are Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 4 / 23 detailed in Fig 2 and Table 1. Note that for races longer than 100m, runners start the race in the curved part. The starting positions are therefore adjusted in order to have the same total distance for all lanes (‘staggered start’). Note that each runner is assumed to run at a distance of 30cm from the inner limit of the lane. This is how the radius for the circular parts is set in order to obtain a 400m distance for lane 1. Then the width of each lane is 1.22m. This leads to different radii of curvature Rk(s) depending on the lane k and the distance s run on the lane since the start. On the straight part, 1/Rk(s) = 0. For more details on the value of Rk(s) according to the track, we refer to the Appendix “Track Shape Details”. We want to point out that at the junction between the circular and straight parts, the runner will experience a discontinuity in the centrifugal force. This force is 0 on the straight part and Fig 1. Illustration of the forces on the runner. https://doi.org/10.1371/journal.pone.0221572.g001 Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 5 / 23 can reach a value of the order of 2.5N per kilo on the circular part (since v * 10m/s and R * 40m), which is about one quarter of the weight. We will see on the numerical simulations that this may lead to an acceleration of the runner when reaching the straight part of the track. One could think that it would be better to build a track where the curvature goes smoothly from 0 to the value of the matched circle so that the runner experiences a continuous variation of his centrifugal force. This type of curve, known as a clothoid, is used for instance for railways and roads. The simulations in Section “Numeri- cal simulations for a single runner” indicate that the final time is actually larger on a clothoid, because the smooth transition leads to a smaller radius for the circular part, therefore a larger centrifugal force. One of the main results of our simulations is that the tracks with shorter straights lead to better performances (see Section “Effect of different track shapes”). Final model for a single runner race. The optimal problem is to minimize T = y(d) with y(s), f(s), e(s) solving (6) and (7), σ being given by (3), with the bounded control df ds    0:015 ð13Þ and the force constraint coming from (12) f 2ðsÞ þ 1 _y4ðsÞR2 kðsÞ  f 2 M ð14Þ where the curvature radius Rk(s) is prescribed according to the lane k and the track shape, see the Appendix “Track shape Details”. We use the convention Rk(s) = + 1 on a straight. Table 1. Track parameters. Track Straight Circle Standard 84.39m (36.50m, 180˚) Track Straight Circle 1 Circle 2 Double Bend 1 79.996m (34.00m, 2 × 70˚) (51.543m, 40˚) Double Bend 2 98.52m (24.00m, 2 × 60˚) (48.00m, 60˚) https://doi.org/10.1371/journal.pone.0221572.t001 Fig 2. Shape for standard track (left) and and double bend 2 track (right). https://doi.org/10.1371/journal.pone.0221572.g002 Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 6 / 23 Finally, introducing a state variable for the inverse of speed z(s) = 1/v(s), the optimal control problem for a single runner is ðOCPÞ1 min yðdÞ; _yðsÞ ¼ zðsÞ; s 2 ½0; dŠ; yð0Þ ¼ 0; _zðsÞ ¼ z2ðsÞ=t basis for runners and the focus on chasing produces better running economy [29–31]. This psychological effect is indeed acknowledged by runners (sometimes called “rabbit effect”) and can allegedly have an effect as high as 1 second per 400m lap [31]. The differential equations for y1 and y2 are therefore €y1ðsÞ ¼ multiply the interaction term F by an characteristic function Iη defined by IZðsÞ ¼ 0 if an overtaking has occured on ½s Final model for a 2-runner race. We define Ti = yi(d) and F from (18). The optimal con- trol problem becomes: ðOCPÞ2 minð minðT1; T2Þ þ kwðT1 þ T2ÞÞ; _yiðsÞ ¼ ziðsÞ; s 2 ½0; dŠ; yið0Þ ¼ 0; zð0Þ ¼ _yið0Þ ¼ 1=v0 i ; i ¼ 1; 2; _eiðsÞ ¼ sðeiðsÞÞziðsÞ Single runner on a straight track We start with a simple straight 200m race to illustrate the effect of parameters fM, τ, and e0. We take as reference athlete A1 of Table 2. The corresponding speed and force profiles are shown with black lines in Fig 4. The velocity increases to its peak value vm * fM τ and then decreases. The runner does not have enough energy to run the whole duration of the race at maximal force. The propulsive force starts at its maximal value fM, then decreases at the constant rate |df/ds|max. The time at which the force begins to decrease depends on the values of fM and e0. Indeed, increasing e0 does not change the beginning of the race but allows to run longer at f = fM. On the other hand, increasing fM increases the peak velocity but does not change much the second part of the race. Finally, increasing τ has a more uniform effect and increases the velocity for the whole duration of the race. Single runner on a standard curved track We simulate the same runner on the so-called standard track, i.e. 115.61m half circle of radius 36.80m followed by a 84.39m straight. Fig 4 shows the race profiles obtained for the inner and outer lanes (respectively 1 and 8), and the straight race. The time splits for 50–100m, 100– 150m and 150–200m are indicated in the figure: we have chosen the parameters for A1 so that they match the order of magnitude of time splits for athletes in World Championships. The velocity profiles of the curved track are quite different from the straight track: 1. the runner starts slower because of the curvature: even though he puts his maximal propul- sive force at the start, part of it is used to counterbalance the centrifugal force, resulting into a lower effective force and a lower velocity f 2 M  f 2 init þ ðv0Þ 4 R2 k : Fig 4. Single runner A1 on a standard track, lanes 1 and 8, and straight track. Force in N/kg vs distance on the right graph. Speed vs distance on the left graph, with the constant speeds given by Eq (21) in dashed lines. Time splits for 50–100m, 100–150m and 150–200m are indicated. https://doi.org/10.1371/journal.pone.0221572.g004 Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 11 / 23 2. in the middle part of the race, the maximal propulsive force is reached and we can derive from (1) and (12) the relation between f and v: v ¼ f t and f 2 M ¼ f 2 þ f 4t4 R2 k ð20Þ with Rk the curvature radius on lane k. We can compare this formula with our simulations: on the straight vs = fM τ, while from (20), the velocity in the middle of the race on lane k is v2 s ¼ v2 k þ v4 kt2 R2 k ; that is v2 k ¼ Effect of different track shapes Now we study the effect of different track shapes defined in Fig 2: standard with 84.39m straight (STD), double bend 1 with 80m straight (DB1), double bend 2 with 100m straight (DB2), and two modified standard tracks with smoothed curvature, including clothoid junc- tions of 10m (CL1) and 30m (CL2). For the clothoid tracks, we choose a straight of 84.39m as the standard track. As explained in the Appendix, the length of the junctions provides the radius of the circle and the angle, which are respectively 33.32m and 164˚ for (CL1) and 29.95m and 118˚ for (CL2). For each track shape, we simulate the race on the inner and outer lanes (1 and 8). The results are summarized in Table 3, with the races for runner A1 (on lane 5) shown in Fig 6. Ref- erence athlete A1 has a difference of 0.17s between the best (DB1 track, lane 8) and worst (DB2, lane 1) case. As mentioned previously, runner A2 with a very high force fM = 13 is almost unaffected by the curvature, with times varying only between 20.32s and 20.36s. Yet, the DB2 track is still worse than the others. Conversely, athlete A3, with a lower force fM = 6.5, is more affected, with 1.01s between the best and worst cases. The DB2 is his worst track and his best performance is on the standard track. Our results show a time difference between inner and outer lanes ranging from 0.02s for the standard track to 0.15s for the worst double bend track. This is consistent with [6] who also finds the double bend track to be the worst, using a simplified model based on constant mean Fig 5. Single runner A3 on a standard track, lanes 1 and 8, and straight track. Force in N/kg vs distance on the right graph. Speed vs distance on the left graph, with the constant speeds given by Eq (21) in dashed lines. Force and velocity increase when the centrifugal force disappears. https://doi.org/10.1371/journal.pone.0221572.g005 Table 3. Times for different runners and track shapes. runner shape: STD DB1 DB2 CL1 CL2 Straight A1 lane 1 20.48 20.49 20.62 20.50 20.56 20.43 A1 lane 8 20.46 20.45 20.47 20.46 20.49 20.43 A2 lane 1 20.32 20.33 20.36 20.33 20.34 20.31 A2 lane 8 20.32 20.32 20.32 20.32 20.33 20.31 A3 lane 1 20.72 20.80 21.55 20.80 21.03 20.32 A3 lane 8 20.54 20.55 20.66 20.57 20.63 20.32 https://doi.org/10.1371/journal.pone.0221572.t003 Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 13 / 23 velocity and curvature. Let us point out that in the next section, we will study a two-runner race where the effect of the lane becomes more pronounced: we find a larger difference between the best and worst mean time per lane. Focusing on runner A1 in Fig 6, we analyze more closely the effect of the track shape and lane: • DB1 is the quickest track for the outside lane, though it is very close to STD. • The standard track has the smallest difference between lanes. • DB2 is the slowest track, from 0.01s on the outside lane to 0.14s on the inner lane. When on the outer radius of curvature 24m, the velocity significantly decreases. • CL1 is quite close to DB1 and STD, though a little slower. CL2 is slower than DB2 on the outer lanes, although not as bad in terms of difference between lanes. It may seem surprising that the tracks with smoothed curvature do not perform better than the ones with a discontinuous curvature. This comes from the fact that the clothoid junction actually results in a smaller radius for the circular part, and thus a greater curvature. The lon- ger the clothoid junction, the more pronounced the effect, and the slower the times. To conclude the single runner races, it appears that the track with the shortest straight is the quickest track for strong athletes in outer lanes. The standard track shape is the one with the best race times overall, and also the smallest time gap between the inner and outer lanes. On the opposite, the double bend with the long 100m straight (DB2) yields the worst times overall, and the highest gap between the inner and outer lanes. These conclusions seem consistent with runners’ feelings though there is no study yet of what the ideal shape of track would be for a specific runner. Numerical simulations for two runners We move to the simulations for two-runner races, combining the interaction effect with the curvature effect previously studied for the single runner case. Firstly, we study races with two runners competing in adjacent lanes, to see the effect of the interaction. Then we compute the mean times corresponding to all possible races of a runner versus himself and find that the best lanes are indeed the center ones. Races on different lanes and illustration of the interaction effect We perform simulations for the optimal control problem (OCP2) for two runners, combining the interaction effect with the centrifugal force. We first set A1 to be the runner on each lane 1 and 2. Fig 6. Effect of the track shape on the race time for Runner A1 vs lane number (left graph). Speed profile for lane 5 (right graph) with zoom. The tracks are standard (STD), double bend 1 (DB1) with short straights, double bend2 with long straights (DB2), and curves with short and long clothoid junctions (CL1 and CL2). https://doi.org/10.1371/journal.pone.0221572.g006 Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 14 / 23 We recall that if A1 runs alone, his time on lane 1 is 20.485s and on lane 2 20.480s, so of course because of the centrifugal effect, lane 2 is quicker. Due to staggered starts, as soon as we set the interaction, the runner on lane 1 benefits from the interaction at the beginning of the race. First, we set the interaction term γ = 0.04 but with no inhibition η = 0. The results are illustrated in Fig 7, with the velocity profile in lane 1 on the left and the interaction for each runner and relative distance on the right. When the relative distance is negative, the runner in lane 1 is behind. So in this case, the runner in lane 2 wins the race and they overtake each other twice: lane 1 starts behind because of the staggered starts, benefits from interaction and overtakes at 50m; then lane 2 benefits from interaction right away and is able to overtake again at 150m. Then they are on the straight, very close to each other, lane 1 benefits from interaction and is ready to overtake again but loses in the end by 0.04s. Then in Fig 8, the interaction term is set at γ = 0.04, and the inhibition frame is η = 20m. This means that the positive interaction is disabled when a runner is overtaken in the previous 20m of race. Fig 8 shows the speed profile (left graph) and interaction / inhibition terms (right Fig 7. Race A1 vs himself at lanes 1-2, with interaction γ = 0.04 and a frame η = 0 that is no inhibition. Left graph: speed profile and time splits of the runner at lane 1. Right graph: distance gap and interaction term for both runners. The sign change of the distance gap corresponds to the overtaking. Lane 2 wins by 0.04s. https://doi.org/10.1371/journal.pone.0221572.g007 Fig 8. Race A1 vs himself at lanes 1-2, with interaction γ = 0.04 and a frame η = 20m for the inhibition. Left graph: speed profile and time splits of the runner at lane 1. Right graph: distance gap and interaction term for both runners. The sign change of the distance gap corresponds to the overtaking. Lane 1 wins by 0.27s. https://doi.org/10.1371/journal.pone.0221572.g008 Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 15 / 23 graph). Compared to the race without inhibition in Fig 9, we observe a different behaviour with only one overtaking and the runner on the inside winning by 0.27s. Note that since we optimize the whole race, there is no reason for the race with inhibition to coincide with the race without inhibition, even before any overtaking occurs. We observe that the inhibition (on the right graph) correctly detects the overtaking and suppresses the interaction accordingly. This prevents the overtaken runner at lane 2 to keep up (and eventually catch up) with the one at lane 1, as we see that the distance gap increases after the overtaking. In the race without inhi- bition, the overtaken runner was benefiting from the interaction right away, which allowed him to catch up and take the lead back. With inhibition, the runner on lane 1 manages to win the race, though he is on a disadvantageous lane. In the full race, of course, the runner in lane 2 has a neighbour on the other side which changes the total result. There are cases where, though the inhibition η = 20, there are still two overtakings. We study for instance the 5-4 race, with the speed and force profile of the runner at lane 5 shown in Fig 9. Without interaction (γ = 0), lane 5 wins without any overtakings, with final time 22.47s. With interaction (γ = 0.04 and η = 20), lane 5 still wins after 2 overtakings, with final time 22.23s. At the start, the runner on lane 4 benefits from the interaction due to the runner at lane 5 being ahead (staggered start). He catches up then overtakes the outer runner, who in turn gains the interaction, catches up and overtakes the inner runner again. At the end the inner runner, being behind, has the interaction again and is catching up with the outer runner, but too late. We have also made simulations with runner A1 vs runner A2, and though runner A2 is stronger in force, on some lanes, runner A1 can benefit from interaction to be able to win. We point out that the interaction parameters can be runner dependent since some may be very sensitive to this effect and others much less. Mean time per lane In a real race, there are eight runners, however our model is only for two. Therefore, we simu- late a set of races with two identical runners, the first on a fixed lane, the second on each possi- ble other lane. We define T k1;k2 1 to be the time for the winner in the race between two identical runners A1 in lanes k1 and k2. We want to compute the average performance at lane i as the Fig 9. Race A1 vs himself at lanes 5-4, with interaction γ = 0.04 and inhibition η = 20m. Left graph: speed profile and time splits of the runner at lane 5. Right graph: distance gap and interaction term for both runners. The sign changes of the distance gap correspond to the 2 overtakings. https://doi.org/10.1371/journal.pone.0221572.g009 Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 16 / 23 mean time T i ¼ 1 7 X j¼1::8;j6¼i T k1¼i;k2¼j 1 : First, we compute the times T i;j 1 : the best times in j for each i are indicated in Table 4. In Fig 10, we have plotted the times for i = 1, 5, and 8. The best times are obtained for the maximal interaction, namely with the second runner on an adjacent lane. For runner A1 on lane 5, his Table 4. Athlete A1 at lane k running against himself at lane k − 1. Interaction γ = 0.04 with inhibition η = 20m. Race time and gain with respect to solo race time. lane 2 3 4 5 6 7 8 solo time 20.480 20.475 20.471 20.467 20.464 20.461 20.459 2-runner time 20.300 20.292 20.283 20.276 20.270 20.264 20.259 time gain 0.180 0.183 0.1880 0.191 0.194 0.197 0.200 https://doi.org/10.1371/journal.pone.0221572.t004 Fig 10. Times for athlete A1 at lanes 1,5,8, running against himself. https://doi.org/10.1371/journal.pone.0221572.g010 Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 17 / 23 best performance is obtained with a neighbor on lane 4 rather than 6. We recall that the model includes a lateral attenuation for the interaction, which is 0 when runners are more than 3 lanes apart. If we compare the best time for each case, it is decreasing with the lane. We show in Fig 11 the mean times T i obtained for runner A1 against himself, with an inter- action weight γ = 0.04 and η = 20 when he runs on each lane i. If we look at the overall perfor- mance then lane 5 is the best, followed by lane 6, 4, 7, 3, 8, 2 and lane 1 is by far the worst. We compare with the solo case (γ = 0) where of course the outside lane is the quickest. The results are nicely consistent with the IAAF rules for the lane drawn. Indeed, according to the IAAF rules [9], starting lanes are drawn in three lots: • a first draw is made for the four best runners in the center lanes 3, 4, 5 and 6. • a second draw is made for the next two runners between the outer lanes 7 and 8. • a last draw is made between the runners with the lowest performance to get the inside lanes 1 and 2. Fig 11. Mean times per lane for runner A1 when in lane i vs himself in all other lanes. Without interaction (γ = 0), and with interaction γ = 0.04, η = 20. Lane performance (sorted by mean time): T 5 < T 6 < T 4 < T 7 < T 3 < T 8 < T 2 < T 1. Gap T 1 Nevertheless, we find that the inside lanes 1,2 are a real disadvantage, the more so as if the runners are not as strong. In [4] the authors recall some average time data for Olympics 1996 and 2000, and World Championship 2001: they indicate an average time gap of 0.16s between inside lanes 1 and 2 and outside lanes 7 and 8. We obtain a smaller gap of 0.047s, which may be due to the fact that we consider identical runners in our simulations, while in actual races the athletes in the out- side lanes were supposedly stronger than those in the inside lanes. Conclusion In this paper, we have studied how the geometry of the track and the psychological interaction between runners affect performances. We have introduced an optimal control model taking into account the centrifugal force as a limiting factor for the maximal propulsive force. We couple this with a new model describing the positive interaction exerted by a runner close ahead and the delay to benefit from it after being overtaken. We carry out numerical simula- tions for several runner profiles on different track shapes. The results indicate that the track with the shortest straights is the quickest for strong athletes in outside lanes. The so-called standard track (two straights and half circles) yields the best performances overall. The double bend tracks with longer straights (DB2) are significantly slower. In particular running on lane 1 on the DB2 track appears to be an overwhelming disadvantage. Furthermore, the combination of the centrifugal and interaction effects leads to the center lanes being the most favorable, followed by the outside lanes, with the inside lanes being the worst. These results fit very well with the IAAF rules for lane draws, which follow this prefer- ence order. Appendix: Track shape details Note that each runner is assumed to run at a distance of 30cm from the inner limit of the lane. This is how the radius of the circular parts is set in order to obtain a 400m distance on lane 1. Standard track The standard track is made up of a circular half-circle of length lc = 115.61m followed by a straight of 84.39m, for a total distance of 200m, which yields R1 ¼ lc=p ¼ 36:80: Since the runner is assumed to be 30cm away from the boundary of the lane, the radius of construction is R1 − 0.3. We denote by Rk the radius for the runner on lane k. Since the width of a lane is 1.22m, the radius at lane k is Rk ¼ R1 þ 1:22ðk the same total distance. This yields the starting angle y0 k ¼ 1:22ðk straight part and lc the length on the circle, we find lc R þ 2φðlÞ ¼ p: Since the total angle for one clothoid is 2φðlÞ ¼ l=R, this equation leads to lc þl ¼ Rp. Moreover ls þ lc þ 2l ¼ d where d is the distance of the race, that is 200m in our case, thus Rp ¼ d On a curved track, this approximation is adjusted by projecting the two runners on a median circle, while also taking into account the staggered start on different lanes: rðsÞ ¼ ðy2ðsÞ 14. Mathis F. The Effect of Fatigue on Running Strategies. SIAM review. 1989; 31(2):306–309. https://doi. org/10.1137/1031054 15. Quinn MD. The effects of wind and altitude in the 400-m sprint. Journal of sports sciences. 2004; 22(11- 12):1073–1081. https://doi.org/10.1080/02640410410001730016 16. Woodside W. The optimal strategy for running a race (a mathematical model for world records from 50 m to 275 km). Mathematical and computer modelling. 1991; 15(10):1–12. https://doi.org/10.1016/0895- 7177(91)90086-M 17. Lemercier S, Jelic A, Kulpa R, Hua J, Fehrenbach J, Degond P, et al. Realistic following behaviors for crowd simulation. In: Computer Graphics Forum. vol. 31. Wiley Online Library; 2012. p. 489–498. 18. Bonnans F, Martinon P, Giorgi D, Grelard V, Maindrault S, Tissot O. BOCOP—A toolbox for optimal control problems;. 19. Morin JB, Slawinski J, Dorel S, Couturier A, Samozino P, Brughelli M, et al. Acceleration capability in elite sprinters and ground impulse: Push more, brake less? Journal of biomechanics. 2015; 48(12):3149–3154. https://doi.org/10.1016/j.jbiomech.2015.07.009 PMID: 26209876 20. Hoogkamer W, Kipp S, Frank JH, Farina EM, Luo G, Kram R. A Comparison of the Energetic Cost of Running in Marathon Racing Shoes. Sports Medicine. 2018; 48(4):1009–1019. https://doi.org/10.1007/ s40279-017-0811-2 PMID: 29143929 21. Hanon C, Thomas C. Effects of optimal pacing strategies for 400-, 800-, and 1500-m races on the _ VO2 response. Journal of sports sciences. 2011; 29(9):905–912. https://doi.org/10.1080/02640414.2011. 562232 PMID: 21547833 22. Spencer MR, Gastin PB. Energy system contribution during 200-to 1500-m running in highly trained ath- letes. Medicine & Science in Sports & Exercise, (2001), 33(1), 157–162. https://doi.org/10.1097/ 00005768-200101000-00024 23. Gastin PB. Energy system interaction and relative contribution during maximal exercise. Sports medi- cine, (2001), 31(10), 725–741. https://doi.org/10.2165/00007256-200131100-00003 PMID: 11547894 24. Aftalion A, Despaigne LH, Frentz A, Gabet P, Lajouanie A, Lorthiois MA, et al. How to identify the physi- ological parameters and run the optimal race. MathS In Action. 2016; 7:1–10. https://doi.org/10.5802/ msia.9 25. Samozino P, Rabita G, Dorel S, Slawinski J, Peyrot N, Saez de Villarreal E, et al. A simple method for measuring power, force, velocity properties, and mechanical effectiveness in sprint running. Scandina- vian Journal of Medicine & Science in Sports. 2016; 26(6):648–658. https://doi.org/10.1111/sms.12490 26. The Physics Of Running. https://wwwreal-world-physics-problemscom/physics-of-runninghtml;. 27. Appert-Rolland C, Hilhorst HJ, Aftalion A. Nash equilibrium in a stochastic model of two competing ath- letes. Journal of Statistical Mechanics: Theory and Experiment. 2018; 2018(5):053401. https://doi.org/ 10.1088/1742-5468/aabc7e 28. Hilhorst H, Appert-Rolland C. Mixed-strategy Nash equilibrium for a discontinuous symmetric N-player game. Journal of Physics A: Mathematical and Theoretical. 2018; 51(9):095001. https://doi.org/10. 1088/1751-8121/aaa883 29. Crews DJ. Psychological state and running economy. Medicine & Science in Sports & Exercise. 1992;. https://doi.org/10.1249/00005768-199204000-00014 30. Stones M. Running under conditions of visual input attenuation. International Journal of Sport Psychol- ogy. 1980; 11(3):172–179. 31. Pugh LGE. The influence of wind resistance in running and walking and the mechanical efficiency of work against horizontal or vertical forces. The Journal of physiology, 1971, 213(2), 255–276. https:// doi.org/10.1113/jphysiol.1971.sp009381 PMID: 5574828 32. Gollmann L, Maurer H. Theory and applications of optimal control problems with multiple time-delays. Journal of Industrial and Management Optimization. 2014; 10(2):413–441. https://doi.org/10.3934/jimo. 2014.10.413 33. Boccia A, Vinter RB. The Maximum Principle for Optimal Control Problems with Time Delays. IFAC- PapersOnLine. 2016; 49(18):951–955. https://doi.org/10.1016/j.ifacol.2016.10.290 34. Guinn T. Reduction of delayed optimal control problems to nondelayed problems. Journal of Optimiza- tion Theory and Applications. 1976; 18(3):371–377. https://doi.org/10.1007/BF00933818 35. Trelat, E., Private communication. 36. Morton RH. Statistical effects of lane allocation on times in running races. Journal of the Royal Statistical Society: Series D (The Statistician), 1997, 46(1), 101–104. Optimizing running a race on a curved track PLOS ONE | https://doi.org/10.1371/journal.pone.0221572 September 5, 2019 23 / 23

Optimizing running a race on a curved track.

09-05-2019

Aftalion, Amandine,Martinon, Pierre

eng

PMC2996571

144 International Journal of Ayurveda Research | July-September 2010 | Vol 1 | Issue 3 Effects of Withania somnifera (Ashwagandha) and Terminalia arjuna (Arjuna) on physical performance and cardiorespiratory endurance in healthy young adults Jaspal Singh Sandhu, Biren Shah, Shweta Shenoy, Suresh Chauhan, G. S. Lavekar, M. M. Padhi Department of Sports Medicine and Physiotherapy, Guru Nanak Dev University, Amritsar, Punjab - 143 005, India ORIGINAL ARTICLE Address for correspondence: Dr. Jaspal Singh Sandhu, Department of Sports Medicine and Physiotherapy, Guru Nanak Dev University, Amritsar, Punjab - 143 005, India. E-mail: [email protected] Submission Date: 19-04-10 Accepted Date: 09-09-10 DOI: 10.4103/0974-7788.72485 IntRoductIon There is renewed interest in traditional medicines because of a perception of lower incidence of side effects. The World Health Organization (WHO) estimates that 80 percent of the world's population presently uses herbal medicines for some aspect of primary health care.[1] Several medicinal plants have been described to be beneficial for cardiac ailments in Ayurveda - the origin of Indian system of Medicine.[2] Withania somnifera (WS), also known as Ashwagandha, Indian ginseng, or winter cherry, has been an important herb in the Ayurvedic and indigenous medical systems for over 3000 years. The roots of the plant are categorised as Rasayanas, and described to promote health and longevity by augmenting defenses against disease, arresting the ageing process, revitalizing the body in debilitated conditions and thus creating a sense of wellbeing.[3] Withania somnifera contains A B S T R A C T Introduction: Several medicinal plants have been described to be beneficial for cardiac ailments in Ayurveda like Ashwagandha and Arjuna. Ashwagandha-categorised as Rasayanas, and described to promote health and longevity and Arjuna primarily for heart ailments. coronary artery disease, heart failure, hypercholesterolemia, anginal pain and can be considered as a useful drug for coronary artery disease, hypertension and ischemic cardiomyopathy. Objective: There are no scientific clinical studies showing effect of both these drugs on exercise performance after regular administration when given as supplements The present study was therefore designed and performed to assess the effects of Withania somnifera (Ashwagandha) and Terminalia arjuna (Arjuna) individually and as a combination on maximum velocity, average absolute and relative Power, balance, maximum oxygen consumption (VO2 max) and blood pressure in humans. Materials and Methods: Forty normal healthy. Subjects (either sex, mean age 20.6 ± 2.5yrs and mean Body Mass Index 21.9 ± 2.2) were recruited after written informed consent was obtained. Institutional Ethics Committee permission was also obtained. Thirty participants were assigned to experimental group of which 10 received standardized root extracts of Withania somnifera, 10 received standardized bark extract of Terminalia arjuna and the rest of the 10 received standardized root extract of Withania somnifera in addition to bark extract of Terminalia arjuna both. Both the drugs were given in the form of capsules (dosage 500mg/day for both the drugs). Ten participants received placebo (capsules filled with flour). All the subjects continued the regimen for 8 weeks. All variables were assessed before and after the course of drug administration. Observations: Our study showed that Withania somnifera increased velocity, power and VO2 max whereas Terminalia arjuna increased VO2 max and lowered resting systolic blood pressure. When given in combination, the improvement was seen in all parameters except balance and diastolic blood pressure. Conclusion: Withania somnifera may therefore be useful for generalized weakness and to improve speed and lower limb muscular strength and neuro-muscular co-ordination. Terminalia arjuna may prove useful to improve cardio-vascular endurance and lowering systolic blood pressure. Both drugs appear to be safe for young adults when given for mentioned dosage and duration. Key words: Absolute and relative power, balance, blood pressure, maximum oxygen consumption (VO2 max), Terminalia arjuna, velocity, Withania somnifera International Journal of Ayurveda Research | July-September 2010 | Vol 1 | Issue 3 145 alkaloids (withanine, withasomnin) and steroidal lactones and glycosides also called as withanoloids and sitoindosides and the extract of Withania somnifera has analgesic, mildly sedative, anti-inflammatory and anabolic activities,[4] and it is useful in stress, strain, fatigue, pain, skin diseases, diabetes, gastrointestinal disease, rheumatoid arthritis, and epilepsy,[5] chronic fatigue syndrome[6] and even during pregnancy without any side effects.[7] It is also used as a general tonic, to increase energy and improve health and longevity.[4] Withania somnifera human studies suggest that, it may promote growth in children and improve hemoglobin level, red blood cell count, and physical performance in adults.[4] Terminalia arjuna is widely used in both Ayurvedic and Unani Systems of medicine, primarily for heart ailments. Terminalia arjuna Wight and Arn. is a deciduous and evergreen tree, standing 20–30m above ground level and belongs to the Combretaceae family.[8,9] It is described as an alexteric, stryptic, tonic, and anthelmintic agent and is also useful in treatment of fractures, ulcers, heart diseases, biliousness, urinary discharges, asthma, tumours, leucoderma, anaemia, excessive perspiration[8] etc and its bark is useful in the treatment of coronary artery disease, heart failure, hypercholesterolemia, anginal pain[10] and can be considered as a useful drug for coronary artery disease, hypertension and ischemic cardiomyopathy.[11-13] Terminalia arjuna has also cardioprotective property,[14] antiviral activity against HSV-2[14,15] and efficiency as potent antioxidant preventing LDL cholesterol oxidation.[16,17] There are no scientific clinical studies showing effect of herbal drugs on exercise performance after regular administration when given as supplements. This study was conducted to explore the effects of these two plants on physical and cardiovascular performance in healthy young adults. MateRIals and Methods Research design The present study was designed to be a randomized controlled, parallel group, single blinded study. Sample Forty healthy individuals of either sex (22 males and 18 females), with a mean age of 20.6 ± 2.5 (aged between 18 to 25 years) years and BMI 21.9 ± 2.2 kg/m² (ranged between 18 to 25) from the population of Guru Nanak Dev University campus volunteered for the study. The sample size was calculated by online ‘Java applets for power and sample size’ software,[18] keeping power of the study at 95%. The subjects were randomly assigned into four groups using the chit in a box method. Group I (n=10): Withania somnifera group, Group II (n=10): Terminalia arjuna Group, Group III (n=10): Withania somnifera and Terminalia arjuna group, Group IV (n=10): Placebo (control) group. Subjects were unaware of which group they were in and which drug they were to receive. It was thus a single blinded study, where all the subjects were completely unaware of drugs which they were going to consume. Selection of subjects College going young adults with age between 18 and 25 years were screened. To avoid confounding effects, we included only those individuals who were free from any lower limb injury within past six months, those whose BMI was between 18 and 25 and who had not participated in regular exercises in gym from past 6 months or more. Individuals who were engaged in regular strenuous physical activity, suffering from chronic illness or had undergone major surgery recently, were suffering from any cardiovascular, musculoskeletal or neurological condition and people with history of alcohol abuse or were under medication of other drugs were excluded. Variables for effect The following variables were assessed before and after drug administration under supervision and while ensuring safety of the subjects: 1. Kinematic Measuring System (KMS) ™ was used to measure maximum velocity. This instrument contains four cameras and they were placed at specific distance and at regular intervals to measure velocity. The participants were asked to sprint and at each phase of camera, the velocity was noted. The maximum velocity was calculated as maximum distance travelled at any phase of camera per second. 2. The same instrument (KMS) was used to measure average absolute and average relative power of the lower limbs. During 10 vertical jumps both the values were derived from the contact mat (automatically calculated kinematic values) and the body mass was calculated by associated software. Absolute power (W) = body mass × gravity × jump height / (contact time/2); Relative power = power (W)/ body mass. 3. A 20-second wobble board test (Kinematic) was performed, and a software program was used to calculate a balance ratio (contact with floor to no contact time). A metal plate connected to the computer hardware was placed under the wobble board. When the perimeter of the wobble board made contact with the metal plate, the duration and frequency (during the 20-second test) of contact was recorded by the software. Subjects received an orientation session for the balance board on a separate day, as well as 1–2 practice attempts on the day of testing. 4. Computer controlled Vista Turbo Trainer™ machine was used for evaluating breath by breath gas exchange kinetics. Peak maximum oxygen consumption (ml/kg/min) was measured by using software ‘Turbofit’ version – 5.04. 5. Sphygmomanometer was used to measure systolic and diastolic blood pressure. Resting blood pressure was taken Sandhu et al.: Withania somnifera and Terminalia arjuna on physical performance and cardiorespiratory endurance 146 International Journal of Ayurveda Research | July-September 2010 | Vol 1 | Issue 3 in consideration. 6. Weighing machine (auto-inc) and kinanthropometric rod were used to measure body mass (kg) and vertical height (meter) to calculate Body Mass Index (BMI). Procedure The study was approved by the Institutional Medical Ethics Committee of Guru Nanak Dev University, Amritsar. Prior to the start of data collection, participants were explained about the drugs and previous research supporting the effectiveness on physical performance and possible side effects due to overdose. Only then the subjects who volunteered to participate in the study were recruited. A written informed consent was taken from each participant prior to recruitment. Only those subjects whose BMI was less than 25[19,20] were recruited. Test drugs Withania somnifera was used in the form of a standardized aqueous root extract and Teminalia arjuna was in the form of aqueous bark extracts. The drugs were obtained from Central Council for Research in Ayurveda and Siddha (CCRAS), Delhi, India. Both the drugs were filled in 500mg gelatin capsules. They were stored in air tight containers and in room temperature below 30ºC throughout the experiment. Both drugs were given in the dose of 1 capsule/day orally for 8 weeks. The compliance of the participant to study drug was ensured as the researcher personally administered the drug to the subjects over the period of 8 weeks. All variables mentioned above were measured before and after 8 weeks of drug administration in Isotonic and VO2 max lab and KMS lab in Department of Sports Medicine and Physiotherapy, Guru Nanak Dev University, Amritsar Monitoring of subjects All subjects were healthy college going young adults with moderately active life style. The subjects were instructed to follow the usual routine without any excess physical exertion or exercises throughout the duration of experiment. All the subjects consumed the same meals given in the hostel mess throughout the procedure and were requested to have meals within specified mess time, when the researcher was present and personally administered the drugs. Volunteers were asked to consume the drug 1 hour after the day meal to maintain uniformity of the drug administration of Withania somnifera and Terminalia arjuna. Though the subjects were informed about possible side effects of the drugs in high dosage, subjects were also asked to report immediately if they feel any side-effect of the drugs but none of them felt any kind of the side-effect. Sandhu et al.: Withania somnifera and Terminalia arjuna on physical performance and cardiorespiratory endurance Statistical analysis The data was analyzed for statistical significance by using the Statistical Package for Social Sciences (SPSS 17.0) software. The student‘t’ test and one way ANOVA were used to analyze the data for the level of significance. The related‘t’ test was used to find intragroup and ANOVA was used to find intergroup differences in pre and post protocol. For all analysis, the P value used for statistical significance was 0.05. All results are expressed as mean ± standard deviation. Results After 8 weeks treatment with Withania somnifera, maximum oxygen consumption increased significantly from 13.54±2.46 to 14.47±2.28 (P=0.005). Similarly, the maximum velocity increased from 5.37±0.75 to 5.53±0.70 (P=0.005), the average Table 1: Effects of Withania somnifera Parameters Withania somnifera Mean ± SD P value Max velocity Pre test 5.37±0.75 0.005 Post test 5.53±0.70 Avg absolute power Pre test 711.90±221.62 0.002 Post test 774.79±247.42 Avg relative power Pre test 11.10±3.17 0.007 Post test 12.22±3.40 Balance Pre test 0.84±0.34 0.412 Post test 0.93±0.33 VO2 max Pre test 13.54±2.46 0.000 Post test 14.47±2.28 Systolic blood pressure Pre test 120.20±3.58 0.591 Post test 119.80±3.19 Diastolic blood pressure Pre test 78.40±3.10 0.443 Post test 78.80±2.70 Table 2 : Effects of Terminalia arjuna Parameters Terminalia arjuna Mean ± SD P value Max velocity Pre test 5.19±0.80 0.180 Post test 5.15±0.81 Avg absolute power Pre test 656.20±220.78 0.024 Post test 680.00±232.51 Avg relative power Pre test 10.29±2.56 0.671 Post test 10.34±2.59 Balance Pre test 0.83±0.33 0.82 Post test 0.84±0.28 VO2 max Pre test 14.34±2.94 0.000 Post test 15.04±2.76 Systolic blood pressure Pre test 123.00±2.87 0.000 Post test 117.80±1.48 Diastolic blood pressure Pre test 78.80±2.35 1.000 Post test 78.80±1.69 International Journal of Ayurveda Research | July-September 2010 | Vol 1 | Issue 3 147 absolute power from 711.90±221.62 to 774.79±247.42 (P=0.002) and average relative power from 11.10±3.17 to 12.22±3.40 (P=0.007). However, there was no significant improvement in balance and blood pressure. Table 1 summarizes these results. The volunteers receiving s treatment with Terminalia arjuna demonstrated significant increase in maximum oxygen consumption capacity from 14.34±2.94 to 15.04±2.76. The systolic blood pressure fell significantly from 123.00±2.87 to 117.80±1.48 mmHg. The average absolute power also increased significantly from 656.20±220.78 to 680.00±232.51 (P=0.024). None of the other parameters showed significant change. This data is summarized in Table 2. Table 3 shows comparison of variables before and after drug administration in group III (Witahnia somnifera and Terminalia arjuna). A significant improvement was seen in average absolute power from 793.61±286.00 to 883.49±274.00 (P=0.000), average relative power from 11.10±3.78 to 12.22±3.69 (P=0.000), maximum oxygen consumption from 16.58±4.70 to 17.70±4.51 (P=0.000), maximum velocity from 5.12±0.86 to 5.21±0.89. The systolic blood pressure fell from 123.40±3.13 to 118.00±2.49 (P=0.000). In comparison, 8 weeks of regular administration of placebo to the control group showed no significant changes in any of the variables [Table 4]. Table 5 shows intergroup comparison of maximum velocity, average absolute power, average relative power, maximum oxygen consumption, and systolic as well as diastolic blood pressure after 8 weeks of drug administration. A significant reduction in resting systolic blood pressure was seen in only group II when groups were compared with each other. ANOVA followed by Post Hoc Multiple Scheffe Range Test after completion of drug dosage showed that group II (Terminalia arjuna) significantly effective (F= 5.757, P= 0.003) in reducing systolic blood pressure [Table 5]. There is no statistically significant difference found in any other parameters when the all four groups were compared with each other. dIscussIon The present study was aimed to assess the effects of Withania somnifera and Terminalia arjuna singly and in combination Withania somnifera and Terminalia arjuna on physical performance and endurance in healthy young adults after an eight week therapy. Maximum velocity, average power (absolute and relative) and balance were measured as physical performance parameters and maximum oxygen consumption and blood pressure as were measured as endurance parameters. Sandhu et al.: Withania somnifera and Terminalia arjuna on physical performance and cardiorespiratory endurance Table 3: Withania somnifera and Terminalia arjuna Parameters Withania somnifera + Terminalia arjuna Mean ± SD P value Max velocity Pre test 5.12±0.86 0.004 Post test 5.21±0.89 Avg absolute power Pre test 793.61±286.00 0.000 Post test 883.49±274.00 Avg relative power Pre test 11.10±3.78 0.000 Post test 12.22±3.69 Balance Pre test 0.72±0.31 0.922 Post test 0.72±0.28 VO2 max Pre test 16.58±4.70 0.000 Post test 17.70±4.51 Systolic blood pressure Pre test 123.40±3.13 0.000 Post test 118.00±2.49 Diastolic blood pressure Pre test 78.60±3.53 0.619 Post test 78.20±1.48 Table 4: Effects of placebo Parameters Placebo Mean ± SD P value Max velocity Pre test 5.30±0.70 0.462 Post test 5.54±0.75 Avg absolute power Pre test 718.29±280.37 0.258 Post test 726.82±279.96 Avg relative power Pre test 10.77±3.36 0.556 Post test 10.84±3.16 Balance Pre test 0.92±0.35 0.974 Post test 0.92±0.29 VO2 Max Pre test 16.02±2.91 0.825 Post test 16.06±2.54 Systolic blood pressure Pre test 121.80±3.58 0.798 Post test 121.60±1.84 Diastolic blood pressure Pre test 79.40±2.99 0.780 Post test 79.60±2.07 Both, the maximum velocity and average power represent short term aerobic activity whereas VO2 max represents long term aerobic and cardiovascular endurance. Balance is an ability to maintain Centre of Gravity (COG) within the base of support with minimal postural sway. It requires integration of inputs from multiple senses. Ayurveda is a rich heritage of herbal practices describing medicinal and nutritional uses of more than 600 plants in seventy books. Many plants have ergogenic effects, with no or very less side effects. Ginseng is known as an adaptogen, which means it increases resistance to physical, chemical, and biological stress and builds energy and general vitality.[21] Withania somnifera is considered to be the “Indian” ginseng.[22] We found that Withania somnifera improved the 148 International Journal of Ayurveda Research | July-September 2010 | Vol 1 | Issue 3 Sandhu et al.: Withania somnifera and Terminalia arjuna on physical performance and cardiorespiratory endurance physical performance and strength parameters in our study after 8 weeks of regular consumption (500mg/day). Singh et al.[7] have described use of Withania somnifera in chronic fatigue syndrome. It helps in delaying onset of fatigue and thus increasing the time for exhaustion and maintaining the power for relatively longer period. In our study, the maximum velocity, average absolute and relative power increased by 2.9%, 8.8% and 10.1% respectively following drug administration compared to the placebo group. Arman et al.,[23] reported that Withania somnifera (improves endurance performance (time to exhaustion) at a moderate intensity of 65% VO2 max, in untrained healthy individuals. In the present study, we found that following 8 weeks of administration of Withania somnifera maximum oxygen consumption capacity increased by 6.8% at moderate intensity but no significant change was seen in balance and resting blood pressure. Terminalia arjuna is a cardio protective drug and is used in ayurveda since centuries for its cardiotonic properties. The present study shows that there is significant improvement in average absolute power of lower limbs by 3.6%. Bharani et al, observed significant improvement in the duration of treadmill exercise in stable angina patients who received Terminalia arjuna when given 500 mg/day for one week.[24] In our study, we found an increase in maximum oxygen consumption capacity by 4.9% after treatment. In animal studies, Ghoshal et al.[25] reported an increased heart rate and force of contraction in cardiac muscles in isolated rats. Shrivastava et al. found a dose dependant fall in blood pressure in rats when Terminalia arjuna bark was given in aqueous form, intravenously.[26] According to Colabawala (1951), the drug is known to have no significant effect on heart rate, blood pressure and cardiac output in healthy volunteers but causes an increase in cardiac output and blood pressure and a decrease in heart rate in patients with a failing heart.[27] Contradicting this statement, in our study we found that, there is significant decrease in systolic blood pressure by 4.2% when compared with placebo group [group IV] but no significant improvement was seen in diastolic blood pressure in healthy young adult volunteers following 8 weeks of Terminalia arjuna bark extract consumption. When Withania somniferaand and Terminalia arjuna were given in combination in group III, all parameters showed significant improvement except balance and diastolic blood pressure. The maximum velocity, average absolute power, average relative power, VO2 max and systolic blood pressure improved by 1.8%, 11.3%, 10.1%, 6.8% and 4.4% respectively in its group when compared with placebo group [group IV]. When results between groups were compared the group which was given both Terminalia arjuna and Withania somnifera (group III) was the most effective in reducing systolic blood pressure (4.37%), which is highest significant reduction in systolic blood pressure between groups followed by group II (4.22%) that consumed only Terminalia arjuna. There is no significant difference were seen for any other parameters Without training or excessive physical exertion, Terminalia arjuna was found to be effective in reducing resting systolic blood pressure in healthy young adults. The maximum velocity was found to be improved the most in the Withania somnifera treated group followed by the group that received both Withania somnifera and Terminalia arjuna. Average absolute power was found to be improved most in the Withania somnifera and Terminalia arjuna group, followed by Withania somnifera group and Terminalia arjuna Table 5: Intergroup comparison of all parameters (One way ANOVA) Sum of Squares df Mean Square F Sig. (p) Max Velocity Between groups 0.853 3 0.284 0.456 0.714 Within groups 22.425 36 0.623 Total 23.278 39 Avg Abs Power Between groups 228120.9 3 76040.29 1.132 0.349 Within groups 2418566 36 67182.39 Total 2646687 39 Avg Rel Power Between groups 44.083 3 14.694 1.403 0.258 Within groups 377.145 36 10.476 Total 421.227 39 VO2 max (ml/kg) Between groups 60.229 3 20.076 2.025 0.128 Within groups 356.909 36 9.914 Total 417.138 39 Systolic blood pressure Between groups 94.8 3 31.6 5.757 0.003 Within groups 197.6 36 5.489 Total 292.4 39 Dialostic blood pressure Between groups 9.9 3 3.3 0.796 0.504 Within groups 149.2 36 4.144 Total 159.1 39 International Journal of Ayurveda Research | July-September 2010 | Vol 1 | Issue 3 149 Sandhu et al.: Withania somnifera and Terminalia arjuna on physical performance and cardiorespiratory endurance group respectively. Withania somnifera and Terminalia arjuna were equally effective in improving relative power of the lower limbs. The maximum oxygen consumption capacity was effectively increased in those subjects, who were given Withania somnifera and Terminalia arjuna in combination followed by those who were given just Terminalia arjuna. The present study was limited to an 8 week period on healthy young adults. The future research should focus on longer treatment duration, dose finding as well as gender specific effects of the drugs. Further studies are also required to assess whether the drugs can improve other physical parameters and to see the effectiveness in elite sports persons so that in future these drugs can be given as ergogenic elements. Withania somnifera may therefore be useful for generalized weakness and to improve speed and lower limb muscular strength and neuro-muscular co-ordination. Terminalia arjuna may prove useful to improve cardio-vascular endurance and lowering systolic blood pressure. Both drugs appear to be safe for young adults when given for mentioned dosage and duration. RefeRences 1. Available from: http://en.wikipedia.org/wiki/Herbalism [Last accessed on 2009 Jul 7]. 2. Shatvalekar S. AtharvaVeda Samhita. 1st ed. Balsad, Maharashtra, India: Suyadhyay Mandal Kila Pardee publication; 1943. p. 25. 3. Weiner MA, Weiner J. Ashwagandha (Indian ginseng). Herbs that Heal. Mill Valley, CA: Quantum Books; 1994. p. 70-2. 4. Mishra LC, Singh BB, Dagenais S. Scientific basis for the therapeutic use of Withania somnifera (ashwagandha): A review. Altern Med Rev 2000;5:334-46. 5. Prakash J, Gupta SK, Dinda AK. Withania somnifera root extract prevents DMBA-induced squamous cell carcinoma of skin in Swiss albino mice. Nutr Cancer 2002;42:91-7. 6. Singh A, Naidu PS, Gupta S, Kulkarni SK. Effect of natural and synthetic antioxidants in a mouse model of chronic fatigue syndrome. J Med Food 2002;5:211-20. 7. Sharma S, Dahanukar S, Karandikar SM. Effects of long-term administration of the roots of ashwagandha and shatavari in rats. Indian Drugs 1985;29:1339. 8. Chopra RN, Ghosh S. Terminalia arjuna: Its chemistry, pharmacology and therapeutic action. Indian Med Gazette 1929;64:70-3. 9. Caius JS, Mhaskar KS, Isaacs M. A comparative study of the driedbarks of the commoner Indian species of genus Terminalia. Indian Med Res Memoirs 1930;16:51-75. 10. Miller AL. Botanical influences on cardiovascular disease. Alterne Med Rev 1998;3:422-31. 11. Bhatia J. Study of the possible cardioprotective role of Terminalia Arjuna in experimental animals and its clinical usefulness in coronary artery disease. MD (Pharmacology) thesis, India: University of Delhi; 1998. 12. Takahashi S, Tanaka H, Hano Y, Ito K, Nomura T, Shigenobu K. Hypotensive effects in rats of hydrophyllic extract from Terminalia Arjuna containing tannin-related compounds. Phytother Res 1997;1:424-7. 13. Jain V, Poonia A, Agarwal RP, Panwar RB, Kochar DK, Mishra SN. Effect of Terminalia arjuna in patients of angina pectoris (A clinical trial). Indian Med Gazette (New Series) 1992;36:56-9. 14. Karthikeyan K, Bai BR, Gauthaman K, Sathish KS, Devaraj SN. Cardioprotective effect of the alcoholic extract of Terminalia arjuna bark in an in vivo model of myocardial ischemic reperfusion injury. Life Sci 2003;73:2727-39. 15. Cheng HY, Lin CC, Lin TC. Antiherpes simplex virus type 2 activity of casuarinin from the bark of Terminalia arjuna Linn. Antiviral Res 2002;55:447-55. 16. Pathak SR, Upadhya L, Singh RN. Effect of Terminalia arjuna on lipid profile of rabbit fed hypercholesterolemic diet. Int J Crude Drug Res 1990;28:48-51. 17. Khanna AK, Chander C, Kapoor NK. Terminalia arjuna: An Ayurvedic cardiotonic regulates lipid metabolism in hyperlipidemic rats. Phytother Res 1996;10:663-5. 18. Java applets for power and sample size software, Availabke from: http://www.stat.uiowa.edu/~rlenth/Power/[Last accessed on 2009 Jul 07]. 19. Executive summary of clinical guidelines on identification, evaluation, and treatment of overweight, obesity in adults. Arch Intern Med 1998;158:1855-67. 20. Sjolie AN. Low back pain in adolescent is associated with poor hip mobility and high body mass index. Scand J Med Sci Sports 2004;14:168-75. 21. Available from: http://herbal-powers.com/ginseng.html [Last accessed on 2009 Jul 07]. 22. Available from: http://en.wikipedia.org/wiki/Withania_ somnifera [Last accessed on 2009 Jul 07]. 23. Arman K, Sandhu JS. Effects of ashwagandha on strength, endurance performance and stress: Unpublished thesis, Department of Sports Med and Physiotherapy, GNDU, Amritsar, India; 2007. 24. Bharani A, Ganguli A, Mathur LK, Jamra Y, Raman PG. Efficacy of Terminalia arjuna in chronic stable angina: A double-blind, placebo-controlled, crossover study comparing Terminalia arjuna with isosorbide mononitrate. Indian Heart J 2002;54:170-5. 25. Ghoshal LM. Terminalia arjuna. Ph.D. thesis, Calcutta, India: Calcutta University; 1909. 26. Srivastava RD, Dwivedi S, Sreenivasan KK, Chandrashekhar CN. Cardiovascular effects of Terminalia species of plants. Indian Drugs 1992;29:144-9. 27. Colabawalla HM. An evaluation of the cardiotonic and other properties of Terminalia arjuna. Indian Heart J 1951;3: 205-30. Source of Support: Nil, Conflict of Interest: None declared.

Effects of Withania somnifera (Ashwagandha) and Terminalia arjuna (Arjuna) on physical performance and cardiorespiratory endurance in healthy young adults.

[]

Sandhu, Jaspal Singh,Shah, Biren,Shenoy, Shweta,Chauhan, Suresh,Lavekar, G S,Padhi, M M

eng

PMC9941268

Vol.:(0123456789) 1 3 European Journal of Applied Physiology (2023) 123:573–583 https://doi.org/10.1007/s00421-022-05084-1 ORIGINAL ARTICLE Modeling lactate threshold in young squad athletes: influence of sex, maximal oxygen uptake, and cost of running Sanghyeon Ji1  · Sebastian Keller1,2,3  · Lukas Zwingmann1  · Patrick Wahl1 Received: 5 July 2022 / Accepted: 20 October 2022 / Published online: 21 November 2022 © The Author(s) 2022, corrected publication 2022 Abstract Purpose This study aimed to investigate: 1. The influence of sex and age on the accuracy of the classical model of endurance performance, including maximal oxygen uptake ( ̇VO2peak ), its fraction (LT2%), and cost of running (CR), for calculating running speed at lactate threshold 2 (vLT2) in young athletes. 2. The impact of different CR determination methods on the accuracy of the model. 3. The contributions of ̇VO2peak , LT2%, and CR to vLT2 in different sexes. Methods 45 male and 55 female young squad athletes from different sports (age: 15.4 ± 1.3 years; ̇VO2peak : 51.4 ± 6.8 mL ⋅ kg−1 ⋅ min−1 ) performed an incremental treadmill test to determine ̇VO2peak , LT2%, CR, and vLT2. CR was assessed at a fixed running speed (2.8 m ⋅ s−1 ), at lactate threshold 1 (LT1), and at 80% of ̇VO2peak , respectively. Results Experimentally determined and modeled vLT2 were highly consistent independent of sex and age (ICC ≥ 0.959). The accuracy of vLT2 modeling was improved by reducing random variation using individualized CR at 80% ̇VO2peak (± 4%) compared to CR at LT1 (± 7%) and at a fixed speed (± 8%). 97% of the total variance of vLT2 was explained by ̇VO2peak , LT2%, and CR. While ̇VO2peak and CR showed the highest unique (96.5% and 31.9% of total R2 , respectively) and common (– 31.6%) contributions to the regression model, LT2% made the smallest contribution (7.5%). Conclusion Our findings indicate: 1. High accuracy of the classical model of endurance performance in calculating vLT2 in young athletes independent of age and sex. 2. The importance of work rate selection in determining CR to accurately predict vLT2. 3. The largest contribution of ̇VO2peak and CR to vLT2, the latter being more important in female athletes than in males, and the least contribution of LT2%. Keywords Maximal metabolic steady state · Performance diagnostics · Aerobic capacity · Endurance performance · Running economy · Youth athletes Abbreviations calLT2 Calculated speed corresponding to lactate threshold 2 calLT2fix Calculated speed corresponding to lactate threshold 2 determined using cost of running determined at a fixed speed of 2.8 m ⋅ s−1 calLT2LT1 Calculated speed corresponding to lactate threshold 2 determined using cost of running determined at lactate threshold 1 Communicated by Michael I. Lindinger. Sanghyeon Ji, Sebastian Keller have contributed equally to this work. * Sanghyeon Ji [email protected] Sebastian Keller [email protected] Lukas Zwingmann [email protected] Patrick Wahl [email protected] 1 Department of Exercise Physiology, German Sport University, Cologne, Germany 2 German Research Centre of Elite Sport, German Sport University, Cologne, Germany 3 Department of Molecular and Cellular Sports Medicine, Institute of Cardiovascular Research and Sports Medicine, German Sport University, Cologne, Germany 574 European Journal of Applied Physiology (2023) 123:573–583 1 3 calLT280% Calculated speed corresponding to lactate threshold 2 determined using cost of running determined at 80% of maximal oxygen uptake C Cost of movement CR Cost of running CRfix Cost of running determined at a fixed speed of 2.8 m ⋅ s−1 CRLT1 Cost of running determined at lactate threshold 1 CR80% Cost of running determined at 80% of maximal oxygen uptake LT1 Lactate threshold 1 (first rise in blood lactate concentration) LT2% Fractional utilization of maximal oxygen uptake at lactate threshold 2 ICC Intra-class correlation coefficient LT2 Lactate threshold 2 ̇VO2peak Maximal oxygen uptake ̇VO2 Oxygen uptake vLT2 Running speed corresponding to lactate thresh- old 2 Introduction Endurance performance depends on a complex interplay of various metabolic and mechanical determinants (Joyner and Coyle 2008). While aerobic capacity, i.e., maximal oxy- gen uptake ( ̇VO2peak ) and its fraction at a disproportion- ate increase in the speed-lactate curve that can be sustained over a longer period (LT2%) (Farrell et al. 1979; McLaughlin et al. 2010; Støa et al. 2020) represent the main metabolic determinants, movement economy (or energy cost of move- ment [C]) depends on the proportion of mechanical power output that contributes to progression (the higher the propor- tion, the more economical the locomotion) (Minetti 2004). Together, these three parameters have been shown to accu- rately predict endurance performance according to formula (1) (Joyner 1991; McLaughlin et al. 2010). For example, McLaughlin et al. (2010) found that in well-trained distance runners, 95.4% of the variation in 16-km running time could be explained by these variables. Recently, Støren et al. (2014) and Støa et al. (2020) used the same equation to model the work rate corresponding to lactate threshold 2 (LT2) in cycling and running, respec- tively. Representing the highest work rate that still elicits a metabolic steady state, LT2 depicts an important parameter for endurance exercise prescription (indicating the upper boundary of the heavy intensity domain) and performance (1) Endurance performance = LT2% ⋅ ̇VO2peak C . prediction (showing high correlations with endurance per- formance) (Faude et al. 2009). Interestingly, in both dis- ciplines, Støren et al. (2014) and Støa et al. (2020) found a strong dependence of LT2 on ̇VO2peak and C as well as a high agreement between calculated and measured work rates corresponding to LT2. Based on these observations, they concluded that training prescription may focus either on improving ̇VO2peak for example using high intensity interval training or on improving C for example by implementing maximal strength training (Støa et al. 2020). Therefore, this model could represent a way to individualize exercise pre- scription for endurance training. So far, however, the model applies only to adult well- trained to elite cyclists (Støren et al. 2014) and to adult rec- reational to elite long-distance runners (Støa et al. 2020), but not to other sport disciplines or age groups. Especially for young well-trained athletes, the model could represent an option to individualize training prescriptions based on the physiological profiles and thus use limited training time as efficiently as possible (Støa et al. 2020). However, it must first be investigated whether the model is dependent on age, since only adult athletes have been studied so far (Støa et al. 2020; Støren et al. 2014). Further, potential dif- ferences between sexes need to be considered, as sex spe- cific prerequisites such as body composition could influence physiological characteristics such as aerobic capacity (Bes- son et al. 2022). In addition, the influence of the methods used to calculate C (model predictor) and LT2 (criterion) are unknown. With regard to LT2, there are a large number of studies that have investigated the agreement of different methods to determine LT2 with the underlying physiologi- cal concept of a maximal metabolic steady state but have yielded heterogeneous results [e.g., Faude et al. (2009)]. Regarding the LT2 determination method used by Støren et al. (2014) and Støa et al. (2020) (i.e., warm-up blood lactate concentration + 2.3 mmol ⋅ L−1 ), to the best of our knowledge, no systematic analysis of validity showing the absolute level of agreement with maximal metabolic steady state has been published. Therefore, re-calculating the model with a threshold concept that validly represents maximal metabolic steady state seems warranted. Regarding model predictors, ̇VO2peak and LT2% repre- sent physiologically well-defined constructs, albeit depend- ent on test protocol and determination method (Faude et al. 2009; Midgley et al. 2007), whereas determination of cost of running (CR) is still strongly debated (Barnes and Kilding 2015; Lundby et al. 2017). This controversy is mainly related to the question of whether or not CR is independent of the running speed and therefore different external work rates have been studied (Iaia et al. 2009; Jones and Doust 1996; Lacour and Bourdin 2015; Svedenhag and Sjödin 1994). In contrast to the common approach of measuring oxygen uptake ( ̇VO2 ) at a fixed submaximal speed (e.g., 12 km ⋅ h−1 ) 575 European Journal of Applied Physiology (2023) 123:573–583 1 3 to ensure achievement of a metabolic steady-state (Barnes and Kilding 2015), Støa et al. (2020) assessed ̇VO2 at a fix percentage (i.e., 70%) of ̇VO2peak calculated from the linear relationship between submaximal running speeds and the corresponding ̇VO2 values. However, apart from the fact that the linearity of the ̇VO2 response to exercise is controver- sial (DiMenna and Jones 2009), from a physiological point of view both assessment methods bear the risk to obtain heterogeneous metabolic responses due to inter-individual variability. For example, Scharhag-Rosenberger et al. (2010) reported a large variability in blood lactate response at a work rate corresponding to 60% and 75% of ̇VO2peak . Simi- larly, measuring ̇VO2 at a fixed submaximal speed (e.g., 12 km ⋅ h−1 ) will most likely elicit different metabolic responses in differently trained individuals. Since such divergent inter- nal metabolic responses may impact inter-individual com- parability, assessing CR at a distinct submaximal metabolic anchor such as the first rise in blood lactate levels (LT1) might be a more individualized option. Due to the potential impact of the cited methodological aspects as well as participant characteristics including sport- ing background, age, and sex on the accuracy of the model, the aims of the present study were to investigate: 1. The accuracy of the model, originally applied to adult runners by Støa et al. (2020), in young athletes of different disciplines depending on age and sex. In contrast to Støa et al. (2020), a validated threshold concept was used as a criterion. 2. The impact of different methods to determine CR on the accuracy of the model. 3. The influence of LT2%, ̇VO2peak , and CR on the running speed at LT2 (vLT2) in young athletes depend- ing on sex. Materials and methods Participants The study sample consisted of young squad athletes from the federal state of North Rhine-Westphalia, Germany (n = 248). All of them participated in regular training and official competitions in various sports (including endurance type individual sports and team, racket, as well as combat sports) on regional to national levels. All athletes gave their assent, and informed consent was obtained from their parents or legal guardians. The experimental procedures were approved by the local ethics committee (approval number 67/2020) and was conducted in accordance with the Declaration of Helsinki. To ensure validity and comparability, only data that met the following criteria were included for further analysis: (a) age < 19 years; (b) exhaustion (see below for determination of ̇VO2peak ); (c) valid determination of LT1 and LT2 using the modified maximal deviation method (see below); (d) number of stages completed dur- ing the incremental step test > three (e) running speed at 80% of ̇VO2peak ≥ 2.4 m ⋅ s−1 (i.e., within the speed range used in the incremental test). If athletes had multi- ple performance diagnostic visits, only data from the first visit were used. A total of 100 athletes (45 males and 55 females) were finally included in the present study. The athlete’s anthropometric characteristics and disciplines are presented in Table 1 and Fig. 1, respectively. Procedures In this cross-sectional study, all athletes completed an incre- mental step test to determine endurance performance as part of a larger performance check-up for young squad athletes at a local performance diagnostic center between January 2018 and January 2022. All tests were performed under constant laboratory conditions on a treadmill (h/p/cosmos, saturn® 250/100, Traunstein, Germany) with an incline of 1% simulating air resistance. Following a two-minute resting measurement in standing position, the initial speed was set to 2.4 m ⋅ s−1 and increased by 0.4 m ⋅ s−1 every 5 min to ensure attainment of metabolic steady state conditions. Between the stages, short resting periods (30 s) were allowed for capillary blood sampling (20 μL ) and tests were performed until volitional exhaustion. Throughout the test, breathing gases (Metalyzer®3B; Cortex Biophysik GmbH, Leipzig, Germany) and heart rate (Polar H7 Sensor; Polar Electro, Kempele, Finland) were recorded every second and averaged over 30 s. The spirom- eter was calibrated weekly with a reference gas (5% CO2 and 15% O2) and before each test with ambient air and with a 3-L syringe, according to the manufacturer’s specifications. Immediately after the test, blood lactate concentrations were determined (Biosen C-line; EKF Diagnostic Sales, Magde- burg, Germany). Parameters Blood lactate concentrations during the incremental step test were plotted against running speed and then fitted by a third- order polynomial function. vLT2 was identified as the point on the lactate performance curve that yielded the maximal perpendicular distance to a straight line formed by the peak lactate point and by the point of the first rise in blood lactate concentration at which the slope of the fitted lactate curve equaled 1.00 (LT1). This method has recently been shown to be a valid estimate of maximal lactate steady state in running (Zwingmann et al. 2019). 576 European Journal of Applied Physiology (2023) 123:573–583 1 3 ̇VO2peak corresponded to the highest measured 30-s moving average of ̇VO2 during the test. Exhaustion was verified using the following criteria (Midgley et al. 2007): respiratory exchange ratio ≥ 1.10, heart rate ≥ 95% of age predicted maximum, blood lactate concentration ≥ 8 mmol ·  L–1, and volitional exhaustion. All spirometric data were averaged over the last third of each stage to ensure that a steady state was achieved at least in the submaximal stages (i.e., ≤ vLT2) (Whipp and Was- serman 1972). LT2% was determined by dividing ̇VO2 corresponding to LT2 by ̇VO2peak. To determine CR, ̇VO2 at three different work rates was divided by the respective running speeds: (1) a fixed run- ning speed of 2.8 m ⋅ s−1 (CRfix); (2) the running speed corresponding to LT1 (CRLT1); and (3) the running speed corresponding to 80% of ̇VO2peak as calculated from the linear regression from running speed and ̇VO2 ( R2 ≥ 0.92) (CR80%). Since extrapolation outside the measured values would have been necessary for several participants to deter- mine 70% of ̇VO2peak in accordance with Støa et al. (2020), 80% was chosen instead being always within the measuring range (except for two participants, which were excluded, see above) and giving the same CR according to Helgerud et al. (2009). Independent of the work rate used, CR was Table 1 Descriptive anthropometric and physiological characteristics (mean ± standard deviation) of the participants ̇VO2peak maximal oxygen uptake, CR cost of running, CRfix CR determined at a fixed speed of 2.8 m ⋅ s−1 , CRLT1 CR determined at lactate threshold 1, CR80% CR determined at 80% of ̇VO2peak , LT2% fractional uti- lization of ̇VO2peak , vLT2 running speed at lactate threshold 2, calLT2fix calculated vLT2 determined using CRfix, calLT2LT1 calculated vLT2 determined using CRLT1, calLT280% calculated vLT2 determined using CR80% Variable All (N = 100) Males (N = 45) Females (N = 55) p (m vs. f) Anthropometrics  Age [y] 15.4 ± 1.3 15.8 ± 1.4 15.2 ± 1.1 0.013  Height [cm] 172.8 ± 9.1 175.2 ± 10.5 170.9 ± 7.3 0.019  Body mass [kg] 62.3 ± 11.8 63.6 ± 14.6 61.2 ± 8.9 0.303 ̇VO2peak  [mL ⋅ min−1] 3182 ± 639 3561 ± 735 2871 ± 304 < 0.001  [mL ⋅ kg−1 ⋅ min−1] 51.4 ± 6.8 56.4 ± 5.9 47.3 ± 4.3 < 0.001 Oxygen CR  CRfix [mL · kg−1 · m −1] 0.223 ± 0.020 0.232 ± 0.021 0.216 ± 0.017 < 0.001  CRLT1 [mL · kg−1 · m −1] 0.222 ± 0.020 0.230 ± 0.020 0.216 ± 0.018 < 0.001  CR80% [mL · kg−1 · m −1] 0.222 ± 0.018 0.229 ± 0.018 0.216 ± 0.016 < 0.001 Energy CR  CRfix[J ⋅ kg−1 ⋅ m−1] 4.84 ± 0.44 5.02 ± 0.46 4.70 ± 0.37 < 0.001  CRLT1 [ J ⋅ kg−1 ⋅ m−1] 4.82 ± 0.43 4.98 ± 0.43 4.68 ± 0.38 < 0.001  CR80% [ J ⋅ kg−1 ⋅ m−1] 4.84 ± 0.40 5.00 ± 0.39 4.71 ± 0.35 < 0.001 Lactate threshold   LT2% [%] 87.0 ± 2.6 87.0 ± 2.9 87.0 ± 2.4 0.999  vLT2 [ m ⋅ s−1] 3.38 ± 0.39 3.61 ± 0.41 3.19 ± 0.24 < 0.001  calLT2fix [ m ⋅ s−1] 3.34 ± 0.38 3.54 ± 0.42 3.18 ± 0.25 < 0.001  calLT2LT1 [ m ⋅ s−1] 3.36 ± 0.41 3.58 ± 0.44 3.18 ± 0.27 < 0.001  calLT280% [ m ⋅ s−1] 3.36 ± 0.39 3.58 ± 0.43 3.18 ± 0.24 < 0.001 Fig. 1 Number of athletes included in the study, separated by sex and sport discipline 577 European Journal of Applied Physiology (2023) 123:573–583 1 3 specified as oxygen cost in mL · kg−1 · m −1 and as energy cost in J ⋅ kg−1 ⋅ m−1 using the respiratory exchange ratio (Péronnet and Massicotte 1991) to take into account poten- tial differences in substrate use (Barnes and Kilding 2015). In addition to CR, minute ventilation was determined at the three different running speeds. Based on Eq. 1 (Støa et al. 2020; Støren et al. 2014), vLT2, as an indicator of endurance performance, was cal- culated using individual ̇VO2peak , LT2%, and each of three differently determined values of CR (using CRfix: calLT2fix, using CRLT1: calLT2LT1, using CR80%: calLT280%). Statistical analysis The Statistical Package for the Social Sciences (version 27.0, IBM SPSS, Chicago, IL) was used for statistical analysis. All results were interpreted as significant for 훼 = 0.05 . For all data, normal distribution and homogeneity of variance were verified using the Shapiro-Wilk test and Levene’s test, respectively. Differences between male and female athletes were determined using independent sample t-tests. Correlations between physiological parameters (model predictors and criteria) were determined using Pearson’s correlation coefficient r. These were interpreted as follows: < 0.30 = negligible, 0.30–0.50 = low, 0.50–0.70 = moderate, 0.70–0.90 = high, and > 0.90 = very high (Mukaka 2012). Intra-class correlation coefficients (ICC) were calculated based on a single measure absolute agreement, two-way mixed model, to examine the agreement between the meth- ods for determining CR and between vLT2 and the modeled threshold estimates. According to Koo and Li (2016), the degree of agreement was interpreted as follows: < 0.50 = poor, 0.50–0.75 = moderate, 0.75–0.90 = good, and > 0.90 = excellent. In addition, a Bland-Altman analysis was per- formed to assess the concordance between vLT2 and the modeled threshold estimates. Multiple regression analysis using bi-directional step- wise selection procedure (criteria: probability of F to enter ≤ 0.05, probability of F to remove ≥ 0.10) was performed to estimate the association between vLT2 (dependent variable) and the three physiological variables ̇VO2peak , LT2%, and CR (independent variables). In addition, sex and age were included as independent variables to assess the influence of these anthropometric variables on the accuracy of the model. To better understand the regression model, we addition- ally assessed the contributions of each predictor (independ- ent variable) to the regression R2 using a commonality analysis using R [R Core Team (2021), Version 4.1.2, yhat package] (Nathans et al. 2012; Ray-Mukherjee et al. 2014). In this way, it can be determined how much variance in the dependent variable is uniquely explained by a single predic- tor, independent of all other predictors (unique effects) and how much variance in the dependent variable is shared by a combination of the predictors (common effects). Further, negative commonality coefficients may indicate improved overall predictive power of the model associated with the suppressor variable, which removes some irrelevant vari- ance or error in other variable(s) in the common effect, thus increasing the variance contributions of other variable(s) to the regression R2 (Nathans et al. 2012; Pandey and Elliott 2010). All analyses were performed both for the whole sam- ple and for the male and female subsets. Results Table 1 summarizes the characteristics of the athletes. Both in absolute and relative terms, male athletes had a higher ̇VO2peak compared to female athletes ( p < 0.001 ). Regard- less of the method used (in terms of both running speed and expression), male athletes showed poorer CR ( p < 0.001 ) but better vLT2 ( p < 0.001 ) than female athletes. No sex differ- ence was found for LT2% ( p = 0.999 ). The running speeds corresponding to LT1 were 3.00 ± 0.33 m ⋅ s−1 and 2.70 ± 0.18 m ⋅ s−1 , and those corresponding to 80% ̇VO2peak were 3.29 ± 0.39 m ⋅ s−1 and 2.92 ± 0.22 m ⋅ s−1 in male and female athletes, respectively. In addition, minute ventilation was significantly higher in male than in female athletes at all submaximal running speeds used for CR assessment (at LT1: 77.2 ± 14.6 L ⋅ min−1 vs. 63.2 ± 8.3 L ⋅ min−1 , at 80% of ̇VO2peak : 89.3 ± 19.8 L ⋅ min−1 vs. 70.9 ± 8.8 L ⋅ min−1 , p < 0.001 ), except at a running speed of 2.8 m ⋅ s−1 (108.6 ± 35.3 L ⋅ min−1 vs. 97.4 ± 26.0 L ⋅ min−1 , p = 0.069). The mean differences along with limits of agreement between calLT2fix, calLT2LT1, and calLT280% vs. vLT2 are shown in the Bland-Altman plots (Fig. 2) and in Table 2. All modeled thresholds showed excellent concordance with vLT2 (Table 2). Since CR80% provided the model with the highest accu- racy due to the smallest limits of agreement (i.e., calLT280% ± 4% vs. and calLT2fix ± 7% and ± 8%, respectively), it was used for further regression analyses. The relationships between vLT2 and the three physiological variables were presented both for the whole sample and separated by sex in Table 3. Entering ̇VO2peak , LT2%, and CR80% into the multiple linear regression model was able to explain 97%, 97%, and 95% of the variance in vLT2 for all, male, and female ath- letes, respectively (Table 3). Based on the selection criteria in the stepwise selection procedure, sex and age were not included in the final multiple regression model. According to the commonality analysis, ̇VO2peak showed the highest unique contribution to the total regression R2 followed by CR80% and LT2% regardless of the analyzed 578 European Journal of Applied Physiology (2023) 123:573–583 1 3 subset. Regarding the common effects, all sets of predictors showed a negative commonality coefficient, indicating the presence of suppression effects. The most noticeable sup- pression effect was observed between ̇VO2peak and CR80%, which was at – 31.6% in the whole sample. This was more pronounced in the female (– 51.7%) compared to the male subset (– 19.5%). The detailed results of the commonality analyses are depicted in Fig. 3. Discussion The present study investigated the accuracy of the classi- cal model with the physiological variables (i.e., ̇VO2peak , LT2%, and CR), applied in previous studies (McLaughlin et al. 2010; Støa et al. 2020; Støren et al. 2014), in deter- mining vLT2 as an indicator of endurance performance in young squad athletes of different disciplines, ages, and sexes. We found an excellent accordance between the mod- eled and experimentally determined vLT2 (ICC ≥ 0.959) with a very low systematic bias (mean difference ± limits of agreement ≤ 0.07 ± 0.32 m ⋅ s−1 ) independent of sex and age, supporting the applicability of calLT2 to assess aero- bic endurance performance also in young athletes. Fur- thermore, the accuracy of vLT2 modeling was improved, when CR was determined by individualized approaches (especially CR80%) rather than at a fixed speed (i.e., CRfix) (Fig. 2). According to the stepwise regression and com- monality analyses, ̇VO2peak is the most important factor for vLT2 in both sexes, followed by CR, whereas LT2% has the least influence (Fig. 3). In the present investigation, CR determined by differ- ent methods showed very high similarity with each other (Table 1). However, the application of individualized approaches, particularly CR80%, further improved the accuracy of the model for calculating vLT2 as reflected by the lower variation (i.e., limits of agreement) compared to the other methods assessing CR (Table 2 and Fig. 2). Espe- cially in a heterogeneous sample as in our study, measur- ing ̇VO2 at a fixed running speed (i.e., CRfix) results in different metabolic responses (e.g., percent utilization of ̇VO2peak and substrate utilization) and thus impairs the Table 2 Mean difference (± limits of agreement) and intra-class cor- relation coefficients (ICC) between running speed at lactate thresh- old 2 (vLT2) vs. calculated lactate threshold 2 using the oxygen cost of running at a fixed running speed of 2.8 m ⋅ s−1 (calLT2fix), at lac- tate threshold 1 (calLT2LT1), and at 80% of maximal oxygen uptake (calLT280%) CI confidence interval All (N = 100) Males (N = 45) Females (N = 55) Mean difference [ m ⋅ s−1] ICC (95% CI) Mean difference [ m ⋅ s−1] ICC (95% CI) Mean difference [ m ⋅ s−1] ICC (95% CI) calLT2fix – 0.04 ± 0.26 0.968 (0.950–0.979) – 0.07 ± 0.32 0.955 (0.908–0.977) – 0.02 ± 0.18 0.962 (0.934–0.978) calLT2LT1 – 0.02 ± 0.22 0.979 (0.968–0.986) – 0.03 ± 0.25 0.976 (0.956–0.987) – 0.01 ± 0.20 0.959 (0.929–0.976) calLT280% – 0.02 ± 0.12 0.993 (0.988–0.996) – 0.02 ± 0.13 0.992 (0.983–0.996) – 0.02 ± 0.11 0.986 (0.975–0.992) Fig. 2 Bland-Altman Plots: differences between calculated running speed a at lactate threshold 2 determined by oxygen cost of running at a fixed running speed of 2.8 m ⋅ s−1 (calLT2fix), b at lactate thresh- old 1 (calLT2LT1), and c at 80% of maximal oxygen uptake (calLT280%) vs. running speed at lactate threshold 2 (vLT2) determined by modi- fied maximal deviation method. The individual data of male (N = 45) and female (N = 55) athletes are presented by blue cycles and red triangles, respectively; the solid line indicates mean difference; the dashed lines indicate the limits of agreement (mean ± 1.96 standard deviation); the dotted line represents the fitted linear regression 579 European Journal of Applied Physiology (2023) 123:573–583 1 3 inter-individual comparability of CR. This might have contributed to the lower accuracy of calLT2fix compared to the other methods (i.e., larger random variation). To compensate for inter-individual variability in metabolic response, we suggested to assess CR at a specific submaxi- mal metabolic anchor, i.e., LT1. However, incorporating CRLT1 in the model (i.e., calLT2LT1) did not considerably improve its accuracy in calculating vLT2 as compared to CR80%. This result can be explained in part by the fact that the running speed at LT1 (2.84 ± 0.30 m ⋅ s−1 ) was similar to 2.8 m ⋅ s−1 (running speed for CRfix) and significantly lower than that at 80% of ̇VO2peak (3.09 ± 0.36 m ⋅ s−1 ). Although it is now widely established that CR is independ- ent of the respective running speed (Di Prampero et al. 2009; Shaw et al. 2014), there are still conflicting findings indicating an increased or decreased CR with the running speed (Iaia et al. 2009; Jones and Doust 1996; Lacour and Bourdin 2015; Svedenhag and Sjödin 1994). Assessing CR at high relative running speed is likely to represent typical movement patterns (e.g., stride length and frequency) and to consider the individual metabolic demand, especially near to LT2. In this regard, it can be speculated that the running speed at 80% of ̇VO2peak may have better repre- sented the inter-individual variation in running energetics and/or mechanics near LT2 compared to the running speed at LT1, thereby increasing the accuracy of the computa- tional model. This is supported by the previous suggestion of Lacour and Bourdin (2015) that athletes’ performance Table 3 Model summary resulting from stepwise multiple regression analyses using the modeled running speed corresponding to lactate thresh- old 2 as dependent variable and classical physiological parameters as independent variables Beta beta coefficient, SE standard error, Std. Beta standardized beta coefficient, r correlation coefficient to the running speed at lactate threshold 2, R2 coefficient of determination, Adj. R2 adjusted coefficient of determination, ̇VO2peak maximal oxygen uptake, CR80% cost of running deter- mined at 80% of ̇VO2peak , LT2% fractional utilization of ̇VO2peak Model Variable Beta SE Std. Beta t p r R2 Adj. R2 F p Analysis 1: All (N = 100)  1 (Constant) 1.033 0.176 5.877 < 0.001 0.650 0.646 181.73 < 0.001 ̇VO2peak 0.046 0.003 0.806 13.481 < 0.001  2 (Constant) 2.894 0.153 18.890 < 0.001 0.899 0.897 431.55 < 0.001 ̇VO2peak 0.063 0.002 1.113 29.378 < 0.001 Oxygen CR80% – 12.409 0.802 – 0.586 – 15.471 < 0.001  3 (Constant) – 0.388 0.226 – 1.718 0.089 0.971 0.970 1086.10 < 0.001 ̇VO2peak 0.065 0.001 1.142 56.078 < 0.001 0.806 Oxygen CR80% – 14.491 0.449 – 0.685 – 32.244 < 0.001 – 0.003 LT2% 0.042 0.003 0.283 15.585 < 0.001 0.175 Analysis 2: Males (N = 45)  1 (Constant) 0.625 0.389 1.606 0.116 0.581 0.571 59.63 < 0.001 ̇VO2peak 0.053 0.007 0.762 7.722 < 0.001  2 (Constant) 3.076 0.325 9.473 < 0.001 0.876 0.871 149.03 < 0.001 ̇VO2peak 0.062 0.004 0.888 15.949 < 0.001 Oxygen CR80% – 12.827 1.280 – 0.558 – 10.024 < 0.001  3 (Constant) – 0.407 0.369 – 1.105 0.276 0.967 0.965 402.69 < 0.001 ̇VO2peak 0.062 0.002 0.895 30.818 < 0.001 0.762 Oxygen CR80% – 14.858 0.694 – 0.646 – 21.396 < 0.001 – 0.359 LT2% 0.045 0.004 0.313 10.643 < 0.001 0.170 Analysis 3: Females (N = 55)  1 (Constant) 1.588 0.277 5.737 < 0.001 0.390 0.379 33.92 < 0.001 ̇VO2peak 0.034 0.006 0.625 5.824 < 0.001  2 (Constant) 2.806 0.191 14.665 < 0.001 0.813 0.806 112.92 < 0.001 ̇VO2peak 0.064 0.004 1.173 14.946 < 0.001 Oxygen CR80% – 12.143 1.121 – 0.851 – 10.836 < 0.001  3 (Constant) – 0.177 0.274 – 0.649 0.519 0.949 0.946 318.67 < 0.001 ̇VO2peak 0.065 0.002 1.189 28.831 < 0.001 0.625 Oxygen CR80% – 14.330 0.617 – 1.004 – 23.208 < 0.001 – 0.094 LT2% 0.039 0.003 0.396 11.725 < 0.001 0.283 580 European Journal of Applied Physiology (2023) 123:573–583 1 3 may be more accurately predicted from CR determined at high relative running speed. Regarding the criterion, Støren et al. (2014) and Støa et al. (2020) used a threshold concept with a warm-up level of blood lactate concentration plus a fixed absolute value (2. 3 mmol ⋅ L−1 ) to determine LT2 as an estimate of maximal metabolic steady state. Although such an approach has been repeatedly used in previous studies as an established indica- tor of endurance performance [e.g., Helgerud et al. (1990, 2009)], there is no explicit study assessing its systematic bias and absolute agreement compared to the underlying physi- ological concept of a maximal metabolic steady state apart from unpublished work [i.e., Helgerud et al. (1990)], which is crucial for ensuring the validity of a threshold concept (Faude et al. 2009). In addition, the increase in blood lactate concentrations of a certain fixed value might not always be equally meaningful, since it is highly affected by various factors (e.g., test protocols, training- and nutrition-status) (Svedahl and MacIntosh 2003). In contrast, we applied a mathematical model for determining inflection points as a determination criteria for LT2, which is based on blood lactate kinetics rather than absolute concentrations (Zwing- mann et al. 2019). Even though the use of mathematical models for LT2 determination has been criticized by some authors because of the lacking physiological fundamental (Janeba et al. 2010), its validity for estimating the maximal lactate steady state has been verified by systematic analyses (Jamnick et al. 2018; Zwingmann et al. 2019). In the present study, the stepwise regression analysis showed that 97% of the total variance (males: 97%, females: 95%) in vLT2 in young squad athletes of different disciplines was explained by ̇VO2peak , LT2%, and CR, supporting that these are the three primary physiological factors influencing aerobic endurance performance (McLaughlin et al. 2010). The single most important determinant of vLT2 independent of sex was ̇VO2peak , which is in accordance with previous research (McLaughlin et al. 2010; Støa et al. 2020; Støren et al. 2014), emphasizing the importance of aerobic energy supply during prolonged weight-bearing exercise such as running. Therefore, especially in heterogeneous samples as in the present study, endurance performance is strongly related to ̇VO2peak . Likewise, the lower values observed in the young female compared to the male athletes are in accordance with previous studies and are likely related to body composition (i.e., greater percentage of body fat) and oxygen carrying capacity (i.e., lower hematocrit levels) (Besson et al. 2022). Based on the commonality analysis, CR, being the second major factor influencing vLT2, seems to act as a suppressor which purifies the “irrelevant” variance of other independent variables (i.e., negative common effects), and thus improves their contribution to the regression model. In particular, the pronounced suppression effect of CR in combination with ̇VO2peak (– 19.5% to – 51.7% of total R2 ) can emphasize the crucial role of the interaction between maximal aero- bic capacity and movement economy for endurance per- formance (Joyner and Coyle 2008). Indeed, in the present investigation, CR separately exhibited only negligible to low correlation with vLT2 (r = – 0.359 to – 0.003), but its incorporation into the regression model in addition to ̇VO2peak resulted in a significantly improved R2 with a sig- nificant (standardized) beta weight in both the whole group and the male and female subgroups (see Table 3). Further- more, the product of ̇VO2peak and CR (expressing maximal aerobic speed) has been demonstrated to be a very good predictor for 16-km time trial performance ( R2 = 0.94) and vLT2 ( R2 = 0.85) in competitive runners (McLaughlin et al. 2010; Støa et al. 2020). Interestingly, the unique (56.3% vs. 37.9%) and common effects (– 51.7% vs. – 19.5%) of CR on the total R2 were distinctly higher in female than in male athletes implying the greater impact of CR in determining endurance performance in females. Further, female athletes showed lower CR values independent of the determination method. This is an interesting finding which contributes to the debate whether or not there are sex differences regarding CR(Besson et al. 2022). While some authors argue that sex differences in CR disappear when expressed as relative inten- sities (i.e., as percentage of ̇VO2peak or lactate threshold) (Fletcher et al. 2013; Helgerud et al. 1990), our data indicate that, at least in young athletes, females exhibit lower values whether expressed as absolute (CRfix) or relative (CR80% Fig. 3 Graphical summary of commonality analyses for the modeled running speed at lactate threshold 2 (LT2) within all (N = 100), male (N = 45) and female (N = 55) athletes. The percentage contribution of each unique predictor to the total regression effect (i.e., R2 ) is pre- sented by the black filled arrows; the dashed lines and external solid lines represent the common effects of two and all predictors in R2 , respectively. ̇VO2peak : maximal oxygen uptake; CR80%: cost of run- ning determined at 80% of ̇VO2peak ; LT2%: fractional utilization of ̇VO2peak at LT2. 581 European Journal of Applied Physiology (2023) 123:573–583 1 3 and CRLT1) values.  Besides differences in anthropometric dimensions (e.g., body height, see Table 1), the lower CR in female athletes might be due to neuromuscular character- istics of the lower extremities such as lower-body stiffness or Achilles moment arm length as shown for well-trained distance runners (Barnes et al. 2014). Furthermore, since we observed significantly higher minute ventilation at the speed corresponding to 80% of ̇VO2peak in male compared to female athletes, it can be assumed that the differences in CR may be attributed, at least in part, to increased demands of breathing. Thus, it has been shown earlier that running economy was improved by training induced decrease in min- ute ventilation (Franch et al. 1998). Nonetheless, it should be noted that our study did not take into account hormonal changes related to the phase of the menstrual cycle in female athletes, which is known to affect CR (Besson et al. 2022; Dokumacı and Hazır 2019). Besides sex-specific differences in CR, it is unclear why CR had a greater impact on vLT2 determination in young female athletes than in males. However, beside the afore- mentioned factors (e.g., anthropometric, neuromuscular, and cardio-respiratory), there are sex-specific differences related to substrate oxidation during exercise and muscle tissue characteristics (e.g., proportion of type I muscle fib- ers and muscle capillarization), that could affect male and female athletes differently in terms of submaximal energy metabolism (Besson et al. 2022). These differences may partly explain why CR appears to have a stronger influence on endurance performance in young female compared to male athletes, similar to the previous findings of Støa et al. (2020) in adult runners. In this context, future longitudinal studies might investigate whether female athletes can actu- ally profit more than male athletes from an improvement in CR. In contrast to ̇VO2peak and CR, we found no sex differ- ence in LT2% (87.0 ± 2.9% vs. 87.0 ± 2.4%) with a very low inter-individual variation (coefficient of variation = 3%). These results are in line with previous investigations indicating no difference in LT2% between well-trained male and female runners (McLaughlin et al. 2010) as well as between elite, national, and recreational runners (Støa et al. 2020). Further, the LT2%-values in the previous stud- ies with adult recreational and elite runners (72–93%) are in a similar range to those in the current study (80–94%). Taken together, it seems plausible to assume that LT2% does not vary substantially depending on the aerobic endur- ance level. Moreover, the minor contribution of LT2% to the regression model for determining vLT2 in the present study (Fig. 3) provides further support for the assumption that LT2% is not a major factor affecting aerobic endurance per- formance (McLaughlin et al. 2010; Støa et al. 2020; Støren et al. 2014). Nonetheless, further studies need to investigate whether long-term adaptations of LT2% lead to altered aero- bic endurance performance (i.e., vLT2) in young athletes of different disciplines. Due to the high accuracy in estimating vLT2, the pro- posed model allows to draw conclusions about the limiting factors (mainly ̇VO2peak and CR) of endurance performance of young athletes of various disciplines and both sexes, and may therefore guide future training design. Since total time to improve endurance capacity in technically or tactically demanding sports is limited especially in young athletes with a restricted schedule, this time needs to be utilized as effi- ciently as possible. Thus, depending on the individual physi- ological prerequisites, training prescription may focus either on improving ̇VO2peak for example using high intensity interval training or on improving CR for example by imple- menting explosive- and maximal-strength training (Røn- nestad and Mujika 2013; Støa et al. 2020). Future studies should use the predictors to model endurance performance longitudinally, e.g., over one or multiple seasonal cycles to examine whether training induced changes in the physiologi- cal predictors actually lead to the intended changes in endur- ance performance. Conclusion In conclusion, the classical model using ̇VO2peak , CR, and LT2% to determine vLT2 is also suitable for assessing endur- ance performance in young squad athletes of different dis- ciplines, ages, and sexes. The accuracy of the model was further improved by an individual determination of CR (in particular CR80%). ̇VO2peak and CR were found to have the most important contributions in determining vLT2. While in young female athletes, the impact of CR on endurance per- formance appeared to be greater than that in male athletes, LT2% was generally found to have the least impact on vLT2 determination. Acknowledgements The authors would like to thank Anja Habbig, Paulina Heumann, Till Kämpfer, Simon Kohne, Christian Manunzio, Inga Schifferdecker, Aldo Sommer, and Sarah Strütt for their enthusi- astic contribution to data collection from 2018 to 2022. Funding Open Access funding enabled and organized by Projekt DEAL. Declarations Conflicts of interest The authors declare no conflict of interest. No funding was received to assist with the preparation of this manuscript. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, 582 European Journal of Applied Physiology (2023) 123:573–583 1 3 provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. References Barnes KR, Kilding AE (2015) Running economy: measurement, norms, and determining factors. Sports Med Open. https:// doi. org/ 10. 1186/ s40798- 015- 0007-y Barnes KR, Mcguigan MR, Kilding AE (2014) Lower-body determi- nants of running economy in male and female distance runners. J Strength Cond Res 28(5):1289–1297. https:// doi. org/ 10. 1519/ jsc. 00000 00000 000267 Besson T, Macchi R, Rossi J, Morio CYM, Kunimasa Y, Nicol C, Millet GY (2022) Sex differences in endurance run- ning. Sports Med 52(6):1235–1257. https:// doi. org/ 10. 1007/ s40279- 022- 01651-w DiMenna FJ, Jones AM (2009) “linear’’ versus “nonlinear’’ vo2 responses to exercise: reshaping traditional beliefs. J Exerc Sci Fit 7(2):67–84. https:// doi. org/ 10. 1016/ S1728- 869X(09) 60009-5 Di Prampero PE, Salvadego D, Fusi S, Grassi B (2009) A simple method for assessing the energy cost of running during incre- mental tests. J Appl Physiol 107(4):1068–1075. https:// doi. org/ 10. 1152/ jappl physi ol. 00063. 2009 Dokumacı B, Hazır T (2019) Effects of the menstrual cycle on running economy: oxygen cost versus caloric cost. Res Q Exerc Sport 90(3):318–326. https:// doi. org/ 10. 1080/ 02701 367. 2019. 15998 00 Farrell PA, Wilmore JH, Coyle EF, Billing JE, Costill DL (1979) Plasma lactate accumulation and distance running performance. Med Sci Sports 11(4):338–344 Faude O, Kindermann W, Meyer T (2009) Lactate threshold con- cepts: how valid are they? Sports Med 39(6):469–490. https:// doi. org/ 10. 2165/ 00007 256- 20093 9060- 00003 Fletcher JR, Pfister TR, MacIntosh BR (2013) Energy cost of running and achilles tendon stiffiness in man and woman trained runners. Physiol Rep 1(7):e00178. https:// doi. org/ 10. 1002/ phy2. 178 Franch J, Madsen K, Djurhuus MS, Pedersen PK (1998) Improved running economy following intensified training correlates with reduced ventilatory demands. Med Sci Sports Exerc 30(8):1250– 1256. https:// doi. org/ 10. 1097/ 00005 768- 19980 8000- 00011 Helgerud J, Ingjer F, Strømme SB (1990) Sex differences in perfor- mance-matched marathon runners. Eur J Appl Physiol Occup Physiol 61(5–6):433–439. https:// doi. org/ 10. 1007/ bf002 36064 Helgerud J, Støren Ø, Hoff J (2009) Are there differences in running economy at different velocities for well-trained distance run- ners? Eur J Appl Physiol 108(6):1099–1105. https:// doi. org/ 10. 1007/ s00421- 009- 1218-z Iaia FM, Hellsten Y, Nielsen JJ, Fernström M, Sahlin K, Bangsbo J (2009) Four weeks of speed endurance training reduces energy expenditure during exercise and maintains muscle oxidative capacity despite a reduction in training volume. J Appl Physiol 106(1):73–80. https:// doi. org/ 10. 1152/ jappl physi ol. 90676. 2008 Jamnick NA, Botella J, Pyne DB, Bishop DJ (2018) Manipulating graded exercise test variables affects the validity of the lactate threshold and v.o2peak. PLoS One 13(7):e0199794. https:// doi. org/ 10. 1371/ journ al. pone. 01997 94 Janeba M, Yaeger D, White R, Stavrianeas S (2010) The dmax method does not produce a valid estimate of the lactate thresh- old. J Exerc Physiol Online 13(4):50–57 Jones AM, Doust JH (1996) A 1% treadmill grade most accurately reflects the energetic cost of outdoor running. J Sports Sci 14(4):321–327. https:// doi. org/ 10. 1080/ 02640 41960 87277 17 Joyner MJ (1991) Modeling: optimal marathon performance on the basis of physiological factors. J Appl Physiol 70(2):683–687. https:// doi. org/ 10. 1152/ jappl. 1991. 70.2. 683 Joyner MJ, Coyle EF (2008) Endurance exercise performance: the physiology of champions. J Physiol 586(1):35–44. https:// doi. org/ 10. 1113/ jphys iol. 2007. 143834 Koo TK, Li MY (2016) A guideline of selecting and reporting intra- class correlation coefficients for reliability research. J Chiropr Med 15(2):155–163. https:// doi. org/ 10. 1016/j. jcm. 2016. 02. 012 Lacour J-R, Bourdin M (2015) Factors affecting the energy cost of level running at submaximal speed. Eur J Appl Physiol 115(4):651–673. https:// doi. org/ 10. 1007/ s00421- 015- 3115-y Lundby C, Montero D, Gehrig S, Hall UA, Kaiser P, Boushel R, Madsen K (2017) Physiological, biochemical, anthropometric, and biomechanical influences on exercise economy in humans. Scand J Med Sci Sports 27(12):1627–1637. https:// doi. org/ 10. 1111/ sms. 12849 McLaughlin JE, Howley ET, Bassett DR, Thompson DL, Fitzhugh EC (2010) Test of the classic model for predicting endurance running performance. Med Sci Sports Exerc 42(5):991–997. https:// doi. org/ 10. 1249/ mss. 0b013 e3181 c0669d Midgley AW, McNaughton LR, Polman R, Marchant D (2007) Cri- teria for determination of maximal oxygen uptake. Sports Med 37(12):1019–1028. https:// doi. org/ 10. 2165/ 00007 256- 20073 7120- 00002 Minetti AE (2004) Passive tools for enhancing muscle-driven motion and locomotion. J Exp Biol 207(8):1265–1272. https:// doi. org/ 10. 1242/ jeb. 00886 Mukaka M (2012) Statistics corner: a guide to appropriate use of correlation in medical research. Malawi Med J 24(3):69–71 Nathans LL, Oswald FL, Nimon K (2012) Interpreting multiple linear regression: a guidebook of variable importance. Pract Assess Res Eval 17(1):9. https:// doi. org/ 10. 7275/ 5fex- b874 Pandey S, Elliott W (2010) Suppressor variables in social work research: ways to identify in multiple regression models. J Soc Soc Work Res 1(1):28–40. https:// doi. org/ 10. 5243/ jsswr. 2010.2 Péronnet F, Massicotte D (1991) Table of nonprotein respiratory quotient: an update. Can J Sport Sci 16(1):23–29 R Core Team (2021) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria Ray-Mukherjee J, Nimon K, Mukherjee S, Morris DW, Slotow R, Hamer M (2014) Using commonality analysis in multiple regressions: a tool to decompose regression effects in the face of multicollinearity. Methods Ecol Evol 5(4):320–328. https:// doi. org/ 10. 1111/ 2041- 210X. 12166 Rønnestad BR, Mujika I (2013) Optimizing strength training for run- ning and cycling endurance performance: a review. Scand J Med Sci Sports 24(4):603–612. https:// doi. org/ 10. 1111/ sms. 12104 Scharhag-Rosenberger F, Meyer T, Gäßler N, Faude O, Kindermann W (2010) Exercise at given percentages of VO2max: hetero- geneous metabolic responses between individuals. J Sci Med Sport 13(1):74–79. https:// doi. org/ 10. 1016/j. jsams. 2008. 12. 626 Shaw AJ, Ingham SA, Folland JP (2014) The valid measure- ment of running economy in runners. Med Sci Sports Exerc 46(10):1968–1973. https:// doi. org/ 10. 1249/ mss. 00000 00000 000311 Støa EM, Helgerud J, Rønnestad BR, Hansen J, Ellefsen S, Støren Ø (2020) Factors influencing running velocity at lactate threshold in 583 European Journal of Applied Physiology (2023) 123:573–583 1 3 male and female runners at different levels of performance. Front Physiol. https:// doi. org/ 10. 3389/ fphys. 2020. 585267 Støren Ø, Rønnestad BR, Sunde A, Hansen J, Ellefsen S, Helgerud J (2014) A time-saving method to assess power output at lactate threshold in well-trained and elite cyclists. J Strength Cond Res 28(3):622–629. https:// doi. org/ 10. 1519/ JSC. 0b013 e3182 a73e70 Svedahl K, MacIntosh BR (2003) Anaerobic threshold: the concept and methods of measurement. Can J Appl Physiol 28(2):299–323. https:// doi. org/ 10. 1139/ h03- 023 Svedenhag J, Sjödin B (1994) Body-mass-modified running economy and step length in elite male middle-and long-distance runners. Int J Sports Med 15(06):305–310. https:// doi. org/ 10. 1055/s- 2007- 10210 65 Whipp BJ, Wasserman K (1972) Oxygen uptake kinetics for various intensities of constant-load work. J Appl Physiol 33(3):351–356. https:// doi. org/ 10. 1152/ jappl. 1972. 33.3. 351 Zwingmann L, Strütt S, Martin A, Volmary P, Bloch W, Wahl P (2019) Modifications of the dmax method in comparison to the maxi- mal lactate steady state in young male athletes. Phys Sportsmed 47(2):174–181. https:// doi. org/ 10. 1080/ 00913 847. 2018. 15461 03 Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Modeling lactate threshold in young squad athletes: influence of sex, maximal oxygen uptake, and cost of running.

11-21-2022

Ji, Sanghyeon,Keller, Sebastian,Zwingmann, Lukas,Wahl, Patrick

eng

PMC6615608

RESEARCH ARTICLE A three-criteria performance score for rats exercising on a running treadmill Juan Gabriel Rı´os-Kristja´nsson1☯, David Rizo-Roca1☯, Karen Mist Kristja´nsdo´ttir1,2, Cristian Andre´s Nu´ñez-Espinosa1,3, Joan Ramon Torrella1, Teresa Pagès1, Gine´s ViscorID1* 1 Department of Cell Biology, Physiology & Immunology, Faculty of Biology, University of Barcelona, Barcelona, Spain, 2 Department of Biotechnology and Chemical Engineering, Aarhus University School of Engineering, Aarhus N, Denmark, 3 School of Medicine, University of Magallanes, Casilla, Punta Arenas, Chile ☯ These authors contributed equally to this work. * [email protected] Abstract In this study, we propose a novel three-criteria performance score to semiquantitatively clas- sify the running style, the degree of involvement and compliance and the validity of electric shock count for rats exercising on a treadmill. Each score criterion has several style-marks that are based on the observational registry of male Sprague-Dawley rats running for 4–7 weeks. Each mark was given a score value that was averaged throughout a session-registry and resulting in a session score for each criterion, ranging from “0” score for a hypothetical “worst runner”, to score “1” for a hypothetical “perfect runner” rat. We found significant differ- ences throughout a training program, thus providing evidence of sufficient sensitivity of this score to reflect the individual evolution of performance improvement in exercise capacity due to training. We hypothesize that this score could be correlated with other physiological or metabolic parameters, thus refining research results and further helping researchers to reduce the number of experimental subjects. Introduction A significant amount of scientific literature on exercise sciences is based on experimental research involving laboratory rats, being motor-driven treadmill running, voluntary wheel running and swimming the most common types of exercise [1–6]. Although a marked differ- ence has been reported on the exercise capacity of rats in relation to different strains [7–9], in most published work using treadmill training the mention of problems with the rats whilst car- rying out the training sessions is absent [10–13]. This could lead to the assumption that all or much of the rats have been trained in the same extent by the end of a training session within each study. However, the natural proneness of rat for short voluntary running bursts [14] sug- gests that, if given the choice, it would run with recurrent breaks; unless, perhaps, if running away from a perceived danger. Thus, the expectancy of good habituation to continuous run- ning on a treadmill, throughout a relatively long exercise session, might not be realistic from a PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 1 / 14 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Rı´os-Kristja´nsson JG, Rizo-Roca D, Kristja´nsdo´ttir KM, Nu´ñez-Espinosa CA, Torrella JR, Pagès T, et al. (2019) A three-criteria performance score for rats exercising on a running treadmill. PLoS ONE 14(7): e0219167. https://doi. org/10.1371/journal.pone.0219167 Editor: Clemens Fu¨rnsinn, Medical University of Vienna, AUSTRIA Received: February 19, 2019 Accepted: June 18, 2019 Published: July 9, 2019 Copyright: © 2019 Rı´os-Kristja´nsson et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the manuscript and its Supporting Information files. Funding: This study was supported by DEP2010- 22205-C02-01 (GV) and DEP2013-48334-C2-1-P (GV & JRT) grants from the Plan Nacional I+D+i (The Spanish Ministry of Economy and Competitiveness). JGRK was supported by a predoctoral grant (BES-2011-044293) from The Spanish Ministry of Economy and Competitiveness. The funders had no role in study biological point of view, especially at the beginning of a training period. Obviously, the rats can still be gradually trained to run, but the fact the rats finished a training session does not guarantee that all of them did the same training volume or equivalent intensity. Few published experiments reported on the different performance between individual rats when running. For instance, Arnold et al. (2014) described a protocol to select aged rats for an exercise protocol, arguing that not all animals are equally prone to run in a treadmill [15], while in a paper from Ferraresso et al. (2012) the authors specifically stated that they differentiated animals that ran voluntarily and animals that refused to ran in order to distribute the rodents among the differ- ent experimental groups [16]. However, this kind of information is usually omitted in experi- mental papers. Many treadmills for rats promote running by means of a light electric shock stimulus via the touch of a metal grid. A monitoring apparatus processes the input of the electric shock count with the related accumulating duration; useful to estimate how successfully a rat has been running. However, this setup has two potential problems: 1) the equipment cannot differ- entiate between a rat that is continuously touching the grid and e.g. a piece of faeces stuck on it; 2) when rats run on the treadmill, a marked majority tend to avoid running in many differ- ent ways. Fig 1 (marks B-N) describes evasive behaviours, which compromise compliance, to avoid running or electrical shock punishment [17]. Thus, the experimenter needs to be aware of these possible artefacts when using a treadmill. Accepting the likelihood of all rats not running equally, or being unequally prone to run on a treadmill, leads to the question of how different qualitative running traits can be represented and how to quantify that. Assumption of equal performance and compliance for the same workload of all the animals in a training protocol is unrealistic. Indeed, if the training itself is important, also the training type is important; e.g. continuous vs. interval training or forced vs. spontaneous exercise. Subsequently, the fundamental questions should address the manner how the rats diverge in training and how this is reflected in the data, and address a systematic approach to determine those deviations. Here we propose a system based on marks aimed to semiquantitatively assess the rat run- ning style on a treadmill, and quantify this assessment with a score value, eventually leading to an average score. Thus, the main goal of this work is to develop and apply a semiquantitative scoring system to assess rat running compliance to an exercise protocol in a motorized tread- mill. This score allows to “classify” the rats, which can serve as a critical factor for further phys- iological data interpretation. The score could also be used to compare the treadmill running performance of rats between different experimental conditions. Moreover, from an ethical point of view, and according to principles of the Three Rs (reduction, refinement and replace- ment), it is necessary to apply data tools that allow researchers to obtain the best from the experimental animals with which they are working. A vast majority of researchers simply dis- card animals that do not show ability to run on the treadmill, but the selected group might not be representative of the “normal” animal population. On the other hand, it certainly becomes an ethical issue if the selection of the experimental animals relies on discarding them, rather than refining the number of rats in the experiment from the onset. Material and methods Animals All procedures were performed in accordance with the internal protocols of our laboratory, authorized by the University of Barcelona’s Ethical Committee for Animal Experimentation and ratified, in accordance with current Spanish legislation, by the Departament de d’Agricul- tura, Ramaderia, Pesca i Alimentacio´ of Generalitat de Catalunya (file #8784). The score Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 2 / 14 design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. developed and presented in this article derived as a secondary observational registry from 130 male Sprague Dawley rats (strain: RjHan:SD, Janvier Labs, France). The animals were used in a primary experiment on the study of the recovery of induced skeletal muscle damage in the hind limbs of trained rats [18–21] in the framework of a research project approved by the Fig 1. Schematic setup of the AEY performance score. The top section explains the setup of a treadmill diagram, followed by three sections for each score criteria: A, E and Y. The score criteria-sections have subsections (from left to right): (1) Mark, a registry note for the described style; (2) Diagram, a schematic drawing (if applicable); (3) Description, the characteristics are described; and (4) Score, the determined score value for further calculations. https://doi.org/10.1371/journal.pone.0219167.g001 Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 3 / 14 Institutional Experimental Animal Ethics Committee. The experimental protocol required sev- eral weeks of previous conditional exercise training in a running treadmill and 1 day of exhaustive downhill exercise followed by different recovery interventions. All the rats used were males and started the training program at 6 weeks of age (body weight: 154 ± 8 g) in order to fulfil the requirements of the primary study, were housed at maximum 3 animals per cage (215×465×145 mm) and were fed standard diet (15-mm diameter granulates) and water ad libitum. The facility’s room temperature and relative humidity ranged between 20–25˚C and 45–55% respectively. All the rats were regularly checked for stress signs judging from their physical appearance and body weight and received treatment fulfilling the National and Euro- pean directives for the care of animal uses for scientific purposes [22]. Instrumental An encased five lane treadmill and its accompanying treadmill controller (LE 8710, Harvard Apparatus, United States) with an adjustable plane (5˚/ramp-setting) from −15˚ to 25˚ was used to carry out the exercise training. An adjustable electric shock stimulator, ranging from 0.2–2.0 mA, discharged when animal contacted the metal grids behind the back end of the treadmill belt. The treadmill encasing, separating the lanes, with front and back wall air holes for each lane, enabled uncontrolled airflow inside it. For each lane, the monitor on the tread- mill controller displayed the number of electric shocks generated, the accumulated time of electric shocks and the calculated distance considering the set velocity and the time lapsed, deducting the time spent receiving a shock. Speed adjustments affected collectively all the five lanes of the treadmill belt. Furthermore, the experimenters encouraged the rats to run with a light push or a sound to minimise the experience of excessive electric shocks, especially during the first days of habituation to exercise on the running treadmill. The three-criteria performance score The score’s style code is based in three domains, each giving their own score value. These 3 cri- teria/scores are: 1) the running “attitude” (A), 2) a clue of the “endurance” (E), and 3) the gross “yield” (Y) provided by the electric shock count. Henceforward, the three-criteria perfor- mance score can be summarized as the AEY-score. More specifically, the criteria for the A- score concerns the physical positioning and actions of the rat whilst running (or trying to avoid running). The E-score concerns the position of the rat’s tail as an indicator of tiredness based on the effort to hold it up to avoid receiving an electric shock from the grid. The Y-score addresses the potential artefacts and problems considering the digital representation of the electric shock count (the sole insight usually considered of the rat performance in the running treadmill). The development of the score was carried out in 2 stages: (a) the initial stage (55 rats) where different situations and characteristics were documented and the criteria for the scores were defined; and (b) the quantification stage (75 rats) where the defined criteria were systematically registered for each rat throughout every training session. Fig 1 presents illustrative style diagrams and text explanations relative to the three domains of the score [A;E;Y] along with its corresponding marks and score values used for registering and quantifying during a training session. The experimental objective was to train the rats towards their best potential, recording the individual trends along the way. If it was needed and physically possible, the experimenters interfered with a rat’s unwanted styles by outlasting its determination. Relatively low treadmill speed (in the first days of habituation as part of the protocol) partially led to unfavourable running styles in some rats; hindering the development a continuous “proper” running (style A). Rats that demonstrated style B and E had no major Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 4 / 14 issues with exercising and needed to be lightly pushed or pulled back onto four legs. Styles C, H and I represented rats that normally had high shock counts and needed an extra physical push to achieve a favourable positioning on the belt. Style D came across as a display of tired- ness (or a sense of outwitting) by periodically stopping until almost touching (or fully touch- ing) the shock grid and then quickly moving to the front of the treadmill. In styles F and G, the rat cleverly discovered that, by lightly touching its lane-sidewalls, enough resistance was pro- vided to slide on the treadmill belt without moving towards the grid. Among styles E, F and G, the rat was equally inactive regarding the exercise; hence, all styles were later assigned the same quantification score, leaving the AEY-score open to take on board similar circumstances with- out major changes. Style J represented distracted or stressed rats (not necessarily “bad run- ners”) seeming susceptible to disrupting stimuli outside the treadmill. Styles K, M and N, normally occurred when rats were exercising to exhaustion; with style M prevailed on a down- ward slope. Style L is specific to treadmills where the belt does not cover the edge (e.g. within the two peripheral lanes of a multi-lane treadmill; a design flaw to be solved). In our case, a style L-rat supported the fore and hind limbs of the same body side onto the non-moving part, whilst the other limbs moved. The occurrence of style L was further minimised by rotating the lane positions between sessions during the first days of the training protocol. Quantifying the assessment Each domain of the score had a maximal value of “1” that was considered for the “ideal run- ner” rat. The running styles or situations that were considered unfavourable would reduce that value towards the score floor of “0” for each part of the score. Thus, the more a trait compro- mised the animal’s compliance to the exercise protocol (to run continuously at 0.45 m/s for 30 min, see Exercise Protocols below), the more negative score was assigned to that trait. For the A-score, the rat was given the roof value of “1” by default. Only one style was con- sidered to represent the “ideal runner” (style A) and reduced this “1” by the value of “0” (i.e. no effect). Meanwhile, the other styles, having unfavourable traits of various degree, were given negative values (Fig 1, score column). Many of the unfavourable traits were visually dif- ferent, although a similar negative value was assigned. For the E-score (Fig 1), the value was applied directly with only four possible style assigna- tions: between the most favourable, with the value of “1”, and the most unfavourable, with the value of “0.25”. The value of “0” was only applied if no registry was obtained during the registry period. The Y-score was assigned the default value of “1” if no observation was registered. The 4 major types of observations (Fig 1), generally followed by a detailed description in the registry notes, were considered all equally influential and assigned the same value of “–0.15”. Each type was only counted once per registry period giving the lowest possible score of “0.4”. The Y- score therefore, served as an indicator for the underlying incidences. However, during registra- tion, if some incidence was considered having a major effect on the monitor-readings, or con- tinuously reoccurred, the score could be assigned with a “0” with the representative mark (Fig 1), to differentiate numerically from the lowest possible calculated score of “0.4”. Averaging out the scores The running styles of each rat were registered with the corresponding marks for each of the particular score values (Fig 1) during all registry periods of a training session. As each registry period (5 min) could contain various styles registered, the notion of a dominant style (the underlined mark in the registry) was considered as having occurred at least twice (in A and E- Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 5 / 14 domains). A representative registry form used in the semiquantitative assessment, with three different 5-min registry periods, is displayed in Fig 2 (Step 1). For each domain of the score criteria (A, E and Y), within each individual registry period, the score values were calculated (Fig 2, Step 2) as: Aresulting in a training-session score for each rat (Fig 2, Step 3). The average training-session scores were further calculated to represent periods of several sessions and the standard devia- tions of these training-session scores were used to evaluate the variability between different rats of the same experimental group. Representation of the AEY score The score should be represented as [A;E;Y], where a theoretical best rat would have [1;1:1] and the theoretical worst rat [0;0;0]. The score should be accompanied by detailed information about the training setup, including: (i) defined experimental period or duration; (ii) training day-frequency per week; (iii) training sessions per day, type of session and resting period between sessions (if applicable); and (iv) session time length (t), set speed (v), treadmill slope angle (θ). The average score can be calculated to represent different temporary stages, i.e.: a training day, a week of training, a general training program score, or a more extended-overall score such as a rat life-span score. Exercise protocols All exercise sessions started with a 5-min warmup (not included in the session time length), to gradually reach a target speed. The rats trained for 29 days over the period of 7 weeks, training 5 days/week (except when noted below), never training on the weekends. They trained either 1 session/day (always around 09:00 h) or 2 session/day (latter session around 17:00 h), as contin- uous-training sessions. Week 1 consisted of three consecutive training days before the weekend, 1 session/day (except 2 session/day the last day), with the gradual changes of: v = 0.30–0.34 m/s, t = 10–25 min/session (θ = 0˚). Week 2 consisted of 2 session/day with the gradual changes of: v = 0.35– 0.45 m/s, t = 30–32 min/session (θ = 0˚). Week 3 and 4 consisted of 2 session/day with v = 0.45 m/s, t = 30 min/session (θ = 0˚). Day D (downhill protocol) consisted of 2 sessions with v = 0.55 m/s, t = until exhaustion (90 min) and 45 min/session (former and latter, respectively) and θ = −15˚. This day was the 2nd day of week 5, were the 1st day was a resting day. Weeks 5, 6 and 7 consisted of 1 session/day (around 13:00 h) at v = 0.30 m/s, t = 15 min and θ = +5˚, but only for one of the experimental groups during the muscle injury recovery period. Week 7 consisted of only two consecutive training days straight after the weekend. Statistical analysis Data in Figs 3 and 4 are represented as box plots. The box represents the interquartile range and shows the first and the third quartiles, which are separated by the median. Black triangles repre- sent the mean. Whisker end points represent the standard deviation, and the black circles repre- sent outlier values. Scores between Week 1 –Week 4 and between Week 5 –Week 7 were statistically compared using a Kruskal-Wallis test followed by Dunn’s multiple comparison post hoc test. The Mann–Whitney U test was used to compare morning vs. afternoon scores within each week and to compare Week 4 vs. D. Statistical significance was considered when P<0.05. Results Fig 3 presents the average score and standard deviation obtained throughout the entire train- ing period, where each week, albeit with different count of training sessions, is separated with at least a weekend of rest. The A-score significantly increases after the first week showing a progressive lower dispersion trend. From day D until week 7, a similar pattern is observed, Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 7 / 14 although starting from lower score values. Throughout weeks 1 to 4, the E-score follows an opposite trend to the A-score, reflected in significant decreasing mean values. Day D shows a marked and significant drop in mean values and the greatest deviation. Thereafter, the average level increases with a simultaneous deviation decrease. The Y-score appears as quite stable throughout the entire experimental period with a significant drop in the average at day D accompanied by increased deviation and, furthermore, a slight non-significant decrease on week 7 along with a greater deviation. Fig 4 focuses on the first four weeks (training period), each separated by a weekend-rest, where the two daily training sessions (at morning and afternoon) are displayed separately. Whilst the overall weekly tendencies are similar in all three parts of the score, when comparing Fig 3. AEY Score along an exercise program. A box plot representation of the weekly average values for the 3 parts of the rat AEY performance score throughout a 7-week treadmill-training programme. On week 1 and 2 the rats (N = 75) carried out a gradual preconditioning training in preparation for a steady training on week 3 and 4. On week 5, 6 and 7 the rats (N = 27, 12 and 6, respectively) carried out light rehabilitation exercise after one day of downhill exhaustion exercise (N = 65) marked specifically as day D when commencing week 5. Statistically significant differences are indicated as follows:  vs. Week 1; # vs. Week 2; † vs. Week 4; ‡ vs. Week 5. One, two, and three repeated symbols correspond to P<0.05, P<0.01, and P< 0.001, respectively. https://doi.org/10.1371/journal.pone.0219167.g003 Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 8 / 14 morning and afternoon mean values, the magnitude of the tendencies varies in A- and E- scores. The A-score of morning session is significantly lower than the score of afternoon ses- sion in the first two weeks. The opposite occurs with E-score, where the afternoon session-val- ues are significantly lower over the 4 weeks. The Y-score shows a marked difference between the morning and the afternoon sessions. Discussion General considerations The main goal of this paper was to develop a semiquantitative tool for an extensive and sensi- tive assessment of the exercise performance on a running treadmill- of laboratory rats. It is easy to design a scoring system only based on “good runners” rats, but running ability is highly Fig 4. AEY Score comparison at two different daytimes. A box plot representation of the morning and afternoon session-average values for the 3 parts of the rat AEY performance score for 75 rats throughout the first four weeks of a 7-week treadmill-exercise programme. On week 1 and 2 the rats carried out a gradual preconditioning training in preparation for a steady training on week 3 and 4. Statistically significant differences are indicated as follows:  vs. Morning within the same week. One, two, and three repeated symbols correspond to P<0.05, P<0.01, and P< 0.001, respectively. https://doi.org/10.1371/journal.pone.0219167.g004 Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 9 / 14 variable. Surprisingly, in most published work, problems during habituation to run and per- formance heterogeneity among trained rats are not mentioned, maybe because most of the animals finally run relatively well or because researchers discard “bad runners”, which finally are used for sedentary or control groups. In any case, it is a fact that the contingency of being a “good runner rat” can be categorised. From our experience, some rats like running from the very beginning of the experiment and they cope well with the run as the speed is increased. Other rats appear putting much more attention on avoiding the shock grid than on forward running, and others continuously touch the grid and need a long time to become continuous runners. The question posed here is whether the score itself can semiquantitatively differenti- ate between these types of running rats and if AEY score correlates to the qualitative differ- ences observed when they are monitored during the training sessions. Whilst the criteria within the A-score have been quantified in accordance with an impact on the theoretical best style of running (mark A), the simple underlying mathematical design does not allow an average score to be directly back-traced to the dominant style/mark regis- tered. That sort of evaluation requires looking at the actual registry and using different sta- tistical approach. All the accumulative average-calculations render the back tracing-style estimation improbable. On the other hand, as the E-score focuses exclusively on the tail’s posi- tioning as a sign of tiredness, its marks are more descriptive and correlate better with the actual score value they give. The assessment for the Y-score is different since it serves rather as a qual- ity control index for the data output from the treadmill monitoring apparatus. Application of the AEY-score in different exercise protocols We repeatedly noted that the rats caught up differently with the exercise on the running tread- mill during the first training sessions. In Fig 3, this progression is reflected in the greatest devi- ation in the A-score on week 1, which progressively decreases until week 4. Simultaneously, the E-score demonstrates an increasing variation, suggesting a dissociation between the “skill” to achieve a continuous running style and the endurance of the animal. Even more, it can be hypothesized that as rats learn to run more continuously, they are more prone to fatigue prob- ably because 1) the real ran distance increases; and 2) rat metabolism and anatomy is prepared for short voluntary running bursts instead of continuous speed [14]. The exercise sessions at day D are completely different; being a downhill running-to-exhaustion protocol, designed to induce muscle damage instead of a continuous running at constant slope and speed. During that protocol, a more A-score variety took place, which was reflected in a lower average A- score with greater underlying variation; and the E-score dropped significantly, as expected in a protocol designed to be extenuating. These results indicate that the AEY score is indeed able to discriminate between good and bad performance: during regular, moderate training (weeks 3 and 4) most animals perform well, obtaining scores close to 1. On the other hand, at day D the performance significantly decreased as a consequence of the exhaustive downhill running pro- tocol. Indeed, a significant increase in the plasmatic concentration of muscle damage biomark- ers creatine kinase-MM and myoglobin was found in a subgroup of animals sacrificed 24 h after day D [23], suggesting that the alterations in the AEY score had a physiological basis. Conversely, during the subsequent weeks (active recovery period from week 5 to week 7), the exercise protocol was very light. However, despite the reduced speed and duration of these sessions, rats exhibited significantly lower A-score the first week after the downhill running- to-exhaustion protocol. This A-score decrease could be associated to muscle damage [23], which could hinder rat’s ability to exercise continuously due to compromised muscle function, soreness or pain. Thus, these results suggest that this tool is sensitive enough to discern between healthy and animals with eccentric exercise-induced muscle damage. Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 10 / 14 In general, it is likely that motor control memory plays a role in the success of a good A- score; frequent similar sessions might improve it whilst less frequent or different sessions do not. Perhaps the training time in relation to circadian and internal biorhythms is also a factor of considerable importance. Comparing the morning and afternoon sessions (Fig 4), there is a significant difference in the A-score in the first two weeks. Indeed, since rats are nocturnal ani- mals, the lower A-score obtained in morning sessions it is not surprising. Nocturnal animals are more prone to perform physical activity (search of food, mating, exploration) during cre- puscular and night hours, which could explain their better performance during the afternoon sessions. It could be hypothesized that the presence of more artefacts, faeces and urine drops contacting the metal grid (reflected by the lower Y-score) during the morning sessions could be due to a physiological stress response [24] as a consequence of the intrinsic circadian bio- rhythm disruption in these animals. Furthermore, some learning effect and habituation to the motorized treadmill could contribute to the better afternoon scores. Conversely, due to a rela- tively short resting period before the afternoon session, the E-score (representing tiredness) decreased. Finally, as can be observed in Fig 3, the Y-score remained constant regardless of the degree of habituation to the treadmill (week 1 vs. week 4) and the intensity and duration of the exer- cise, which ranged from exhaustive (D) to light (weeks 5–7). Thus, parameters directly related only to electric shocks count should not be used in rat performance and compliance evalua- tion, but could serve as a good quality control indicator of training session. Limitations and advantages Probably the most adverse aspect of the proposed score is the non-automated assessment, although it gets easier with practice. There are many calculations involved, albeit simple, and setting them up in a software programme such as Microsoft Excel is one way to work automat- ically through them. The main benefits, besides the pinpointing of potential outliers, are that the method provides a systematic and sensitive way to compare rats in different treatments or experimental conditions and even among different studies. Still, the assessment leading up to an individual AEY-score will always be subjected to some bias. In terms of future develop- ments, the next step for the AEY score would be to correlate it with physiological representa- tive parameters of exercise performance, such as VO2 measurements, in a similar way to the widely used Borg’s scale for rating perceived exertion in humans [25,26]. This could either reinforce our style-score correlation or shed a new light on it. Perhaps it is difficult and unreal- istic to get a %VO2max correlation with specific running styles, but it could correlate with our 0–1 scale in either or both the A- and E-score. Moreover, this and other physiological measure- ments would be needed to further assess the validity of the proposed tool. Additionally, this score system is open enough to be applied to different strains and age. However, because we do not have tested this tool in these cases, one should empirically check if the AEY-score needs any adjustments, especially regarding the running style. For instance, it is plausible that healthy but aged rats would be unable to keep running with an A-style for a given speed and duration, or that an specific running behaviour would be more prevalent in certain conditions. In these cases, experimenters could decide to modify the scores and classifications to better fit their experimental conditions. Conclusion As an inexpensive tool relying on the experimenter’s observational capacity, we have suggested the three-part AEY score is a good method to assess and describe the laboratory rat perfor- mance capacity during exercise on a running treadmill in a semiquantifiable manner. This Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 11 / 14 would be useful for considering and correlating the sample dispersion with other physiological or anatomical parameters and, moreover, to obtain the best from the animals in question, thus contributing to promoting the principles of the 3Rs (Replacement, Reduction and Refinement) in the use of experimental animals. Supporting information S1 File. RawDataAEYScore. This database contains the registered data and calculations for the exercise style of each rat on a running treadmill according to the descriptions in Fig 1 and as reflected in the sheet 1 labelled as “Styles”. The procedure was applied in four distinct phases of a higher scale project designed for the study the effects of different interventions for skeletal muscle recovery after injury induced by forced eccentric exercise in trained rats. This phases were: a two weeks habituation to treadmill exercise period (sheet 2 “Habit”), a four weeks of exercise training period (sheet 3 “Training”), a day of two sessions (4 hours of break) of down- hill exercise until exhaustion (sheet 4 “DH Day”), and a rehabilitation period of 21 days (sheet 5”Rehab”). Pooled total data are available in a whole dataset (sheet 6 “Pool”). (XLSX) Acknowledgments This study was supported by DEP2010-22205-C02-01 and DEP2013-48334-C2-1-P grants from the Plan Nacional I+D+i (The Spanish Ministry of Economy and Competitiveness). The authors are grateful to Elizabeth Radley, Inês Ferreira, Sergi Montero, Daniel Ramos, Maria Tsimpidi, Sara Martı´nez, Marien Espino, Carla Lo´pez-Grado, Alexandra Giraldo and Fer- nando Garzo´n for their technical cooperation. The authors declare no conflict of interest. Author Contributions Conceptualization: Juan Gabriel Rı´os-Kristja´nsson, Karen Mist Kristja´nsdo´ttir. Data curation: Juan Gabriel Rı´os-Kristja´nsson, David Rizo-Roca, Cristian Andre´s Nu´ñez- Espinosa. Formal analysis: Juan Gabriel Rı´os-Kristja´nsson, Karen Mist Kristja´nsdo´ttir. Funding acquisition: Gine´s Viscor. Investigation: David Rizo-Roca, Karen Mist Kristja´nsdo´ttir, Cristian Andre´s Nu´ñez-Espinosa, Joan Ramon Torrella, Teresa Pagès, Gine´s Viscor. Methodology: Juan Gabriel Rı´os-Kristja´nsson, David Rizo-Roca, Karen Mist Kristja´nsdo´ttir, Cristian Andre´s Nu´ñez-Espinosa. Project administration: Gine´s Viscor. Resources: Teresa Pagès, Gine´s Viscor. Software: Juan Gabriel Rı´os-Kristja´nsson, Karen Mist Kristja´nsdo´ttir. Supervision: Juan Gabriel Rı´os-Kristja´nsson, Joan Ramon Torrella, Teresa Pagès, Gine´s Viscor. Validation: Cristian Andre´s Nu´ñez-Espinosa, Joan Ramon Torrella. Visualization: David Rizo-Roca, Gine´s Viscor. Writing – original draft: Juan Gabriel Rı´os-Kristja´nsson, Karen Mist Kristja´nsdo´ttir. Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 12 / 14 Writing – review & editing: David Rizo-Roca, Cristian Andre´s Nu´ñez-Espinosa, Joan Ramon Torrella, Teresa Pagès, Gine´s Viscor. References 1. Manchado-Gobatto F de B, Gobatto CA, Contarteze RVL, Papoti M, De Araujo GG, De Mello MAR. Determination of critical velocity and anaerobic capacity of running rats. J Exerc Physiol Online. 2010; 13: 40–50. Available: http://eds.a.ebscohost.com.ezproxy.liv.ac.uk/eds/pdfviewer/pdfviewer?sid= 2132f67c-5487-4a00-a900-745e96c0faf2@sessionmgr4003&vid=2&hid=4102 2. Aguiar e Silva MA, Vechetti-Junior IJ, Nascimento AF do, Furtado KS, Azevedo L, Ribeiro DA, et al. Effects of swim training on liver carcinogenesis in male Wistar rats fed a low-fat or high-fat diet. Appl Physiol Nutr Metab. 2012; 37: 1101–1109. https://doi.org/10.1139/h2012-129 PMID: 22957766 3. Chavanelle V, Sirvent P, Ennequin G, Caillaud K, Montaurier CC, Morio B, et al. Comparison of oxygen consumption in rats during uphill (concentric) and downhill (eccentric) treadmill exercise tests. J Sport Sci Med. 2014; 13: 689–694. Available: http://jssm.org/abstresearch.php?id=jssm-13-689.xml 4. Real CC, Garcia PC, Britto LRG, Pires RS. Different protocols of treadmill exercise induce distinct neu- roplastic effects in rat brain motor areas. Brain Res. Elsevier; 2015; 1624: 188–198. https://doi.org/10. 1016/j.brainres.2015.06.052 PMID: 26232571 5. Rios JL, Boldt KR, Mather JW, Seerattan RA, Hart DA, Herzog W. Quantifying the Effects of Different Treadmill Training Speeds and Durations on the Health of Rat Knee Joints. Sport Med—open. 2018; 4: 15. https://doi.org/10.1186/s40798-018-0127-2 PMID: 29610999 6. Boldt KR, Rios JL, Joumaa V, Herzog W. Force properties of skinned cardiac muscle following increas- ing volumes of aerobic exercise in rats. J Appl Physiol. 2018; 125: 495–503. https://doi.org/10.1152/ japplphysiol.00631.2017 PMID: 29722623 7. Barbato JC, Koch LG, Darvish A, Cicila GT, Metting PJ, Britton SL. Spectrum of aerobic endurance run- ning performance in eleven inbred strains of rats. J Appl Physiol. 1998; 85: 530–536. https://doi.org/10. 1152/jappl.1998.85.2.530 PMID: 9688730 8. Koch LG, Green CL, Lee AD, Hornyak JE, Cicila GT, Britton SL. Test of the principle of initial value in rat genetic models of exercise capacity. Am J Physiol Integr Comp Physiol. 2005; 288: R466–R472. 9. Ways JA, Smith BM, Barbato JC, Ramdath RS, Pettee KM, DeRaedt SJ, et al. Congenic strains confirm aerobic running capacity quantitative trait loci on rat chromosome 16 and identify possible intermediate phenotypes. Physiol Genomics. 2007; 29: 91–7. https://doi.org/10.1152/physiolgenomics.00027.2006 PMID: 17179209 10. Griffiths WJ. The occurrence of convulsion in rats in the absence of auditory stimulation. Annu Anim Psychol. 1954; 4: 1–6. https://doi.org/10.2502/janip1944.4.1 11. Gerald MC. Effects of (+)-amphetamine on the treadmill endurance performance of rats. Neuropharma- cology. 1978; 17: 703–704. https://doi.org/10.1016/0028-3908(78)90083-7 PMID: 692827 12. Moraska A, Deak T, Spencer RL, Roth D, Fleshner M. Treadmill running produces both positive and negative physiological adaptations in Sprague-Dawley rats. Am J Physiol Regul Integr Comp Physiol. 2000; 279: R1321–9. https://doi.org/10.1152/ajpregu.2000.279.4.R1321 PMID: 11004000 13. Sakakima H, Yoshida Y, Sakae K, Morimoto N. Different frequency treadmill running in immobilization- induced muscle atrophy and ankle joint contracture of rats. Scand J Med Sci Sport. 2004; 14: 186–192. https://doi.org/10.1111/j.1600-0838.2004.382.x PMID: 15144359 14. Squibb RE, Collier GH, Squibb RL. Effect of treadmill speeds and slopes on voluntary exercise in rats. J Nutr. 1977; 107: 1981–4. https://doi.org/10.1093/jn/107.11.1981 PMID: 908955 15. Arnold JC, Salvatore MF. Getting to compliance in forced exercise in rodents: a critical standard to eval- uate exercise impact in aging-related disorders and disease. J Vis Exp. MyJoVE Corporation; 2014; https://doi.org/10.3791/51827 PMID: 25178094 16. Ferraresso RLP, de Oliveira RB, Macedo DV, Nunes LAS, Brenzikofer R, Damas D, et al. Interaction between overtraining and the interindividual variability may (not) trigger muscle oxidative stress and car- diomyocyte apoptosis in rats. Oxid Med Cell Longev. Hindawi Limited; 2012; 2012: 935483. https://doi. org/10.1155/2012/935483 PMID: 22848785 17. Tekin D, Dursun AD, Fic¸icilar H. The body weight performance relationship of rats on treadmill running. J Exerc Physiol. 2005; 11: 44–55. 18. Nu´ñez-Espinosa C, Ferreira I, Rı´os-Kristja´nsson JG, Rizo-Roca D, Garcı´a Godoy MD, Rico LG, et al. Effects of intermittent hypoxia and light aerobic exercise on circulating stem cells and side population, after strenuous eccentric exercise in trained rats. Curr Stem Cell Res Ther. 2015; 10: 132–9. https://doi. org/10.2174/1574888X09666140930130048 PMID: 25266982 Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 13 / 14 19. Nu´ñez-Espinosa C, Douziech A, Rı´os-Kristja´nsson JG, Rizo D, Torrella JR, Pagès T, et al. Effect of intermittent hypoxia and exercise on blood rheology and oxygen transport in trained rats. Respir Physiol Neurobiol. 2014; 192: 112–7. https://doi.org/10.1016/j.resp.2013.12.011 PMID: 24373840 20. Rizo-Roca D, Bonet JB, I´nal B, Rı´os-Kristja´nsson JG, Pagès T, Viscor G, et al. Contractile Activity Is Necessary to Trigger Intermittent Hypobaric Hypoxia-Induced Fiber Size and Vascular Adaptations in Skeletal Muscle. Front Physiol. Frontiers; 2018; 9: 481. https://doi.org/10.3389/fphys.2018.00481 PMID: 29780328 21. Rizo-Roca D, Rı´os-Kristja´nsson JG, Nu´ñez-Espinosa CA, Ascensão A, Magalhães J, Torrella JR, et al. A semiquantitative scoring tool to evaluate eccentric exercise-induced muscle damage in trained rats. Eur J Histochem. 2015; 59. https://doi.org/10.4081/ejh.2015.2544 PMID: 26708179 22. European Council, European Parliament. Caring for animals aiming for better science [Internet]. Official Journal of the European Union European Union; 2010 p. 2010/63/EU. Available: http://ec.europa.eu/ environment/chemicals/lab_animals/pdf/guidance/directive/en.pdf 23. Rizo-Roca D, Rı´os-Kristja´nsson JG, Nu´ñez-Espinosa C, Santos-Alves E, Gonc¸alves IO, Magalhães J, et al. Intermittent hypobaric hypoxia combined with aerobic exercise improves muscle morphofunctional recovery after eccentric exercise to exhaustion in trained rats. J Appl Physiol. 2017; 122: 580–592. https://doi.org/10.1152/japplphysiol.00501.2016 PMID: 27765844 24. Hall CS. Emotional behavior in the rat. I. Defecation and urination as measures of individual differences in emotionality. J Comp Psychol. 1934; https://doi.org/10.1037/h0071444 25. Borg G. Psychophysical scaling with applications in physical work and the perception of exertion. Scand J Work Environ Heal. 1990; 16: 55–58. https://doi.org/10.5271/sjweh.1815 26. Borg E, Kaijser L. A comparison between three rating scales for perceived exertion and two different work tests. Scand J Med Sci Sport. 2006; 16: 57–69. https://doi.org/10.1111/j.1600-0838.2005.00448.x PMID: 16430682 Performance score for running rats PLOS ONE | https://doi.org/10.1371/journal.pone.0219167 July 9, 2019 14 / 14

A three-criteria performance score for rats exercising on a running treadmill.

07-09-2019

Ríos-Kristjánsson, Juan Gabriel,Rizo-Roca, David,Kristjánsdóttir, Karen Mist,Núñez-Espinosa, Cristian Andrés,Torrella, Joan Ramon,Pagès, Teresa,Viscor, Ginés

eng

PMC7379642

Supplement Table 4. Change in VO2max (ml·min-1·kg-1) from 1995-1997 to 2016-2017 in relation to length of education. L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 Year n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change 95-97 731 2.74 (0.12) Ref 36.2 (1.30) Ref 3 216 2.80 (0.15) Ref 38.5 (1.55) Ref 627 2.84 (0.13) Ref 39.9 (1.45) Ref 98-99 880 2.64 (0.14) -3,5% 34.5 (2.09) -4,6% 4 416 2.72 (0.16) -2,7% 37.1 (1.67) -3,7% 1 247 2.83 (0.13) -0,2% 39.0 (1.64) -2,2% 00-01 1 543 2.63 (0.12) -4,2% 35.0 (1.50) -3,3% 8 398 2.80 (0.13) -0,1% 36.9 (1.59) -4,2% 2 604 2.71 (0.17) -4,5% 37.6 (2.19) -5,7% 02-03 2 572 2.43 (0.15) -11,5% 32.5 (1.84) -10,3% 15 551 2.66 (0.14) -5,0% 35.5 (1.56) -7,8% 4 506 2.74 (0.13) -3,6% 37.9 (1.51) -4,9% 04-05 3 625 2.55 (0.14) -7,1% 33.8 (1.68) -6,6% 24 312 2.67 (0.13) -4,7% 35.6 (1.46) -7,5% 9 483 2.72 (0.13) -4,1% 37.8 (1.52) -5,3% 06-07 3 909 2.57 (0.14) -6,3% 33.5 (1.60) -7,4% 25 167 2.67 (0.13) -4,7% 35.4 (1.39) -8,1% 9 443 2.75 (0.13) -3,3% 37.9 (1.36) -4,9% 08-09 4 171 2.51 (0.12) -8,2% 32.7 (1.64) -9,6% 28 057 2.69 (0.13) -3,9% 35.4 (1.33) -8,1% 11 251 2.78 (0.13) -2,1% 38.3 (1.43) -4,0% 10-11 3 626 2.55 (0.12) -7,0% 32.9 (1.38) -9,2% 24 837 2.68 (0.13) -4,2% 35.0 (1.38) -9,1% 10 714 2.80 (0.13) -1,5% 38.3 (1.41) -4,1% 12-13 4 384 2.47 (0.11) -9,7% 31.9 (1.42) -11,8% 34 838 2.67 (0.12) -4,8% 34.7 (1.35) -9,8% 18 024 2.75 (0.13) -3,3% 37.9 (1.46) -5,0% 14-15 4 053 2.45 (0.12) -10,6% 31.5 (1.41) -13,1% 35 047 2.63 (0.12) -5,9% 34.2 (1.26) -11,2% 16 484 2.71 (0.12) -4,6% 37.2 (1.34) -6,7% 16-17 2 446 2.43 (0.12) -11,4% 31.6 (1.22) -12,8% 23 341 2.63 (0.12) -6,2% 34.1 (1.30) -11,5% 10 774 2.71 (0.12) -4,5% 37.1 (1.27) -7,0% ≤9 years 10-12 years ≥12 years

Decline in cardiorespiratory fitness in the Swedish working force between 1995 and 2017.

11-15-2018

Ekblom-Bak, Elin,Ekblom, Örjan,Andersson, Gunnar,Wallin, Peter,Söderling, Jonas,Hemmingsson, Erik,Ekblom, Björn

eng

PMC7379642

Supplement Table 10. Test for equility of variance in unstandardized mean (SD) of relative VO2max in the first five years (1995-1999) and and the last five years (2013-2017) in relation to sub-groups of sex, age and educational level. Age-group Education length Year Mean SD F-value p-value 18-34 years <12 years 1995-1999 42.6 9.7 2013-2017 39.9 10.2 2.37 0.124 ≥12 years 1995-1999 46.1 10.3 2013-2017 42.8 10.2 0.38 0.538 35-49 years <12 years 1995-1999 36.9 9.2 2013-2017 34.8 9.2 0.06 0.803 ≥12 years 1995-1999 38.4 8.8 2013-2017 38.3 9.7 8.09 0.004 50-74 years <12 years 1995-1999 31.2 7.8 2013-2017 30.0 7.8 1.19 0.274 ≥12 years 1995-1999 33.8 7.8 2013-2017 32.8 8.3 3.16 0.076 Age-group Education length Year Mean SD F-value p-value 18-34 years <12 years 1995-1999 44.1 10.3 2013-2017 39.6 9.9 6.29 0.012 ≥12 years 1995-1999 46.0 10.5 2013-2017 43.2 10.5 0.45 0.501 35-49 years <12 years 1995-1999 37.4 8.9 2013-2017 34.3 8.8 0.05 0.822 ≥12 years 1995-1999 38.8 8.7 2013-2017 38.7 9.6 9.78 0.002 50-74 years <12 years 1995-1999 32.5 7.5 2013-2017 30.3 7.6 0.46 0.499 ≥12 years 1995-1999 34.4 7.5 2013-2017 33.2 8.3 4.05 0.044 Women Men Levene's Test for Equality of Variances Levene's Test for Equality of Variances

Decline in cardiorespiratory fitness in the Swedish working force between 1995 and 2017.

11-15-2018

Ekblom-Bak, Elin,Ekblom, Örjan,Andersson, Gunnar,Wallin, Peter,Söderling, Jonas,Hemmingsson, Erik,Ekblom, Björn

eng

PMC8048782

Ann. N.Y. Acad. Sci. ISSN 0077-8923 ANNALS OF THE NEW YORK ACADEMY OF SCIENCES Special Issue: Annals Reports Original Article Instructed versus spontaneous entrainment of running cadence to music tempo Edith Van Dyck, Jeska Buhmann, and Valerio Lorenzoni IPEM, Ghent University, Ghent, Belgium Address for correspondence: Edith Van Dyck, IPEM, Ghent University, De Krook, floor +4, Miriam Makebaplein 1, 9000 Ghent, Belgium. [email protected] Matching exercise behavior to musical beats has been shown to favorably affect repetitive endurance tasks. In this study, our aim was to explore the role of spontaneous versus instructed entrainment, focusing on self-paced exercise of healthy, recreational runners. For three 4-min running tasks, 33 recreational participants were either running in silence or with music; when running with music, either no instructions were given to entrain to the music, or participants were instructed to match their running cadence with the tempo of the music. The results indicated that less entrainment occurred when no instruction to match the exercise with the musical tempo was provided. In addition, similar to the condition without music, lower speeds and shorter step lengths were observed when runners were instructed to match their running behavior to the musical tempo when compared with the condition without such instruction. Our findings demonstrate the impact of instruction on running performance and stress the importance of intention to entrain running behavior to musical beats. Keywords: music; running; movement; entrainment; auditory–motor coupling Introduction Music and exercise—the combination of the two is often regarded as a great match, a viewpoint that has important implications for the sports and exer- cise domain. It is, for instance, hard to imagine a gym deprived of loud, energetic music motivated by the conviction that music boosts performance. In sports and exercise research, this hypothesis has been tested repeatedly, demonstrating that music is indeed capable of increasing exercise intensity and endurance,1–5 stimulating rhythmic movement,4 distracting from fatigue and discomfort,6,7 prompt- ing and altering mood states,7 spurring motivation,8 inducing arousal,9 relieving stress,6 and evoking a sense of power and producing power-related cog- nition and behavior.10 Music to which performance can be synchronized in particular was shown to extend endurance and increase exercise intensity.11 The process underlying this particular type of auditory–motor coupling is commonly referred to as entrainment. It entails a match of musical tempi with exercising tempi, locked in a particular period relationship (e.g., running tempo matching musi- cal tempo) and resulting in regular corporeal pat- terns. Performance boosting effects of entrained music are rooted in the ability of motor-to-music entrainment to reduce the metabolic cost of exer- cise by enhancing neuromuscular or metabolic efficiency.12,13 Owing to the absence of timely adjustments within the kinetic pattern and an increase in the level of relaxation resulting from the precise expectancy of the forthcoming movement, regular corporeal patterns demand less energy to imitate.14 Hence, by employing music that can be corporeally emulated, a point of reference is estab- lished that is able to attract and thus entrain recur- ring motor patterns.13,15 Entrainment and its related benefits were shown to be particularly useful for repetitive endurance tasks, such as walking, running, rowing, and doi: 10.1111/nyas.14528 91 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. Instructed and spontaneous music-entrained running Van Dyck et al. cycling.11,12,16–18 In addition, improvements in endurance performance proved to be most appar- ent at low-to-moderate exercise intensities.6,19 This is largely explained by Rejeski’s parallel processing hypothesis, which states that as exercise intensity increases, physiological cues (e.g., heart and respira- tion rates) predominate.20 Thus, when the exercise becomes too strenuous, perception of neural exer- tion signals coming from the muscles, joints, and cardiopulmonary systems increases, resulting in an attention shift toward the painful and/or fatiguing effects of the exercise.20–24 Most previous research on the effects of music- to-motor coupling on exercise and sports focused on instructed (or imposed, intended) entrainment (e.g., see Refs. 9, 11, 16, 25, and 26). In this case, the exerciser is explicitly instructed to match his/her exercise behavior to a musical beat or pulse. How- ever, entrainment can also occur spontaneously, or when the exerciser is not instructed to match his/her behavior to the music. Although less research has been performed regarding spontaneous (or unin- structed/unintended) entrainment, some have indi- cated that humans indeed possess a natural pre- disposition to respond to rhythmical qualities of music.27–29 Yet, spontaneous entrainment of one tempo with another is only believed to occur when the strength of the coupling is able to overcome pos- sible contrasts in the natural movement period or tempo. The difference between the period of the music and that of the exercise, thus, should not exceed a specific range, referred to as the entrain- ment basin.28,30–32 It remains rather unclear whether these different approaches could result in divergent effects on performance output, as research combining both instructed and spontaneous entrainment is sparse. However, some research on walking behavior did compare both approaches, stressing the limitations of spontaneous entrainment.29,33 Moreover, when the required intensity to match the walking behav- ior to the beats proved too large, it was shown that intentional entrainment with an active cognitive control mechanism was required in order to obtain movement-to-music coupling.29 In our study, the aim was to further explore possible differences between instructed and spon- taneous entrainment by focusing on a repetitive endurance exercise, namely running. We used a within-subjects design to investigate possible con- trasts in the effects of both approaches (comple- mented with a baseline condition without music, serving as a point of reference) on a selection of key outcome measures. As music was indicated to be of greater benefit to untrained or recreationally- active individuals than to those who are highly trained,34,35 and since this group is heavily repre- sented in current society, recreational runners were targeted here. Intrinsically, the goal of our study was to provide outcomes that might prove to be of inter- est to a large population of exercisers and valuable to future research on music and exercise. Method Participants To establish sample size, power analysis for a repeated-measures design was conducted using G∗Power 3.1.9.2.36 On the basis of a small effect size, with alpha set at 0.05 and power at 0.90, it was estimated that about 32 participants would be required. Thirty-three healthy adult participants (18 females/15 males) took part in the study. The test group consisted of recreational runners with an average age of 34.21 years (SD = 8.17), a mean body mass of 62.49 kg (SD = 12.95), and an aver- age height of 1.70 m (SD = 0.11), who reported being fit enough to run comfortably for at least 30 min without feeling exhausted. Only a minority (36.36%) had received musical training. On aver- age, musically trained participants had 9.50 years (SD = 12.03) of musical experience and were edu- cated in music schools (37.50%) or conservato- ries (6.25%), through private lessons (37.50%), self- education (18.75%), or a combination of the above. All participants reported running regularly, with varying degrees of frequency (66.67% reported run- ning multiple times a week; 30.30% about once a week; and 3.03% about once a month). Of all par- ticipants, 51.52% reported generally running with- out music, 39.39% typically trained with music, and 9.09% ran both with and without musical accom- paniment. Fisher’s exact test showed no signifi- cant association between participants’ sex and their musical background (χ2(1) = 0.16, P = 0.73) or their habit to run to music (χ2(2) = 0.92, P = 0.69). Ethics statement The study was approved by the Ethics Commit- tee of the Faculty of Arts and Philosophy at Ghent University, Belgium, and all procedures followed 92 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Van Dyck et al. Instructed and spontaneous music-entrained running were in accordance with the Declaration of Helsinki. In addition, all participants signed a form to declare that they participated voluntarily; that they had received sufficient information concerning the tasks, procedures, and technologies used; that they had the opportunity to ask questions; and that they were aware of the fact that running movements were measured for scientific and educational purposes only. Stimulus For all participants and conditions, the same music track was played to control for possible effects of musical characteristics. However, since recreational running tempi generally vary between 130 and 200 steps per minute (SPM), the track was required to efficiently deal with substantial tempo variations. Yet, to minimize the degree of tempo-stretching, a stimulus with an original tempo of about 165 beats per minute (BPM) was selected. Furthermore, to facilitate the activating character, clearly audible beats were mandatory as a stable tempo through- out the entire track.37 Finally, to further facilitate the imperceptibility of the tempo-stretching, a track was selected with low to no appearances in national and international music charts, that is, one that was unfamiliar to (most of) the participants. Familiarity with the stimulus was further checked in a postques- tionnaire, with 87.88% stating to not know the track at all, 6.06% reporting to have possibly recognized the track, 0% stating to know the track, and 6.06% to being indecisive. Taking the above-described cri- teria into account, the song International Dateline by Ladytron (2005), with an original tempo of 168 BPM, was selected. As the duration of the track did not cover the complete length of a condition (i.e., 4 min), the chorus part in the middle of the song was copied and repeated at the end of the track (using Audacity software, see http://audacity. sourceforge.net) when it had to undergo substantial tempo increases (up to 200 BPM). Beats were auto- matically detected using BeatRoot38 and manually checked afterward. Apparatus Participants were equipped with two iPods (4th generation); one attached to each ankle. Using the Sensor Monitor Pro application on the iPods, data from the iPod accelerometers and gyroscopes were streamed wirelessly at 100 Hz to a 7′′ tablet (Pana- sonic Roughpad FZ-M1) running Windows 8.1. The tablet was strapped to a backpack, together with a sonar (MaxBotix LV-MaxSonar-EZ: MB1010) pointing to the right of the runner and con- nected to the tablet through a Teensy 3.1 micro- controller. Twenty-nine 1.9-m vertical marker rods were placed on the right side of the running track (289 m) with a spacing of 9.97 meters. The sonar and the rods were used to calculate the runners’ speed in a postprocessing phase. The wireless connection between the tablet and iPods was provided through a Wi-Fi router (TP- Link M5360), firmly strapped to the backpack, ensuring reliable communication between the iPods and the tablet. On the tablet, Max/MSP from Cycling74’ was running together with a patch specifically designed to read out the sensor data, implement the different conditions, and store the data. The audio output was provided through Sennheiser HD 215 headphones. Music tempo was manipulated using the MAX/MSP elastic∼ object by Simon Adcock, which allows for tempo alter- ations of ±/–100% of the original music tempo without pitch modifications. Procedure The experiments took place at an indoor track- and-field site (Flanders Sports Arena, Ghent, Bel- gium). Participants were equipped with the iPods, headphones, and a backpack containing the tablet, sonar, and Wi-Fi router. They were asked to run four times for 4 minutes. No information was dis- tributed concerning the real purpose of the exper- iment and all participants ran solo. After each 4-min running session, a break of at least 5 min was introduced to enable them to recover suffi- ciently. During the break, the participant was asked to take sufficient rest. After he/she expressed feel- ing approximately as fit as at the start of the exper- iment, the participant initiated the following run- ning session. Between sessions, participants were asked to fill out the Borg Rating of Perceived Exer- tion (RPE) Scale39,40 and indicate how heavy the effort had been during the exercise, ranging from 6 (“no exertion at all”) to 20 (“maximal exertion”). In addition, they rated the level of physical enjoyment of the previously performed exercise on the 8-item version of the Physical Activity Enjoyment Scale (PACES),41,42 a single-factor 7-point Likert scale to assess the level of enjoyment during physical activ- ity in adults across exercise modalities. 93 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Instructed and spontaneous music-entrained running Van Dyck et al. In the training session, participants were asked to run at their self-paced cadence without musical accompaniment. This session was included to warm up and get acquainted with the running track and was not taken into account in the analysis. In the first running session (no music condition), no music was played and participants were asked to run at their self-paced cadence. Next, a familiarization task took place where the participants first listened to the music track without moving to it. In the sec- ond session (uninstructed condition), participants exercised at their self-paced cadence again, this time accompanied by music with a tempo matching their cadence assessed during the last 120 steps taken in the previous condition.a During the third session (instructed condition), the same stimulus was pre- sented, yet participants were instructed to “match their running cadence with the tempo of the music.” As exercise behavior might be influenced by a fore- seen completion of the task (e.g., speeding up near the end of the experiment), a fourth and final con- dition was added, in which participants were asked to run at their self-pace cadence once more with musical accompaniment. This condition was merely implemented to control for confounding effects of anticipated task completion and was not taken into account in the analysis. At the end of the experiment, participants filled out a questionnaire on personal background, music education, and sports training. In addi- tion, participants’ perception of their personal level of movement-to-music matching behavior in the instructed condition was assessed, as well as to what extent they generally tend to match their cadence to musical tempi outside the experimental setting. Data analysis To test the effect of the specific condition on run- ning behavior, the following features were calcu- lated: cadence, speed, step length, tempo entrain- ment, mean relative phase angle (rPA), and resul- tant vector length (RVL). Before the calculation of all features, the initial 60 s of each run was discarded to avoid a start-up effect. The final 30 s of each run was ignored as well, to eliminate altered running aThe final 40 steps were excluded in order to disregard possible alterations in cadence due to the anticipation of the end of the condition. behavior due to the anticipated ending (e.g., slow- ing down or speeding up). Movement features were calculated as follows. Cadence (SPM). Running cadence was calculated in real time using the acceleration data acquired by the iPods. A change in the movement direction of the leg, detected by the gyroscope, was identified as a step. The tempo intervals of eight consecutive steps of the same leg were used to calculate cadence (SPM) in a moving average manner. Step length (m). Step length was calculated in real time as the distance measured from the heel print of one foot to the heel print of the other foot. Speed (km/h). The distance measurements pro- vided by the sonar were used in a postprocessing phase to evaluate running speed. When the runner passed along the rods, placed on the right side of the track, a distance minimum was detected. Through computation of the time between the minima, that is, between the rods, average speed was determined. The analog signal was sampled at 250 Hz and digi- tized using the Teensy microcontroller. Tempo entrainment (%). Another measure con- sisted of the percentage of tempo-entrained steps during the conditions with music. A step taken in a tempo sufficiently close to the music tempo (max- imum of 1% difference between SPM and BPM) at that specific moment is regarded as a tempo- entrained step. The tempo entrainment score is the percentage of tempo-entrained steps out of the total number of steps. Mean rPA (degrees). The mean rPA is a measure of the timing of the footfall relative to the closest beat and can be expressed as either a positive (foot- fall after the beat) or a negative (footfall before the beat) angle in degrees. The rPA of 0° refers to a foot- fall that is exactly timed on the beat, and an angle of 180° refers to a footfall that is timed precisely in between two beats. Such an rPA can be calculated for each step with the following equation (St refers to the time of a step, B1 refers to the time of the beat before the step, and B2 refers to the time of the beat after the step) φ = 360∗ St − B1 B2 − B1, 94 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Van Dyck et al. Instructed and spontaneous music-entrained running after which the circular mean of all rPAs can be calculated.43 The mean of all rPAs is only of inter- est if there is a sufficient amount of consistency in entrainment, which is expressed by the RVL (see below). Therefore, only mean rPA values that cor- respond to RVL values of ≥0.75 are considered in the analysis. RVL (value from 0 to 1). The RVL expresses the coherence or stability of the rPA over time.44 If the distribution of the rPA over time is narrow (when all phase angles are clustered around the mean), it leads to a high RVL (maximum value 1), which indicates highly consistent entrainment. In the case of a broad or multimodal rPA over time, RVL is low (minimum value toward 0), indicating no auditory–motor cou- pling or entrainment with the music. Addition- ally, participants were divided into entrainers and nonentrainers using a cutoff of ≥0.75, based on pre- vious research (e.g., see Refs. 29 and 37). For all movement features, a 3 × 2 mixed-design ANOVA with condition as within-subjects factor (no music, uninstructed, and instructed) and sex as a between-subject factor was performed. The no music condition could not be taken into account for features depending on musical parameters (e.g., tempo entrainment, rPA, and RVL). An indepen- dent samples t-test was performed to check for the effects of musical training on tempo entrain- ment and RVL, while one-way ANOVA was exe- cuted to check for differences between participants reporting to habitually run with, without, or both with and without music. Friedman’s ANOVA was employed to check for differences in PACES and BORG RPE-scale ratings between conditions and one-way ANOVA was used to examine the poten- tial effects of perceived alignment on tempo entrain- ment. Results Cadence A significant main effect of condition was revealed, F(2,62) = 18.12, P < 0.001. Contrasts showed that running cadence was significantly lower in the no music condition (M = 168.78; SE = 1.53) compared with the other conditions: uninstructed (M = 170.89; SE = 1.54), F(1,31) = 42.61, P < 0.001, η2 = 0.58, and instructed (M = 170.30; SE = 1.45), F(1,31) = 13.07, P = 0.001, η2 = 0.30. No significant difference was found between the instructed and uninstructed conditions, F(1,31) = 3.31, P = 0.08, η2 = 0.10. A significant main effect of sex was obtained as well, revealing higher cadence rates for females (M = 173.34; SE = 1.86) than males (M = 165.97; SE = 2.03), F(1,31) = 7.18, P = 0.009, η2 = 0.18. There was no significant effect of the condition × sex interaction, F(2,62) = 0.31, P = 0.73 (Fig. 1A). Step length We obtained a significant main effect of condi- tion, F(2,62) = 13.43, P < 0.001, demonstrating larger step lengths in the uninstructed condition (M = 1.12; SE = 0.04) compared with the no music (M = 1.09; SE = 0.03), F(1,31) = 18.41, P < 0.001, η2 = 0.37, and the instructed one (M = 1.08; SE = 0.03), F(1,31) = 21.25, P < 0.001, η2 = 0.41. No significant difference was obtained between the no music and instructed conditions, F(1,31) = 3.01, P = 0.09, η2 = 0.09. There was a significant main effect of sex, reveal- ing larger step lengths for males (M = 1.21; SE = 0.04) compared with their female counterparts (M = 1.00; SE = 0.04), F(1,31) = 14.39, P = 0.001, η2 = 0.31. In addition, a significant effect of the condition × sex interaction was obtained, indicating that males make larger differences in step length between the uninstructed and instructed condition compared with their female counterparts, F(2,62) = 4.88, P = 0.01 (Fig. 1B). Speed There was a significant main effect of condition, F(2,62) = 15.21, P < 0.001, demonstrating faster running behavior in the uninstructed condition (M = 11.44; SE = 0.35) compared with the no music (M = 11.02; SE = 0.32), F(1,31) = 40.40, P < 0.001, η2 = 0.56, and the instructed (M = 10.96; SE = 0.29), F(1,31) = 20.86, P < 0.001, η2 = 0.41, conditions. No significant difference was found between the no music and instructed conditions, F(1,31) = 0.43, P = 0.52, η2 = 0.01. There was a significant main effect of sex, reveal- ing higher speed levels for males (M = 12.07; SE = 0.42) compared with their female counterparts (M = 10.37; SE = 0.38), F(1,31) = 8.87, P = 0.006, η2 = 0.22. A significant effect of the condition × sex inter- action was obtained as well, showing that larger speed differences between the uninstructed and 95 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Instructed and spontaneous music-entrained running Van Dyck et al. Figure 1. Results on cadence (A), step length (B), and speed (C) data of male and female runners, for the no music, uninstructed, and instructed conditions. Data presented are the mean ± SE. instructed conditions were made by males than by females, F(2,62) = 4.36, P = 0.02 (Fig. 1C). Tempo entrainment Tempo entrainment proved to be significantly higher in the instructed (M = 0.62; SD = 0.28) com- pared with the uninstructed condition (M = 0.51; SD = 0.30), F(1,31) = 7.68, P = 0.009, η2 = 0.20. There was no significant main effect of sex, F(1,31) = 1.36, P = 0.25, η2 = 0.04, nor was there a significant effect of the condition × sex interaction, F(1,31) = 0.07, P = 0.79, η2 = 0.003 (Fig. 2A). Mean rPA (rPA) A significant difference was found for mean rPA, indicating footfalls more closely match musical beats in the instructed condition (M = −33.75; SD = 36.63) compared with the uninstructed one (M = −47.69; SD = 42.16), F(1,7) = 7.32, P = 0.03, η2 = 0.52. No significant main effect of sex, F(1,31) = 0.14, P = 0.72, η2 = 0.02, or of the condition × sex inter- action, F(1,31) = 0.03, P = 0.87, η2 = 0.004, was obtained. (Fig. 2B). RVL The RVL was shown to be significantly higher in the instructed (M = 0.69; SD = 0.30) compared with the uninstructed condition (M = 0.49; SD = 0.34), F(1,31) = 17.18, P < 0.001, η2 = 0.13. No significant main effect of sex, F(1,31) = 1.55, P = 0.22, η2 = 0.05, or of the condition × sex inter- action, F(1,31) = 0.63, P = 0.44, η2 = 0.0004, was found (Fig. 2C). Musical experience Significantly higher levels of tempo entrainment and RVL were obtained for musically trained par- ticipants (tempo entrainment: M = 0.68, SD = 0.16; and RVL: M = 0.73, SD = 0.20), compared with their untrained counterparts (tempo entrainment: M = 0.49, SD = 0.29, t(31) = −2.12, P = 0.04, η2 = 0.13; and RVL: M = 0.51, SD = 0.30, t(31) = −2.33, P = 0.03, η2 = 0.15). No significant effects for either of these parame- ters (tempo entrainment, F(2,30) = 0.17, P = 0.85; and RVL, F(2,30) = 0.002, P = 0.99) were found between participants who reported habitually run- ning with or without music, or those who indicated 96 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Van Dyck et al. Instructed and spontaneous music-entrained running Figure 2. Results on tempo entrainment (A), mean relative phase angle (B), and resultant vector length (C) data of male and female runners, for both the uninstructed and instructed conditions. Data presented are the mean ± SE. both running with and without musical accompani- ment. PACES A significant main effect of the ratings was obtained, χ2(2) = 15.53, P < 0.001, and Wilcoxon tests were used to follow up on this finding. After Bonfer- roni correction, it appeared that ratings for the no music condition (Mdn = 4.63) were signifi- cantly lower compared with the other two con- ditions: uninstructed, (Mdn = 5.00), z = −3.77, P < 0.001, η2 = 0.23; and instructed, (Mdn = 5.00), z = −2.61, P = 0.009, η2 = 0.10. No significant differences were found between the uninstructed and instructed conditions, z = −0.97, P = 0.33, η2 = 0.01. Borg RPE No significant change in Borg RPE-scale ratings over the conditions was obtained, χ2(2) = 0.29, P = 0.87. Perception of alignment Of all participants, none of them answered nega- tively when asked if they believed they had aligned their running movements with the music in the instructed condition; 15.15% replied that they did not know whether they did so or not; 21.21% answered that, at times, they indeed aligned with the music; and 63.64% reported to have aligned their running cadence with the music tempo most of the time. We checked for differences in auditory– motor coupling between these three groups of par- ticipants. However, no significant effect of per- ceived alignment was found for tempo entrainment, F(2,30) = 1.10, P = 0.35, rPA, F(2,16) = 0.24, P = 0.79, or RVL, F(2,30) = 1.12, P = 0.34. When asked about their entrainment behavior in daily life, 39.39% of all participants did not know whether they entrain their running behavior to the perceived music or generally do not run to music; 3.03% of them reported not entraining to music in daily life; 45.45% disclosed to occasionally entrain- ing with the music; and 12.12% pointed out gener- ally entraining to music while running. Discussion In this study, we aimed to investigate the dif- ference between instructed and uninstructed (or 97 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Instructed and spontaneous music-entrained running Van Dyck et al. spontaneous) entrainment of running to music. Our findings showed that instruction resulted in a significant increase in the level of movement-to- music entrainment when compared with the same running task without instruction. When the run- ners were not instructed to align their movements with the musical beats, only 33.33% of the partici- pants spontaneously entrained to the stimulus. Yet, when instructions to “match running cadence with musical tempo” were given, 57.58% of the runners entrained with the beats. Such results are rather surprising, since in walking research usually larger population ratios display spontaneous entrainment to musical stimuli in tempi close to preferred exer- cising paces (e.g., 40%, see Ref. 33; about 50%, see Ref. 37; and about 60%, see Ref. 29). Moreover, stud- ies on walking also revealed higher levels of entrain- ment with instruction (e.g., 74%, see Ref. 33; and up to 93%, see Ref. 29). However, running is a more strenuous effort and involves different biomechan- ics; although it is a natural extension of walking, running involves increased velocities, joint range of motion, forces, muscle activity, joint moments, and joint powers as compared with walking. Thus, run- ning stresses the mechanics of the body to a greater extent, as such also increasing the risk of related injury.45 Even though, on average, entrainment was lower compared with previous walking research, a similar difference between spontaneous and instructed entrainment was demonstrated for runners, with higher levels of tempo entrainment as well as RVL (an alternative measure of entrainment) when instructed to match exercise behavior to musical beats. In addition, despite the fact that footfall instances generally preceded musical beats, rPAs decreased with instruction, indicating a closer match to beat occurrences. However, even when instructions to match running behavior to music were given, entrainment frequency remained rather low, a finding similar to previous work showing low movement-to-music coupling frequencies after participants were instructed to adapt movements of the entire body to the music.46 These results support the hypothesis that matching movements to musical beats may not be a simple, low-level task; entrainment in itself may be cognitively demanding,46 most particularly for individu- als who have difficulty perceiving the beat in music.33,47–49 Besides entrainment, also cadence, step length, and speed—three key performance measures of running—were scrutinized. All three features proved to increase in the uninstructed condition when compared with the condition without musical accompaniment. This is in line with the idea that music is capable of increasing exercise intensity and endurance.1–5 Although the precise mechanisms through which music can boost performance still require further investigation, this effect might be (partly) explained by the propensity of music to heighten arousal.34,50,51 In the instructed con- dition, a similar increase in cadence occurred. However, compared with the silent condition, step length and speed did not significantly change when instructed to run to the beat. These results are consistent with previous results on walking behavior, demonstrating that instructing partic- ipants to move to the beat elicited slower and shorter strides than when instruction was absent.33 They are also in accordance with earlier findings indicating that, when compared with stride-based pacing, step-based pacing leads to more stable auditory–motor coordination in both walking and running.52 Consequently, although a number of studies demonstrated that auditory–motor cou- pling improved performance in motor tasks,11,12,18 our findings suggest that entrainment as such does not necessarily speed up recreational runners or lengthen their steps, as this seems to depend on the presence/absence of instruction. The fact that instructed entrainment did not lead to an increase in speed and step length, whereas spontaneous entrainment did, might be related to the idea that instruction results in more goal-directed behavior, as such directing the focus to the achievement of entrainment and suppressing possible arousal effects caused by the auditory accompaniment. The combination of our results on entrainment as well as cadence, step length, and speed indeed corresponds with cognitive motor learning models, suggesting that explicit instruction in motor control contexts may lead to more intentional behavior and promote greater deliberate control of movement compared with baseline and, in turn, disrupt movement in line with the conscious processing hypothesis.33,53 This hypothesis is applied to healthy populations, such as the recreational runners studied here. How- ever, it might not hold for specific gait-disordered populations, since previous clinical work indicated 98 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Van Dyck et al. Instructed and spontaneous music-entrained running that rhythmic auditory cues can, for instance, help Parkinson’s disease patients to take faster and longer (as well as less variable) strides, even when instructed to entrain.54 Although rPAs decreased with instruction (implying a closer match of footsteps and beat instances), negative angles were exhibited both with and without instruction to entrain to the music. As a negative rPA implies footfalls to occur before the beat, a prediction error minimization process occurred; runners presumably relied on anticipatory mechanisms, which allowed them to predict the beats and coordinate their own antici- pated actions with these predictions.55 This idea is supported by previous research revealing positive correlations between prediction/tracking ratios and the acuity of auditory imagery for timing.56 It has been suggested that the formation of auditory images largely relies on working memory.57–59 Moreover, activation of the corresponding brain areas was observed during auditory imagery.60 Although music is believed to distract from feel- ings of fatigue and discomfort,6,7 self-rated per- ceived physical fatigue did not change over condi- tions. This is possibly the result of the short duration of the running tasks, in combination with the low- to-moderate intensity of the exercise, as such not prompting significant feelings of fatigue or exhaus- tion. Yet, levels of physical enjoyment did improve in the presence of a musical stimulus, which is in accordance with the general idea that music can alter mood states and stimulate motivation.7,8 As corporeal coupling to musical stimuli can support the feeling of agency,61 further igniting motivational components,29 we did expect to obtain increased levels of physical enjoyment in the instructed con- dition compared with the uninstructed one. Yet, no such effects were found, suggesting that instruction as such did not influence runners’ enjoyment of the exercise. Since some previous research provided (direct or indirect) proof to indicate that women are more responsive to musical stimuli than men,8,28,62,63 run- ners’ sex was taken into account in the analysis. In contrast with such evidence, our results did not reveal differences between men and women regard- ing entrainment behavior. However, larger differ- ences in speed and step length between the unin- structed and instructed conditions were exhibited for male runners, possibly indicating that they were more responsive to the instruction. Yet, this is a matter of some speculation and other factors might have been at play as well. On average, women were, for instance, shown to exercise more often to music than men, as well as to prefer other music styles,64 and experience different affects and levels of moti- vation while doing so.8,63 An effect of musical training was retrieved, demonstrating higher levels of tempo entrainment and increased RVLs for musically trained runners compared with their untrained counterparts. As such, musical experience might be suggested to facilitate auditory–motor coupling, which is in con- sonance with previous finger-tapping research indi- cating greater synchronization accuracy for musi- cians than nonmusicians; musicians synchronized more flexibly while tapping, while nonmusicians showed greater temporal rigidity.65 The observed decreased ability of nonmusically trained individ- uals to entrain to musical beats might result from weaker auditory–motor integration.66,49 Findings in cognitive psychology also suggested that success- ful adaptation to stimuli is mediated by the level of regularity in the specific environment (i.e., mak- ing it more predictable) and the opportunity to have obtained sufficient practice in such a setting.67 This would thus imply that individuals who obtained more musical practice would adapt more efficiently to a regular (thus predictable) beat, which was indeed confirmed by our results. In our current study, the type of entrainment (or synchronization) refers to the period match- ing of two (or more) dynamical systems. Although most research on the alignment of running and walking behavior and musical beats focused on period matching, we could also have opted to target phase-locking (footfall instances occurring in phase with the musical beats). However, as research indi- cated that footfall instances of running and walking behavior usually occur before or after the beat of the music (e.g., see Refs. 28, 29, and 37), tempo entrain- ment (or tempo synchronization) was studied here. A within-subjects design was selected to control for a wide range of features previously indicated to possibly impact auditory–motor coupling, such as biomechanical characteristics of the individual subjects,45 preferred running pace,19 age,8 training level,35,34 and music preference.1,64 As a result, the order of the conditions could not be counterbal- anced; however, measures were taken to circumvent 99 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Instructed and spontaneous music-entrained running Van Dyck et al. potential associated effects. To prevent confound- ing effects of exhaustion and fatigue, only runners who reported to be fit to run comfortably for at least 30 min without feeling exhausted were invited. In addition, participants were asked to take suffi- cient rest (and were required to pause for at least 5 min) between running tasks. Moreover, reported fatigue was analyzed, demonstrating no differences between conditions. We also aimed to control for possible effects of familiarity with the musical stim- ulus through the inclusion of a familiarization task before the first running session with musical accom- paniment. It should be stressed that this study focused on recreational runners running at a self-paced tempo. However, as less-trained exercisers were shown to depend to a greater extent on the positive feeling states generated by music, while trained exercisers generally tend to focus on the tasks and specifics of their training,34,35 current findings might not be applicable to more professionally trained run- ners. Moreover, when studying higher levels of run- ning intensity, different results might be obtained. When high workloads are undertaken, the exer- ciser’s attention could be shifted toward the painful or fatiguing effects of the exercise,20–24 which might result in lower levels of entrainment with the musi- cal beats. Overall, this study demonstrates the impact of instruction on running performance. Compared with a similar running task without instruction, results showed higher levels of tempo entrainment, lower speeds, and shorter step lengths of recre- ational runners when instructed to match exer- cise with musical tempo. Our results are especially relevant to recreational runners, as their perfor- mance might be mediated through intentionality. We would, however, expect that instruction might not impact a runner’s entrainment basin. Previ- ously, recreational runners were shown to sponta- neously adapt their running cadence up to 2% of their baseline cadence to tempo changes in music.28 As instruction did not seem to impact running cadence in the current study, we would expect a similar entrainment basin both with and without instructions to adapt to the musical beats. However, this is a matter of some speculation and would ben- efit from further study. Finally, our findings might prove to be interesting to trainers and researchers as well, since the desired exercise output might, at least to a certain extent, depend on what exercis- ers/participants were exactly asked to do. As larger step lengths can negatively impact loading of the lower extremity joints,68–70 instruction might also prove its value in the light of prevention and treat- ment of common running-related injuries. Author contributions E.V.D., J.B., and V.L. participated in research design and conducted experiments. E.V.D. and J.B. per- formed data analysis. E.V.D. wrote the first draft. J.B. and V.L. edited or contributed to the writing of the manuscript. Competing interests The authors declare no competing interests. References 1. Priest, D.L. & C.I. Karageorghis. 2008. A qualitative investi- gation into the characteristics and effects of music accompa- nying exercise. Eur. Phy. Educ. Rev. 14: 347–366. 2. Edworthy, J. & H. Waring. 2006. The effects of music tempo and loudness level on treadmill exercise. Ergonomics 49: 1597–1610. 3. Rendi, M., A. Szabo & T. Szabó. 2008. Performance enhance- ment with music in rowing sprint. Sport Psychol. 22: 175– 182. 4. Atkinson, G., D. Wilson & M. Eubank. 2004. Effects of music on work-rate distribution during a cycling time trial. Int. J. Sports Med. 25: 611–615. 5. Maddigan, M.E., K.M. Sullivan, I. Halperin, et al. 2019. High tempo music prolongs high intensity exercise. PeerJ 8: e6164. 6. Yamashita, S., K. Twai, T. Aktmoto, et al. 2006. Effects of music during exercise on RPE, heart rate and the autonomic nervous system. J. Sports Med. Phys. Fitness 46: 425–430. 7. Shaulov, N. & D. Lufi. 2009. Music and light during indoor cycling. Percept. Mot. Skills 108: 597–607. 8. Priest, D.L., C.I. Karageorghis & N.C.C. Sharp. 2004. The characteristics and effects of motivational music in exercise settings: the possible influence of gender, age, frequency of attendance, and time of attendance. J. Sports Med. Phys. Fit- ness 44: 77–86. 9. Lim, H.B.T., C.I. Karageorghis, L.M. Romer, et al. 2014. Psychophysiological effects of synchronous versus asyn- chronous music during cycling. Med. Sci. Sports Exerc. 46: 407–413. 10. Hsu, D.Y., L. Huang, L.F. Nordgren, et al. 2015. The music of power: perceptual and behavioral consequences of powerful music. Soc. Psychol. Personal Sci. 6: 75–83. 11. Terry, P.C., C.I. Karageorghis, A. Mecozzi Saha, et al. 2012. Effects of synchronous music on treadmill running among elite triathletes. J. Sci. Med. Sport 15: 52–57. 12. Karageorghis, C.I., D. Mouzourides, D.L. Priest, et al. 2009. Psychophysical and ergogenic effects of synchronous music during treadmill walking. J. Sport Exerc. Psychol. 31: 18–36. 100 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Van Dyck et al. Instructed and spontaneous music-entrained running 13. Kenyon, G.P. & M.H. Thaut. 2003. Rhythm-driven optimiza- tion of motor control. Recent Res. Dev. Biomech. 1: 29–47. 14. Smoll, F.L. & R.W. Schultz. 1982. Accuracy of motor behav- ior in response to preferred and nonpreferred tempos. J. Hum. Mov. Stud. 8: 123–138. 15. Rossignol, S. & G. Melvill-Jones. 1976. Audiospinal influ- ences in man studied by the H-reflex and its possible role in rhythmic movement synchronized to sound. Electroen- cephalogr. Clin. Neurophysiol. 41: 83–92. 16. Bood, R.J., M. Nijssen, J. van der Kamp, et al. 2013. The power of auditory–motor synchronization in sports: enhancing running performance by coupling cadence with the right beats. PLoS One 8: e70758. 17. Bacon, C.J., T.R. Myers & C.I. Karageorghis. 2012. Effect of music-movement synchrony on exercise oxygen consump- tion. J. Sports Med. Phys. Fitness 52: 359–365. 18. Simpson, S.D. & C.I. Karageorghis. 2006. The effects of syn- chronous music on 400-m sprint performance. J. Sports Sci. 24: 1095–1102. 19. Karageorghis, C.I. & D.L. Priest. 2012. Music in the exer- cise domain: a review and synthesis (Part II). Int. Rev. Sport Exerc. Psychol. 5: 44–66. 20. Rejeski, W.J. 1985. Perceived exertion: an active or passive process? J. Sport Exerc. Psychol. 7: 371–378. 21. Nethery, V.M. 2002. Competition between internal and external sources of information during exercise: influence on RPE and the impact of the exercise load. J. Sports Med. Phys. Fitness 42: 172–178. 22. Tenenbaum, G. 2005. The study of perceived and sustained effort: concepts, research findings, and future directions. In Handbook of Research on Applied Sport Psychology. D. Hack- fort, J. Duda & R. Lidor, Eds.: 335–349. Morgantown: Fitness Information Technology. 23. Razon, S., I. Basevitch, W. Land, et al. 2009. Perception of exertion and attention allocation as a function of visual and auditory conditions. Psychol. Sport Exerc. 10: 636–643. 24. Hutchinson, J.C. & G. Tenenbaum. 2007. Attention focus during physical effort: the mediating role of task intensity. Psychol. Sport Exerc. 8: 233–245. 25. Styns, F., L. van Noorden, D. Moelants, et al. 2007. Walking on music. Hum. Mov. Sci. 26: 769–785. 26. Mendonça, C., M. Oliveira, L. Fontes, et al. 2014. The effect of instruction to synchronize over step frequency while walking with auditory cues on a treadmill. Hum. Mov. Sci. 33: 33–42. 27. Large, E.W. 2000. On synchronizing movements to music. Hum. Mov. Sci. 19: 527–566. 28. Van Dyck, E., B. Moens, J. Buhmann, et al. 2015. Spon- taneous entrainment of running cadence to music tempo. Sports Med. Open 1: 15. 29. Moumdjian, L., B. Moens, E. Vanzeir, et al. 2019. A model of different cognitive processes during spontaneous and inten- tional coupling to music in multiple sclerosis. Ann. N.Y. Acad. Sci. 1445: 27–38. 30. Lopresti-Goodman, S.M., M.J. Richardson, P.L. Silva, et al. 2008. Period basin of entrainment for unintentional visual coordination. J. Mot. Behav. 40: 3–10. 31. Schmidt, R.C. & M.J. Richardson. 2008. Dynamics of Inter- personal Coordination. Berlin: Springer-Verlag. 32. Strogatz, S.H. 1994. Nonlinear Dynamic and Chaos: with Applications to Physics, Biology, Chemistry, and Engineering. Cambridge: Perseus Books. 33. Leow, L.A., K. Waclawik & J.A. Grahn. 2018. The role of attention and intention in synchronization to music: effects on gait. Exp. Brain Res. 236: 99–115. 34. Brownley, K.A., R.G. McMurray & A.C. Hackney. 1995. Effects of music on physiological and affective responses to graded treadmill exercise in trained and untrained runner. Int. J. Psychophysiol. 19: 193–201. 35. Mohammadzadeh, H., B. Tartibiyan & A. Ahmadi. 2008. The effects of music on the perceived exertion rate and per- formance of trained and untrained individuals during pro- gressive exercise. Facta Univ. Ser. Phys. Educ. Sport 6: 67–74. 36. Faul, F., E. Erdfelder, A.-G. Lang, et al. 2007. G∗ power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav. Res. Methods 39: 175–191. 37. Buhmann, J., F. Desmet, B. Moens, et al. 2016. Spontaneous velocity effect of musical expression on self-paced walking. PLoS One 11: e0154414. 38. Dixon, S. 2007. Evaluation of the audio beat tracking system BeatRoot. J. New Music Res. 36: 39–50. 39. Ajzen, I. & M. Fishbein. 1977. Attitude–behavior relations: a theoretical analysis and review of empirical research. Psy- chol. Bull. 84: 888–918. 40. Borg, G. 1998. Borg’s Perceived Exertion and Pain Scales. Champaign, IL: Human Kinetics. 41. Kendzierski, D. & K.J. DeCarlo. 1991. Physical activity enjoyment scale: two validation studies. J. Sport Exerc. Psy- chol. 13: 50–64. 42. Mullen, S.P., E.A. Olson, S.M. Phillips, et al. 2011. Measur- ing enjoyment of physical activity in older adults: invari- ance of the Physical Activity Enjoyment Scale (PACES) across groups and time. Int. J. Behav. Nutr. Phys. Act. 8: 1–9. 43. Berens, P. 2009. CircStat: a MATLAB toolbox for circular statistics. J. Stat. Softw. 31: 10. 44. Mormann, F., K. Lehnertz, P. David, et al. 2000. Mean phase coherence as a measure for phase synchronization and its application to the EEG of epilepsy patients. Physica D 144: 358–369. 45. Ounpuu, S. 1994. The biomechanics of walking and run- ning. Clin. Sports Med. 13: 843–863. 46. Burger, B., M.R. Thompson, G. Luck, et al. 2014. Hunting for the beat in the body: on period and phase locking in music- induced movement. Front. Hum. Neurosci. 8: 903. 47. Fujii, S. & G. Schlaug. 2013. The Harvard Beat Assessment Test (H-BAT): a battery for assessing beat perception and production and their dissociation. Front. Hum. Neurosci. 7: 771. 48. Launay, J., M. Grube & L. Stewart. 2014. Dysrhythmia: a specific congenital rhythm perception deficit. Front. Psychol. 5: 18. 49. Sowinski, J. & S. Dalla Bella. 2013. Poor synchronization to the beat may result from deficient auditory-motor mapping. Neuropsychologia 51: 1952–1963. 50. Becker, N., S. Brett, C. Chambliss, et al. 1994. Mellow and frenetic antecedent music during athletic performance of 101 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences Instructed and spontaneous music-entrained running Van Dyck et al. children, adults, and seniors. Percept. Mot. Skills 79: 1043– 1046. 51. Karageorghis, C.I., K. Drew & P. Terry. 1996. Effects of pretest stimulative and sedative music on grip strength. Per- cept. Mot. Skills 83: 1347–1352. 52. Nijs, A., M. Roerdink & P.J. Beek. 2020. Cadence modulation in walking and running: pacing steps or strides? Brain Sci. 10: 273. 53. Masters, R.S.W. 1992. Knowledge, knerves, and know-how: the role of explicit versus implicit knowledge in the break- down of a complex motor skill under pressure. Br. J. Psychol. 83: 343–358. 54. Hausdorff, J.M. 2007. Gait dynamics, fractals and falls: find- ing meaning in the stride-to-stride fluctuations of human walking. Hum. Mov. Sci. 26: 555–589. 55. Keller, P.E. 2008. Joint action in music performance. In Enacting Intersubjectivity: A Cognitive and Social Per- spective to the Study of Interactions. F. Morganti, A. Carassa & G. Riva, Eds.: 205–221. Amsterdam: IOS Press. 56. Pecenka, N. & P.E. Keller. 2009. The relationship between auditory imagery and musical synchronization abilities in musicians. In Proceedings of the 7th Triennial Conference of European Society for the Cognitive Sciences of Music. J. Louhivuori, T. Eerola, S. Saarikallio et al., Eds.:409–414. Jyväskylä: University of Jyväskylä. 57. Baddeley, A. & R. Logie. 1992. Auditory imagery and work- ing memory. In Auditory Imagery. D. Reisberg, Ed.: 179–197. Hillsdale: Erlbaum. 58. Smith, J., M. Wilson & D. Reisberg. 1995. The role of subvo- calization in auditory imagery. Neuropsychologia 33: 1433– 1454. 59. Hubbard, T.L. 2010. Auditory imagery: empirical findings. Psychol. Bull. 136: 302–329. 60. Aleman, A. & M. van’t Wout. 2004. Subvocalization in auditory–verbal imagery: just a form of motor imagery? Cogn. Process. 5: 228–231. 61. Fritz, T.H., S. Hardikar, M. Demoucron, et al. 2013. Musical agency reduces perceived exertion during strenuous phys- ical performance. Proc. Natl. Acad. Sci. USA 110: 17784– 17789. 62. Cole, Z. & H. Maeda. 2015. Effects of listening to preferen- tial music on sex differences in endurance running perfor- mance. Percept. Mot. Skills 121: 390–398. 63. Karageorghis, C.I., D.L. Priest, L.S. Williams, et al. 2010. Ergogenic and psychological effects of synchronous music during circuit-type exercise. Psychol. Sport Exerc. 11: 551– 559. 64. Hallett, R. & A. Lamont. 2016. Music use in exercise: a ques- tionnaire study. Media Psychol. 20: 658–684. 65. Scheurich, R., A. Zamm & C. Palmer. 2018. Tapping into rate flexibility: musical training facilitates synchronization around spontaneous production rates. Front. Psychol. 9: 458. 66. Pfordresher, P.Q. & S. Brown. 2007. Poor-pitch singing in the absence of “tone deafness.” Music Percept. 25: 95–115. 67. Klein, G. 2017. Sources of Power: How People Make Decisions. Cambridge, MA: MIT Press. 68. Taunton, J.E., M.B. Ryan, D.B. Clement, et al. 2002. A retro- spective case–control analysis of 2002 running injuries. Br. J. Sports Med. 36: 95–101. 69. Ferber, R., B. Noehren, J. Hamill, et al. 2010. Competitive female runners with a history of iliotibial band syndrome demonstrate atypical hip and knee kinematics. J. Orthop. Sports Phys. Ther. 40: 52–58. 70. Noehren, B., I. Davis & J. Hamill. 2007. ASB clinical biome- chanics award winner 2006 prospective study of the biome- chanical factors associated with iliotibial band syndrome. Clin. Biomech. 22: 951–956. 102 Ann. N.Y. Acad. Sci. 1489 (2021) 91–102 © 2020 The Authors. Annals of the New York Academy of Sciences published by Wiley Periodicals LLC on behalf of New York Academy of Sciences

Instructed versus spontaneous entrainment of running cadence to music tempo.

11-18-2020

Van Dyck, Edith,Buhmann, Jeska,Lorenzoni, Valerio

eng

PMC4363791

Research Article Physical Activity Enhances Metabolic Fitness Independently of Cardiorespiratory Fitness in Marathon Runners M. J. Laye,1,2 M. B. Nielsen,1 L. S. Hansen,1 T. Knudsen,1 and B. K. Pedersen1 1Centre of Inflammation and Metabolism, Rigshospitalet, Department of Biomedical Sciences, University of Copenhagen, 2200 Copenhagen, Denmark 2The Buck Center for Research on Aging, Novato, CA 94945, USA Correspondence should be addressed to M. J. Laye; [email protected] Received 3 February 2015; Accepted 12 February 2015 Academic Editor: Francisco Blanco-Vaca Copyright © 2015 M. J. Laye et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. High levels of cardiovascular fitness (CRF) and physical activity (PA) are associated with decreased mortality and risk to develop metabolic diseases. The independent contributions of CRF and PA to metabolic disease risk factors are unknown. We tested the hypothesis that runners who run consistently >50 km/wk and/or >2 marathons/yr for the last 5 years have superior metabolic fitness compared to matched sedentary subjects (CRF, age, gender, and BMI). Case-control recruitment of 31 pairs of runner-sedentary subjects identified 10 matched pairs with similar VO2max (mL/min/kg) (similar-VO2max). The similar-VO2max group was compared with a group of age, gender, and BMI matched pairs who had the largest difference in VO2max (different-VO2max). Primary outcomes that defined metabolic fitness including insulin response to an oral glucose tolerance test, fasting lipids, and fasting insulin were superior in runners versus sedentary controls despite similar VO2max. Furthermore, performance (velocity at VO2max, running economy), improved exercise metabolism (lactate threshold), and skeletal muscle levels of mitochondrial proteins were superior in runners versus sedentary controls with similar VO2max. In conclusion subjects with a high amount of PA have more positive metabolic health parameters independent of CRF. PA is thus a good marker against metabolic diseases. 1. Introduction High levels of physical activity (PA) and cardiorespiratory fitness (CRF) are independently associated with a low risk for many chronic diseases [1] and mortality [2]. While related, the distinction between these CRF and PA is critical. CRF integrates many different physiological systems into a single measure of function. On the other hand PA is comprised of any body movements, which may not necessarily lead to improvements in CRF. However, in general metareviews and independent studies indicate that the protective effect of CRF is greater when compared to PA, in most [3–5], but not all cases [6]. One factor in the discrepancy between the relative health effects for CRF versus PA is the level of precision in measurements of CRF versus PA. Assessment of CRF in large epidemiological studies is typically assessed by a treadmill or ergometer test [3], while PA is assessed by a questionnaire [7]. Indeed, the correlation between reported PA using the international physical activity questionnaire and CRF as measured by a treadmill test ranged from 0.24 to 0.32 in 3 reviewed studies [8]. Furthermore, the range and types of PA vary dramatically, while CRF consists of a single well characterized number. CRF is a powerful predictor of early mortality inde- pendent of PA levels, BMI, or other risk factors [9, 10]. CRF also varies largely within sedentary populations [11]. For instance in a population of 1707 men aged 20–49 the difference between 20th and 80th percentile is more than 30% (36.8–48.5 mL/kg/min) [11]. Likewise the change in CRF to a standardized exercise training program varies largely from −4.7% to +58.0 [12]. While large cross-sectional studies suggest higher CRF are associated with lower levels in biomarkers for metabolic disease other intervention studies show that improvements in CRF are only weakly correlated with improved biomarkers [13]. VO2max in sedentary population of monozygotic twins is highly heritable (77%) even with correction for various Hindawi Publishing Corporation Disease Markers Volume 2015, Article ID 806418, 11 pages http://dx.doi.org/10.1155/2015/806418 2 Disease Markers anthropometric measures [14]. Similarly, gains in CRF fol- lowing 20 weeks of endurance exercise training (3 d/wk at 55– 75% VO2max) are heritable (47%) and highly variable, with some individuals showing no response in VO2max [15]. The variability of the response to endurance training is in part explained by differences in age, sex, race, and initial VO2max [12, 16, 17], an explanation that is not universally found or contributes in totality to the variation of gains in CRF [18]. On the other hand changes in PA can be accomplished through behavioural strategies that are not subject to the same variability that CRF is. Furthermore levels of PA are associated with changes in disease risk. In general, results from the Aerobics Center Longitudinal Study suggest that within groups of individuals who have similar CRF, overall health (cardiovascular health and cancer) is better in indi- viduals with higher PA [6]. One specific example is that PA, independent of CRF, is atheroprotective by improving lipoprotein subclass distribution, postprandial lipoprotein metabolism, inflammation, and endothelial function [19–23]. Conversely, physical inactivity increases the relative risk for at least 35 pathological and clinical conditions [1]. Remarkably, physical inactivity as defined by sitting time is a risk factor for premature death, a number of chronic diseases and pathologies [24]. For example, 20 days of bed rest [25] or 14 days of reduced step count (and thus increased sitting) [26] reduce CRF 27% and 7%, respectively. However, it remains difficult to isolate the independent roles of PA, inactivity, and CRF in health parameters as most studies are not designed to explicitly control for CRF, while changing just PA. In the present study, we hypothesized that PA would improve metabolic fitness independent of CRF. We there- fore sought to identify two groups of people who were closely matched with regard to age, gender, and CRF, but who differed markedly with regard to their PA level. We recruited an endurance-trained group (consisting of recre- ational marathon runners) and a control group, matched for BMI, age, and gender, and tested various metabolic health parameters (blood lipids, glucose tolerance, and body composition) as our primary outcome. The endurance group fulfilled at least one of two inclusion criteria: (1) they had had an average training volume of at least 50 kilometers per a week for at least the past 5 years or (2) they completed at least 10 marathons within the past 5 years including 2 within the last 14 months. Because of the known variability in CRF it was our aim to compare a subgroup (runners versus controls) with similar-VO2max (similar-VO2max) and a subgroup (runners versus controls) with different-VO2max (different-VO2max). 2. Methods 2.1. Recruitment of Subjects. Subjects were recruited through newspaper adverts and emails listserves in 2008-2009 from the greater Copenhagen metro area. Marathon runners ful- filled at least one of two inclusion criteria: (1) having an average training volume of at least 50 kilometers per week for at least the past 5 years, as determined by questionnaire or (2) having ran at least 10 marathons in the past 5 years including 2 within the last 14 months. After a marathon runner had completed the physiological testing a sedentary subject was recruited to match the BMI, age, and sex of the marathon runner. Sedentary subjects were limited to individuals who did not obtain more than 1 hour of structured exercise per week. Exclusion criteria prior to inclusion in the study for both groups included chronic diseases, pregnancy within the last 3 months, abuse of alcohol, use of cigarettes, or use of performance enhancing drugs. In total, 31 pairs of subjects were recruited. No subjects were excluded after testing. The purpose of the study, possible risks, and discomforts were explained to the subjects before written consent was obtained. The study was approved by the Local Ethical Committee of Copenhagen and Frederiksberg and was in accordance with the Declaration of Helsinki. Subjects underwent two days of testing. On day 1, subjects arrived to the lab in a fasted condition and underwent a physical examination during which blood pressure, resting heart rate (HR), and blood were taken for standard laboratory measurements including blood lipids. Following the examination subjects underwent a muscle biopsy. Briefly, muscle biopsies from vastus lateralis were obtained at rest using the percutaneous Bergstrom needle method with suction under local anaesthesia, using 3–5 mL of 20 mg mL−1 lidocaine (SAD, Denmark Copenhagen). Muscle tissue was immediately frozen in liquid nitrogen and stored at −80∘C until further analysis. Following the muscle biopsy, subjects underwent a dual-energy X-ray absorptiometry (DEXA) measurement and an oral glucose tolerance test, which was administered at ∼10:00 AM. 2.2. Separation of Subjects into “Similar-VO2𝑚𝑎𝑥” and “Differ- ent-VO2𝑚𝑎𝑥”. After all subjects had undergone the standard testing, each runner-sedentary pair of subjects was ranked by their relative difference in VO2max (mL/kg/min). Runners VO2max ranged from −5% to 64% higher than their sedentary pair. We separated the cohort into thirds and focused our analysis on the 10 pairs of marathon runner and sedentary controls with closest VO2max (−5% to 15%, mean 5% higher, and absolute difference 2.2 mL/kg/min) and the 10 pairs with the largest difference in VO2max (27% to 64%, mean of 44% high, and absolute difference 17.3 mL/kg/min). We refer to these two groups as similar-VO2max and different-VO2max, respectively. 2.3. Body Composition. Whole body fat and fat-free tissue mass measurements were performed using a dual-energy X-ray absorptiometry (DXA) scanner (Lunar Prodigy, GE Healthcare, WI, Madison USA, software v. 8.8). 2.4. Oral Glucose Tolerance Test. Subjects underwent a three- hour oral glucose tolerance test (OGTT). Within 2 minutes, the subject ingested a drink containing 75 g glucose (Dextrose Anhydre, Roquette Freres, France) dissolved in 300 mL of water. Venous blood samples were collected from an antecu- bital venous catheter before and 10, 20, 30, 60, 90, 120, 150, and 180 minutes after ingestion of the solution. Blood was drawn into tubes (Vacuette, Serum clot activator and Sodium Fluoride/Potassium Oxalate, Hettich, Labinstruments APS) for determination of glucose and insulin to each time point. Disease Markers 3 Blood samples were analyzed for standard biochemical mea- surements by the biochemical department of Rigshospitalet. 2.5. Exercise Testing. The treadmill (Runrace, Technogym, Italy) test consisted of a lactate threshold and maximal oxygen consumption (VO2max) portion, which took less than 20 minutes. Indirect calorimetry measurements were collected throughout the test (Quark b2, CosMed, Rome, Italy). For marathon runners the test began at 3 km⋅h−1 slower than their current marathon pace. The controls started at 6-7 km⋅h−1. Every third minute we increased the speed 1 km⋅h−1, the third minute of which VO2 was averaged until lactate threshold was reached. After each stage subjects stepped off the treadmill for the 15 s required to obtain a blood sample, after which they began the next stage while the measurement was conducted. Blood samples were collected within 15 sec of the end of the stage by wiping sweat, ethanol cleaning, and drying of a nonlanced finger. Lactate was measured by a handhold lactate device (Lactate Scout, EKF Senslab GmbH, Leipzig, Germany), which requires 0.5 𝜇L whole blood and 10 seconds for a measurement. The workload at which the concen- tration of blood lactate reached 4.0 mmol/mL or began to increase exponentially was selected as lactate threshold. We performed posttest analysis to ensure that the stage at which lactate threshold was reached showed agreement with the ventilatory threshold. Without a rest subjects began the VO2max test at the speed lactate threshold was reached. The speed was increased 1 km⋅h−1 every minute until the participant was unable to keep up with speed and/or increasing the workload no longer increased VO2 (mL/L). All subjects reached exhaustion or plateaued VO2max and had an RER > 1.10. Heart rate was recorded during the whole test. 2.6. Immunoblotting. Immunoblots were completed as pre- viously reported [27]. Briefly, skeletal muscle biopsies were weighed and homogenized using a Tissuelyser (Qiagen) (50 mM Tris⋅HCl, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 5 mM NaP, and 0.2% Ipegal-CA-630) supple- mented with complete protease inhibitor cocktail (Roche) and phosphatase inhibitors (Sigma). Protein concentrations were measured with the Bradford assay [28]. Equal amounts of proteins were subjected to SDS-PAGE using Invitrogen 8% precast gels and an I-blot dry transfer machine according to the manufacturer’s instructions. Each GEL contained either the different-VO2max group or the similar-VO2max group with pairs loaded in adjacent lanes. 30 𝜇g of protein was loaded in each well. Polyvinylidene fluoride membranes were probed with primary antibodies at the following concentrations: MnSOD (#06-984; Upstate) 1 : 2000, GPX1 (number 3206; Cell Signaling) 1 : 5000, HSP72 (SPA-810F; Gentaur) 1 : 1000, COXIV-3E11 (number 4850; Cell Signaling) 1 : 2000, GLUT4 (PA1-1065; Thermo Fischer) 1 : 1000, VEGF (sc-152; Santa Cruz) 1 : 500, and MHCIIa (number 3403; Cell Signaling) Detection of primary antibodies was performed using either a mouse (Pierce) or rabbit (Dako) peroxidase-conjugated IgG, and protein signals were visualized using FEMTO-enhanced chemiluminescence and a Bio-Rad Chemidoc XRS imager. Equal protein loading and transfer was verified by beta- tubulin signal and total lane reactive brown signal, which stains for total protein. Quantification of the immunoblots was done using Image J (National Institutes of Health, Bethesda, MD, http://rsb.info.nih.gov/ij/) and corrected for total signal on each blot to correct across blots. 2.7. Statistics. All statistics were performed in Graph Pad (version 5.00 for Windows, GraphPad Software, San Diego, California, USA, http://www.graphpad.com/). Unless noted in the text, a two-way ANOVA with marathon runner/control and similar-VO2max/different-VO2max as the two factors was performed. If either factor showed significance post hoc analysis was done by Bonferroni correction with significance set at 𝑃 < 0.05. 3. Results 3.1. Matching of Healthy Controls and Marathon Runners. A total of 40 subjects, 20 runners, and 20 sedentary runners were included for analysis to take advantage of similarities or differences in VO2max. The subjects were further divided into similar-VO2max (𝑛 = 10 for runners and sedentary) and different-VO2max (𝑛 = 10 for runners and sedentary). Basic anthropometrics can be found in Table 1, indicating that the experimental design resulted in well matched subjects by age (𝑃 = 0.97 for interaction) and body weight (𝑃 = 0.76 for interaction). The similar-VO2max group runners and sedentary subjects had similar body fat percentages (𝑡-test), while the different-VO2max group runners had significantly lower body fat than the sedentary group. 3.2. Separation of Paired Subjects Based on Their VO2𝑚𝑎𝑥 Similarity. As expected runners as a group had a higher VO2max than healthy controls (Figure 1, 𝑃 < 0.0001). In ranking VO2max difference between one marathon runners and one sedentary subject the 31 differences ranged from −5% to 64%, with one-third of the pairs showing a 10% or less difference in VO2max. In this subgroup, which we call similar-VO2max, runners had on average only a 5.2% higher VO2max (range of −5% to 15% for runners/controls, Figure 1, 𝑃 = 0.22 by 𝑡-test). The similar-VO2max group allowed us to compare a group of individuals with the same CRF, but different levels of PA. At the other extreme in the 10 pairs of subjects with the largest difference in VO2max, which we call different-VO2max, runners had on average a 43% higher VO2max (range of 27%–64% higher, Figure 1, 𝑃 < 0.001). The different-VO2max subgroup allowed us to examine the degree and magnitude to which high CRF and high PA were associated with better metabolic health. 3.3. Runners Metabolic Profile Was Improved Relative to Healthy Controls. To assess the overall metabolic health of the subjects, blood lipids and oral glucose tolerance tests were conducted. Fasting levels of insulin, total cholesterol, LDL, and triglycerides were all significantly lower, and HDL was significantly higher in the runners versus controls (Table 1). Similarly, the insulin response during an OGTT was 4 Disease Markers Table 1 𝑛 Age BMI Fat-free mass (kg) Fat % Fasting glucose (mM) Fasting insulin (pmol) Total cholesterol (mM) HDL (mM) LDL (mM) TAGs (mM) Similar-VO2max Controls 10 (9 M, 1 F) 46.8 ± 2.9 (32–61) 23.6 ± 0.5 (21.0–26.9) 60.5 ± 3.2 19.0 ± 1.9 (10.6–27.6) 5.9 ± 0.1 35 ± 5 5.5 ± 0.3 1.5 ± 0.1 3.5 ± 0.2 1.13 ± 0.17 Runners 10 (9 M, 1 F) 46.2 ± 2.7 (31–60) 23.8 ± 0.6 (21.7–25.7) 58.4 ± 2.1 19.1 ± 2.2# (8.3–26.7) 6.1 ± 0.1 29 ± 5# 4.9 ± 0.3# 1.7 ± 0.2# 2.9 ± 0.3# 0.90 ± 0.15# Different-VO2max Controls 10 (7 M, 3 F) 42.5 ± 3.1 (32–64) 23.4 ± 0.6 (20.8–26.1) 57.7 ± 3 26.7 ± 2.2 (17.0–36.6) 6.2 ± 0.2 38 ± 5 5.7 ± 0.4 1.5 ± 0.1 3.4 ± 0.3 1.28 ± 0.23 Runners 10 (7 M, 3 F) 42.4 ± 3.1 (29–62) 23.0 ± 0.5 (21.4–27.4) 60.3 ± 2.7 14.9 ± 1.9#∧ (6.9–24.0) 5.8 ± 0.2 25 ± 4# 4.6 ± 0.2#∧ 1.9 ± 0.1# 2.5 ± 0.2# 0.76 ± 0.08# All data presented as mean ± SEM, with the range of values within parentheses. # indicates significant effect of marathon running. ∧ indicates post hoc difference in runner versus control within either similar or divergent group. M = males, F = females, BMI = body mass index, HDL = high density lipoproteins, LDL = low density lipoproteins, and TAGs = triglycerides. Disease Markers 5 30 40 50 60 70 Controls Runners All subjects VO 2max (mL/kg/min) ∗ (a) Similar-VO2max Different-VO2max Controls Runners Controls Runners Separated groups 30 40 50 60 70 VO2max (mL/kg/min) ∗∗ (b) Figure 1: VO2max (mL/kg/min) was determined in marathon runners and controls who were individually matched for age, BMI, and gender with one runner. Runners as a group had a significantly higher VO2max than their sedentary paired control (Panel (a)) (𝑛 = 31/group, 𝑃 < 0.0001, 𝑡-test). When we stratified the pairs by difference in VO2max between each runner and their control the 10 pairs with the most similar- VO2max did not significantly differ in VO2max (Panel (b)) (𝑛 = 10, 𝑃 = 0.22, 𝑡-test). In 10 pairs with the most different-VO2max the runners had a significantly higher VO2max (𝑛 = 10, 𝑃 < 0.0001, 𝑡-test). Table 2 VO2max (mL/kg/min) VO2max (mL/kg FFM/min) Predicted VO2max VVO2max (k/h) Lactate threshold (k/h) RE (mL O2/kg/km) Resting HR (bpm) Similar-VO2max Controls 46.3 ± 1.6 58.2 ± 1.6 39.2 ± 2.4 13.2 ± 0.6 10.1 ± 0.5 234.2 ± 6.9 58 ± 2 Runners 48.5 ± 1.2# 62.5 ± 1.4# 51.6 ± 4.4#∧ 15.7 ± 0.5#∧ 13.0 ± 0.6#∧ 205.7 ± 8.8#∧ 50 ± 3# Different-VO2 max Controls 40.2 ± 2.2 54.2 ± 1.8 36.9 ± 3.5 11.8 ± 0.6 8.2 ± 0.5 224.4 ± 3.5 59 ± 3 Runners 57.5 ± 2.2#∧ 69.9 ± 2.2#∧ 57.3 ± 4.0#∧ 17.2 ± 0.9#∧ 14.5 ± 0.5#∧ 210.8 ± 5.1#∧ 47 ± 3#∧ All data presented as mean ± SEM. # indicates significant effect of marathon running. ∧ indicates post hoc difference in runner versus control within either similar or divergent group. VO2max = ventilation of maximum oxygen consumption, FFM = fat-free mass, VVO2max = velocity at VO2max, k/h = kilometers per hour, RE = running economy, HR = heart rate, and bpm = beats per a minute. significantly lower in both the similar-VO2max and different- VO2max (Figures 2(b) and 2(d)), while the glucose response to an OGTT did not differ (Figures 2(a) and 2(c)). Together these data indicate that in subjects matched for CRF high levels of PA are necessary for improved metabolic risk factors. 3.4. Runners Markers of Performance Are Better Than Healthy Controls Despite Similar-VO2𝑚𝑎𝑥. VO2max is not necessarily a predictor of performance so we examined additional markers of exercise performance to see whether runners perform better regardless of their VO2max. The velocity at VO2max and speed at lactate threshold was higher in the runners from both the similar-VO2max and different-VO2max groups (Table 2). What may account for the higher velocity at VO2max, but not absolute VO2max in the runners from the similar-VO2max group was a difference in running economy. The similar- VO2max group, but not the different-VO2max, uses less relative oxygen at a given speed compared to their matched sedentary control group. 3.5. Runners Skeletal Muscle Has Higher Antioxidant and Oxidative Enzyme Content Relative to Healthy Controls. In addition to systemic measures of exercise capacity such as VO2max, velocity and VO2max, and lactate threshold we wanted to examine whether skeletal muscle markers for antioxidant and mitochondrial content were higher regard- less of VO2max in vastus lateralis skeletal muscle biopsies of runners (Figure 3). The skeletal muscle of marathon runners had significantly higher protein content of the antioxidant enzymes glutathione peroxidase 1 (GPX1) and mitochondrial superoxide dismutase (mnSOD), but not heat shock protein 72 (HSP70) (Figure 3). The difference in GPX1 (Figure 3(a)) was higher in runners from both the similar- VO2max and different-VO2max group. Similarly, cytochrome oxidase subunit 4 (COXIV), a mitochondrial enzyme in electron transport chain, was higher in skeletal muscle from the runners in both the similar-VO2max and different-VO2max group (Figure 3(f)). However, runners and sedentary subjects had no differences in protein content glucose transporter 4 6 Disease Markers Similar-VO2max 0 0.0 2.5 5.0 7.5 10.0 30 60 90 120 150 180 Time (min) Glucose (mmol) (a) Similar-VO2max 0 0 100 200 300 400 500 30 60 90 120 150 180 Time (min) Insulin (pmol) ∗ ∗ PA P < 0.0001 (b) Different-VO2max 0 30 60 90 120 150 180 Time (min) 0.0 2.5 5.0 7.5 10.0 Glucose (mmol) Control Runner (c) Different-VO2max 0 100 200 300 400 500 Insulin (pmol) 0 30 60 90 120 150 180 Time (min) Control Runner PA P = 0.0004 (d) Figure 2: Oral glucose tolerance tests were conducted on marathon runners and sedentary paired controls with similar-VO2max ((a, b), 𝑛 = 10) and with different-VO2max ((c, d), 𝑛 = 10). While no differences in plasma glucose were seen between runners and sedentary controls regardless of VO2max stratification, runners regardless of VO2max stratification had a significantly lower insulin response (𝑃 < 0.0005 for both groups, One-Way Repeated Measures ANOVA). (Figure 3(g)), myosin heavy chain isoform IIa (Figure 3(d)), or vascular endothelial growth factor (Figure 3(e)). Thus, high levels of PA, rather than whole body VO2max, are asso- ciated with higher levels of specific markers of antioxidant capacity and mitochondrial content in skeletal muscle. 4. Discussion The present case-control study clearly shows that a high amount of physical activity (PA) is associated with benefits on metabolic health parameters independent of cardiores- piratory fitness (CRF). Conversely a higher level of CRF without high levels of PA was associated with lower levels of metabolic fitness. Our measures of metabolic fitness included improved blood lipid profile, lower insulin response to an OGTT, and increased skeletal muscle mitochondrial markers. Furthermore, runners within the similar-VO2max had supe- rior exercise performance as measured by velocity at VO2max, lactate threshold, and submaximal running economy. It is well accepted that PA and CRF reduce the risk of cardiovascular disease, diabetes, and all-cause mortality [29– 31]. However, the relative contributions of PA and CRF to the reduction in risk are less well known and few studies have been designed to pre hoc separate the effects of PA and CRF from each other. Still analyses of larger cohorts have in cross-sectional and longitudinal studies attempted to separate the effects of CRF and PA on health. For instance, with regard to mortality, the associations of PA and CRF were examined separately and in combination in a cohort of relatively healthy 20–82-year-old men and women [32]. Subjects with high levels CRF had lower mortality, while no association between PA and mortality was found. Conversely, data from another study showed higher levels of both PA and CRF associated with reduced risk factors for cardiovascular disease in a cohort of randomly selected 20–65-year-old Swedish men and women [33]. Furthermore, in attempt to separate the effects of CRF and PA, the author’s analysis showed that subjects with a low CRF, but who were physically Disease Markers 7 0 20 40 60 ∗ ∗ GPX1 AU (mean intensity normalized to 𝛽-tubulin ± SEM) Controls Runners Similar-VO2max Different-VO2max (a) 0 20 40 60 80 ∗ MnSOD AU (mean intensity normalized to 𝛽-tubulin ± SEM) Controls Runners Similar-VO2max Different-VO2max (b) 0 20 40 60 ∗ ∗ COXIV3E11 AU (mean intensity normalized to 𝛽-tubulin ± SEM) Controls Runners Similar-VO2max Different-VO2max (c) 0 50 10 20 30 40 MHCIIa AU (mean intensity normalized to 𝛽-tubulin ± SEM) Controls Runners Similar-VO2max Different-VO2max (d) 0 50 10 20 30 40 VEGF AU (mean intensity normalized to 𝛽-tubulin ± SEM) Controls Runners Similar-VO2max Different-VO2max (e) Similar-VO2max Different-VO2max 0 10 20 30 40 HSP70 AU (mean intensity normalized to 𝛽-tubulin ± SEM) Controls Runners (f) Figure 3: Continued. 8 Disease Markers 0 50 10 20 30 40 Controls GLUT4 Runners AU (mean intensity normalized to 𝛽-tubulin ± SEM) Similar-VO2max Different-VO2max (g) C R C R GPX1 MnSOD HSP70 𝛽-tubulin 𝛽-tubulin 𝛽-tubulin COXIV GLUT4 VEGF MHCIIa VO2max VO2max Similar- Different- (h) Figure 3: Protein levels as measured by western blot of traditional training markers in runners (solid bars) or pair sedentary controls (open bars) in resting vastus lateralis biopsies. Glutathione peroxidase 1 (a) and manganese superoxide dismutase (MnSOD) (b) were significantly higher in runners (𝑛 = 10, 𝑃 < 0.001, 2-way ANOVA). Mitochondrial complex IV (COXIV) (c) was significantly higher in runners (𝑛 = 10, 𝑃 < 0.001, 2-way ANOVA). Post hoc significance as determined by Bonferroni corrected 𝑡-test is indicated by a ∗(𝑃 < 0.05). Myosin heavy chain IIa (d), vascular endothelial growth factor (VEGF) (e), heat shock protein 70 (HSP70) (f), nor glucose transporter 4 (GLUT4) (g) did not differ between groups. Proteins of interest are normalized to 𝛽-tubulin as a loading control that did not differ between groups. Representative blots are shown in (h), with C indicating control and R indicating runner. activity, had a 50% reduction in risk factors. However, the design of the study only groups subjects as only fit/unfit active/nonactive, which is a less robust method to match groups as we employed in the current study. While useful for large cohorts the dichotomization of groups by fit/unfit and active/nonactive may not best represent the complex interaction between CRF and PA. The finding that the highly PA marathon runners had a more beneficial lipid profile than controls, independent of CRF, is in accordance with previous findings in “at risk groups” [34]. In a classic randomized-controlled trial by Kraus et al. [20] they determined the effect of the amount and intensity of training in 111 sedentary, overweight men and women with mild-to-moderate dyslipidemia. The study design randomly assigned subjects to a control group (no exercise training) or to 8 months of physical training entail- ing either high-amount-high-intensity exercise (32 km jog- ging/week at 65–80% of peak oxygen uptake (VO2max)), low- amount-high-intensity exercise (19 km jogging/week at 65– 80% percent of VO2max), or low-amount-moderate-intensity exercise (19 km walking/week at 40–55% of VO2max). Regard- less of which exercise group they were assigned to all subjects showed improved lipoprotein profile (lower triglycerides, lower vLDL, and lower vLDL particle size) relative to the control group. However, which exercise group subjects were assigned to did make a difference in whether they improved CRF. The group that performed low-amount-high-intensity exercise improved CRF more than the low-amount-low- intensity group. Thus, both groups improved lipid profiles to a similar degree but improved CRF to a different level. In healthy subjects a change in PA without a change in CRF can lead to improved metabolic fitness. For instance, subjects who skied an average of 342 min/day on a 32 day cross country skiing trip across Greenland had a decrease in maximal CRF but still had an improvement in circulating lipoproteins [35]. While both of these studies were interventions lasting less than 9 months, we were interested in whether a similar finding would occur in long-term runners versus long-term sedentary subjects. Indeed we found that sedentary and run- ners with similar levels of CRF had different lipoprotein and insulin responses to an OGTT test. One potential explanation for this separation in CRF and metabolic health is the type of exercise the marathon runners do. Many of the subjects perform only low-intensity long duration types of exercise. Disease Markers 9 This is in agreement with Houmard et al. [36] who showed that insulin sensitivity was improved more by an exercise training programme with high volume than high intensity. Insulin sensitivity is highly dependent on skeletal muscle specific adaptations, which may differ in response to PA or exercise training that increases CRF. Some types of PA, such as low intensity or limited to small amount of tissue, may lead to significant local changes important for metabolic health without improved whole body CRF. Adaptation specific to the working muscle is illustrated in a number of studies on one-legged exercise training, which utilize a small fraction of whole body muscle mass result in important local adaptations in trained muscle such as increased capillarization, muscle glycogen content, mitochondrial activity, insulin sensitivity, and transcription of metabolic genes [37–40]. Vollaard et al. [41] find these local changes in muscle metabolism are associated with changes in maximal work capacity more so than VO2max. Furthermore, an increase in VO2max with a PA intervention is not even necessary to improve muscle mitochondrial content. For instance, despite not improving CRF cross country, skiers sill improved skeletal muscle mitochondrial [35]. Similarly, in our study it was the high levels of PA, and not a high CRF, that was associated with higher protein levels of the mitochondrial protein COXIV and antioxidant enzyme mnSOD. The exact physiological, genetic, and environmental contributions that determine individuals VO2max vary tremendously and are not completely known. We were surprised that so many (one-third) of the marathon runners did not have substantially higher VO2max than their sedentary match subject. A number of twin studies (reviewed in [1]) have determined a typical genetic contribution for the baseline physical fitness characteristics and responses to standardized aerobic and strength training programs of approximately 50%. The largest single study examining genetic and exercise training interactions with health outcomes is the HERITAGE Family Study [16], which suggested significant genetic variation in the interaction between VO2max and risk factors such as resting systolic blood pressure, fasting plasma HDL-cholesterol, triglycerides, and insulin. The HERITAGE study showed that between 8% and 13% of exercise intervention participants have worse off risk factors following 20 weeks to 6 months of exer- cise training programs [42]. Some specific genes that may contribute to these differences in response to exercise have been identified already. An ongoing study listed in 2005 that 165 autosomal gene entries were found to alter physical performance and health parameters, some of which are only present in response to PA, such as variants in AKT [43], perilipin [44], and FTO [45]. However, the HERITAGE study did not examine what proportion of the inactive groups developed worse risk factors. Other studies which include an inactive control group show worsening of 12 different risk factors including 3 of the 4 examined above (blood pressure was not measured) following 6 months of inactivity [46]. Our current cross-sectional study design only allows us to make association and future studies that trained subjects over a period of time or followed the same subjects over a number of years could provide information about the relative contribution of environment and genetics. We believe our study has distinctions over previous studies. First is that we selected runners that have been running for at least 5 years in a fairly consistent manner, either completing two marathons per a year or having averaged more than 50 km/week over the past 5 years. This is sub- stantially more exercise for a much longer duration than the large clinical intervention studies such as the HERITAGE and STRRIDE studies [15, 46]. Second, we recruited everyday ath- letes rather than elites, which have been studied at older ages previously. For instance, Trappe et al. [47] examined the per- formance abilities of elite octotarian lifelong athletes versus healthy untrained men. Octotarian athletes had VO2max and max workloads that were 80% and 51% higher than a matched sedentary group. While our subjects spanned a much larger age range the subjects most runners had significantly higher VO2max and max workload (running speed). Excluding the runners in the similar-VO2max group (𝑟2 = 0.04) VO2max and max workload were highly correlated (𝑟2 = 0.48–0.80), a relationship that is also present in octotarian athletes and sedentary controls. In general the runners in the similar- VO2max group had higher max workloads than their VO2max would predict. This discrepancy between max workload and VO2max is likely due to the superior running economy it the similar-VO2max group compared to their sedentary controls. Together the data suggests that multiple types of physiological adaptations may occur to result in an improved work capacity following years of high levels of PA. There are a number of limitations in the present study that are worth discussing. First, although within our similar- VO2max group there was no statistical difference between the runners and the sedentary group there was a 5% rela- tive (2.2 mL/kg/min absolute) higher-VO2max that may have contributed to some of the metabolic benefits seen in the runners. However, runners in both the similar-VO2max and different-VO2max had similar improvements in metabolic fitness, arguing against the relative difference in VO2max being important for metabolic health. Second, a larger subject population number would have allowed for more advanced statistical analysis that could model the variables as contin- uous and more precisely determine the relative contribution of VO2max to different metabolic phenotypes. However, given our difficult inclusion criteria for the runner group and difficulty in finding BMI and age matched sedentary controls, it would still remain difficult to recruit hundreds of subjects. Similarly because the training that the runners did was not controlled there are likely a number of individual differences within the runner group that may contribute positively or negatively to the findings seen. Finally, while we observed improved metabolic health of the runners in our group, the metabolic health of the sedentary group remained within clinically normal ranges. However, even within the normal range of clinical variables, small variations in lipoproteins [48, 49] and blood pressure [50] are associated with varied outcomes and risks for cardiovascular disease and death. In summary the present cross-sectional case-control study suggests that a high amount of PA independent of CRF 10 Disease Markers is associated with positive benefits on a number of metabolic health parameters as well as markers of performance. Conflict of Interests No conflict of interests is reported. Acknowledgments The Centre of Inflammation and Metabolism (CIM) is supported by a grant from the Danish National Research Foundation (no. 02-512-55) and is part of the UNIK Project: Food, Fitness & Pharma for Health and Disease, supported by the Danish Ministry of Science, Technology, and Innovation. M. J. Laye was supported by a postdoctoral grant from the Danish Ministry of Science, Technology, and Innovation and additional support was provided by Augustinus Fonden. The authors wish to thank Miss Hanne Villumsen and Miss Ruth Rovsing for their technical support and assistance. They also thank the subjects for taking part in their study. M. J. Laye and B. K. Pedersen formulated the study design and wrote the paper. M. J. Laye takes responsibility for the integrity of the data. M. B. Nielsen, L. S. Hansen, and T. Knudsen conducted the clinical studies. M. J. Laye planned the analyses and performed laboratory assays and analyzed the data. All authors were involved in editing and approval of the final version of the paper. References [1] F. W. Booth, C. K. Roberts, and M. J. Laye, “Lack of exercise is a major cause of chronic diseases,” Comprehensive Physiology, vol. 2, no. 2, pp. 1143–1211, 2012. [2] R. D. Telford, “Low physical activity and obesity: causes of chronic disease or simply predictors?” Medicine and Science in Sports and Exercise, vol. 39, no. 8, pp. 1233–1240, 2007. [3] M. Fogelholm, “Physical activity, fitness and fatness: relations to mortality, morbidity and disease risk factors. A systematic review,” Obesity Reviews, vol. 11, no. 3, pp. 202–221, 2010. [4] D. C. Lee, X. Sui, and S. N. Blair, “Does physical activity ameliorate the health hazards of obesity?” British Journal of Sports Medicine, vol. 43, no. 1, pp. 49–51, 2009. [5] J. W. R. Twisk, H. C. G. Kemper, and W. Van Mechelen, “The relationship between physical fitness and physical activity dur- ing adolescence and cardiovascular disease risk factors at adult age. The Amsterdam Growth and Health Longitudinal Study,” International Journal of Sports Medicine,, vol. 23, supplement 1, pp. S8–S14, 2002. [6] S. N. Blair, Y. Cheng, and J. Scott Holder, “Is physical activity or physical fitness more important in defining health benefits?” Medicine and Science in Sports and Exercise, vol. 33, no. 6, pp. S379–S399, 2001. [7] P. Ferrari, C. Friedenreich, and C. E. Matthews, “The role of measurement error in estimating levels of physical activity,” American Journal of Epidemiology, vol. 166, no. 7, pp. 832–840, 2007. [8] P. H. Lee, D. J. Macfarlane, T. H. Lam, and S. M. Stewart, “Valid- ity of the international physical activity questionnaire short form (IPAQ-SF): a systematic review,” International Journal of Behavioral Nutrition and Physical Activity, vol. 8, article 115, 2011. [9] J. Myers, M. Prakash, V. Froelicher, D. Do, S. Partington, and J. Edwin Atwood, “Exercise capacity and mortality among men referred for exercise testing,” The New England Journal of Medicine, vol. 346, no. 11, pp. 793–801, 2002. [10] S. N. Blair, H. W. Kohl III, R. S. Paffenbarger Jr., D. G. Clark, K. H. Cooper, and L. W. Gibbons, “Physical fitness and all-cause mortality: a prospective study of healthy men and women,” The Journal of the American Medical Association, vol. 262, no. 17, pp. 2395–2401, 1989. [11] C. Y. Wang, W. L. Haskell, S. W. Farrell et al., “Cardiorespiratory fitness levels among us adults 20-49 years of age: findings from the 1999-2004 national health and nutrition examination survey,” The American Journal of Epidemiology, vol. 171, no. 4, pp. 426–435, 2010. [12] W. M. Kohrt, M. T. Malley, A. R. Coggan et al., “Effects of gender, age, and fitness level on response of VO2max to training in 60– 71 yr olds,” Journal of Applied Physiology, vol. 71, no. 5, pp. 2004– 2011, 1991. [13] J. H. Wilmore, J. S. Green, P. R. Stanforth et al., “Relationship of changes in maximal and submaximal aerobic fitness to changes in cardiovascular disease and non-insulin-dependent diabetes mellitus risk factors with endurance training: the Heritage Family Study,” Metabolism: Clinical and Experimental, vol. 50, no. 11, pp. 1255–1263, 2001. [14] R. Fagard, E. Bielen, and A. Amery, “Heritability of aerobic power and anaerobic energy generation during exercise,” Jour- nal of Applied Physiology, vol. 70, no. 1, pp. 357–362, 1991. [15] C. Bouchard, P. An, T. Rice et al., “Familial aggregation of VO(2max) response to exercise training: results from the HERITAGE family study,” Journal of Applied Physiology, vol. 87, no. 3, pp. 1003–1008, 1999. [16] C. Bouchard and T. Rankinen, “Individual differences in response to regular physical activity,” Medicine & Science in Sports & Exercise, vol. 33, no. 6, supplement, pp. S443–S452, 2001. [17] A. J. Hautala, T. H. M¨akikallio, A. Kiviniemi et al., “Cardio- vascular autonomic function correlates with the response to aerobic training in healthy sedentary subjects,” The American Journal of Physiology—Heart and Circulatory Physiology, vol. 285, no. 4, pp. H1747–H1752, 2003. [18] K. O. Bennett, C. J. Billings, M. R. Molis, and M. R. Leek, “Neu- ral encoding and perception of speech signals in informational masking,” Ear and Hearing, vol. 33, no. 2, pp. 231–238, 2012. [19] J. M. R. Gill and D. Malkova, “Physical activity, fitness and cardiovascular disease risk in adults: interactions with insulin resistance and obesity,” Clinical Science, vol. 110, no. 4, pp. 409– 425, 2006. [20] W. E. Kraus, J. A. Houmard, B. D. Duscha et al., “Effects of the amount and intensity of exercise on plasma lipoproteins,” The New England Journal of Medicine, vol. 347, no. 19, pp. 1483–1492, 2002. [21] J. M. R. Gill and A. E. Hardman, “Exercise and postprandial lipid metabolism: an update on potential mechanisms and interactions with high-carbohydrate diets (review),” Journal of Nutritional Biochemistry, vol. 14, no. 3, pp. 122–132, 2003. [22] A. M. W. Petersen and B. K. Pedersen, “The anti-inflammatory effect of exercise,” Journal of Applied Physiology, vol. 98, no. 4, pp. 1154–1162, 2005. [23] A. Maiorana, G. O’Driscoll, C. Cheetham et al., “The effect of combined aerobic and resistance exercise training on vascular function in type 2 diabetes,” Journal of the American College of Cardiology, vol. 38, no. 3, pp. 860–866, 2001. Disease Markers 11 [24] N. Owen, G. N. Healy, C. E. Matthews, and D. W. Dunstan, “Too much sitting: the population health science of sedentary behavior,” Exercise and Sport Sciences Reviews, vol. 38, no. 3, pp. 105–113, 2010. [25] B. Saltin, G. Blomqvist, J. H. Mitchell, R. L. Johnson Jr., K. Wildenthal, and C. B. Chapman, “Response to exercise after bed rest and after training,” Circulation, vol. 38, no. 5, pp. 71– 78, 1968. [26] R. H. Olsen, R. Krogh-Madsen, C. Thomsen, F. W. Booth, and B. K. Pedersen, “Metabolic responses to reduced daily steps in healthy nonexercising men,” The Journal of the American Medical Association, vol. 299, no. 11, pp. 1261–1263, 2008. [27] C. J. Green, M. Pedersen, B. K. Pedersen, and C. Scheele, “Elevated NF-𝜅B activation is conserved in human myocytes cultured from obese type 2 diabetic patients and attenuated by AMP-activated protein kinase,” Diabetes, vol. 60, no. 11, pp. 2810–2819, 2011. [28] M. M. Bradford, “A rapid and sensitive method for the quanti- tation of microgram quantities of protein utilizing the principle of protein dye binding,” Analytical Biochemistry, vol. 72, no. 1-2, pp. 248–254, 1976. [29] F. J. Apullan, M. G. Bourassa, J.-C. Tardif, A. Fortier, M. Gayda, and A. Nigam, “Usefulness of self-reported leisure-time physical activity to predict long-term survival in patients with coronary heart disease,” The American Journal of Cardiology, vol. 102, no. 4, pp. 375–379, 2008. [30] A. Rosengren and L. Wilhelmsen, “Physical activity protects against coronary death and deaths from all causes in middle- aged men. Evidence from a 20-year follow-up of the primary prevention study in Goteborg,” Annals of Epidemiology, vol. 7, no. 1, pp. 69–75, 1997. [31] K.-T. Khaw, R. Jakes, S. Bingham et al., “Work and leisure time physical activity assessed using a simple, pragmatic, validated questionnaire and incident cardiovascular disease and all- cause mortality in men and women: the European Prospective Investigation into Cancer in Norfolk prospective population study,” International Journal of Epidemiology, vol. 35, no. 4, pp. 1034–1043, 2006. [32] S. D. Stovitz, “Contributions of fitness and physical activity to reducing mortality,” Clinical Journal of Sport Medicine, vol. 22, no. 4, pp. 380–381, 2012. [33] E. Ekblom-Bak, M.-L. Hell´enius, ¨O. Ekblom, L.-M. Engstr¨om, and B. Ekblom, “Independent associations of physical activity and cardiovascular fitness with cardiovascular risk in adults,” European Journal of Cardiovascular Prevention and Rehabilita- tion, vol. 17, no. 2, pp. 175–180, 2010. [34] B. K. Pedersen and B. Saltin, “Evidence for prescribing exercise as therapy in chronic disease,” Scandinavian Journal of Medicine and Science in Sports, vol. 16, no. 1, pp. 3–63, 2006. [35] J. W. Helge, R. Damsgaard, K. Overgaard et al., “Low-intensity training dissociates metabolic from aerobic fitness,” Scandina- vian Journal of Medicine and Science in Sports, vol. 18, no. 1, pp. 86–94, 2008. [36] J. A. Houmard, C. J. Tanner, C. A. Slentz, B. D. Duscha, J. S. McCartney, and W. E. Kraus, “Effect of the volume and intensity of exercise training on insulin sensitivity,” Journal of Applied Physiology, vol. 96, no. 1, pp. 101–106, 2004. [37] B. Kiens, B. Essen-Gustavsson, N. J. Christensen, and B. Saltin, “Skeletal muscle substrate utilization during submaximal exer- cise in man: effect of endurance training,” Journal of Physiology, vol. 469, pp. 459–478, 1993. [38] H. Pilegaard, B. Saltin, and D. P. Neufer, “Exercise induces transient transcriptional activation of the PGC-1alpha gene in human skeletal muscle,” Journal of Physiology, vol. 546, part 3, pp. 851–858, 2003. [39] H. Pilegaard, G. A. Ordway, B. Saltin, and P. D. Neufer, “Tran- scriptional regulation of gene expression in human skeletal muscle during recovery from exercise,” The American Journal of Physiology—Endocrinology and Metabolism, vol. 279, no. 4, pp. E806–E814, 2000. [40] F. Dela, J. J. Larsen, K. J. Mikines, T. Ploug, L. N. Petersen, and H. Galbo, “Insulin-stimulated muscle glucose clearance in patients with NIDDM: effects of one-legged physical training,” Diabetes, vol. 44, no. 9, pp. 1010–1020, 1995. [41] N. B. J. Vollaard, D. Constantin-Teodosiu, K. Fredriksson et al., “Systematic analysis of adaptations in aerobic capacity and submaximal energy metabolism provides a unique insight into determinants of human aerobic performance,” Journal of Applied Physiology, vol. 106, no. 5, pp. 1479–1486, 2009. [42] C. Bouchard, S. N. Blair, T. S. Church et al., “Adverse metabolic response to regular exercise: is it a rare or common occur- rence?” PLoS ONE, vol. 7, no. 5, Article ID e37887, 2012. [43] J. A. Mckenzie, S. Witkowski, A. T. Ludlow, S. M. Roth, and J. M. Hagberg, “AKT1 G205T genotype influences obesity-related metabolic phenotypes and their responses to aerobic exercise training in older Caucasians,” Experimental Physiology, vol. 96, no. 3, pp. 338–347, 2011. [44] N. T. Jenkins, J. A. McKenzie, C. M. Damcott, S. Witkowski, and J. M. Hagberg, “Endurance exercise training effects on body fatness, VO2max HDL-C subfractions, and glucose tolerance are influenced by a PLIN haplotype in older Caucasians,” Journal of Applied Physiology, vol. 108, no. 3, pp. 498–506, 2010. [45] T. Rankinen, T. Rice, M. Teran-Garcia, D. C. Rao, and C. Bouchard, “FTO genotype is associated with exercise training- induced changes in body composition,” Obesity, vol. 18, no. 2, pp. 322–326, 2010. [46] C. A. Slentz, J. A. Houmard, and W. E. Kraus, “Modest exercise prevents the progressive disease associated with physical inac- tivity,” Exercise and Sport Sciences Reviews, vol. 35, no. 1, pp. 18– 23, 2007. [47] S. Trappe, E. Hayes, A. Galpin et al., “New records in aerobic power among octogenarian lifelong endurance athletes,” Jour- nal of Applied Physiology, vol. 114, no. 1, pp. 3–10, 2013. [48] S. Lewington, G. Whitlock, R. Clarke et al., “Blood cholesterol and vascular mortality by age, sex, and blood pressure: a meta- analysis of individual data from 61 prospective studies with 55, 000 vascular deaths,” The Lancet, vol. 370, no. 9602, pp. 1829– 1839, 2007. [49] R. Huxley, S. Lewington, and R. Clarke, “Cholesterol, coronary heart disease and stroke: a review of published evidence from observational studies and randomized controlled trials,” Seminars in vascular medicine, vol. 2, no. 3, pp. 315–323, 2002. [50] S. Lewington, R. Clarke, N. Qizilbash, R. Peto, and R. Collins, “Age-specific relevance of usual blood pressure to vascular mortality: a meta-analysis of individual data for one million adults in 61 prospective studies,” The Lancet, vol. 360, no. 9349, pp. 1903–1913, 2002.

Physical activity enhances metabolic fitness independently of cardiorespiratory fitness in marathon runners.

03-03-2015

Laye, M J,Nielsen, M B,Hansen, L S,Knudsen, T,Pedersen, B K

eng

PMC10099854

146 | Scand J Med Sci Sports. 2023;33:146–159. wileyonlinelibrary.com/journal/sms Received: 22 May 2022 | Revised: 9 August 2022 | Accepted: 28 September 2022 DOI: 10.1111/sms.14251 O R I G I N A L A R T I C L E Aerobic high- intensity intervals are superior to improve V̇O2max compared with sprint intervals in well- trained men Håkon Hov1,2 | Eivind Wang2,3 | Yi Rui Lim4 | Glenn Trane5 | Magnus Hemmingsen4 | Jan Hoff1,6 | Jan Helgerud1,4 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2022 The Authors. Scandinavian Journal of Medicine & Science In Sports published by John Wiley & Sons Ltd. 1Myworkout, Medical Rehabilitation Clinic, Trondheim, Norway 2Faculty of Health Sciences and Social Care, Molde University College, Molde, Norway 3Department of Psychosis and Rehabilitation, Psychiatry Clinic, St. Olavs University Hospital, Trondheim, Norway 4Department of Circulation and Medical Imaging, Faculty of Medicine and Health Sciences, Norwegian University of Science and Technology, Trondheim, Norway 5Physical Education, Sports Science and Outdoor Education, NORD University, Bodø, Norway 6Department of Physical Medicine and Rehabilitation, St. Olav's University Hospital, Trondheim, Norway Correspondence Håkon Hov, Myworkout, Medical Rehabilitation Clinic, Ingvald Ystgaards veg 23, 7047 Trondheim, Norway. Email: hakon.hov@treningsklinikken. no Funding information Norges Forskningsr aring;d Maximal oxygen uptake (V̇O2max) may be the single most important factor for long- distance running performance. Interval training, enabling high intensity, is forwarded as the format that yields the largest increase in V̇O2max. However, it is uncertain if an optimal outcome on V̇O2max, anaerobic capacity, and running performance is provided by training with a high aerobic intensity or high over- all intensity. Thus, we randomized 48 aerobically well- trained men (23 ± 3 years) to three commonly applied interval protocols, one with high aerobic intensity (HIIT) and two with high absolute intensity (sprint interval training; SIT), 3× week for 8 weeks: (1) HIIT: 4 × 4 min at ~95% maximal aerobic speed (MAS) with 3 min active breaks. (2) SIT: 8 × 20 s at ~150% MAS with 10 s passive breaks. (3) SIT: 10 × 30 s at ~175% MAS with 3.5 min active breaks. V̇O2max increased more (p < 0.001) following HIIT, 4 × 4 min (6.5 ± 2.4%, p < 0.001) than SIT, 8 × 20 s (3.3 ± 2.4%, p < 0.001) and SIT, 10 × 30 s (n.s.). This was accompanied by a larger (p < 0.05) increase in stroke volume (O2- pulse) following HIIT, 4 × 4 min (8.1 ± 4.1%, p < 0.001) compared with SIT, 8 × 20 s (3.8 ± 4.2%, p < 0.01) and SIT, 10 × 30 (n.s.). Anaerobic capacity (maximal accumulated oxygen deficit) increased following SIT, 8 × 20 s (p < 0.05), but not after HIIT, 4 × 4 min, nor SIT, 10 × 30 s. Long- distance (3000- m) endurance performance increased (p < 0.05– p < 0.001) in all groups (HIIT, 4 × 4 min: 5.9 ± 3.2%; SIT, 8 × 20 s: 4.1 ± 3.7%; SIT, 10 × 30 s: 2.2 ± 2.2%), with HIIT increasing more than SIT, 10 × 30 s (p < 0.05). Sprint (300- m) performance exhibited within- group increases in SIT, 8 × 20 s (4.4 ± 2.0%) and SIT, 10 × 30 s (3.3 ± 2.8%). In conclusion, HIIT improves V̇O2max more than SIT. Given the importance of V̇O2max for most endurance performance scenarios, HIIT should typically be the chosen interval format. K E Y W O R D S aerobic power, anaerobic capacity, HIIT, MAOD, running economy, running performance, SIT, Tabata  | 147 HOV et al. 1 | INTRODUCTION Maximal oxygen uptake (V̇O2max) may be considered the single most important predictor for long- distance endur- ance performance.1– 3 Furthermore, such events are also influenced by other physiological factors involved in aer- obic energy processes, that is, running economy (CR) and lactate threshold (LT),1 as well as a contribution from an- aerobic metabolism.4 However, the capacity to produce energy derived from anaerobic sources is limited,5 and when whole- body performance persists more than 75 s the majority of energy utilized originates from aerobic sources,6 a proportion which increases to ~90% when the event lasts ~10 min.4,7 Given the great importance for long- distance endur- ance performance, a critical question is which training modality may yield the most potent V̇O2max improve- ments. Of duration, frequency and intensity, the latter is forwarded as particularly important to increase V̇O2max.8,9 To achieve high intensity, training can be organized as intervals separated by recovery periods, in which metab- olites accumulated during the intervals can be removed, or accumulation at least alleviated. Aerobic high- intensity interval training (HIIT), applying intervals of 3– 5  min, is one well- documented format to effectively improve V̇O2max in healthy individuals8,10 and various patient populations.11,12 The rationale for this design is that a high overload on oxygen transporting organs may only be achieved after 1– 2  min because of sluggish oxygen kinetics,13 and that in the other end of the spectrum fa- tiguing processes sets an upper limit to the length of the interval, likely around 8– 12 min.14 Consequently, intervals should be between these limits, and towards the lower end (e.g., 4 min) if repeated several times. The intensity in this format typically elicits 90– 95% of maximal heart rate (HRmax) within 2– 3 min, which corresponds to an inten- sity of ~95% of maximal aerobic speed (MAS) and ~90% V̇O2max.10 However, of notice, HIIT may also be organized as series of shorter intervals (e.g., 30 s) if they are inter- spersed by short breaks (e.g., 15 s) in which V̇O2 do not drop significantly, and thus, enabling a high aerobic inten- sity over the course of several intervals (i.e., accumulated time spent ≥90% of V̇O2max).8,10 Supramaximal sprint interval training (SIT) is another intermittent format that is advocated for effective improve- ments in V̇O2max and endurance performance. SIT is exe- cuted at high absolute intensities, often ≥150% of MAS.15,16 However, since fatigue occurs rapidly, the aerobic intensity is not necessarily accordingly high because of the sluggish V̇O2 kinetics. Again, this feature may be somewhat ma- nipulated by the work/rest ratio of a protocol (i.e., short recovery periods may limit a drop in V̇O2 during breaks and enable a higher aerobic intensity).17 Indeed, SIT with short recovery periods (~10 s) commonly improve V̇O2max in moderately and well- trained individuals,16,18 while studies are conflicting regarding the capability of SIT with long recovery periods (~3 min) to increase V̇O2max in healthy and aerobically well- trained individuals.15,19,20 On the contrary, the very high overall intensity applied in SIT protocols may be important for improving the attributes limiting anaerobic capacity, which in well- trained subjects may be related to intramuscular ion handling.19,21 For HIIT and SIT, there is a trade- off between inten- sity and volume, and they may both be organized with recovery periods ranging from a few seconds (~10  s) to several minutes. It is, by definition, the intensity (i.e., work output) that distinguishes HIIT and SIT from each other. The very high absolute intensity during SIT (≥150% of MAS) necessitates short intervals, and its potential to accumulate a high metabolic volume at ≥90% of V̇O2max compared with HIIT (~95% of MAS) is therefore impeded. Considering that SIT protocols often are conducted until absolute exhaustion at a severe work output, the volume of work conducted during a SIT- session must be limited. It is therefore, in a practical manner, not possible to match commonly applied protocols of HIIT and SIT for total work without drastically reducing the volume of HIIT protocols. Where, in the latter case, HIIT cannot be per- formed as intended. Given the great importance of V̇O2max for long- distance endurance performance, studies investigating which in- terval training format may yield the largest increases in this crucial factor are warranted. High intensity appears to be imperative to achieve an optimal outcome, but direct comparisons between interval protocols with high aerobic or very high overall intensity, like HIIT and SIT, on V̇O2max are scarce. Thus, the aim of this study was to compare the effects of three commonly applied interval formats, one HIIT protocol, one SIT protocol with short recovery pe- riods, and one SIT protocol with long recovery periods, on V̇O2max in aerobically well- trained men. Furthermore, to comprehensively outline how of these protocols affect running performance and its physiological determinants, we also compared the effects on anaerobic capacity, CR, LT, relevant hematological parameters and long- distance and sprint running performance. A high aerobic intensity during exercise, tailored to overload oxygen transporting organs, may be essential for enhancing V̇O2max,10 while a very high absolute intensity may more favor anaerobic ca- pacity improvements.22 Accordingly, we hypothesized that (1) HIIT, carried out as 4 × 4 min at ~95% MAS with 3 min active recovery periods, would improve V̇O2max more than the two SIT protocols, carried out as 8 × 20 s until absolute exhaustion (~150% MAS) with 10 s passive recovery pe- riods, and 10 × 30 s of maximal effort (~175% MAS) with 3.5 min active recovery periods, respectively, (2) Both SIT 148 | HOV et al. protocols would improve anaerobic capacity more than HIIT, (3) HIIT would improve long- distance (3000 m) en- durance performance more than both SIT protocols while sprint (300 m) endurance performance would exhibit the reverse result. 2 | METHODS 2.1 | Subjects Forty- eight healthy non- smoking males volunteered to participate in the present study. Females were not invited to participate to ensure homogeneity of physiological fac- tors and baseline training status. The subjects were aero- bically well- trained and relatively accustomed to treadmill running, but not specialized runners nor engaged in long- distance or sprint running competitions or train- ings. They were randomized into three training groups: HIIT 4 × 4  min, SIT 8 × 20 s, or SIT 10 × 30 s (Figure  1). A V̇O2max ≥ 50 ml kg−1 min−1 and whole- body endurance training at least once per week were set as inclusion cri- teria. Subjects were excluded if they had a compliance to the training interventions of <80%. Subject characteristics are given in Table 1. The study was carried out in accord- ance with the Declaration of Helsinki. Participants were informed with a written consent, and the Institutional Review Board of the Norwegian University of Science and Technology approved the protocol. 2.2 | Study timeline After randomization, and within 2 weeks before the in- tervention period, the subjects met three times in the laboratory where two of them were for metabolic testing and the last one for a blood draw. Additionally, the par- ticipants met once at an indoor track and field arena. All subjects had at least 1 day of rest preceding each of the test days (see below). The tests were repeated in the same order for each individual post intervention. The training interventions all consisted of three sessions per week for 8 weeks. 2.3 | Testing procedures 2.3.1 | Test day 1 (V̇O2max, running economy and lactate threshold) The motorized treadmill (Woodway PPS 55 Sport, Germany) was set at 3° inclination throughout test day 1 and 2. Hence, all measurements of the relation- ship between velocity and pulmonary oxygen uptake (V̇O2) (e.g., CR, LT, MAS) was collected at this incline. Following 10  min of warm- up, the participants pro- ceeded into 5- min stages of running at 1 km h−1 increas- ing velocities to determine LT. At least three stages had to be completed. Heart rate (HR) and V̇O2 was continu- ously measured throughout the test using a HR moni- tor (Polar Electro Oy, Finland) and a Cortex Metamax II (Cortex Biophysik GmbH, Leipzig, Germany), respec- tively. Blood was drawn from the fingertip following warm- up and each stage and analyzed using a Biosen C- line lactate analyzer (EKF- diagnostic GmbH, Germany). LT was defined as the V̇O2, HR, or velocity associated with a blood lactate concentration ([la−]b) 1.5 mM above the lowest measured [la−]b.8 CR was assessed as an aver- age of the V̇O2 measurements the last 30 s at 7 km h−1, and visual inspection to control that a steady state had FIGURE 1 Flow chart of the study.  | 149 HOV et al. been achieved was conducted. The CR stage was imple- mented in the LT protocol, and a [la−]b measurement as- sured that 7 km h−1 was below LT. Subsequent to the CR and LT procedure, subjects walked for about 5 min be- fore performing an incremental V̇O2max- test. Starting at an intensity ≥ LT, the velocity was increased by 1 km h−1 every minute until exhaustion, resulting in the test last- ing 4– 7 min. Verbal encouragement was given during the last minutes of the V̇O2max- test. A capillary blood sample was drawn within 1 min after termination of the test to measure [la−]b. The highest recorded HR was re- garded as HRmax. V̇O2max was calculated as the highest 30- s average V̇O2 and maximal O2 pulse was calculated as V̇O2max/HRmax. The presence of a plateau in V̇O2 de- spite increased workload or ventilation (V̇E), combined with either a [la−]b above 8 mM and/or a respiratory ex- change ratio above 1.10 were used as V̇O2max criteria.23 Additionally, V̇O2max- values from the incremental pro- tocol were confirmed during the second test day.24 If ei- ther 30- s average V̇O2 and/or HR reached higher values during the second test day, these values were used as V̇O2max and/or HRmax. Empirically, V̇O2 does not increase proportional to body mass (Mb) but with an exponent of approximately 0.75.25 Consequently, V̇O2max, CR expressed as V̇O2, and V̇O2 at LT should be scaled with Mb raised to the power of 0.75 (ml kg−0.75 min−1). Both stroke volume and anaerobic capacity (absolute volumes), as well as O2 pulse (volume per time unit divided by frequency), should be scaled with body mass raised to the power of 1.26 2.3.2 | Test day 2 (maximal accumulated oxygen deficit and V̇O2max verification) A linear regression was established between V̇O2 and velocity, using at least three submaximal measurements from test day 1 and a Y- intercept of 5.0 ml kg−1 min−1 (representing standing resting metabolism). MAS was defined as the velocity corresponding to a subjects' V̇O2max, according to his linear regression. Anaerobic capacity was measured as maximal accumulated oxygen deficit (MAOD) based on the simplified procedure nr. 3 in Medbø et al.5 Test day 2 started with a 15- min warm- up at ~70% of HRmax, including 2 × 10 s at 120 ± 10% of MAS, which was the intensity for the upcoming supramaximal bout. The warm- up procedure was followed by 10 min of rest and a [la−]b measurement to ensure low [la−]b prior to the su- pramaximal bout. Subjects received verbal instructions to run until absolute exhaustion, without revealing the target duration of 2– 3 min. If the target duration was missed by ±15 s, the test was repeated on a separate day. Data from the supramaximal bout were used to calculate MAOD and verify V̇O2max from test day 1. Additionally, peak rate of in- crease in V̇O2 was measured as the mean rate (ml kg−1 s−1) during the steepest 60- s period. Total accumulated oxygen cost (in VO2) of the supra- maximal bout was estimated as a theoretical value by ex- trapolating the linear relationship between submaximal V̇O2 and velocity to the supramaximal intensity of the test, giving an estimated oxygen cost per unit of time equiva- lent to 120 ± 10% of V̇O2max. The actual accumulated VO2 during this bout was measured, and MAOD was then cal- culated as: However, since the relationship between V̇O2 and ve- locity might be slightly curvilinear, total accumulated ox- ygen cost was also calculated applying the velocity during the supramaximal bout (−7 km h−1) raised to the power of 1.05, based on Equation 1 in Hill and Vingren:27 (1) Estimatedtotaloxygencost − measuredaccumulatedVO2 (2) O2 cost = O2 costat7kmh−1 + [a(velocity−7kmh−1)1.05] HIIT 4 × 4 min (n = 10) SIT 8 × 20 s (n = 12) SIT 10 × 30 s (n = 9) Pre Pre Pre Age (year) 23 ± 2 23 ± 2 24 ± 4 Height (cm) 178 ± 5 180 ± 5 184 ± 6* Body mass (kg) 75.2 ± 6.5 75.2 ± 11.0 81.0 ± 8.1 V̇O2max (ml kg−1 min−1) 62.1 ± 4.8 64.0 ± 6.1 63.1 ± 5.3 Note: Data are presented as mean ± SD. 4 × 4 min, 4 × 4 min running at ~95% of maximal aerobic speed (MAS) interspersed by 3 min active recovery; 8 × 20 s, 8 × 20 s exhaustive running at ~150% of MAS interspersed by 10 s passive recovery; 10 × 30 s, 10 × 30 maximal running (average of ~175% MAS) interspersed by 3.5 min active recovery; V̇O2, oxygen uptake. *Significantly different from 4 × 4 min at baseline (p ≤ 0.05). TABLE 1 Subjects' descriptive data. 150 | HOV et al. Stored oxygen bound to myoglobin and hemoglobin constitutes about 9% of the MAOD and was not corrected for in the calculation.5 2.3.3 | Test day 3 (long- distance and sprint running performance) Performance tests were conducted on a banked 200- m indoor track and field. 10 min of individual warm- up, in- cluding 2– 4 short sprints, preceded a sprint running test of 300 m. After the sprint test, subjects rested for 30 min before the long- distance running test of 3000 m, of which the last 10 min were dedicated to another warm- up. The sprint test was performed as an interval start with sub- jects in random order while the long- distance test was performed as mass start with up to 10 participants. Time was measured manually using a stopwatch and rounded to the nearest tenth of a second for the sprint test and to the nearest second for the long- distance test. The subjects received verbal encouragement during both tests. 2.3.4 | Test day 4 (hematological parameters) Fasting venous blood samples were drawn from the ante- cubital area. Bicarbonate were analyzed using a Siemens Advia Chemistry XPT (Siemens Healthliners, Germany). Erythrocytes, hemoglobin, mean corpuscular volume, mean corpuscular hemoglobin, and hematocrit were an- alyzed using a Sysmex XN (Sysmex Corporation, Kobe, Japan). 2.4 | Training interventions Subjects were instructed to refrain from other high- intensity endurance training during the study. However, subjects were encouraged to continue as usual with other physical activities (e.g., soccer, handball, hiking). For all interventions, treadmills (Gymleco LTX200, Sweden) were set at ~3° inclination and the warm- up consisted of running at ~70% of HRmax for 10 min. Additionally, for the SIT groups, 2– 3 supramaximal bouts of 10– 15 s near the interval training intensity were included in the warm- up. 2.4.1 | HIIT 4 × 4 min The HIIT group performed 4 intervals of 4 min duration at ~95% of MAS, aiming to elicit 90– 95% of HRmax within 3 min of each interval.8 The intervals were separated by 3 min of active recovery at an intensity corresponding to 70% of HRmax, and finally 3 min of cool- down at the same intensity ended the sessions. Throughout the intervention period, treadmill velocity was regularly adjusted to reach the target HR within 3 min of every interval. Including warm- up and cool- down, the HIIT 4 × 4 min protocol lasted 38 min. 2.4.2 | SIT 8 × 20 s Consisted of ~8 × 20- s intervals at ~150% of MAS sepa- rated by 10  s of passive rest, aiming to exhaust the subject during the eighth or ninth interval. If a ninth in- terval was completed, the velocity was increased in the following training session. Every subject had one- to- one follow- up and received verbal encouragement during all intervals, ensuring that absolute exhaustion was at- tained. Including the warm- up and a 10- min cool- down at an intensity corresponding to 70% of HRmax, the SIT 8 × 20 s protocol lasted ~25 min. Originally, this proto- col was reported to be carried out at ~170% of MAS.16 However, a pilot study in our laboratory revealed that subjects were exhausted before the seventh interval at this intensity and had to jump off the treadmill during the fourth to sixth interval before the allotted time of 20 s had passed. Therefore, an intensity of ~150% of MAS was chosen for the first training session. Thereafter, perfor- mance during the previous training session determined the intensity. 2.4.3 | SIT 10 × 30 s The protocol was carried out in accordance with Skovgaard et al.,15 consisting of 10 × 30 s running inter- vals of maximal effort separated by active rest periods of 3.5 min at <70% of HRmax. The starting workload during the first session was calculated to represent each sub- jects' average workload from their 300- m performance. The intensity within a training session was, when nec- essary to endure 30 s, gradually reduced from interval to interval since the fatiguing intensity of a 30- s maxi- mal sprint cannot be maintained for 10 consecutive bouts. The average interval intensity during a training session was ~175% of MAS. During all intervals, every subject had one- to- one follow- up and received verbal encouragement, ensuring that the intensity was maxi- mal during every single interval. 3 min of cool down, at an intensity corresponding to ≤70% of HRmax, were added at the end of each session, giving a total duration of 49 min.  | 151 HOV et al. 2.5 | Statistical analysis All statistical analyses were conducted using IBM SPSS Statistics 27 software (IBM Corp., USA). Figures were created using GraphPad Prism 9.0 (GraphPad Software, USA). In all cases, p ≤ 0.05 were used as the level of sig- nificance. V̇O2max and MAOD data were tested for nor- mality using QQ- plots and the Shapiro– Wilk test, and the assumptions of normal distribution were met. Two- way ANOVAs were used to investigate differences between groups, and Tukey's WSD post hoc analysis was used when appropriate. Differences within groups were ana- lyzed using paired samples t- tests. Results are presented as mean ± SD in tables and mean ± SE in figures. 3 | RESULTS 3.1 | Withdrawal and compliance to training Of the 48 subjects randomized to the three training groups, nine withdrew before the interventions started (Figure  1). During the training period, two subjects dropped out due to injuries not related to the study, four subjects withdrew without giving any reason, and two subjects dropped out because they were not able to commit to the SIT 10 × 30 s protocol (Figure 1). Of the 24 training sessions planned, the compliance was 23 ± 1 (98 ± 3%) for HIIT 4 × 4 min, 23 ± 1 (95 ± 6%) for SIT 8 × 20 s, and 21 ± 2 (89 ± 7%) for SIT 10 × 30 s, respectively. All participants included in the analysis completed the intervention in accordance with their respective protocol and accomplished at least 20 of the 24 sessions (>83%). In the SIT 8 × 20 s group, subjects on average conducted 7.7 ± 0.4 intervals per training session. Examples of typi- cal HR and V̇O2 responses during the three exercise in- terventions are shown in Figure 2. 3.2 | Maximal oxygen uptake and oxygen pulse HIIT 4 × 4 min and SIT 8 × 20 s exhibited within- group increases (p < 0.01) in V̇O2max and maximal O2 pulse, while SIT 10 × 30 s did not (Figure  3; Table  2). The in- creases in V̇O2max (ml kg−0.75 min−1 and ml kg−1 min−1) and maximal O2 pulse (ml kg−1  beat−1) were larger in HIIT 4 × 4 min compared with both SIT groups (p < 0.05, Figure  3; Table  2). There was no difference between V̇O2max from test day 1 and V̇O2peak from test day 2 in any of the groups. 3.3 | Maximal accumulated oxygen deficit The calculation of MAOD is illustrated in Figure 4. SIT 8 × 20 s exhibited a 11.6 ± 15.6% within- group increase (p < 0.05) from pre- to posttest in MAOD (ml kg−1) while no such increase was observed for HIIT 4 × 4 min or SIT 10 × 30 s (Table  3). This was also apparent as a larger (p < 0.05) increase in MAOD in SIT 8 × 20 s compared with HIIT 4 × 4 min (Table 3). 3.4 | Long- distance and sprint running performance HIIT 4 × 4 min, SIT 8 × 20 s, and SIT 10 × 30 s improved 3000- m time trial by 5.9 ± 3.2%, 4.1 ± 3.7% and 2.2 ± 2.2%, respectively (p < 0.05, Figure 3; Table 2), and the increase following HIIT 4 × 4 min was larger (p < 0.05) than SIT 10 × 30 s. SIT 8 × 20 s and SIT 10 × 30 s exhibited within- group improvements (p < 0.01) in the 300- m time trial by 4.4 ± 2.0% and 3.3 ± 2.8%, respectively, while no such im- provement was seen following HIIT 4 × 4 min (Figure 3; Table  2). No between- groups differences were observed for the performance on 300- m (Figure 3; Table 2). 3.5 | Hematological variables HIIT 4 × 4 min increased (p < 0.01) bicarbonate concen- tration by 6.9 ± 4.0%. The bicarbonate concentration increased more (p < 0.01) following HIIT 4 × 4 min com- pared with SIT 8 × 20 s and SIT 10 × 30 s (Table 4). 3.6 | Noteworthy correlations Post training, V̇O2max, MAS, and velocity at LT were as- sociated with (p < 0.001) long- distance (3000- m) running performance (V̇O2max (ml kg−1 min−1): r = −0.80; V̇O2max (ml kg−0.75  min−1): r  =  −0.74; MAS: r  =  −0.82; velocity at LT: r  =  −0.87). Sprint running performance (300- m) post training was associated with V̇O2max measured as ml kg−1 min−1 (r = −0.43, p < 0.05) and ml kg−0.75 min−1 (r = −0.49, p < 0.01), MAS (r = −0.41, p < 0.05), velocity at LT (r = −0.42, p < 0.05), MAOD measured as ml kg−1 (r  =  −0.53, p < 0.01), and long- distance running perfor- mance (r = 0.56, p < 0.001). The change in 3000- m perfor- mance from pre- to posttest were associated with change in MAS (r = −0.45, p = 0.012) and velocity at LT (r = −0.46, p = 0.009). No other parameters were associated with the changes in long- distance or sprint running performance. 152 | HOV et al. 4 | DISCUSSION V̇O2max is a crucial indicator for endurance performance, and it may be effectively improved through high- intensity interval training. It is, however, elusive which interval training format that yields an optimal outcome. Therefore, we sought to compare the effects of three popular and well- documented protocols, one with high (HIIT) and two with very high (SIT) intensity, on V̇O2max. We also sought to in- vestigate the protocols' effect on anaerobic capacity as well as long- distance and sprint endurance performance. Our main findings were that HIIT 4 × 4 min increased V̇O2max more than the two SIT protocols, while SIT 8 × 20 s also improved V̇O2max more than SIT 10 × 30 s. Furthermore, HIIT 4 × 4 min enhanced long- distance endurance per- formance more than SIT 10 × 30 s, while SIT 8 × 20 s in- creased anaerobic capacity more than HIIT 4 × 4 min. Our findings imply that HIIT should be the interval format of choice if the objective is to improve V̇O2max and aerobic endurance performance. 4.1 | HIIT, SIT, and V̇O2max improvements The present study shows that HIIT 4 × 4 min is more effec- tive than SIT with short (8 × 20 s) and long (10 × 30 s) recov- ery periods for improving V̇O2max, of which the first finding is novel and the second is in line with Laursen et al.28 The improvement in V̇O2max following HIIT 4 × 4 min in the current study was ~0.3% per training session, and this is in accordance with what has previously been documented for aerobically trained men.8 As expected, the improve- ment was somewhat smaller in comparison with what may be expected for less trained individuals.29 Although the increase was lower than HIIT 4 × 4 min, the SIT 8 × 20 s group exhibited an increase in V̇O2max, in accordance with previous studies of comparable subjects.16,18 In accordance with our hypothesis, aerobic intensity (i.e., accumulated time spent ≥90% V̇O2max), and not over- all intensity (% of MAS), seems paramount for enhanc- ing V̇O2max. Indeed, in line with previous research,30– 32 FIGURE 2 Representative examples of the three exercise interventions. Dotted line (- - - ) represents heart rate whereas solid line (– ) represents oxygen uptake. Notice how the heart rate typically relates to oxygen uptake during the three interval formats. Gray area represents ≥90% of maximum. (A) HIIT 4 × 4 min running at ~95% of maximal aerobic speed (MAS) interspersed by 3 min active recovery. (B) SIT 8 × 20 s exhaustive running at ~150% of MAS interspersed by 10 s passive recovery. (C) SIT 10 × 30 s maximal running (average of ~175% MAS) interspersed by 3.5 min active recovery. During these training sessions, accumulated time ≥90% of V̇O2max was 7 min (A), 1.5 min (B), and 0 min (C).  | 153 HOV et al. Figure 2 illustrates this point as HIIT 4 × 4 min results in several minutes at a high aerobic intensity. In comparison, despite a high HR, only about 1– 2 min appears to be per- formed at this aerobic intensity during a SIT 8 × 20 s ses- sion and no time at all during SIT 10 × 30 s. Therefore, even though V̇O2max may be elicited by a SIT 8 × 20 s training ses- sion,17 the oxygen transporting system is not highly taxed for a long period during this intervention. Interestingly, it can also be seen in Figure 2 that the short recovery periods following SIT 8 × 20 s prevented a drop in V̇O2 and HR, and thus appearing, physiologically, as a single interval. It is also noteworthy that during SIT with longer recov- ery periods, such as 10 × 30 s, the decrease in V̇O2 during recovery is so large that the interval length is not sufficient to reach a high aerobic intensity in the following interval, likely because of the relatively slow V̇O2 kinetics (Table 3). In accordance with this notion, no change in V̇O2max was observed following the SIT 10 × 30 s intervention, and it was different from both the other training groups. This finding is in line with most previous studies investigating similar SIT interventions conducted on endurance trained runners (55– 63 ml kg−1  min−1),19,20 but in contrast to Skovgaard et al.15 Despite that 30- s SIT with long recovery periods may effectively increase V̇O2max in unfit popula- tions,33 this protocol appears to be an inadequate stimulus to improve V̇O2max for males with a baseline V̇O2max ex- ceeding 55 ml kg−1 min−1. Exercise at ~95% of MAS can be maintained for sev- eral minutes, while an intensity ≥150% of MAS necessi- tates very short intervals because fatigue occurs rapidly. This limits the capability of SIT protocols to accumulate as large volumes as HIIT protocols are designed to achieve, and any attempt to match for total work between such protocols would be futile. Essentially, one cannot com- bine a very high work output (≥150% of MAS) with a vol- ume associated with less intense exercise (e.g., ≥ 10 min). However, it should be noted that the total work during the SIT protocols (excluding warm- up, breaks, and cool- down) were 29% (8 × 20 s) and 63% (10 × 30 s) of the total work during HIIT 4 × 4 min. The superior improvement in V̇O2max following HIIT 4 × 4 min in the current study was likely due to a greater overload on oxygen transporting organs from air to mi- tochondria. Although no single factor limits V̇O2max, improvements following HIIT 4 × 4 min have previously been largely attributed to increases in heart stroke vol- ume.8,34 Indeed, in support of this, indicating an im- proved heart stroke volume, HIIT 4 × 4 min increased maximal O2 pulse (ml kg−1 beat−1) (8%) and decreased submaximal HR (9%) at 7 km h−1 more than both SIT protocols in the present study. Albeit, an increased arterio- venous oxygen difference cannot be excluded as a contributing component. However, following pre- vious HIIT 4 × 4 min interventions with healthy young FIGURE 3 Percentage change in V̇O2max (A), O2 pulse (B), 3000- meter running performance, (C) and 300- m running performance (D) from pre- to posttest. 4 × 4 min, 4 × 4 min running at ~95% of maximal aerobic speed (MAS) interspersed by 3 min active recovery; 8 × 20 s, 8 × 20 s exhaustive running at ~150% of MAS interspersed by 10 s passive recovery; 10 × 30 s, 10 × 30 s maximal running (average of ~175% MAS) interspersed by 3.5 min active recovery. Data presented as mean and standard error of the mean. Significant different change from pre- to posttest within group (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001), compared to 10 × 30 s (ap ≤ 0.05, aaap ≤ 0.001), compared to 8 × 20 s (bp ≤ 0.05, bbbp ≤ 0.001). 154 | HOV et al. TABLE 2 Data from pre- and posttests of V̇O2max, running economy, lactate threshold and running performance. HIIT 4 × 4 min (n = 10) SIT 8 × 20 s (n = 12) SIT 10 × 30 s (n = 9) Pre Post Pre Post Pre Post V̇O2max L min−1 4.66 ± 0.35 4.92 ± 0.30*** aa 4.76 ± 0.36 4.95 ± 0.49*** a 5.10 ± 0.47 5.13 ± 0.49 ml kg−1 min−1 62.1 ± 4.8 66.1 ± 5.0*** aaa,bbb 64.0 ± 6.1 66.0 ± 4.9*** 63.1 ± 5.3 64.1 ± 5.5 ml kg−0.75 min−1 182.6 ± 12.8 194.0 ± 12.7*** aaa,b 187.4 ± 11.1 193.8 ± 8.9*** 189.0 ± 13.9 191.4 ± 14.5 V̇E (L min−1) 151.7 ± 13.2 156.2 ± 10.8 150.7 ± 14.2 154.4 ± 17.4 153.2 ± 16.7 159.7 ± 15.7 RER 1.14 ± 0.06 1.14 ± 0.04 1.12 ± 0.05 1.14 ± 0.04 1.13 ± 0.04 1.11 ± 0.04 [La−]b (mM) 13.2 ± 1.6 13.1 ± 1.5 13.5 ± 1.8 13.7 ± 3.0 13.7 ± 1.9 13.3 ± 3.1 HRmax (beats min−1) 203 ± 5† 199 ± 6** 195 ± 7 194 ± 7 197 ± 11 195 ± 8 Maximal O2 Pulse ml beat−1 23.0 ± 1.9 24.7 ± 1.6*** a 24.4 ± 2.2 25.6 ± 3.0** 25.9 ± 2.9§ 26.3 ± 2.7 ml kg−1 beat−1 0.307 ± 0.025 0.332 ± 0.026*** a,b 0.328 ± 0.031 0.340 ± 0.023** 0.321 ± 0.030 0.328 ± 0.027 MAS (km h−1) 13.2 ± 1.6 14.7 ± 1.6*** aa,b 13.3 ± 1.4 13.9 ± 1.1** 13.5 ± 1.6 13.8 ± 1.9 Running Economy L min−1 2.63 ± 0.34 2.52 ± 0.28* 2.69 ± 0.41 2.67 ± 0.42 2.87 ± 0.37 2.74 ± 0.43 ml kg−1 min−1 34.9 ± 2.7 33.8 ± 2.0 35.8 ± 1.8 35.4 ± 2.5 35.4 ± 1.5 34.0 ± 2.9 ml kg−0.75 min−1 102.7 ± 8.6 99.3 ± 6.4 105.3 ± 6.4 104.2 ± 7.8 106.1 ± 6.4 101.8 ± 10.0 HR (beats min−1) 149 ± 17 135 ± 15*** aa,bb 140 ± 10 140 ± 12 143 ± 12 144 ± 10 [La−]b (mM) 1.7 ± 0.6 1.4 ± 0.4 2.0 ± 0.7 1.8 ± 0.6 1.8 ± 0.7 1.5 ± 0.6 Lactate Threshold L min−1 3.40 ± 0.31 3.56 ± 0.34 3.60 ± 0.32 3.69 ± 0.39 3.87 ± 0.54§ 3.86 ± 0.42 ml kg−1 min−1 45.4 ± 3.9 47.8 ± 4.7* 48.5 ± 4.5 49.2 ± 4.0 47.7 ± 4.5 48.1 ± 3.0 ml kg−0.75 min−1 133.6 ± 10.7 140.5 ± 13.0* 142.2 ± 8.8 144.4 ± 8.0 143.1 ± 14.1 143.8 ± 8.7 % V̇O2max 73 ± 5 73 ± 5 76 ± 3 75 ± 3 76 ± 5 75 ± 3 HR (beats min−1) 177 ± 8 170 ± 10* aa,b 172 ± 7 171 ± 7 173 ± 11 175 ± 9 % HRmax 87 ± 3 86 ± 4 88 ± 3 88 ± 3 88 ± 2 90 ± 2 vLT (km h−1) 9.4 ± 1.4 10.2 ± 1.3** 9.7 ± 1.1 10.0 ± 0.9 9.6 ± 1.1 10.0 ± 1.0** [La−]b (mM) 3.0 ± 0.5 2.8 ± 0.3a 3.3 ± 0.5 3.5 ± 0.7 2.9 ± 0.4 3.2 ± 0.4 Time Trial 300- m (s) 46 ± 3 45 ± 3 45 ± 2 43 ± 2** 45 ± 5 43 ± 5** 3000- m (s) 709 ± 58 668 ± 62*** a 714 ± 75 684 ± 60** 711 ± 66 695 ± 63* Body mass (kg) 75.2 ± 6.5 74.8 ± 6.1 75.2 ± 11.0 75.7 ± 11.2 81.0 ± 8.1 80.4 ± 8.6 Note: Data are presented as mean ± SD. 4 × 4 min, 4 × 4 min running at ~95% of maximal aerobic speed (MAS) interspersed by 3 min active recovery; 8 × 20 s, 8 × 20 s exhaustive running at ~150% of MAS interspersed by 10 s passive recovery; 10 × 30 s, 10 × 30 s maximal running (average of ~175% MAS) interspersed by 3.5 min active recovery; V̇O2max, maximal oxygen uptake; V̇E, pulmonary ventilation; RER, respiratory exchange ratio; [La−]b, blood lactate concentration; HR, heart rate; O2 pulse, oxygen pulse; MAS, maximal aerobic speed; vLT, velocity at lactate threshold. Significant different change from pre- to posttest; within group (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001), compared to 10 × 30 s (ap ≤ .05, aap ≤ 0.01, aaap ≤ 0.001), compared to 8 × 20 s (bp ≤ 0.05, bbp ≤ 0.01, bbbp ≤ 0.001). Significantly different at baseline compared to; 4 × 4 min (§p ≤ 0.05), 8 × 20 s (†p ≤ 0.05).  | 155 HOV et al. men (V̇O2max ≥ 50 ml kg−1  min−1), the arterio- venous oxygen difference has been documented to remain unchanged.8,34 4.2 | Running economy and lactate threshold Aside V̇O2max, CR and LT are two other important fac- tors determining aerobic endurance performance. In the present study, CR was improved by 4% (L min−1) follow- ing HIIT 4 × 4 min while no change was observed follow- ing the SIT protocols. Lack of adaptations in CR may be explained by the subjects being relatively accustomed to treadmill running at baseline combined with the low vol- ume of training, especially following the shorter SIT pro- tocols. However, contrary to the present study, enhanced CR following SIT with long recovery periods have previ- ously been demonstrated in aerobically trained males.19,20 This discrepancy with previous studies may be attributed to methodological differences, that is, the velocity during the CR- test. LT as a percentage of V̇O2max was not altered in any of the groups, a finding in line with other studies including above averagely trained subjects.8,35 Therefore, the pres- ent investigation is in agreement with existing literature and the suggestion by Sjodin and Svedenhag,36 that im- provements in LT as a percentage of V̇O2max do not occur in already aerobically trained subjects. This implies, be- cause LT as a percentage of V̇O2max remains unaltered, that increased V̇O2 and velocity at LT is expected when V̇O2max increase. 4.3 | HIIT, SIT, and anaerobic capacity In the current study, anaerobic capacity, measured as MAOD, increased more after SIT 8 × 20 s compared with HIIT 4 × 4 min, and no changes were observed following HIIT 4 × 4 min or SIT 10 × 30 s. As HIIT 4 × 4 min is de- signed to enable a high aerobic intensity with minimal anaerobic contribution, it is unsurprising that MAOD re- mained unchanged following training with this interval format. SIT protocols are, in contrast to the HIIT 4 × 4 min format, typically designed to also overload the anaerobic energy system. Thus, the improved MAOD following SIT 8 × 20 s in the present study was expected and in line with previous studies.16,18 However, the finding that SIT 10 × 30 s did not increase MAOD was against our hypothesis. This finding is novel as, to the best of our knowledge, no other studies have investigated how a SIT intervention with maximal effort and long recovery breaks (≥3 min break) affects MAOD. Albeit, it has been reported improved an- aerobic performance following similar protocols.19,37 The explanation for the two SIT protocols’ different effect on MAOD in the current study is likely the different length of the recovery periods separating the supramaximal in- tervals. Although more energy is released from anaerobic sources during SIT 10 × 30 s compared with SIT 8 × 20 s, both in absolute terms and relative to the accumulated time of intervals, the anaerobic capacity was likely more challenged during the latter. Indeed, previous literature has demonstrated that MAOD is regularly reached during SIT 8 × 20 s, but not during SIT 4 × 30 s separated by 2 min recovery.17 Our study suggests that the percentage of MAOD attained during exercise is a better estimate for a protocols' potential to improve MAOD rather than the total quantity of anaerobic energy released. Interestingly, this has striking similarity to the established principle for aerobic training; that aerobic intensity (% of V̇O2max) is more important than volume (accumulated VO2) for im- proving V̇O2max.8,9 4.4 | HIIT, SIT, and long- distance endurance performance All three training groups in the current study improved long- distance endurance performance. Recognizing the greater aerobic energy contribution (90– 95%) to an event lasting 11– 12 min,7 it was not surprising that HIIT 4 × 4 min FIGURE 4 Illustration of the calculation of maximal accumulated oxygen deficit (MAOD) for a subject, with a V̇O2max of 65.5 ml kg−1 min−1. The subject ran at 118% of maximal aerobic speed (16.0 km h−1 at 3° inclination) and with a theoretical O2 cost of 76.1 ml kg−1 min−1. During the time to exhaustion of 157 s, the total accumulated O2 cost (white and gray area combined) was calculated to equal 199.2 ml kg−1. The accumulated VO2 (gray area) during the 157 s was 116.2 ml kg−1, giving a MAOD (white area) of 83.0 ml kg−1. 156 | HOV et al. was superior to SIT 10 × 30 s, and exhibited a clear tendency to be better than SIT 8 × 20 s, for improving the long- distance endurance performance (Figure 3C). Supported by the strong correlation between V̇O2max and 3000- meter running performance in the present study (r  =  −0.74), despite our sample's relatively homogenous V̇O2max, it is likely that the enhanced time trial improvement following HIIT 4 × 4 min was mainly a consequence of the increased V̇O2max. However, this cannot be concluded considering that the change in V̇O2max only exhibited a tendency for a correlation with the change in 3000- m performance (r  =  −0.32, p  =  0.08). For SIT 8 × 20 s, the enhanced performance may be explained by changes in both V̇O2max and MAOD. To our knowledge, the latter finding is the first to show how this SIT format affects long- distance time trial performance. Although a smaller improvement than HIIT 4 × 4 min and SIT 8 × 20 s, the contributing factors to improved long- distance endurance performance following SIT 10 × 30 s are elusive, since neither V̇O2max, CR, LT nor anaerobic capacity improved in this group. However, the SIT 10 × 30 s group did improve the rate of increase in V̇O2 during the supramaximal posttest (Table  3). This adaptation should enable a slightly increased ve- locity during the 3000- m time- trial without attaining a larger oxygen debt or [la−]b and may thus explain the result. It has been shown that CR deteriorates when [la−]b is elevated,38 and a faster rate of increase in V̇O2 TABLE 3 Data from pre- and posttests of maximal accumulated oxygen deficit. HIIT 4 × 4 min (n = 10) SIT 8 × 20 s (n = 12) SIT 10 × 30 s (n = 9) Pre Post Pre Post Pre Post MAOD L 6.10 ± 0.59 6.11 ± 0.65 6.26 ± 1.41 7.02 ± 1.67** c 7.07 ± 1.95 6.70 ± 1.42 ml kg−1 83.6 ± 8.2 81.8 ± 6.0 83.2 ± 11.9 92.7 ± 16.8* c 88.0 ± 18.0 83.1 ± 12.9 L (curvilinear) 6.91 ± 0.68 6.89 ± 0.69 6.83 ± 1.28 7.78 ± 1.56** c 8.07 ± 2.17 7.83 ± 1.46 ml kg−1 (curvilinear) 93.5 ± 8.9 92.1 ± 3.7 91.3 ± 12.4 103.3 ± 17.8* c 98.9 ± 20.3 97.0 ± 17.0 Velocity % MAS 121 ± 10 117 ± 5 120 ± 7 123 ± 8c 120 ± 8 126 ± 9* cc Velocity (km h−1) 16.2 ± 1.3 17.3 ± 1.4*** 16.1 ± 1.3 17.3 ± 1.3** 16.2 ± 1.4 17.4 ± 1.5*** Time (s) 162 ± 37 145 ± 18 153 ± 34 150 ± 21 164 ± 21 142 ± 12* V̇O2 response Rate (ml kg−1 s−1) 0.70 ± 0.07 0.77 ± 0.09* b 0.75 ± 0.10 0.77 ± 0.10 0.69 ± 0.09 0.79 ± 0.12** b Sec 90% V̇O2peak 95 ± 11 92 ± 8 92 ± 12 88 ± 9 96 ± 11 85 ± 10* Sec pre- 90% V̇O2peak 86 ± 8* 89 ± 12 83 ± 13** b Note: Data are presented as mean ± SD. 4 × 4 min, 4 × 4 min running at ~95% of maximal aerobic speed (MAS) interspersed by 3 min active recovery; 8 × 20 s, 8 × 20 s exhaustive running at ~150% of MAS interspersed by 10 s passive recovery; 10 × 30 s, 10 × 30 s maximal running (average of ~175% MAS) interspersed by 3.5 min active recovery; MAOD, maximal accumulated oxygen deficit; curvilinear, calculation based on a curvilinear relationship between velocity and oxygen uptake; V̇O2peak, peak oxygen uptake; MAS, maximal aerobic speed; Rate, maximum rate of increase in V̇O2 during 60 continuous seconds; Sec 90% V̇O2peak, seconds to reach 90% of V̇O2peak. Significant different change from pre- to posttest; within group (*p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001), compared to 8 × 20 s (bp ≤ 0.05), compared to 4 × 4 min (cp ≤ 0.05, ccp ≤ 0.01). TABLE 4 Hematological variables HIIT 4 × 4 min (n = 10) SIT 8 × 20 s (n = 12) SIT 10 × 30 s (n = 9) Pre Post Pre Post Pre Post Erythrocytes (1012 L−1) 5.06 ± 0.21 5.14 ± 0.32 5.09 ± 0.31 5.14 ± 0.29 5.25 ± 0.28 5.28 ± 0.32 Hemoglobin (g dl−1) 15.11 ± 0.29 15.20 ± 0.82 15.28 ± 0.48 15.18 ± 0.44 15.47 ± 0.78 15.48 ± 0.79 Hematocrit (%) 44.1 ± 1.7 44.7 ± 2.8 44.4 ± 1.8 45.1 ± 1.9 45.8 ± 1.8 46.2 ± 2.0 MCV (fL) 87 ± 2 87 ± 2 87 ± 3 88 ± 3 87 ± 3 87 ± 3 MCH (pg) 29.9 ± 1.4 29.6 ± 0.9 30.1 ± 1.1 29.6 ± 1.0* 29.5 ± 1.1 29.4 ± 0.9 Bicarbonate (mM) 26.71 ± 0.95 28.57 ± 1.72** aa,bb 28.00 ± 1.41 28.25 ± 1.16 28.56 ± 1.42§ 27.78 ± 1.92 Note: Data are presented as mean ± SD. 4 × 4 min, 4 × 4 min running at ~95% of maximal aerobic speed (MAS) interspersed by 3 min active recovery; 8 × 20 s, 8 × 20 s exhaustive running at ~150% of MAS interspersed by 10 s passive recovery; 10 × 30 s, 10 × 30 s maximal running (average of ~175% MAS) interspersed by 3.5 min active recovery; MCV, mean corpuscular volume; MCH, mean corpuscular hemoglobin. Significant different change from pre- to posttest; within group (*p ≤ 0.05, **p ≤ 0.01), compared to 10 × 30 s (aap ≤ 0.01), compared to 8 × 20 s (bbp ≤ 0.01). Significantly different from 4 × 4 at baseline §p ≤ 0.05.  | 157 HOV et al. may therefore enhance long- distance endurance perfor- mance. Another possible explanation is the lack of fa- miliarization before the time trials, probably affecting tactical and technical aspects. 4.5 | HIIT, SIT, and 300- m sprint endurance performance No differences between groups were observed for sprint endurance performance. Albeit, both SIT- groups exhib- ited within- group improvements. The improved 300- m running performance following SIT 10 × 30 s is in close agreement to previous research demonstrating anaerobic performance improvements in the range of 5– 7% follow- ing comparable protocols.19,37 Furthermore, we are not aware of previous research investigating the effects of SIT 8 × 20 s or HIIT 4 × 4 min on 300- m running performance. However, mean power during a Wingate test increase fol- lowing SIT 8 × 20 s.39 Therefore, although obvious differ- ences exist between 300- m running and a Wingate test, the enhanced sprint endurance performance after SIT 8 × 20 s may be in line with previous research. 4.6 | HIIT, SIT, and hematological variables Bicarbonate concentration, an indicator of buffer capac- ity, increased following HIIT 4 × 4 min compared with both SIT protocols, suggesting a superior buffer capac- ity following HIIT. However, this finding contrasts with previous work,40 and whether bicarbonate concentration regularly increase with HIIT remains to be elucidated. Since neither hematocrit nor the concentration of eryth- rocytes and hemoglobin did change in any of the groups, the increases in V̇O2max cannot be explained by improved oxygen carrying capacity of the blood, in accordance with previous research on well- trained males.8 4.7 | Perspective The present study may guide the public, coaches, and athletes towards selecting the most suitable interval for- mat for aerobically well- trained men, depending on the purpose of prescribing exercise. If the objective is to im- prove V̇O2max, a pivotal parameter for aerobic endurance performance, HIIT protocols such as 4 × 4 min should be recommended. SIT with short recovery breaks, for exam- ple, 8 × 20 s, may be a supplement for enhancing the an- aerobic fraction of such events. Anaerobic capacity, and likely sprint endurance performance, are better enhanced applying SIT 8 × 20 s. A noteworthy difference between HIIT 4 × 4 min and the SIT formats is that the former is not performed with a maxi- mal effort while the latter are. Individuals reach absolute ex- haustion during the SIT protocols, either at the end of each interval (SIT 10 × 30 s) or at the end of the last interval (SIT 8 × 20 s). Anecdotally, we experienced several non- severe ad- verse effects during the SIT interventions, such as nausea, vomiting, and dizziness. Therefore, it should be questioned if the extremely intense and fatiguing nature of SIT is ap- propriate in many populations, such as elderly and patients. 5 | CONCLUSION In conclusion, HIIT 4 × 4 min is superior for increas- ing V̇O2max compared with SIT protocols, carried out as 8 × 20 s and 10 × 30 s. Despite a lower overall intensity dur- ing HIIT 4 × 4 min than SIT, the aerobic intensity is higher during the former. HIIT should be the recommended in- terval format for aerobic performance. ACKNOWLEDGEMENT We thank the subjects for their time, effort, and coopera- tion during this project. The authors declare no conflicts of interest. The results of this study are presented clearly, honestly, and without fabrication falsification, or inappro- priate data manipulation. The study was funded by The Research Council of Norway. FUNDING INFORMATION The study was funded by The Research Council of Norway. CONFLICT OF INTEREST The authors declare no conflicts of interest. DATA AVAILABILITY STATEMENT The data that support the findings of this study are avail- able from the corresponding author upon reasonable request. INFORMED CONSENT Participants were informed with a written consent. ORCID Håkon Hov  https://orcid.org/0000-0003-2976-2079 REFERENCES 1. Pate RR, Kriska A. Physiological basis of the sex difference in cardiorespiratory endurance. Sports Med. 1984;1(2):87- 98. doi:10.2165/00007256- 198401020- 00001 158 | HOV et al. 2. Wagner PD. New ideas on limitations to V̇O2max. Exerc Sport Sci Rev. 2000;28(1):10- 14. 3. Saltin APO. Maximal oxygen uptake in athletes. J Appl Physiol. 1967;23(3):353- 358. doi:10.1152/jappl.1967.23.3.353 4. Joyner MJ, Coyle EF. Endurance exercise performance: the physiology of champions. J Physiol. 2008;586(1):35- 44. doi:10.1113/jphysiol.2007.143834 5. Medbø JI, Mohn AC, Tabata I, Bahr R, Vaage O, Sejersted OM. Anaerobic capacity determined by maximal accumu- lated O2 deficit. J Appl Physiol. 1988;64(1):50- 60. doi:10.1152/ jappl.1988.64.1.50 6. Li Y, Niessen M, Chen X, Hartmann U. Overestimate of relative aerobic contribution with maximal accumulated oxygen defi- cit: a review. J Sports Med Phys Fitness. 2015;55(5):377- 382. 7. Duffield R, Dawson B, Goodman C. Energy system contri- bution to 1500- and 3000- metre track running. J Sports Sci. 2005;23(10):993- 1002. doi:10.1080/02640410400021963 8. Helgerud J, Høydal K, Wang E, et al. Aerobic high- intensity intervals improve V̇O2max more than moderate train- ing. Med Sci Sports Exerc. 2007;39(4):665- 671. doi:10.1249/ mss.0b013e3180304570 9. Wenger HA, Bell GJ. The interactions of intensity, frequency and duration of exercise training in altering cardiorespiratory fitness. Sports Med. 1986;3(5):346- 356. 10. Buchheit M, Laursen PB. High- intensity interval training, solutions to the programming puzzle: part I: cardiopulmo- nary emphasis. Sports Med. 2013;43(5):313- 338. doi:10.1007/ s40279- 013- 0029- x 11. Brobakken MF, Nygård M, Güzey IC, et al. One- year aerobic in- terval training in outpatients with schizophrenia: a randomized controlled trial. Scand J Med Sci Sports. 2020;30(12):2420- 2436. doi:10.1111/sms.13808 12. Helgerud J, Karlsen T, Kim WY, et al. Interval and strength training in CAD patients. Int J Sports Med. 2011;32(1):54- 59. doi:10.1055/s- 0030- 1267180 13. Burnley M, Doust JH, Ball D, Jones AM. Effects of prior heavy exercise on VO(2) kinetics during heavy exercise are related to changes in muscle activity. J Appl Physiol. 2002;93(1):167- 174. doi:10.1152/japplphysiol.01217.2001 14. Billat VL, Morton RH, Blondel N, et al. Oxygen kinetics and modelling of time to exhaustion whilst running at vari- ous velocities at maximal oxygen uptake. Eur J Appl Physiol. 2000;82(3):178- 187. doi:10.1007/s004210050670 15. Skovgaard C, Almquist NW, Bangsbo J. The effect of repeated periods of speed endurance training on performance, running economy, and muscle adaptations. Scand J Med Sci Sports. 2018;28(2):381- 390. doi:10.1111/sms.12916 16. Tabata I, Nishimura K, Kouzaki M, et al. Effects of moderate- intensity endurance and high- intensity in- termittent training on anaerobic capacity and V̇O2max. Med Sci Sports Exerc. 1996;28(10):1327- 1330. doi:10.1097/00005768- 199610000- 00018 17. Tabata I, Irisawa K, Kouzaki M, Nishimura K, Ogita F, Miyachi M. Metabolic profile of high intensity intermittent exercises. Med Sci Sports Exerc. 1997;29(3):390- 395. 18. Ravier G, Dugue B, Grappe F, Rouillon JD. Impressive an- aerobic adaptations in elite karate athletes due to few in- tensive intermittent sessions added to regular karate training. Scand J Med Sci Sports. 2009;19(5):687- 694. doi:10.1111/j.1600- 0838.2008.00807.x 19. Bangsbo J, Gunnarsson TP, Wendell J, Nybo L, Thomassen M. Reduced volume and increased training intensity elevate muscle Na+- K+ pump alpha2- subunit expression as well as short- and long- term work capacity in humans. J Appl Physiol. 2009;107(6):1771- 1780. doi:10.1152/japplphysiol.00358.2009 20. Iaia FM, Hellsten Y, Nielsen JJ, Fernström M, Sahlin K, Bangsbo J. Four weeks of speed endurance training re- duces energy expenditure during exercise and maintains muscle oxidative capacity despite a reduction in training volume. J Appl Physiol. 2009;106(1):73- 80. doi:10.1152/ japplphysiol.90676.2008 21. Iaia FM, Bangsbo J. Speed endurance training is a powerful stimulus for physiological adaptations and performance im- provements of athletes. Scand J Med Sci Sports. 2010;20(Suppl 2):11- 23. doi:10.1111/j.1600- 0838.2010.01193.x 22. Buchheit M, Laursen PB. High- intensity interval training, solu- tions to the programming puzzle. Part II: anaerobic energy, neuromuscular load and practical applications. Sports Med. 2013;43(10):927- 954. doi:10.1007/s40279- 013- 0066- 5 23. Wang E, Solli GS, Nyberg SK, Hoff J, Helgerud J. Stroke volume does not plateau in female endurance athletes. Int J Sports Med. 2012;33(9):734- 739. doi:10.1055/s- 0031- 1301315 24. Poole DC, Jones AM. Measurement of the maximum ox- ygen uptake V̇o(2max): V̇o(2peak) is no longer accept- able. J Appl Physiol. 2017;122(4):997- 1002. doi:10.1152/ japplphysiol.01063.2016 25. Bergh U, Sjödin B, Forsberg A, Svedenhag J. The relationship between body mass and oxygen uptake during running in hu- mans. Med Sci Sports Exerc. 1991;23(2):205- 211. 26. Astrand P- O, Rodahl K. Textbook of Work Physiology: Physiological Bases of Exercise. 3rd ed. McGraw- Hill; 1986:392- 410. 27. Hill DW, Vingren JL. Maximal accumulated oxygen deficit in running and cycling. Appl Physiol Nutr Metab. 2011;36(6):831- 838. doi:10.1139/h11- 108 28. Laursen PB, Shing CM, Peake JM, Coombes JS, Jenkins DG. Interval training program optimization in highly trained en- durance cyclists. Med Sci Sports Exerc. 2002;34(11):1801- 1807. doi:10.1097/00005768- 200211000- 00017 29. Støren Ø, Helgerud J, Sæbø M, et al. The effect of age on the V̇O2max response to high- intensity interval train- ing. Med Sci Sports Exerc. 2017;49(1):78- 85. doi:10.1249/ mss.0000000000001070 30. Falz R, Fikenzer S, Holzer R, Laufs U, Fikenzer K, Busse M. Acute cardiopulmonary responses to strength training, high- intensity interval training and moderate- intensity contin- uous training. Eur J Appl Physiol. 2019;119(7):1513- 1523. doi:10.1007/s00421- 019- 04138- 1 31. Viana RB, Naves JPA, de Lira CAB, et al. Defining the number of bouts and oxygen uptake during the "Tabata protocol" per- formed at different intensities. Physiol Behav. 2018;189:10- 15. doi:10.1016/j.physbeh.2018.02.045 32. Buchheit M, Abbiss CR, Peiffer JJ, Laursen PB. Performance and physiological responses during a sprint interval training session: relationships with muscle oxygenation and pulmonary oxygen uptake kinetics. Eur J Appl Physiol. 2012;112(2):767- 779. doi:10.1007/s00421- 011- 2021- 1 33. Vollaard NBJ, Metcalfe RS, Williams S. Effect of number of sprints in an SIT session on change in V̇O2max: a meta- analysis. Med Sci Sports Exerc. 2017;49(6):1147- 1156. doi:10.1249/ mss.0000000000001204  | 159 HOV et al. 34. Wang E, Næss MS, Hoff J, et al. Exercise- training- induced changes in metabolic capacity with age: the role of central cardiovascular plasticity. Age (Dordr). 2014;36(2):665- 676. doi:10.1007/s11357- 013- 9596- x 35. Sjodin B, Jacobs I, Svedenhag J. Changes in onset of blood lac- tate accumulation (OBLA) and muscle enzymes after training at OBLA. Eur J Appl Physiol Occup Physiol. 1982;49(1):45- 57. 36. Sjodin B, Svedenhag J. Applied physiology of marathon run- ning. Sports Med. 1985;2(2):83- 99. 37. Iaia FM, Thomassen M, Kolding H, et al. Reduced volume but increased training intensity elevates muscle Na+- K+ pump alpha1- subunit and NHE1 expression as well as short- term work capacity in humans. Am J Physiol Regul Integr Comp Physiol. 2008;294(3):R966- R974. doi:10.1152/ajpregu.00666.2007 38. Hoff J, Støren Ø, Finstad A, Wang E, Helgerud J. Increased blood lactate level deteriorates running economy in world class endurance athletes. J Strength Cond Res. 2016;30(5):1373- 1378. doi:10.1519/jsc.0000000000001349 39. Foster C, Farland CV, Guidotti F, et al. The effects of high in- tensity interval training vs steady state training on aerobic and anaerobic capacity. J Sports Sci Med. 2015;14(4):747- 755. 40. Edge J, Eynon N, McKenna MJ, Goodman CA, Harris RC, Bishop DJ. Altering the rest interval during high- intensity interval training does not affect muscle or performance ad- aptations. Exp Physiol. 2013;98(2):481- 490. doi:10.1113/ expphysiol.2012.067603 How to cite this article: Hov H, Wang E, Lim YR, et al. Aerobic high-intensity intervals are superior to improve V̇O2max compared with sprint intervals in well-trained men. Scand J Med Sci Sports. 2023;33:146-159. doi: 10.1111/sms.14251

Aerobic high-intensity intervals are superior to improve V̇O_2max compared with sprint intervals in well-trained men.

11-18-2022

Hov, Håkon,Wang, Eivind,Lim, Yi Rui,Trane, Glenn,Hemmingsen, Magnus,Hoff, Jan,Helgerud, Jan

eng

PMC3136187

Hindawi Publishing Corporation Journal of Nutrition and Metabolism Volume 2011, Article ID 623182, 11 pages doi:10.1155/2011/623182 Research Article Aerobic Exercise Training Adaptations Are Increased by Postexercise Carbohydrate-Protein Supplementation Lisa Ferguson-Stegall, Erin McCleave, Zhenping Ding, Phillip G. Doerner III, Yang Liu, Bei Wang, Marin Healy, Maximilian Kleinert, Benjamin Dessard, David G. Lassiter, Lynne Kammer, and John L. Ivy Exercise Physiology and Metabolism Laboratory, Department of Kinesiology and Health Education, University of Texas at Austin, Austin, TX 78712, USA Correspondence should be addressed to John L. Ivy, [email protected] Received 5 February 2011; Accepted 19 April 2011 Academic Editor: Marta Van Loan Copyright © 2011 Lisa Ferguson-Stegall et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. Carbohydrate-protein supplementation has been found to increase the rate of training adaptation when provided postresistance exercise. The present study compared the effects of a carbohydrate and protein supplement in the form of chocolate milk (CM), isocaloric carbohydrate (CHO), and placebo on training adaptations occurring over 4.5 weeks of aerobic exercise training. Thirty- two untrained subjects cycled 60 min/d, 5 d/wk for 4.5 wks at 75–80% of maximal oxygen consumption (VO2 max). Supplements were ingested immediately and 1 h after each exercise session. VO2 max and body composition were assessed before the start and end of training. VO2 max improvements were significantly greater in CM than CHO and placebo. Greater improvements in body composition, represented by a calculated lean and fat mass differential for whole body and trunk, were found in the CM group compared to CHO. We conclude supplementing with CM postexercise improves aerobic power and body composition more effectively than CHO alone. 1. Introduction It is well established that aerobic exercise training leads to cardiovascular, skeletal muscle, and metabolic adaptations. Cardiovascular adaptations include increased stroke volume and cardiac output, which contributes greatly to increased maximal oxygen consumption (VO2 max) [1, 2]. Skele- tal muscle adaptations include increases in activators of mitochondrial biogenesis such as peroxisome proliferator- activated receptor γ coactivator-1α (PGC-1α), and increased activity of oxidative enzymes such as citrate synthase and succinate dehydrogenase [3–10]. While many investigations have addressed the effects of endurance exercise training on such adaptations, few have examined the role of postexercise nutritional supplementation in facilitating the adaptive pro- cess. The beneficial effects of postexercise supplementation in the form of carbohydrate (CHO) or carbohydrate-protein (CHO+PRO) supplements following an acute exercise bout have been the focus of many investigations. Several studies performed by our laboratory, and others have demonstrated a greater improvement in acute exercise recovery with CHO+PRO supplementation compared to CHO alone [11– 14] or to placebo [15]. Okazaki and colleagues [16] recently compared the effects of a CHO+PRO supplement to a pla- cebo supplement in older male subjects who cycled for 60 min/d, 3 d/wk for 8 wk at 60–75% VO2 peak. They reported a twofold increase in VO2 max in the CHO+PRO group compared to the placebo group [16]. Thus, nutritional supplementation may increase the magnitude of training adaptations compared to the exercise stimulus alone. How- ever, it was not possible to determine from their results if the increase in VO2 max was due to cellular or systemic adaptations. Moreover, their experimental design did not allow for macronutrient specific comparisons, as they did not include a CHO-only supplement. 2 Journal of Nutrition and Metabolism Recently, chocolate milk (CM) has been investigated as a practical and effective CHO+PRO postexercise recovery supplement after aerobic exercise [17–19]. In addition, sev- eral investigations have reported the efficacy of milk-based supplements in increasing protein synthesis [20] and lean mass accrual [19, 21, 22] in response to resistance exercise. However, the effects of aerobic endurance exercise train- ing and nutritional supplementation on body composition changes have not been investigated. Therefore, the purpose of the present study was to inves- tigate training adaptations that occurred after a 4.5 wk aer- obic endurance exercise (cycling) training program when supplementing after each daily exercise session with a CHO+PRO supplement in the form of CM, CHO or placebo. We aimed to determine if nutritional supplementation resulted in a greater increase in VO2 max and skeletal muscle oxidative enzyme activity. We also sought to determine if supplementation resulted in a greater increase in lean mass and a greater decrease in fat mass. Although the exercise training was expected to induce positive adaptations in VO2 max and muscle oxidative capacity, we hypothesized that postexercise CM supplementation would induce a greater extent of adaptations than would occur with CHO or PLA supplementation. We further hypothesized that the CM group would demonstrate greater lean body mass increases and fat mass decreases compared to the CHO and PLA groups. 2. Materials and Methods 2.1. Subjects. Thirty-two healthy, recreationally active but untrained males and females (16 males and 16 females) between 18 and 35 years old completed the study. Subject characteristics are listed in Table 1. In order to be classified as recreationally active but not endurance trained, subjects could not have exercised regularly more than 3 h/wk over the last 2 years, and had VO2 max values of <40 mL/kg/min for females and <45 mL/kg/min for males. Potential subjects who did not meet these criteria were screened out of the study. A total of 36 subjects were admitted to the study. Subjects were separated into three groups matched for age, gender, and body composition and then were randomized into one of the 3 treatment groups. Four subjects voluntarily withdrew due to illness or work scheduling conflicts. Written informed consent was obtained from all subjects, and the study was approved by The University of Texas at Austin Institutional Review Board. 2.2. Research Design. This study followed a randomized, double-blinded, placebo-controlled design. The protocol for the training period is shown in Table 2. The entire protocol period was 7 wk long and consisted of the following: a baseline testing week, the first and second weeks of training, a midpoint testing week in which subject’s VO2 max was reassessed, followed by 2 days of training, training weeks 3 and 4, and a partial week during which the end testing was performed. Subjects reported to the laboratory before the start of their training period on two occasions, once for a baseline biopsy and dual energy X-ray absorbency (DEXA) scan for body composition determination (described below), and again the following day for determination of lactate threshold (LT), maximal oxygen consumption (VO2 max) and maximal workload (Wmax). This same test battery and schedule was repeated at the end of the training period (Table 2). The LT test was performed first, followed by the VO2 max test after a 5-min cool-down between the two tests. These tests were performed on a VeloTron DynaFit Pro cycle ergometer (RacerMate, Seattle, Wash, USA). LT was deter- mined using 5-min stages beginning at 70 Watts (W) for males and 50 W for females. The Watts were increased by 25 W (males) or 20 W (females) each stage for the first 3- 4 stages, followed by increases of 15 W (males) or 10 W (females) for the last 2-3 stages. A drop of blood was collected onto a lactate test strip after a finger stick during the last minute of each stage, and lactate levels were measured using a Lactate Pro LT-1710 lactate analyzer (Arkray, Inc., Minami- ku, Kyoto, Japan). LT was defined as the breakpoint at which lactate levels begin to rise above baseline levels. After the 5-min cool-down in which the subjects pedaled easily and drank water ad libitum, the VO2 max test began. VO2 max was measured using a True One 2400 system (ParvoMedics, Sandy, UT). Subjects breathed through a Hans Rudolph valve, with expired gases directed to a mixing chamber for analysis of oxygen (O2) and carbon dioxide (CO2). Outputs were directed to a computer for calculation of ventilation, O2 consumption (VO2), CO2 production (VCO2), and respir- atory exchange ratio (RER) every 15 s. The protocol for establishing VO2 max consisted of 2 min stages beginning at 125 W for males or 75 W for females. The workload was increased by 50 W (males) or 30 W (females) every 2 min until 275 W and 200 W, respectively. After this point, the workload increased 25 W (males) or 20 W (females) every minute until the subject could not continue to pedal despite constant verbal encouragement. The criteria used to establish VO2 max was a plateau in VO2 with in- creasing exercise intensity and RER > 1.10. Maximum power output in Watts was calculated from the VO2 max test data using the formula adapted from ˚Astrand and Rodahl [23]: Wmax = (VO2 max mL − 300 mL O2)/12.5W/mL O2. (1) The workload for the desired intensity level of the training rides (75% of VO2 max for the first 3.5 weeks and 80% for week 4) was then set as percentages of the Wmax as follows: W = [(VO2 max mL × %VO2 max desired) −300mL O2] / 12.5 W/mL O2. (2) With the exception of determining Wmax for the purposes of setting ride intensity levels, the baseline and end testing con- sisted of the same tests in the same order. During the training weeks, subjects reported to the train- ing laboratory each morning after fasting overnight. All sub- jects began the rides as a group at the same time each day Journal of Nutrition and Metabolism 3 Table 1: Subject characteristics at baseline. All subjects (32) CM (11) CHO (11) PLA (10) Age (y) 22.0 ± 0.5 22.1 ± 0.7 21.3 ± 0.9 22.6 ± 1.0 Weight (kg) 71.7 ± 2.4 70.9 ± 5.1 71.2 ± 3.1 73.2 ± 4.5 Height (cm) 168.6 ± 1.5 169.1 ± 2.3 168.0 ± 2.7 168.8 ± 3.1 VO2 max (L·min−1) 2.6 ± 0.2 2.7 ± 0.3 2.6 ± 0.2 2.6 ± 0.2 VO2 max (mL·kg·min−1) 35.9 ± 1.9 36.8 ± 1.4 35.7 ± 2.2 35.2 ± 2.1 Values are mean ± SE. No significant differences existed between the groups at baseline. Numbers in parentheses indicates subject numbers. Table 2: Protocol for training period. Mon Tue Wed Thurs Fri Sat Sun Baseline LT and VO2 max testing; biopsy; DEXA scan Week 1 (75% VO2 max) 30 min 40 min 50 min 60 min 60 min Rest Rest Week 2 (75% VO2 max) 60 min 60 min 60 min 60 min 60 min Rest Rest Midpoint VO2 max testing 60 min 60 min Rest Rest Week 3 (75% VO2 max) 60 min 60 min 60 min 60 min 60 min Rest Rest Week 4 (80% VO2 max) 60 min 60 min 60 min 60 min 60 min Rest Rest End LT and VO2 max testing; biopsy; DEXA scan (6:00 AM or 7:30 AM), Monday–Friday. After each session, subjects were provided one dose of supplement immediately postexercise and were required to drink it in the laboratory. Subjects were then provided a second dose in an opaque to- go cup with a lid and straw and instructed to drink it 1 h later. They were also instructed to not ingest anything other than water until 1 h after ingesting the second dose. The daily training rides were performed on Kona Dew bicycles (Kona, Ferndale, Wash, USA) mounted on Compu- Trainer stationary trainers (RacerMate, Seattle, Wash, USA) interfaced with MultiRider III software (RacerMate, Seattle, Wash, USA). Six bicycles and CompuTrainers were interfaced with the system to allow for training groups of 6 subjects at one time. The bikes were set up based on each subject’s physical measurements. The CompuTrainers were calibrated each morning. To minimize thermal stress, air was circulated over the subjects with standing floor fans, and water was provided ad libitum. Investigators encouraged the subjects to drink as needed. The first week of training served to get the subjects accus- tomed to cycling for prolonged periods. The first ride was 30 min in duration, the second was 40 min, the third ride, 50 min, and the fourth, 60 min. With the exception of 3 rides the first week, all rides on Monday–Friday were 1 h in duration throughout the training period. Each training ride began with a 10-min warm up at 60% VO2 max, after which the work rate was increased to elicit ∼75% VO2 max for duration of each training ride. At the midpoint, VO2 max was reassessed, and the workloads were adjusted accordingly to keep the subjects exercising at 75% VO2 max for the third week. For the fourth week, the intensity was increased to 80% VO2 max. A 5-min VO2 measurement was performed at the beginning of each week to verify that the workload corresponded to the calculated intensity (%VO2 max) for each subject. The Table 3: Energy and macronutrient content of the supplements. CM CHO PLA CHO, g/100 mL 11.48 15.15 0 PRO, g/100 mL 3.67 0 0 Fat, g/100 mL 2.05 2.05 0 Kcals/100mL 79.05 79.05 0 Ratio of CHO : PRO 3.12 : 1 — — Per 100 mL, CM, chocolate milk; CHO, carbohydrate + fat; PLA, placebo. Wattage calculated for each subject was set by the investi- gators, and subjects were asked to maintain a cadence of ∼70 rpms in order to maintain the Wattage. Subjects were not allowed to shift gears or vary their cadence during the rides. The duration of the training period (4.5 weeks) was chosen, because 4 weeks or less has been shown to be an adequate amount of time to demonstrate VO2 max and oxi- dative enzyme activity changes [13, 16]. 2.3. Experimental Beverages. After each daily session, sub- jects ingested the experimental beverages (CM, CHO, or PLA) immediately and l h postexercise. The CM (Kirkland Organic Low-Fat Chocolate Milk, Costco Inc.) and CHO beverages were isocaloric and contained the same amount of fat. The placebo was an artificially flavored and artificially sweetened supplement that resembled the CHO beverage in taste and appearance but contained no calories. Grape- flavored Kool-Aid was selected for the CHO and PLA treat- ments, because it best matched the dark coloring of the CM treatment visible only through a semiopaque lid on the drink containers. The energy and macronutrient composition of the beverages is shown in Table 3. 4 Journal of Nutrition and Metabolism The amounts of supplement provided were stratified according to body weight ranges. Subjects weighing less than 63.6 kg (140 lbs) received 250 mL per supplement (197.5 kcals each), totaling 500 mL and 395 kcals. Subjects weighing between 63.6 kg (140 lbs) and 77.2 kg (170 lbs) received 300 mL per supplement (237 kcals), totaling 600 mL and 474 kcals. Subjects weighing between 77.2 kg (170 lbs) and 90.9 kg (200 lbs) received 350 mL per supplement (277 kcals), totaling 700 mL and 554 kcals. Subjects weighing over 90.9 kg (200 lbs) received 375 mL per supplement (296.5 kcals), totaling 750 mL and 593 kcals. For the CHO treatment, the amount of carbohydrate (dextrose) and fat (canola oil) matched that provided in the CM as measured for the individual’s weight range. The CM supplement provided an average of 0.94 g carbohydrate, 0.31 g protein, and 0.17 g fat per kg body weight. The CHO supplement provided an average of 1.25 g carbohydrate and 0.17 g fat per kg body weight. 2.4. Diet and Exercise. Subjects were asked to keep their diets and activity levels consistent for the duration of the study (i.e., no significant changes in caloric intake, dietary habits, or activity levels outside of the study’s training sessions). The subjects were instructed to maintain a dietary and activity log for the 2 days prior to their baseline and end biopsies and testing. The subjects were also asked to replicate their diet and activity on the days the logs are kept such that the diet and activity was the same on the 2 days prior to each biopsy session. The self-reported activity was compared for consistency in duration and intensity. The diet logs were analyzed for macronutrient composition and total caloric intake using Nutritionist V Dietary Analysis Software (First Data Bank, Inc, San Bruno, Calif, USA). All subjects complied with the diet and activity requirements. Subjects were instructed to arrive at the laboratory having fasted overnight for 12 h for every exercise session and laboratory visit except for the LT and VO2 max testing sessions. 2.5. Lactate Threshold and VO2 max. These measures were determined at baseline and at the end of the 4.5 wk training period, as shown in Table 2. The protocol for these tests is detailed above. 2.6. Muscle Biopsy Procedure. Muscle biopsies were taken at baseline and the end of the training period, as shown in Table 2. Prior to each biopsy, the subject’s thigh was cleansed with 10% betadine solution and 1.4 mL of 1% Lidocaine Hydrochloride (Elkins-Sinn, Inc., Cherry Hill, NJ) was in- jected to prepare the leg for the muscle biopsy. Approx- imately ∼45–60 mg wet wt of tissue was taken from the vastus lateralis through a 5–8 mm incision made through the skin and fascia, 6 inches from the midline of the thigh on the lateral side and 2.5 inches above the patella. The tissue samples were trimmed of adipose and connective tissue and immediately frozen in liquid nitrogen at −80◦C for subsequent analysis. 2.7. Muscle Tissue Processing. The muscle samples were weighed and cut in half. One half of the tissue sample was used for the determination of citrate synthase and succinate dehydrogenase activity, and the other half for measurement of total PGC-1α content. For the enzymatic analyses, samples were homogenized in ice-cold buffer containing 20 mM Hepes, 2 mM EGTA, 50 mM sodium fluoride, 100 mM potas- sium chloride, 0.2 mM EDTA, 50 mM glycerophosphate, 1 mM DTT, 0.1 mM PMST, 1 mM benzamidine, and 0.5 mM sodium vanadate (pH 7.4) at a dilution of 1 : 10. Homog- enization was performed on ice using 3×5 s bursts with a Caframo RZRl Stirrer (Caframo Limited, Warton, Ontario, Canada). The homogenate was immediately centrifuged at 14,000 g for 10 min at 4◦C, the supernatant aliquoted to storage tubes for each assay and stored at −80◦C. For deter- mination of total PGC-1α content, the tissue samples were homogenized at a dilution of 1 : 10 in a modified RIPA buffer based on a previously described protocol [24] containing: 50 mM Tris-HCL (pH 7.4); 150 mM NaCl (pH 7.4); 1% each Igepal CA-630 and sodium deoxycholate; 1 mM each EDTA (pH 7.4), Na3VO4 (pH 10), NaF, and phenylmethylsulfonyl fluoride; 1 μg/mL each aprotinin, leupeptin, and pepstatin. Homogenization was performed on ice using 4×5 s bursts with a Caframo RZRl Stirrer (Caframo Limited, Warton, Ontario, Canada). The homogenates were sonicated on ice for 10 s and then centrifuged at 5,000 g for 20 min at 4◦C. The supernatant was aliquoted to storage tubes and stored at −80◦C. Protein concentration was determined from the supernatant using a modified version of the Lowry assay [25] for each sample and was measured before each of the assays were performed. PGC-1α Content. Total PGC-1α content was determined by Western blotting. Total α-tubulin content was also determined as a housekeeping protein. Aliquots of homog- enized muscle sample supernatants and standards were slowly thawed over ice and diluted 1 : 1 with sample buffer containing 1.25 M Tris, pH 6.8, glycerol, 20% SDS, 2- mercaptoethanol, 0.25% bromophenol blue solution, and deionized water. Samples containing 70 μg of total protein were separated on 10% polyacrylamide gels by SDS-PAGE for 75 min at 200 V (Bio-Rad Laboratories, Hercules, Calif, USA) After electrophoresis, the gels were electrotransferred using a semi-dry transfer cell (Bio-Rad Laboratories, Hercules, Calif, USA) using 25 V for 18 min to 0.4 μm polyvinylidene fluoride (PVDF) membranes (Millipore, Bedford, Mass, USA). The membranes were blocked in TTBS (TBS, 50 mM Tris, 150 mM NaCl, containing 0.1% Tween-20), and 10% nonfat dry milk for 2 h at room temperature on a rocking platform at medium speed. The membranes were then washed in 1x TTBS 3 times for 5 min each wash. Using the molecular weight markers visible on the membranes as a guide, the membranes were cut at the 75 kD marker. The upper section was used for detection of PGC-1α, and the lower section was used to detect α-tubulin. Each membrane section was incubated overnight at 4◦C on a rocking platform at low speed with antibodies directed against PGC-1α (no. 515667, EMD Calbiotech/Merck KGaA, Darmstadt, Germany), and α-tubulin (no. 2144, Cell Signaling, Danvers, Mass, USA). Journal of Nutrition and Metabolism 5 The antibodies were diluted 1 : 1000 for PGC-1α, and 1 : 900 for α-tubulin in TTBS containing 2% nonfat dry milk. Fol- lowing the overnight incubation, membranes were washed three times with TTBS for 5 min each wash and incubated for 1.5 h with a secondary antibody (goat antirabbit, HRP- linked IgG, no. 7074, Cell Signaling, Danvers, Mass, USA). Dilutions were 1 : 7500 for PGC-1α and 1 : 1000 for α- tubulin. The immunoblots were visualized by enhanced chemiluminescence (Perkin Elmer, Boston, Mass, USA) using a Bio-Rad ChemiDoc detection system, and the mean density of each band was quantified using Quantity One 1- D Analysis software (Bio-Rad Laboratories, Hercules, Calif, USA). A molecular weight ladder (Precision Plus Protein Standard, Bio-Rad) and a rodent internal control standard prepared from insulin-stimulated mixed skeletal muscle were also included on each gel. All blots were compared with the rodent control standard and the values of each sample were represented as a percent of standard for each blot. 2.8. Oxidative Enzymes. Citrate synthase (CS) activity was determined according to the protocol of Srere [26] on the homogenates after further dilution of 1 : 10 (wt/vol) with 0.1 M Tris-HCI and 0.4% Triton X-100 buffer (pH 8.1). The rate of appearance of DTNBwas determined spectrophoto- metrically over 5 min at 412 nm and 37◦C using a Beckman DU 640 spectrophotometer (Fullerton, Calif, USA), and was proportional to CS activity. Succinate dehydrogenase (SDH) activity was measured according to the method of Lowry and Passonneau [27]. The amount of NADH produced during a 5 min incubation time was read on a Varian Cary Eclipse fluorometer with an excitation wavelength of 340 nm and emission wavelength of 450 nm (Varian, Inc., Palo Alto, Calif, USA) and corresponded to SDH activity in the sample. CS and SDH activities were expressed as μmol/g/min protein. 2.9. Body Composition. DEXA (Medical Systems Prodigy, General Electric, Madison, Wis, USA) was used to determine both whole body and regional (trunk and legs) changes in fat mass, and lean mass, as well as bone mineral density (BMD). A three-compartment model design for assessing body composition was used, dividing the body into bone, fat mass, and fat-free mass. The total region percentage of fat mass and lean mass were used to assess the subjects’ body fat and lean mass levels. The trunk region and legs region were used to assess fat and lean mass changes in the trunk and legs independently. The DEXA machine was calibrated each morning prior to subject measurement. Measurements were performed at baseline and at the end of the training period. The same trained technician performed all of the DEXA scans for the entire study. The body composition differentials (Figures 4(a), 4(b), and 4(c)) were calculated according to the formula  LMkgEnd − LMkgBaseline  −  FMkgEnd − FMkgBaseline  = Differential kg . (3) Using this formula, a gain in lean mass, and a loss of fat mass would result in a higher differential value than a loss in lean mass and gain in fat mass, or no change in lean and fat mass. This differential was calculated for whole body as well as regional (trunk and legs) changes (Figures 4(a), 4(b), and 4(c)). Therefore, the whole body differential was calculated as follows, using the CM treatment group values as an example: 1.408kg −  −1.363kg  = 2.771kg. (4) The regional differentials were calculated by the same for- mula using the values from those specific regions. 2.10. Statistical Analyses. Using data in the literature similar to the type of study we proposed, a power analysis was per- formed using G-Power 3.0.10 software (Buchner, Erdfelder and Faul, Dusseldorf University, Germany) for an effect size of 0.3, P < .05, and desired power value of 0.8, using 3 treatment groups. A total sample population of 24 subjects was calculated for an actual power of 0.86 although we collected data on 32 subjects total. VO2 max, LT, muscle enzyme activity, PGC-1α content and body composition measures (lean mass, fat mass, and weight) taken at baseline and end were analyzed using two-way (treatment x time) analysis of variance (ANOVA) for repeated measures. Differences in the baseline and end measurements for VO2 max, as well as forthe body compo- sition differentials were analyzed using a one-way ANOVA. For all measures, post hoc analysis was performed when significance was found using least significant difference (LSD). Differences were considered significant at P < .05. Effect sizes were calculated for VO2 max changes and body composition differentials using the value of Cohen’s d and the effect-size correlation. Data were expressed as mean ± SE. All statistical analyses were performed using SPSS version 16.0 statistical software (SPSS Inc., Chicago, Ill, USA). 3. Results 3.1. VO2 max and Lactate Threshold. Absolute and relative changes in VO2 max are shown in Figure 1. No significant differences existed between the groups at baseline. All treat- ment groups experienced significant increases in absolute and relative VO2 max over the 4.5 wk training period. The change in both absolute and relative VO2 max was significantly greater in the CM group compared to CHO (P < .05; absolute effect size, 0.86; relative effect size, 0.89) and PLA (P < .05; absolute effect size, 0.89; relative effect size, 0.90). The increases in the CHO and PLA groups were not statistically different from each other (Figure 1). LT increased significantly over time in all 3 treatment groups. However, there were no significant differences among the treatments (Table 4). 3.2. Oxidative Enzymes and PGC-1α. No significant treat- ment or treatment by time effects were found for CS, SDH (Figures 2(a) and 2(b)) or PGC-1α (Figure 3). Significant time effects existed for both enzymes in all treatment groups (P < .05). A similar response was found for PGC-1α (P < .05). 6 Journal of Nutrition and Metabolism 0 0.45 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 ∗ Change in absolute VO2 max (L/min) (a) 0 1 2 3 4 5 6 7 ∗ CM CHO PLA Change in relative VO2 max (mL/kg/min) (b) Figure 1: VO2 max changes after 4.5 wks of aerobic endurance training. (a) Change from baseline in absolute VO2 max. (b) Change from baseline in relative VO2 max. Values are mean ± SE. Significant treatment differences: ∗, CM versus PLA and CHO (P < .05). Table 4: Lactate threshold. Baseline End LT (VO2, L/min) CM 1.61 ± 0.16 1.83 ± 0.16† CHO 1.47 ± 0.10 1.67 ± 0.11† PLA 1.53 ± 0.11 1.70 ± 0.13† Values are mean±SE. Significant differences: †, time only (P < .05). 3.3. Body Composition. Changes in body weight, lean mass and fat mass (assessed for whole body, trunk, and legs) are shown in Table 5. Whole-body lean mass increased in all treatment groups, with no treatment differences detected (P < .05). Although whole-body fat mass decreased in all groups, the change was not significant for treatment or time. In the trunk region, a significantly greater gain in lean mass was found in the CM group compared to PLA (P < .05). Trunk region fat mass differences were not significantly different between treatments although a significant time 0 5 10 15 20 25 30 35 CS activity (mmol/g/min) Baseline End Time of assessment † † † (a) 0 1 2 3 4 5 6 7 SDH activity (mmol/g/min) Baseline End Time of assessment † † † CM CHO PLA (b) Figure 2: Oxidative enzyme activity. (a) Citrate synthase activity. (b) Succinate dehydrogenase activity. Biopsies were taken at baseline (before starting the training period) and at the end of the 4.5 wk training period. No significant treatment differences were found. †, significant time effect (P < .05). effect was found for all groups (P < .05). In the legs region, significant time effects were found for lean mass increases and fat mass decreases in all groups (P < .05). The whole body and regional differentials are shown in Figures 4(a), 4(b), and 4(c). The whole body differential and the trunk differential were significantly greater in the CM group compared to CHO (P < .05; effect size for trunk, 0.81; effect size for whole body, 0.82). Whole body and trunk differentials for PLA were not significantly different than those for CM or CHO. The differential for the legs region was not significantly different among the three treatments. No significant treatment or time differences existed for BMD (Table 5). 4. Discussion The most significant finding of the present study was that the increase in VO2 max was significantly greater in the CM group than the CHO or PLA groups. The average increase in absolute VO2 max for the CM group was 12.5% higher Journal of Nutrition and Metabolism 7 0 50 100 150 200 250 Base End Time point Std PGC-1α α-tubulin † † † CM CHO PLA PGC-1α content, percent of standard Figure 3: PGC-1α content before and after 4.5 wks of cycling exercise training. No significant treatment differences were found. †, significant time effect (P < .05). than baseline levels, a twofold improvement over the increase found in the CHO and PLA groups. The average absolute VO2 max (L/min) increase for all subjects and treatment groups combined was 9.2% over the 4.5 wk training period, which is in agreement with other investigations of aerobic training and VO2 max improvements using a similar time period [28, 29]. It has been established that the primary determinants of VO2 max are an increased ability of the cardiovascular system to transport oxygen to the working skeletal muscle, and the improved ability of the muscle to utilize the delivered oxygen. The former is a result of increased stroke volume, which improves cardiac output; the latter is determined by the increases in oxidative enzymes and mitochondrial content [1, 2]. We measured the activity of two key oxidative enzymes that are indicative of muscle oxidative capacity, CS and SDH. Both are found in the mitochondria and are key enzymes of the Krebs cycle, and each has been demonstrated to increase in response to endurance training [3–7, 9, 10, 30].We also measured total protein content of the transcription coactivator peroxisome proliferator-activated receptor γ coactivator-1α(PGC-1α) as a marker for increased mitochondrial biogenesis. PGC-1α is a transcriptional coac- tivator of transcription factor PPARγ, and together, they regulate the expression of genes that encode mitochondrial proteins. An acute bout of exercise or stimulated skeletal muscle contraction induces an increase in both PGC-1α mRNA and protein in skeletal muscle [31–34], and it has been shown that increased PGC-1α activation and total protein amount leads to increased mitochondrial biogenesis [24]. In the present study, we demonstrated that the activity of CS and SDH, and the total protein content of PGC-1α increased significantly in response to 4.5 wks of training. Table 5: Body composition. Baseline End Weight (kg) CM 71.7 ± 5.5 71.7 ± 5.5 CHO 71.4 ± 3.4 71.4 ± 3.4 PLA 73.2 ± 4.5 72.9 ± 4.4 Lean mass, whole body (kg) CM 49.6 ± 4.1 51.0 ± 4.1† CHO 49.4 ± 3.7 50.0 ± 3.74† PLA 47.7 ± 3.7 48.5 ± 3.5† Fat mass, whole body (kg) CM 19.1 ± 2.2 17.7 ± 2.15† CHO 19.0 ± 2.1 18.5 ± 1.9† PLA 22.5 ± 2.6 21.5 ± 2.6† Lean mass, trunk (kg) CM 23.7 ± 1.8 24.6 ± 1.7§ CHO 22.6 ± 1.7 22.7 ± 1.7 PLA 20.9 ± 1.6 21.3 ± 1.6 Fat mass, trunk (kg) CM 11.6 ± 1.6 10.7 ± 1.6† CHO 10.2 ± 1.3 9.6 ± 1.1† PLA 10.7 ± 1.3 10.0 ± 1.3† Lean mass, legs (kg) CM 17.8 ± 1.4 18.3 ± 1.4† CHO 16.7 ± 1.3 17.1 ± 1.3† PLA 15.6 ± 1.2 16.0 ± 1.2† Fat mass, legs (kg) CM 7.0 ± 0.9 6.6 ± 0.9† CHO 6.7 ± 0.8 6.8 ± 0.8† PLA 7.7 ± 0.8 7.5 ± 0.8† Bone mineral density (g/cm2) CM 1.2 ± 0.1 1.2 ± 0.1 CHO 1.2 ± 0.0 1.2 ± 0.0 PLA 1.2 ± 0.0 1.2 ± 0.0 Values are mean ± SE. Significant differences: †, significant time effect; §, CM versus CHO (P < .05). However, no significant treatment differences in these mea- sures were detected. There was a slight but nonsignificant trend for a greater increase in CS and SDH activity in CM compared to CHO and PLA. It may be that the training period was not long enough to detect any potential differ- ences that could emerge in response to chronic nutritional supplementation. Thus, our results suggest that the greater VO2 max improvements with CM supplementation are most likely due to cardiovascular adaptations rather than increases in oxidative enzymes or in mitochondrial biogenesis. As mentioned previously, endurance training leads to an adaptive increase in cardiac output, and this increase is due to augmented stroke volume [1]. While we did not measure these variables in the present study, our results suggest that the significant improvement in VO2 max in the CM group is likely due to increased stroke volume and cardiac output, which is likely due to increased plasma volume. 8 Journal of Nutrition and Metabolism 0 0.5 1 1.5 2 2.5 3 3.5 Lean and fat mass differential, whole body (kilograms) § (a) 0 0.25 0.5 0.75 1 1.25 (kilograms) Lean and fat mass differential, legs (b) 0 0.5 1 1.5 2 2.5 (kilograms) § CM CHO PLA Lean and fat mass differential, trunk (c) Figure 4: Body composition lean and fat mass differentials. (a) Whole body differential. Lean mass (kg) gained and fat mass (kg) lost was used to calculate a whole-body differential to quantify overall body composition changes in response to 4.5 wks of cycling exercise training. (LM) − (FM) = differential. Example: (0.900kg lean mass) − (−0.350 kg fat mass) = 1.250 kg. (b) Lean and fat mass differential for the legs. (c) Lean and fat mass differential for the trunk region. Values are mean ± SE. Significant treatment differences: §, CM versus CHO (P < .05). Plasma volume expansion is a hallmark of aerobic endurance training [35] and is directly associated with increased plasma albumin content. Increased albumin in the plasma causes water to be retained in the vasculature due to increases in the colloid osmotic pressure gradient [36, 37]. Hepatic albumin synthesis has been shown to increase in response to endurance exercise training [38, 39]. Moreover, plasma albumin content was reported increased 23 h after an acute bout of cycling exercise when CHO+PRO supplementation was provided postexercise compared to placebo [40]. These results, along with the findings of the present study, suggest that hepatic albumin synthesis may have been increased to a greater extent in the CM group compared to the CHO or PLA groups and contributed to the significantly greater increase in VO2 max in the CM group. Okazaki and colleagues [16] recently demonstrated that CHO+PRO supplementation provided immediately after daily cycling exercise training in older male subjects increased stroke volume and plasma volume compared to a placebo group. Their subjects cycled for 60 min/d, 3 d/wk for 8 wk at 60–75% VO2 peak and ingested either CHO+PRO or placebo immediately postexercise each session. VO2peak increased 3.3% in the control group and 6.8% in the CHO+PRO group, with significant stroke volume and plasma volume increases only found in the CHO+PRO group [16]. In the present study, we extend the findings of Okazaki and colleagues [16] by demonstrating that the effect of nutritional supplementation on VO2 max increases is nutrient specific. In comparing CM against an isocaloric CHO only supplement and a placebo, we have shown that the Journal of Nutrition and Metabolism 9 increased VO2 max response is not due to simply providing calories postexercise. In the present study, the VO2 max increase in the CHO and PLA groups was not significantly different. Thus, these results suggest that the benefit from a CHO+PRO or CM supplement in improving VO2 max is due to the combined ingestion of carbohydrate and protein. However, we cannot rule out the possibility that a supplement composed of protein alone would not have the same effect. In addition to well-documented increases in VO2 max with training, it is known that lactate threshold improves with endurance exercise training of moderate to high inten- sity [41]. In the current study, LT improved significantly over the 4.5 wks of training although there were no significant treatment differences detected (Table 4). It has been shown that the respiratory capacity of the muscle is the key deter- minant of LT [42]. Given that muscle oxidative enzyme activity and PGC-1α content increased significantly over time without demonstrating a treatment effect, it would be expected that LT would follow a parallel pattern. Therefore, the results suggest that while LT is increased by exercise training in parallel with muscle oxidative capacity, it likewise may not be affected by nutritional supplementation. The other key finding of the present study was that body composition improvements, represented by a calculated lean and fat mass differential, were significantly greater in the CM group than the CHO group. Compared to the CHO group, the CM group lost more fat mass and gained more lean mass measured in the whole body, as well as in the trunk region only (P < .05). While these differentials were also greater for CM compared with PLA, the differences were not significant. It is well established that resistance exercise training induces significant gains in lean mass, whereas endurance exercise training is not associated with large increases in lean mass or gains in muscular strength [43]. A previous inves- tigation comparing the effects of aerobic and aerobic + resistance training showed that the aerobic + resistance group increased lean mass in arm, trunk, and total body regions, and the aerobic only group increased lean mass in trunk region only [44]. However, the aforementioned inves- tigation did not use supplementation. The body composition improvement with CM is also in agreement with the findings of Josse and colleagues [22], who recently demonstrated significantly greater muscle mass accretion, fat mass loss, and strength gains with milk supplementation compared to soy and CHO after a 12-wk resistance training program [22]. Therefore, the findings of our study are in line with what is reported in the literature for exercise mode-dependent body composition changes. As shown in Table 5, all groups demonstrated significant changes over time in whole-body lean mass, trunk fat mass, and legs lean and fat mass, and the CM group demonstrated a significant treatment effect compared to CHO when whole body and regional differentials were calculated (Figures 4(a), 4(b), and 4(c)). The whole body and trunk differentials for PLA were slightly greater than CHO although not sig- nificantly different from either CM or CHO. The lack of difference in the PLA treatment from CHO suggests that a component of the CM treatment facilitated the significant body composition change, since simply supplementing with an energy-containing supplement (CHO) did not have a sig- nificant effect compared to PLA. In fact, a slight, nonsignifi- cant increase in fat mass in the legs region was detected with CHO, whereas fat mass of the legs decreased in CM and PLA during the training period. To our knowledge, no evidence exists in the literature to suggest that postexercise CHO supplementation would mediate this type of change, given that the subjects’ diets were not standardized and controlled during the study. However, this finding further under- scores the difference in supplementing with a CHO+PRO- containing supplement versus calories from CHO alone in facilitating body composition changes. There are two possible explanations for the difference found with the CM treatment compared to CHO: first, the availability of amino acids (AAs) in the milk for anabolism and muscle mass accretion, and second, a fat-loss promoting effect of dairy calcium and protein. It is known that AAs, along with a permissive amount of insulin, are required for muscle protein accretion to occur in response to exercise [45, 46]. The CHO treatment would increase plasma insulin levels and provide glucose as an energy and glycogen-synthesizing substrate, but would provide no AAs for the synthesis of new muscle protein. Thus, AAs availability from the milk proteins whey and casein provided substrate for this adaptive process. In addition, Zemel [47, 48] have shown that the increased consumption of dietary calcium is associated with reduced adiposity and greater weight loss in energy restricted diets. Moreover, the fat and weight loss effects were greater when the source of the dietary calcium was from dairy products rather than a calcium supplement [47, 48]. Additional evidence that the dairy component of the CM treatment likely underlies some of the body composition changes is found in the resistance training study of Hartman and colleagues [21], who demonstrated that fat mass decreased, and lean mass increased, in groups provided either milk, soy, or CHO postexercise but that milk significantly promoted increased hypertrophy compared to soy and CHO [21]. Another well-known benefit of dairy calcium consumption is improved bone mineral density. We did not detect treatment or time differences in BMD (Table 5); however, this is not surprising, given the relatively short duration of the training program and the lack of a resistance training component. Taken together, these data suggest that the dairy component of the CM treatment was instrumental in facilitating the fat mass changes compared to the CHO and PLA groups, while the AAs from milk proteins provided substrate for lean mass accretion in the present study. There are several limitations to the present study. First, the subjects’ normal diets were not controlled nor standard- ized for the majority of the training period. Although the diets were recorded and replicated for 3 days each week as described above, there could have been within and between- subject variations in the amount of protein, calcium, and total caloric intake on the nonrecorded days during the training period which could have influenced the adaptive response. Second, CM contains many other micronutrients and flavonoids in addition to the major macronutrients and calcium. However, the possible effects of these additional 10 Journal of Nutrition and Metabolism components on the training adaptations reported here are not known at this time. Third, the taste and appearance of the three treatments were different. However, the subjects were not aware of what the three treatments were, and since they only ingested the treatment for which they were randomized for the entire study period, they did not taste any of the other treatments. Fourth, we did not match the supplement dosing to each individual’s body weight, but stratified the amount of supplement for each dose according to body weight ranges. We have previously shown that supplementation with ∼1.0 g of CHO and ∼0.3 g of PRO per kg body wt postexercise will substantially increase muscle glycogen synthesis and recovery from exercise [14]. In the present study, providing supplement based on a weight range represented a more realistic and practical approach. Finally, while we propose that the greater increase in VO2 max in the CM group is likely due to albumin synthesis, we did not measure plasma volume or plasma albumin and, therefore, cannot say with certainly that this is the reason for the VO2 max differences. Further investigation is necessary to expand upon these results and elucidate the mechanisms of the greater adaptive response. 5. Conclusion Our results demonstrated that CM supplementation postex- ercise increased the magnitude of VO2 max improvement in response to a 4.5 wk aerobic exercise-training program. Mus- cle oxidative capacity and lactate threshold improved signif- icantly in all treatment groups, with no differences found between treatments. This would suggest that the greater improvement in VO2 max when supplementing with CM as compared with CHO or placebo was cardiovascular rather than cellular in nature. In addition, CM supplementation significantly improved body composition as defined by the combination of an increase in lean mass and a decrease in fat mass compared to CHO. We conclude that CM is an effective postexercise recovery supplement that can induce positive increases in aerobic training adaptations in healthy, untrained humans. Acknowledgments The authors wish to thank the subjects in this study for their time, energy, and dedication. They thank Dr. Dong-Ho Han at Washington University in St. Louis for sharing his expert advice for the measurement of total PGC-1α content. They also thank our colleagues in the Exercise Physiology and Metabolism Laboratory at the University of Texas at Austin for assisting with this study. This project was supported by a grant from the National Dairy Council and the National Fluid Milk Processor Promotion Board. References [1] B. Ekblom, P. O. ˚Astrand, B. Saltin, J. Stenberg, and B. Wall- str¨om, “Effect of training on circulatory response to exercise,” Journal of Applied Physiology, vol. 24, no. 4, pp. 518–528, 1968. [2] B. Saltin, G. Blomqvist, J. H. Mitchell, R. Johnson Jr., K. Wildenthal, and C. Chapman, “A longitudinal study of adap- tive changes in oxygen transport and body composition,” Cir- culation, vol. 38, no. 7, pp. 1–78, 1968. [3] S. L. Carter, C. D. Rennie, S. J. Hamilton, and M. A. Tarnopol- sky, “Changes in skeletal muscle in males and females follow- ing endurance training,” Canadian Journal of Physiology and Pharmacology, vol. 79, no. 5, pp. 386–392, 2001. [4] K. De Bock, W. Derave, B. O. Eijnde et al., “Effect of training in the fasted state on metabolic responses during exercise with carbohydrate intake,” Journal of Applied Physiology, vol. 104, no. 4, pp. 1045–1055, 2008. [5] M. Fernstr¨om, M. Tonkonogi, and K. Sahlin, “Effects of acute and chronic endurance exercise on mitochondrial uncoupling in human skeletal muscle,” Journal of Physiology, vol. 554, no. 3, pp. 755–763, 2004. [6] J. Henriksson and J. S. Reitman, “Time course of changes in human skeletal muscle succinate dehydrogenase and cyto- chrome oxidase activities and maximal oxygen uptake with physical activity and inactivity,” Acta Physiologica Scandinav- ica, vol. 99, no. 1, pp. 91–97, 1977. [7] J. O. Holloszy, “Biochemical adaptations in muscle: effects of exercise on mitochondrial oxygen uptake and respiratory enzyme activity in skeletal muscle,” Journal of Biological Chemistry, vol. 242, no. 9, pp. 2278–2282, 1967. [8] J. O. Holloszy and E. F. Coyle, “Adaptations of skeletal muscle to endurance exercise and their metabolic consequences,” Journal of Applied Physiology, vol. 56, no. 4, pp. 831–838, 1984. [9] J. O. Holloszy, L. B. Oscai, I. J. Don, and P. A. Mol´e, “Mito- chondrial citric acid cycle and related enzymes: adaptive response to exercise,” Biochemical and Biophysical Research Communications, vol. 40, no. 6, pp. 1368–1373, 1970. [10] A. Puntschart, H. Claassen, K. Jostarndt, H. Hoppeler, and R. Billeter, “mRNAs of enzymes involved in energy metabolism and mtDNA are increased in endurance-trained athletes,” American Journal of Physiology, vol. 269, no. 3, pp. C619–C625, 1995. [11] J. M. Berardi, T. B. Price, E. E. Noreen, and P. W. R. Lemon, “Postexercise muscle glycogen recovery enhanced with a car- bohydrate-protein supplement,” Medicine and Science in Sports and Exercise, vol. 38, no. 6, pp. 1106–1113, 2006. [12] E. S. Niles, T. Lachowetz, J. Garfi et al., “Carbohydrate-pro- tein drink improves time to exhaustion after recovery from endurance exercise,” Journal of Exercise Physiology Online, vol. 4, no. 1, pp. 45–52, 2001. [13] M. J. Saunders, M. D. Kane, and M. K. Todd, “Effects of a car- bohydrate-protein beverage on cycling endurance and muscle damage,” Medicine and Science in Sports and Exercise, vol. 36, no. 7, pp. 1233–1238, 2004. [14] M. B. Williams, P. B. Raven, D. L. Fogt, and J. L. Ivy, “Effects of recovery beverages on glycogen restoration and endurance exercise performance,” Journal of Strength and Conditioning Research, vol. 17, no. 1, pp. 12–19, 2003. [15] J. L. Ivy, Z. Ding, H. Hwang, L. C. Cialdella-Kam, and P. J. Morrison, “Post exercise carbohydrate-protein supplementa- tion: phosphorylation of muscle proteins involved in glycogen synthesis and protein translation,” Amino Acids, vol. 35, no. 1, pp. 89–97, 2008. [16] K. Okazaki, T. Ichinose, H. Mitono et al., “Impact of protein and carbohydrate supplementation on plasma volume expan- sion and thermoregulatory adaptation by aerobic training in older men,” Journal of Applied Physiology, vol. 107, no. 3, pp. 725–733, 2009. Journal of Nutrition and Metabolism 11 [17] J. R. Karp, J. D. Johnston,S. Tecklenburg, T. D. Mickleborough, A. D. Fly, and J. M. Stager, “Chocolate milk as a post-exercise recovery aid,” International Journal of Sport Nutrition and Exercise Metabolism, vol. 16, no. 1, pp. 78–91, 2006. [18] K. Pritchett, P. Bishop, R. Pritchett, M. Green, and C. Katica, “Acute effects of chocolate milk and a commercial recovery beverage on postexercise recovery indices and endurance cycling performance,” Applied Physiology, Nutrition and Me- tabolism, vol. 34, no. 6, pp. 1017–1022, 2009. [19] K. Thomas, P. Morris, and E. Stevenson, “Improved endurance capacity following chocolate milk consumption compared with 2 commercially available sport drinks,” Applied Physiol- ogy, Nutrition and Metabolism, vol. 34, no. 1, pp. 78–82, 2009. [20] T. A. Elliot, M. G. Cree, A. P. Sanford, R. R. Wolfe, and K. D. Tipton, “Milk ingestion stimulates net muscle protein synthesis following resistance exercise,” Medicine and Science in Sports and Exercise, vol. 38, no. 4, pp. 667–674, 2006. [21] J. W. Hartman, J. E. Tang, S. B. Wilkinsonet al., “Consumption of fat-free fluid milk after resistance exercise promotes greater lean mass accretion than does consumption of soy or carbohy- drate in young, novice, male weightlifters,” American Journal of Clinical Nutrition, vol. 86, no. 2, pp. 373–381, 2007. [22] A. R. Josse, J. E. Tang, M. A. Tarnopolsky, and S. M. Phillips, “Body composition and strength changes in women with milk and resistance exercise,” Medicine and Science in Sports and Exercise, vol. 42, no. 6, pp. 1122–1130, 2010. [23] P. O. ˚Astrand and K. Rodahl, Textbook of Work Physiology: Physiological Bases of Exercise, McGraw-Hill Book Company, New York, NY, USA, 1977. [24] D. C. Wright, D. H. Han, P. M. Garcia-Roves, P. C. Geiger, T. E. Jones, and J. O. Holloszy, “Exercise-induced mitochondrial biogenesis begins before the increase in muscle PGC-1α expression,” Journal of Biological Chemistry, vol. 282, no. 1, pp. 194–199, 2007. [25] O. H. Lowry, N. J. Rosebrough, A. L. Farr, and R. J. Randall, “Protein measurement with the Folin phenol reagent,” The Journal of Biological Chemistry, vol. 193, no. 1, pp. 265–275, 1951. [26] P. A. Srere, “Citrate synthase,” Methods in Enzymology, vol. 13, pp. 3–5, 1969. [27] H. Lowry and J. V. Passonneau, A Flexible System of Enzymatic Analysis, Academic Press, New York, NY, USA, 1972. [28] S. M. Phillips, H. J. Green, M. A. Tarnopolsky, G. J. Heigen- hauser, and S. M. Grant, “Progressive effect of endurance training on metabolic adaptations in working skeletal muscle,” American Journal of Physiology, vol. 270, pp. E265–E272, 1996. [29] R. J. Spina, M. M. Chi, M. G. Hopkins, P. M. Nemeth, O. H. Lowry, and J. O. Holloszy, “Mitochondrial enzymes increase in muscle in response to 7–10 days of cycle exercise,” Journal of Applied Physiology, vol. 80, no. 6, pp. 2250–2254, 1996. [30] J. P. Morton, L. Croft, J. D. Bartlett et al., “Reduced car- bohydrate availability does not modulate training-induced heat shock protein adaptations but does upregulate oxidative enzyme activity in human skeletal muscle,” Journal of Applied Physiology, vol. 106, no. 5, pp. 1513–1521, 2009. [31] K. Baar, A. R. Wende, T. E. Jones et al., “Adaptations of skeletal muscle to exercise: rapid increase in the transcriptional coac- tivator PGC-1,” FASEB Journal, vol. 16, no. 14, pp. 1879–1886, 2002. [32] M. Goto, S. Terada, M. Kato et al., “cDNA cloning and mRNA analysis of PGC-1 in epitrochlearis muscle in swimming- exercised rats,” Biochemical and Biophysical Research Commu- nications, vol. 274, no. 2, pp. 350–354, 2000. [33] I. Irrcher, P. J. Adhihetty, T. Sheehan, A. M. Joseph, and D. A. Hood, “PPARγ coactivator-1α expression during thy- roid hormone and contractile activity-induced mitochondrial adaptations,” American Journal of Physiology, vol. 284, no. 6, pp. C1669–C1677, 2003. [34] S. Terada, M. Goto, M. Kato, K. Kawanaka, T. Shimokawa, and I. Tabata, “Effects of low-intensity prolonged exercise on PGC- 1 mRNA expression in rat epitrochlearis muscle,” Biochemical and Biophysical Research Communications, vol. 296, no. 2, pp. 350–354, 2002. [35] M. N. Sawka, V. A. Convertino, E. R. Eichner, S. M. Schnieder, and A. J. Young, “Blood volume: importance and adaptations to exercise training, environmental stresses, and trauma/ sickness,” Medicine and Science in Sports and Exercise, vol. 32, no. 2, pp. 332–348, 2000. [36] V. A. Convertino, P. J. Brock, L. C. Keil, E. M. Bernauer, and J. E. Greenleaf, “Exercise training-induced hypervolemia: role of plasma albumin, renin, and vasopressin,” Journal of Applied Physiology, vol. 48, no. 4, pp. 665–669, 1980. [37] C. M. Gillen, R. Lee, G. W. Mack, C. M. Tomaselli, T. Nishiyasu, and E. R. Nadel, “Plasma volume expansion in humans after a single intense exercise protocol,” Journal of Applied Physiology, vol. 71, no. 5, pp. 1914–1920, 1991. [38] K. Nagashima, G. W. Cline, G. W. Mack, G. I. Shulman, and E. R. Nadel, “Intense exercise stimulates albumin synthesis in the upright posture,” Journal of Applied Physiology, vol. 88, no. 1, pp. 41–46, 2000. [39] R. C. Yang, G. W. Mack, R. R. Wolfe, and E. R. Nadel, “Albumin synthesis after intense intermittent exercise in human sub- jects,” Journal of Applied Physiology, vol. 84, no. 2, pp. 584–592, 1998. [40] K. Okazaki, M. Goto, and H. Nose, “Protein and carbohydrate supplementation increases aerobic and thermoregulatory capacities,” Journal of Physiology, vol. 587, no. 23, pp. 5585– 5590, 2009. [41] D. C. Poole and G. A. Gaesser, “Response of ventilatory and lactate thresholds to continuous and interval training,” Journal of Applied Physiology, vol. 58, no. 4, pp. 1115–1121, 1985. [42] J. L. Ivy, R. T. Withers, P. J. Van Handel, D. H. Elger, and D. L. Costill, “Muscle respiratory capacity and fiber type as determinants of the lactate threshold,” Journal of Applied Physiology, vol. 48, no. 3, pp. 523–527, 1980. [43] R. C. Hickson, “Interference of strength development by simultaneously training for strength and endurance,” Euro- pean Journal of Applied Physiology and Occupational Physiol- ogy, vol. 45, no. 2-3, pp. 255–263, 1980. [44] L. M. Pierson, W. G. Herbert, H. J. Norton et al., “Effects of combined aerobic and resistance training versus aerobic training alone in cardiac rehabilitation,” Journal of Cardiopul- monary Rehabilitation, vol. 21, no. 2, pp. 101–110, 2001. [45] G. Biolo, K. D. Tipton, S. Klein, and R. R. Wolfe, “An abundant supply of amino acids enhances the metabolic effect of exercise on muscle protein,” American Journal of Physiology, vol. 273, pp. E122–E129, 1997. [46] G. Biolo, B. D. Williams, R. Y. Fleming, and R. R. Wolfe, “Insulin action on muscle protein kinetics and amino acid transport during recovery after resistance exercise,” Diabetes, vol. 48, no. 5, pp. 949–957, 1999. [47] M. B. Zemel, “Mechanisms of dairy modulation of adiposity,” Journal of Nutrition, vol. 133, no. 1, pp. 252S–256S, 2003. [48] M. B. Zemel, “Role of calcium and dairy products in energy partitioning and weight management,” The American journal of clinical nutrition, vol. 79, no. 5, pp. 907S–912S, 2004.

Aerobic exercise training adaptations are increased by postexercise carbohydrate-protein supplementation.

06-09-2011

Ferguson-Stegall, Lisa,McCleave, Erin,Ding, Zhenping,Doerner Iii, Phillip G,Liu, Yang,Wang, Bei,Healy, Marin,Kleinert, Maximilian,Dessard, Benjamin,Lassiter, David G,Kammer, Lynne,Ivy, John L

eng

PMC7503696

International Journal of Environmental Research and Public Health Article Using Accelerometry for Evaluating Energy Consumption and Running Intensity Distribution Throughout a Marathon According to Sex Carlos Hernando 1,2,* , Carla Hernando 3, Ignacio Martinez-Navarro 4,5, Eladio Collado-Boira 6 , Nayara Panizo 6 and Barbara Hernando 7 1 Sport Service, Jaume I University, 12071 Castellon, Spain 2 Department of Education and Specific Didactics, Jaume I University, 12071 Castellon, Spain 3 Department of Mathematics, Carlos III University of Madrid, 28911 Leganés, Madrid, Spain; [email protected] 4 Department of Physical Education and Sport, University of Valencia, 46010 Valencia, Spain; [email protected] 5 Sports Health Unit, Vithas-Nisa 9 de Octubre Hospital, 46015 Valencia, Spain 6 Faculty of Health Sciences, Jaume I University, 12071 Castellon, Spain; [email protected] (E.C.-B.); [email protected] (N.P.) 7 Department of Medicine, Jaume I University, 12071 Castellon, Spain; [email protected] * Correspondence: [email protected]; Tel.: +34-964-728808 Received: 28 July 2020; Accepted: 25 August 2020; Published: 26 August 2020



 Abstract: The proportion of females participating in long-distance races has been increasing in the last years. Although it is well-known that there are differences in how females and males face a marathon, higher research may be done to fully understand the intrinsic and extrinsic factors affecting sex differences in endurance performance. In this work, we used triaxial accelerometer devices to monitor 74 males and 14 females, aged 30 to 45 years, who finished the Valencia Marathon in 2016. Moreover, marathon split times were provided by organizers. Several physiological traits and training habits were collected from each participant. Then, we evaluated several accelerometry- and pace-estimated parameters (pacing, average change of speed, energy consumption, oxygen uptake, running intensity distribution and running economy) in female and male amateur runners. In general, our results showed that females maintained a more stable pacing and ran at less demanding intensity throughout the marathon, limiting the decay of running pace in the last part of the race. In fact, females ran at 4.5% faster pace than males in the last kilometers. Besides, their running economy was higher than males (consumed nearly 19% less relative energy per distance) in the last section of the marathon. Our results may reflect well-known sex differences in physiology (i.e., muscle strength, fat metabolism, VO2max), and in running strategy approach (i.e., females run at a more conservative intensity level in the first part of the marathon compared to males). The use of accelerometer devices allows coaches and scientific community to constantly monitor a runner throughout the marathon, as well as during training sessions. Keywords: accelerometry; sex; physical activity; running intensity; energy consumption; pacing; marathoners; running economy 1. Introduction Marathons have growth in popularity and therefore in participants worldwide at a record pace [1–3]. However, the increase in the number of female marathoners has been delayed, as compared to male, due to different social and behavioral causes previously pointed out by Joyner and colleagues Int. J. Environ. Res. Public Health 2020, 17, 6196; doi:10.3390/ijerph17176196 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2020, 17, 6196 2 of 14 in 2017 [4]. Although sex ratios are still far of being equivalent (i.e., 18.7% of females from the total number of participants in the Valencia Marathon 2019), female’s participation in marathon races has increased exponentially since Kathy Swizer finished the Boston Marathon in 1967. This situation has encouraged scientific community to study female’s behavior in long-distance races and compare them with males. Studies have been focused on analyzing different factors affecting running performance such as running speed [5–8], pacing [9], physiological traits [4,10], running economy [7,11,12] and predominant type of metabolism used [13–15], as well as physical, biomechanical, psychological and social factors [16–19]. The assessment of physiological parameters affecting running performance has been carried out in lab-based conditions—normally by measuring the volume of expired gases (the gold standard test) [11–13,19]. However, lab-based conditions are far from normal race conditions. Up to now, field-based studies have been focused on estimating the energy consumption throughout a long-distance race by analyzing changes in running speed [7,10,20–23]. The use of portable measurement systems to obtain parameters for estimating energy consumption in real conditions is nowadays a reality [24–29]. In particular, the use of triaxial accelerometry has strongly emerged as a tool that allows the evaluation of a physical activity, in terms of duration, frequency and intensity, performed by an individual in free-living conditions [30–33]. Thus, using the cut-off points previously established for a specific population and/or an activity, the accelerometer output data allowed to indirectly estimate the energy cost of an activity [34–38]. With the aim of monitoring middle-aged recreational marathoners during a marathon using accelerometry-based devices, our research group has established the GENEActiv® cut-off points, under lab-based conditions, for discriminating the six relative-intensity activity levels in female and male marathoners [39]. Once cut-off points were established, we used accelerometer output data for analyzing the running intensity distribution and energy consumption of runners during a marathon race (a free-living condition) [40]. Interestingly, accelerometer output data can also be used for inferring other useful parameters (i.e., running economy of the runner) in real conditions. Since sex was not taken into account in our previous work, this study focused on evaluating several accelerometry-estimated parameters (energy consumption, running intensity distribution and running economy) according to the individual’s sex. The use of accelerometers allowed us to directly and constantly monitor a total of 88 recreational marathon runners (74 males and 14 females) throughout the marathon race. Here, accelerometry- and pace-based data collected from females and males were analyzed separately. In this study, we also pointed out the valuable additional information that accelerometry offers to athletes, coaches and scientific community, as compared to the evaluation of running speed. 2. Materials and Methods 2.1. Sample Set and Data Collection From all participants of the Valencia Fundación Trinidad Alfonso EDP 2016 Marathon (20 November 2016), a total of 103 recreational marathon runners, aged 30 to 45 years, were selected to participate in this study. Eight runners did not start the race and were discarded from our study population. Finally, a total of 88 recreational marathon runners (74 males and 14 females) crossed the finish line of the Valencia Fundacion Trinidad Alfonso EDP 2016 Marathon and thus were analyzed in this study. The entire process of sampling, as well as the weather and track conditions of the race, has been previously described [41]. Details of data collection, processing and analysis have been previously described [40]. Four weeks before the marathon, participants completed a cardiopulmonary test. In this appointment, we also collected anthropometric data, demographics, medical information, training program and competition history. One hour before the marathon, all participants were weighed. During the race, participants wore a GENEActiv accelerometer (Activinsights Ltd., Kimbolton, Cambridgeshire, United Kingdom) Int. J. Environ. Res. Public Health 2020, 17, 6196 3 of 14 on the non-dominant wrist as a watch. Accelerometers were adjusted to record acceleration data at a rate of 85.7 Hz, and data was summarized into Signal Vector Magnitude gravity subtracted (SVMgs) per minute. Recording and processing of acceleration data has been previously explained in detail [39]. All individuals underwent the same testing under the same experimental conditions. Raw data of this study is available in Supplementary File S1. All individuals included in the current study were fully informed and gave their written consent to participate. All experiments were performed in accordance with international guidelines and regulations that govern human research. The research was approved by the Research Ethics Committee of the University Jaume I of Castellon and is enrolled in the ClinicalTrails.gov database (NCT03155633). 2.2. Data Analysis The marathon race was divided into nine sections as previously described [40]. All analyses were performed for each of the nine race sections, as well as for the entire marathon distance. Firstly, the physical effort distribution of each runner throughout the marathon, in terms of relative intensity levels of physical activity, was estimated using accelerometry-based devices and following the methodology previously described by our research group [39]. Different cut-off points were used to discriminate the six relative-intensity activity levels in female and male recreational marathoners (Table 1). Then, we estimated the time of each participant running at each one of the six-relative intensity levels (sedentary, light, moderate, vigorous, very vigorous and extremely vigorous). This estimation was performed for each of the nine race sections and for the whole race (Tables S1 and S2). Table 1. Values established for delineating the six-relative intensity levels of physical activity according to runner’s sex. Reference Values Established for Each Intensity Level in Males by Hernando et al. (2018) Values used for Energy Consumption Estimation Sex Relative-Intensity Levels of Physical Activity # VO2 (mL·kg−1·min−1) METs * %VO2max (mL·kg−1·min−1) VO2 (mL·kg−1·min−1) METs * Males Sedentary X < 10% VO2 < 5.57 METs < 1.59 8.1 4.5 1.29 Light 10% ≤ X <25% 5.57 ≤ VO2 <13.94 1.59 ≤ METs < 3.97 17.5 9.75 2.79 Moderate 25% ≤ X < 45% 13.94 ≤ VO2 < 25.08 3.97 ≤ METs < 7.15 35.0 19.51 5.57 Vigorous 45% ≤ X < 65% 25.08 ≤ VO2 < 36.23 7.15 ≤ METs < 10.33 55.0 30.66 8.76 Very Vigorous 65% ≤ X < 85% 36.23 ≤ VO2 < 47.38 10.33 ≤ METs < 13.54 75.0 41.81 11.94 Extremely Vigorous X ≥ 85% VO2 ≥ 47.38 METs ≥ 13.54 92.5 51.56 14.73 Females Sedentary X < 10% VO2 < 4.82 METs < 1.38 8.1 3.91 1.12 Light 10% ≤ X <25% 4.82 ≤ VO2 <12.07 1.38 ≤ METs < 3.45 17.5 8.44 2.41 Moderate 25% ≤ X < 45% 12.07 ≤ VO2 < 21.72 3.45 ≤ METs < 6.21 35.0 16.89 4.83 Vigorous 45% ≤ X < 65% 21.72 ≤ VO2 < 31.38 6.21 ≤ METs < 8.97 55.0 26.55 7.59 Very Vigorous 65% ≤ X < 85% 31.38 ≤ VO2 < 41.03 8.97 ≤ METs < 11.72 75.0 36.20 10.34 Extremely Vigorous X ≥ 85% VO2 ≥ 41.03 METs ≥ 11.72 92.5 44.65 12.76 Abbreviations: N, number of individuals; VO2max, maximum oxygen consumption; VO2, oxygen consumption; MET, metabolic equivalent task. Each minute of the cardiopulmonary test was classified into one of the six intensity categories of physical activity relative to an individual’s level of cardiorespiratory (VO2max). # X denotes the percentage of an individual’s aerobic capacity (VO2max) used to classify each one of the six relative-intensity categories. * 1 MET = 3.5 mLO2·kg−1·min−1. 1 MET = 1 Kcal·h−1. Next, energy consumption was calculated by using the median %VO2max value of the range delimiting each intensity category in males and females (Table 1). That was applied for all intensity levels except for the sedentary category, in which the standing oxygen cost (4.5 mLO2·kg−1·min−1) was Int. J. Environ. Res. Public Health 2020, 17, 6196 4 of 14 applied as reference value in males [42]. The reference value corresponds to the 8.1% of the maximum oxygen uptake in our males. For females, we used the 8.1% of the maximum oxygen uptake seen in our females as the VO2 standing (3.91 mLO2·kg−1·min−1). Following previous recommendations [21], we considered that one MET is equal to 3.5 mLO2·kg−1·min−1, and one MET is equal to one kcal·kg−1·h−1. As the energy consumption depends on the individual’s body mass, we calculated (i) the calories consumed per kilogram of body weight per minute (kcal·kg−1·min−1), in order to obtain the physical effort intensity [20,21,43]; and (ii) the calories consumed per kilogram of body weight per kilometer (kcal·kg−1·km−1), to infer the running economy of runners [16,44]. Indeed, accelerometry-derived data was also used for estimating the %VO2max maintained during the marathon by each runner, an indicator of the physical effort degree respect to the maximum value [19,42,43,45]. Following the methodology described previously, we also inferred the VO2net and the energy of cost running above standing (Crnet) for each participant included in the study [12,46,47]. These estimations were performed by applying the corresponding reference values for females and males. The split-times in minutes of the marathon sections were recorded for calculating the average running speed of all sections and the whole marathon distance. Then, the average change in speed (ACS) for each segment, related to the average race speed, was calculated. The average change in speed through the whole race was estimated by averaging the ACS values of all sections. The ACS is a valuable measure for assessing maintenance of running pace [9,48]. The squat jump test was performed to measure the runner’s strength before and after the race. Jumping height was estimated using the flight time of the jump, which was measured by a contact platform (Chronojump, Barcelona, Spain) [49]. Individuals were familiarized with the test’s procedure before to carry out it. Before the marathon race, all individuals performed a total of three jumps, and the best jump was recorded. After crossing the finish line, the number of attempts was conditioned by the capacity of each runner to jump (due to muscular fatigue). No more than three attempts were performed per runner, and again only the best jump was recorded. Sex comparisons were performed by calculating the percentage of differences (gap) in all measures between males and females, as previously proposed [7,10]. Briefly, we applied the following formula: Gap = ((Xfemales − Xmales)/Xfemales) × 100. 2.3. Statistical Analysis Statistical analyses were done using the IBM SPSS Statistics v.26 software, and null hypothesis was rejected when the two-sided p-value was lower than 0.05. The Kolgomorov–Smirnov test was used for testing data normality. Since variables were not normally distributed, all statistical analyses were performed by applying non-parametric statistical tests. The Chi-squared test was used for comparing categorical variable between males and females, while the Mann–Whitney U test was applied for comparing quantitative variables between sex groups. The meaningfulness of the outcomes was additionally estimated by inferring its effect size via the calculation of Cohen’s d, as following described [50,51]. Outcomes with values of d lower than 0.5 were considered to have an small relevance; those with d values between 0.5 and 0.8 presented moderate relevance; and those with values greater than 0.8 had large significance [52]. 3. Results A detailed description of individuals included in this study is summarized in Table 2. Sex differences were observed for several physiological traits, such as the body mass index (BMI; p-value = 0.001), the percentage of body fat (p-value = 2.03 × 10−5), the maximum oxygen uptake (VO2max; p-value = 6.59 × 10−6), and the maximum metabolic equivalent of task (MET; p-value = 6.77 × 10−6). Moreover, sex differences were also observed in the sessions of training performed per week (p-value = 0.04) and in the number of marathons completed (p-value = 0.03). Int. J. Environ. Res. Public Health 2020, 17, 6196 5 of 14 Table 2. Population description. Variables Males (N = 74) Females (N = 14) p-Value Physiological characteristics * age 38.58 ± 3.70 39.21 ± 3.14 0.61 BMI 23.15 ± 1.46 21.65 ± 1.93 0.001 % body fat 13.76 ± 3.68 19.94 ± 4.26 2.03 × 10−5 VO2max (mL·kg−1·min−1) 55.55 ± 5.25 48.39 ± 3.60 6.59 × 10−6 maximum METs 15.87 ± 1.50 13.83 ± 1.03 6.77 × 10−6 Training indicators * years of running 6.42 ± 2.89 6.43 ± 2.17 0.99 sessions per week 4.97 ± 0.83 4.50 ± 0.76 0.04 kilometers per week 64.32 ± 13.16 58.93 ± 11.96 0.14 hours per week 7.54 ± 2.57 6.46 ± 1.82 0.16 History as marathoner * marathons finished 3.62 ± 3.11 2.00 ± 2.15 0.03 marathon per year 1.12 ± 0.64 1.00 ± 0.55 0.58 Work intensity # high intensity 9.46% 0.00% 0.44 medium intensity 31.08% 28.57% low intensity 59.46% 71.43% Levels of study # school graduate 4.11% 7.14% 0.72 high school graduate 19.18% 7.14% professional certificate 6.85% 7.14% undergraduate degree 69.86% 78.57% Abbreviations: N, number of samples; BMI, body mass index; SD, standard deviation. * Values are presented as mean ± SD. #. Values are presented as percentage. Mann–Whitney U test was used for comparing quantitative variables among groups. Chi-square test was used for comparing categorical variables among groups. Bold denotes significant results. Using the squat jump test, we measured the level of lower body strength of each runner before and after running the marathon (Table 3). As expected, the basal squat jump height was significantly higher in males compared to females (27.34 ± 4.28 cm versus 23.84 ± 3.82 cm; p-value = 0.007; Cohen’s d = 0.60). No significant sex differences in the squat jump height were observed after crossing the finish line (21.88 ± 6.19 cm versus 20.53 ± 6.72 cm; p-value = 0.300; Cohen’s d = 0.22). Therefore, the lower body strength of females seemed to be less altered by running a marathon, one of the most challenging endurance competitions. Table 3. Comparison of the different variables collected over the whole marathon distance between males and females. Variable Males (N = 74) Females (N = 14) p-Value Cohen’s d Gap Speed (m·min−1) 201.29 ± 17.84 180.96 ± 14.07 1.74 × 10−4 0.87 −11.24% Energy consumed (kcal) 3274.07 ± 599.82 2423.01 ± 239.76 9.32 × 10−7 1.23 −35.12% Relative energy consumed per minute (kcal·kg−1·min−1) 0.21 ± 0.03 0.19 ± 0.02 9.91 × 10−4 0.75 −14.42% Relative energy consumed per kilometer (kcal·kg−1·km−1) 1.07 ± 0.16 1.04 ± 0.11 0.34 0.21 −2.91% Cost running net (Crnet) 4.22 ± 0.69 4.11 ± 0.49 0.38 0.19 −2.82% Percentage of VO2max (%) 80.76 ± 11.51 81.57 ± 7.59 0.68 0.09 1.00% Basal Metabolic Rate (BMR) 12.87 ± 1.84 11.24 ± 1.06 8.29 × 10−4 0.76 −14.47% Marathon time (minutes) 211.28 ± 19.16 234.50 ± 18.46 1.74 × 10−4 0.87 9.90% Squat jump at the start line (cm) 27.24 ± 4.29 23.84 ± 3.82 0.007 0.60 −14.26% Squat jump at the finish line (cm) 21.89 ± 6.19 20.53 ± 6.72 0.30 0.22 −6.62% Average change in speed (%) 5.39 ± 2.62 6.29 ± 2.58 0.15 0.31 14% % of time at sedentary level 0.01 ± 0.12 0.00 ± 0.00 0.66 0.02 NA % of time at light level 0.09 ± 0.5 0.07 ± 0.27 0.81 0.02 −29% % of time at moderate level 3.61 ± 7.40 1.07 ± 1.59 0.11 0.33 −237% % of time at vigorous level 11.58 ± 19.58 4.79 ± 8,11 0.10 0.35 −142% % of time at very vigorous level 30.82 ± 29.33 50.50 ± 30.29 0.02 0.52 39% % of time at extremely vigorous level 53.88 ± 39.29 43.64 ± 34.86 0.27 0.24 −23% Abbreviations: N, number of samples; SD, standard deviation; NA, not available; Gap, percentage of sex differences. Values are presented as mean ± SD. Mann–Whitney U test was used for comparing quantitative variables among groups. Cohen’s d was calculated for inferring the effect size of a variable. Bold denotes significant results. Int. J. Environ. Res. Public Health 2020, 17, 6196 6 of 14 Firstly, accelerometer-derived data was used to determine the running intensity distribution of female and male runners throughout the marathon. In this regard, the time consumed at each intensity level was expressed as the percentage respect to the total time needed for covering each one of the nine race sections and to the marathon time (Figure 1 and Table 3). Int. J. Environ. Res. Public Health 2020, 17, x FOR PEER REVIEW 7 of 15 distribution of time running at the very vigorous intensity level, females seemed to spend a higher percentage of time running at this intensity level than males, with significant differences in the 10–15 km section, the 15-HM section, the HM-25 km section, the 25–30 km section, the 30–35 km section and the entire marathon distance (Figure 1 and Table 3). Figure 1. Bar plots showing the percentage of time performing at each of the six-relative intensity levels of physical activity. Running intensity distribution was analyzed (A–I) in each one of the nine race sections, and (J) for the whole marathon distance. Bars represent the average values for males Figure 1. Bar plots showing the percentage of time performing at each of the six-relative intensity levels of physical activity. Running intensity distribution was analyzed (A–I) in each one of the nine race sections, and (J) for the whole marathon distance. Bars represent the average values for males (orange) and females (blue), and error bars represent the standard deviation of the mean. Mean time spent (±standard deviation) to cover each race sections and the whole race by males (M) and females (F) is showed in the corresponding panel. A Mann–Whitney U test was used for testing sex differences. * p-value < 0.05; ** p-value < 0.01. Int. J. Environ. Res. Public Health 2020, 17, 6196 7 of 14 Time racing at extremely vigorous intensity level was similar in females and males at all race sections and in the entire marathon distance (Figure 1 and Table 3). However, males tended to spend more time running at extremely vigorous intensity than females, except for the last race section (from 40km to the finish line) (61.05 ± 40.69 % versus 62.93 ± 38.08%, respectively). Regarding the distribution of time running at the very vigorous intensity level, females seemed to spend a higher percentage of time running at this intensity level than males, with significant differences in the 10–15 km section, the 15-HM section, the HM-25 km section, the 25–30 km section, the 30–35 km section and the entire marathon distance (Figure 1 and Table 3). Nevertheless, males were more time running at vigorous intensity than females, showing significant differences in the 10–15 km and the 25–30 km race sections (Figure 1). There were sex differences in the percentage of time running at moderate intensity in the 10–15 km and the 15-HM race sections, with males presenting higher values than females. As for the extremely vigorous level of physical intensity, females showed a higher, but not significantly, percentage of time running at moderate intensity than males in the last race section (from 40 km to the finish line) (5.50 ± 20.58% versus 4.41 ± 13.23%, respectively). As expected by the conditions of the activity, the percentage of time running at both sedentary and light intensities was minimum for both males and females. In fact, the time running at the two highest intensity levels (very vigorous and extremely vigorous) represented the 84.95 ± 23.70% of the marathon time for males and the 94.21 ± 9.64% for females (Figure 2). Therefore, running at these high intensities is crucial for achieving the marathon goal time of runners. The differences in absolute percentages denoted that females had a minimum decay rate in their running intensity, while males tended to drop from running at a very high intensity level (mainly at extremely vigorous) to a vigorous intensity level. Int. J. Environ. Res. Public Health 2020, 17, x FOR PEER REVIEW 8 of 15 (orange) and females (blue), and error bars represent the standard deviation of the mean. Mean time spent (± standard deviation) to cover each race sections and the whole race by males (M) and females (F) is showed in the corresponding panel. A Mann–Whitney U test was used for testing sex differences. * p-value < 0.05; ** p-value < 0.01. Nevertheless, males were more time running at vigorous intensity than females, showing significant differences in the 10–15 km and the 25–30 km race sections (Figure 1). There were sex differences in the percentage of time running at moderate intensity in the 10–15 km and the 15-HM race sections, with males presenting higher values than females. As for the extremely vigorous level of physical intensity, females showed a higher, but not significantly, percentage of time running at moderate intensity than males in the last race section (from 40 km to the finish line) (5.50 ± 20.58% versus 4.41 ± 13.23%, respectively). As expected by the conditions of the activity, the percentage of time running at both sedentary and light intensities was minimum for both males and females. In fact, the time running at the two highest intensity levels (very vigorous and extremely vigorous) represented the 84.95 ± 23.70% of the marathon time for males and the 94.21 ± 9.64% for females (Figure 2). Therefore, running at these high intensities is crucial for achieving the marathon goal time of runners. The differences in absolute percentages denoted that females had a minimum decay rate in their running intensity, while males tended to drop from running at a very high intensity level (mainly at extremely vigorous) to a vigorous intensity level. Figure 2. Bar plot showing the percentage of time performing at the two highest intensity levels of physical activity. Bars represent the average values for males (orange) and females (blue). Dots represent each runner included in our study. Mean percentage (±standard deviation) of time performing at these intensity levels is showed in the corresponding panel. A Mann–Whitney U test was used for testing differences between females (F) and males (M). * p-value < 0.05; ** p-value < 0.01. We also compared the evolution of running speed throughout the marathon race (Figure 3A and Table 3). Overall, running speed was higher in males than in females, except for the last section of the race (from 40 km to the finish line) (238.16 ± 51.50 m·min−1 for females versus 227.44 ± 40.16 m·min−1 for males). In fact, the Gap (percentage of the relative sex difference) in running speed was decreasing during the course of the marathon. For the entire marathon distance, the running speed of females was an 11.24% slower than males (Gap = −11.24%). However, a positive Gap value (4.50%) was Figure 2. Bar plot showing the percentage of time performing at the two highest intensity levels of physical activity. Bars represent the average values for males (orange) and females (blue). Dots represent each runner included in our study. Mean percentage (±standard deviation) of time performing at these intensity levels is showed in the corresponding panel. A Mann–Whitney U test was used for testing differences between females (F) and males (M). * p-value < 0.05; ** p-value < 0.01. We also compared the evolution of running speed throughout the marathon race (Figure 3A and Table 3). Overall, running speed was higher in males than in females, except for the last section of the race (from 40 km to the finish line) (238.16 ± 51.50 m·min−1 for females versus 227.44 ± 40.16 m·min−1 for males). In fact, the Gap (percentage of the relative sex difference) in running speed was decreasing Int. J. Environ. Res. Public Health 2020, 17, 6196 8 of 14 during the course of the marathon. For the entire marathon distance, the running speed of females was an 11.24% slower than males (Gap = −11.24%). However, a positive Gap value (4.50%) was obtained in the last race section (from 40 km to the finish line), denoting the faster speed achieved by females compared to males in the last kilometers. Int. J. Environ. Res. Public Health 2020, 17, x FOR PEER REVIEW 9 of 15 obtained in the last race section (from 40 km to the finish line), denoting the faster speed achieved by females compared to males in the last kilometers. Figure 3. Evolution of (A) speed, (B) percentage of average change speed, (C) relative energy consumed per minute, (D) relative energy consumed per kilometer, (E) percentage of maximum oxygen uptake, and (F) cost running above the standing level. Dots represent the average values for males (orange) and females (blue), and error bars represent the standard deviation of the mean. Bars represent the percentage of sex difference (Gap). Mean values (±standard deviation) of each variable analyzed for the whole marathon distance are showed in the corresponding panel. A Mann–Whitney Figure 3. Evolution of (A) speed, (B) percentage of average change speed, (C) relative energy consumed per minute, (D) relative energy consumed per kilometer, (E) percentage of maximum oxygen uptake, and (F) cost running above the standing level. Dots represent the average values for males (orange) and females (blue), and error bars represent the standard deviation of the mean. Bars represent the percentage of sex difference (Gap). Mean values (±standard deviation) of each variable analyzed for the whole marathon distance are showed in the corresponding panel. A Mann–Whitney U test was used for testing sex differences. Bold denotes significant differences in values obtained for the whole marathon distance between females (F) and males (M). Int. J. Environ. Res. Public Health 2020, 17, 6196 9 of 14 Additionally, we evaluated the evolution of ACS (Figure 3B). Results denoted higher changes of speed in females compared to males (6.29 ± 2.58 versus 5.40 ± 2.62, respectively). This observation is mainly caused by the significant increase of running speed made by females in the last section of the race. In fact, ACS values were lower in females compared to males (4.24 ± 1.75 versus 4.37 ± 2.68, respectively) when the ACS was evaluated without taking into account the last race section. That is, running pace of females was more stable than males in the first 40 km of the race and, for that reason, they were able to sprint in the last section of the marathon. Our results denoted significant sex differences in the ACS only in this last section of the race (the unique race section without significantly sex differences in the absolute running speed; Figure 3A). Energy consumption was also compared between females and males. Males consumed 12% to 18% more energy (kilocalories normalized per body mass) per minute than females at all marathon sections (Figure 3C) and in the entire race distance (Table 3). Accelerometry-derived data was also used to estimate the %VO2max sustained throughout the race by each runner, following the methodology previously published by our research group [38]. Females and males consumed a similar %VO2max at the different race sections analyzed (Figure 3E). Besides, very similar overall values were observed in males (80.76 ± 11.51% of VO2max) and females (81.57 ± 7.59% of VO2max) (Table 3). However, it is noted that females needed to maintain a slightly higher %VO2max for running at a lower running intensity level in comparison to males (at very vigorous and at extremely vigorous intensity for females and males, respectively). Running economy was measured by estimating both the energy (kilocalories normalized by body mass) consumed per kilometer and the Cost running above standing (Crnet). As expected for the last section of the race, no differences were observed in running economy between male and female runners, independently of the method used (Figure 3D,F, and Table 3). Females seemed to have better running economy than males in the last race section (from 40 km to the finish line), which is denoted by: (i) females consumed less energy per kilometer than males (0.822 ± 0.187 kcal·kg−1·km−1 versus 0.977 ± 0.210 kcal·kg−1·km−1; p-value = 0.017; Cohen’s d = 0.53), and (ii) females presented a lower Crnet than males (3248 ± 0.794 j·kg−1·m−1 versus 3860 ± 0.875 j·kg−1·m−1; p-value = 0.022; Cohen’s d = 0.51). Maximum Gap values were observed in the last race section for both variables (−18.89% for kcal·kg−1·km−1; and −18.85% for Crnet). In this race section, the negative values of Gap denoted that males consumed more energy per distance (and therefore presented a lower running economy) than females. 4. Discussion This observational study aimed at increasing our understanding on how females and males achieve their marathon goal. In this regard, we focused on analyzing the evolution of several accelerometry- estimated parameters (energy consumption, running intensity distribution, running economy, oxygen uptake), as well as pace-related variables (running speed, average change in speed), throughout a marathon race taking into account runner’s sex. For this purpose, we directly monitored female and male recreational runners throughout the entire marathon distance by using accelerometer-based devices. Moreover, we also collected split times of each runner (provided by the organizers of the Valencia Marathon). Similar to previous studies [4,53], the average running speed was significantly higher in males compared to females. In this study, the difference rate in running speed between males and females was 11.24%. This percentage matched with values obtained in previous studies, which ranged from 8% to 14% [4–6,8,10,17]. However, this percentage of sex difference seems to be lower in elite compared to amateur marathoners. In fact, the female marathon world record (2:14:09, Brigid Kosgei, Chicago 2019) is only 9.32% slower than the male marathon world record (2:01:39, Eliud Kipchoge, Berlin 2018). Sex variability in marathon performance may be explained by the well-known differences in several physiological traits [4,53–55]. Descriptive analyses showed that males presented higher maximum oxygen uptake, body mass index, muscle strength, and lower percentage of body fat than Int. J. Environ. Res. Public Health 2020, 17, 6196 10 of 14 females. Muscle strength has been shown to determine the runner’s ability of displacement and thus running speed. In fact, a lower percentage of sex difference has been observed for swimming speed (6–7%), a physical activity in which the muscle strength component is considerably less crucial for succeeding compared to running [56,57]. However, according to the split times collected of each runner, males were more likely to slow their pace in the last part of the marathon race than females. Males started the marathon running at a 15.53% faster speed than females in the first race section (from start line to 5 km), and this difference rate was decreasing throughout the marathon. In fact, females run at 4.5% faster speed than males in the last marathon section (from 40 km to finish line). Additionally, by analyzing the ACS [9], we were able to confirm that females raced at a more constant pace from the start line to the 40 km as compared to males. This pace strategy may allow females to significantly increase their running speed in the last section of the race, while males were more likely to “hit the wall” in the last kilometers. At this point, we would like to highlight the notably difference in the running speed maintained by males and females in the first section of the race (Gap of −15.53%). This difference may be attributed to the fact that, in races with more than 20,000 participants, runners cannot run the first kilometers freely without difficulty due to the large number of participants and the limited space. This notable sex difference may thus indicate that males started the race being more ambitious, while females adopted a more cautious attitude [54,55,58]. This observation was not previously seen by Nikolaidis and cols (2019) [9], probably because they split the race distance into 10 km sectors and not in 5 km sectors as we did in this work. Having shorter sections allowed us to observe changes in running pace and intensity in greater detail. To further explore sex differences in marathon performance, and taking into account that there is a lack of gold standard for measuring energy consumption in free-living conditions (as a marathon race) [40], we used accelerometer-based devises for estimating the distribution of physical effort throughout the marathon according to runner’s sex [39]. Specifically, we estimated the time running at each of the six-relative intensity levels (sedentary, light, moderate, vigorous, very vigorous and extremely vigorous) in each one of the nine race sections and in the entire marathon. The analysis of physical effort distribution denoted that females tended to race at a lower intensity level than males (females significantly ran more percentage of time at very vigorous intensity than males, who mainly ran at extremely vigorous intensity). In addition, accelerometry-based data allowed us to estimate the energy consumed and the %VO2max sustained per each runner, and afterwards his/her running economy. According to our results, females reported a better efficiency of movement than males in the last section of the marathon (from 40 km to the finish line). That is, a superior energy was demanded by males for running at a given speed the last 2.195 km of the marathon. This may be a consequence of the high physical effort sustained by males in the first part of the marathon, pointing out the importance of controlling physical effort distribution in a marathon race to avoid “hitting the wall”. Running at high intensities has been shown to accelerate glycolytic depletion [55], which may contribute to the decrease of running pace observed in males in the last part of the marathon. Females, however, may use fats as principal energy source maintaining their glycogen stores in muscles thanks to running at less demanding intensities. As stated in lab-based conditions [13], females may present lower respiratory exchange ratio (RER) compared to males, indicating in turn that fat may be the principal fuel source used by females. Future work may be focused on validating accelerometry for RER estimations. Two limitations are noteworthy in the present study. Firstly, we are aware about the low number of females included in our population (15.91%). However, this percentage is even higher than the rate of females, aged 30 to 45 years, finishing the Valencia Marathon in 2016 (13.16%). Higher effort should be done in future studies for increasing the number of females collected. The second limitation is that values of accelerometer-based parameters analyzed in this study were merely estimations. No gold standard method is available yet to perform a direct measurement of VO2 consumed by a runner in free-living conditions. We may assume a plausible maximum error of 10% in our estimations. Int. J. Environ. Res. Public Health 2020, 17, 6196 11 of 14 In summary, thanks to the accelerometry-based and pace-based data collected, this study reveals how female and male middle-aged amateur marathoners face a marathon in terms of pacing, running strategy, running intensity distribution, energy consumption and running economy. The use of accelerometer devices for monitoring runners allowed us to perform an individualized assessment in the context of free-living movement. In general, females showed a good control of physical effort throughout the marathon, while the running intensity distribution and pacing of males were not so well-balanced. Subsequently, an increased decay of running pace in the last part of the marathon was observed for males. Results may reflect well-known sex differences in physiology (i.e., muscle strength, fat metabolism, VO2max), and in running strategy approach (i.e., females run at a more conservative intensity level in the first part of the marathon compared to males). 5. Conclusions Compared to males, females maintained a more stable pace and ran at less demanding running intensities throughout the marathon, limiting the decay of running pace in the last part of the race. Together with previous studies, the results obtained after analyzing a huge number of variables suggest that the steady pacing of females may be because of the following reasons: • Females may manage energy during the race more efficiently than males [7,55]. • Females may make better decisions in terms of pacing strategy than males [6,48,54]. • Females typically use more fat than carbohydrates during endurance exercise compared to males [14,59,60]. • Females tend to preserve muscle strength and have less neuromuscular fatigue than males at the end of the marathon [61]. Supplementary Materials: The following are available online at http://www.mdpi.com/1660-4601/17/17/6196/s1. Table S1: Evaluation of running intensity distribution and estimation of calories consumed by male runners based on accelerometry data. Table S2: Evaluation of running intensity distribution and estimation of calories consumed by male runners based on accelerometry data. File S1: Raw data of the study. Data Availability: All data generated or analyzed during this study are included in this published article (and its Supplementary Materials). Author Contributions: C.H. (Carlos Hernando) and B.H. contributed to conception and design of the study, article drafting and critical revision of the article. C.H. (Carlos Hernando) and C.H. (Carla Hernando) contributed to data curation, analysis and interpretation. C.H. (Carlos Hernando), I.M.-N., E.C.-B. and N.P. contributed to data collection and critical revision of the article. C.H. (Carlos Hernando), I.M.-N. and E.C.-B. contributed to funding acquisition. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by the Fundacion Trinidad Alfonso, the Vithas-Nisa Hospitals group and the Sociedad Deportiva Correcaminos. Acknowledgments: We are grateful to all the stuff involved in the organization of the Valencia Marathon Fundacion Trinidad Alfonso EDP 2016, and all marathoners and volunteers participating in this study. Conflicts of Interest: The authors declare no conflict of interest. References 1. Ahmadyar, B.; Rüst, C.A.; Rosemann, T.; Knechtle, B. Participation and performance trends in elderly marathoners in four of the world’s largest marathons during 2004–2011. SpringerPlus 2015, 4, 465. [CrossRef] [PubMed] 2. Aschmann, A.; Knechtle, B.; Onywera, V.O.; Nikolaidis, P.T. Pacing Strategies in the New York City Marathon—Does Nationality of Finishers Matter? (Request PDF). Asian J. Sports Med. 2018. [CrossRef] 3. Maratón de Valencia Fundación Trinidad Alfonso EDP. Available online: https://www.valenciaciudaddelrunning. com/maraton/ediciones-anteriores-maraton/ (accessed on 18 November 2019). 4. Joyner, M.J. Physiological limits to endurance exercise performance: Influence of sex. J. Physiol. 2017, 595, 2949–2954. [CrossRef] 5. Eichenberger, E.; Knechtle, B.; Rüst, C.A.; Rosemann, T.; Lepers, R. Age and sex interactions in mountain ultramarathon running—The Swiss Alpine Marathon. Open Access J. Sports Med. 2012, 3, 73–80. [CrossRef] Int. J. Environ. Res. Public Health 2020, 17, 6196 12 of 14 6. Nikolaidis, P.T.; Onywera, V.O.; Knechtle, B. Running Performance, Nationality, Sex, and Age in the 10-km, Half-Marathon, Marathon, and the 100-km Ultramarathon IAAF 1999–2015. J. Strength Cond. Res. 2017, 31, 2189–2207. [CrossRef] 7. Senefeld, J.; Smith, C.; Hunter, S.K. Sex Differences in Participation, Performance, and Age of Ultramarathon Runners. Int. J. Sports Physiol. Perform. 2016, 11, 635–642. [CrossRef] [PubMed] 8. Zavorsky, G.S.; Tomko, K.A.; Smoliga, J.M. Declines in marathon performance: Sex differences in elite and recreational athletes. PLoS ONE 2017, 12, e0172121. [CrossRef] 9. Nikolaidis, P.T.; Rosemann, T.; Cuk, I.; Knechtle, B. Performance and Pacing of Age Groups in Half-Marathon and Marathon. Int. J. Environ. Res. Public Health 2019, 16, 1777. [CrossRef] 10. Cheuvront, S.N.; Carter, R.; Deruisseau, K.C.; Moffatt, R.J. Running performance differences between men and women:an update. Sports Med. Auckl. N. Z. 2005, 35, 1017–1024. [CrossRef] 11. Helgerud, J.; Støren, Ø.; Hoff, J. Are there differences in running economy at different velocities for well-trained distance runners? Eur. J. Appl. Physiol. 2010, 108, 1099–1105. [CrossRef] 12. di Prampero, P.E.; Atchou, G.; Brückner, J.C.; Moia, C. The energetics of endurance running. Eur. J. Appl. Physiol. 1986, 55, 259–266. [CrossRef] [PubMed] 13. Loftin, M.; Sothern, M.; Tuuri, G.; Tompkins, C.; Koss, C.; Bonis, M. Gender comparison of physiologic and perceptual responses in recreational marathon runners. Int. J. Sports Physiol. Perform. 2009, 4, 307–316. [CrossRef] [PubMed] 14. Tarnopolsky, M.A.; Atkinson, S.A.; Phillips, S.M.; MacDougall, J.D. Carbohydrate loading and metabolism during exercise in men and women. J. Appl. Physiol. (Bethesda Md. 1985) 1995, 78, 1360–1368. [CrossRef] 15. Barnes, K.R.; Kilding, A.E. Running economy: Measurement, norms, and determining factors. Sports Med. Open 2015, 1, 8. [CrossRef] [PubMed] 16. di Prampero, P.E. Factors limiting maximal performance in humans. Eur. J. Appl. Physiol. 2003, 90, 420–429. [CrossRef] 17. Saunders, P.U.; Pyne, D.B.; Telford, R.D.; Hawley, J.A. Factors Affecting Running Economy in Trained Distance Runners. Sports Med. 2004, 34, 465–485. [CrossRef] 18. Nikolaidis, P.T.; Rosemann, T.; Knechtle, B. A Brief Review of Personality in Marathon Runners: The Role of Sex, Age and Performance Level. Sports (Basel Switz.) 2018, 6, 99. [CrossRef] 19. McLellan, T.M.; Skinner, J.S. Submaximal endurance performance related to the ventilation thresholds. Can. J. Appl. Sport Sci. 1985, 10, 81–87. 20. Ainsworth, B.E.; Haskell, W.L.; Leon, A.S.; Jacobs, D.R.; Montoye, H.J.; Sallis, J.F.; Paffenbarger, R.S. Compendium of physical activities: Classification of energy costs of human physical activities. Med. Sci. Sports Exerc. 1993, 25, 71–80. [CrossRef] 21. Ainsworth, B.E.; Haskell, W.L.; Whitt, M.C.; Irwin, M.L.; Swartz, A.M.; Strath, S.J.; O’Brien, W.L.; Bassett, D.R.; Schmitz, K.H.; Emplaincourt, P.O.; et al. Compendium of physical activities: An update of activity codes and MET intensities. Med. Sci. Sports Exerc. 2000, 32, S498–S504. [CrossRef] 22. Ainsworth, B.E.; Haskell, W.L.; Herrmann, S.D.; Meckes, N.; Bassett, D.R.; Tudor-Locke, C.; Greer, J.L.; Vezina, J.; Whitt-Glover, M.C.; Leon, A.S. 2011 Compendium of Physical Activities: A second update of codes and MET values. Med. Sci. Sports Exerc. 2011, 43, 1575–1581. [CrossRef] 23. Davies, M.J.; Mahar, M.T.; Cunningham, L.N. Running economy: Comparison of body mass adjustment methods. Res. Q. Exerc. Sport 1997, 68, 177–181. [CrossRef] [PubMed] 24. Klepin, K.; Wing, D.; Higgins, M.; Nichols, J.; Godino, J.G. Validity of Cardiorespiratory Fitness Measured with Fitbit Compared to VO2max. Med. Sci. Sports Exerc. 2019. [CrossRef] [PubMed] 25. Sperlich, B.; Aminian, K.; Düking, P.; Holmberg, H.-C. Editorial: Wearable Sensor Technology for Monitoring Training Load and Health in the Athletic Population. Front. Physiol. 2019, 10, 1520. [CrossRef] [PubMed] 26. Kinnunen, H.; Häkkinen, K.; Schumann, M.; Karavirta, L.; Westerterp, K.R.; Kyröläinen, H. Training-induced changes in daily energy expenditure: Methodological evaluation using wrist-worn accelerometer, heart rate monitor, and doubly labeled water technique. PLoS ONE 2019, 14, e0219563. [CrossRef] [PubMed] 27. Tudor-Locke, C.; Aguiar, E.J.; Han, H.; Ducharme, S.W.; Schuna, J.M.; Barreira, T.V.; Moore, C.C.; Busa, M.A.; Lim, J.; Sirard, J.R.; et al. Walking cadence (steps/min) and intensity in 21-40 year olds: CADENCE-adults. Int. J. Behav. Nutr. Phys. Act. 2019, 16, 8. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2020, 17, 6196 13 of 14 28. Suchert, V.; Isensee, B.; Sargent, J.; Weisser, B.; Hanewinkel, R.; Läuft Study Group. Prospective effects of pedometer use and class competitions on physical activity in youth: A cluster-randomized controlled trial. Prev. Med. 2015, 81, 399–404. [CrossRef] 29. Clevenger, K.A.; Molesky, M.J.; Vusich, J.; Montoye, A.H.K. Free-Living Comparison of Physical Activity and Sleep Data from Fitbit Activity Trackers Worn on the Dominant and Nondominant Wrists. Meas. Phys. Educ. Exerc. Sci. 2019, 23, 194–204. [CrossRef] 30. Montoye, H.J.; Washburn, R.; Servais, S.; Ertl, A.; Webster, J.G.; Nagle, F.J. Estimation of energy expenditure by a portable accelerometer. Med. Sci. Sports Exerc. 1983, 15, 403–407. [CrossRef] 31. Smith, M.P.; Horsch, A.; Standl, M.; Heinrich, J.; Schulz, H. Uni- and triaxial accelerometric signals agree during daily routine, but show differences between sports. Sci. Rep. 2018, 8, 15055. [CrossRef] 32. Strath, S.J.; Kaminsky, L.A.; Ainsworth, B.E.; Ekelund, U.; Freedson, P.S.; Gary, R.A.; Richardson, C.R.; Smith, D.T.; Swartz, A.M.; American Heart Association Physical Activity Committee of the Council on Lifestyle and Cardiometabolic Health and Cardiovascular, Exercise, Cardiac Rehabilitation and Prevention Committee of the Council on Clinical Cardiology, and Council. Guide to the Assessment of Physical Activity: Clinical and Research Applications A Scientific Statement from the American Heart Association. Circulation 2013. [CrossRef] [PubMed] 33. Stiles, V.H.; Pearce, M.; Moore, I.S.; Langford, J.; Rowlands, A.V. Wrist-worn Accelerometry for Runners: Objective Quantification of Training Load. Med. Sci. Sports Exerc. 2018, 50, 2277. [CrossRef] [PubMed] 34. Arvidsson, D.; Fridolfsson, J.; Börjesson, M.; Andersen, L.B.; Ekblom, Ö.; Dencker, M.; Brønd, J.C. Re-examination of accelerometer data processing and calibration for the assessment of physical activity intensity. Scand. J. Med. Sci. Sports 2019, 29, 1442–1452. [CrossRef] [PubMed] 35. Esliger, D.W.; Rowlands, A.V.; Hurst, T.L.; Catt, M.; Murray, P.; Eston, R.G. Validation of the GENEA Accelerometer. Med. Sci. Sports Exerc. 2011, 43, 1085–1093. [CrossRef] 36. Menai, M.; van Hees, V.T.; Elbaz, A.; Kivimaki, M.; Singh-Manoux, A.; Sabia, S. Accelerometer assessed moderate-to-vigorous physical activity and successful ageing: Results from the Whitehall II study. Sci. Rep. 2017, 8, 45772. [CrossRef] 37. Welch, W.A.; Bassett, D.R.; Thompson, D.L.; Freedson, P.S.; Staudenmayer, J.W.; John, D.; Steeves, J.A.; Conger, S.A.; Ceaser, T.; Howe, C.A.; et al. Classification accuracy of the wrist-worn gravity estimator of normal everyday activity accelerometer. Med. Sci. Sports Exerc. 2013, 45, 2012–2019. [CrossRef] 38. Migueles, J.H.; Cadenas-Sanchez, C.; Tudor-Locke, C.; Löf, M.; Esteban-Cornejo, I.; Molina-Garcia, P.; Mora-Gonzalez, J.; Rodriguez-Ayllon, M.; Garcia-Marmol, E.; Ekelund, U.; et al. Comparability of published cut-points for the assessment of physical activity: Implications for data harmonization. Scand. J. Med. Sci. Sports 2019, 29, 566–574. [CrossRef] 39. Hernando, C.; Hernando, C.; Collado, E.J.; Panizo, N.; Martinez-Navarro, I.; Hernando, B. Establishing cut-points for physical activity classification using triaxial accelerometer in middle-aged recreational marathoners. PLoS ONE 2018, 13, e0202815. [CrossRef] 40. Hernando, C.; Hernando, C.; Martinez-Navarro, I.; Collado-Boira, E.; Panizo, N.; Hernando, B. Estimation of energy consumed by middle-aged recreational marathoners during a marathon using accelerometry-based devices. Sci. Rep. 2020, 10, 1523. [CrossRef] 41. Bernat-Adell, M.D.; Collado-Boira, E.J.; Moles-Julio, P.; Panizo-González, N.; Martínez-Navarro, I.; Hernando-Fuster, B.; Hernando-Domingo, C. Recovery of Inflammation, Cardiac, and Muscle Damage Biomarkers After Running a Marathon. J. Strength Cond. Res. 2019. [CrossRef] 42. Abe, D.; Fukuoka, Y.; Horiuchi, M. Economical Speed and Energetically Optimal Transition Speed Evaluated by Gross and Net Oxygen Cost of Transport at Different Gradients. PLoS ONE 2015, 10, e0138154. [CrossRef] [PubMed] 43. Byrne, N.M.; Hills, A.P.; Hunter, G.R.; Weinsier, R.L.; Schutz, Y. Metabolic equivalent: One size does not fit all. J. Appl. Physiol. (Bethesda Md. 1985) 2005, 99, 1112–1119. [CrossRef] [PubMed] 44. Lazzer, S.; Taboga, P.; Salvadego, D.; Rejc, E.; Simunic, B.; Narici, M.V.; Buglione, A.; Giovanelli, N.; Antonutto, G.; Grassi, B.; et al. Factors affecting metabolic cost of transport during a multi-stage running race. J. Exp. Biol. 2014, 217, 787–795. [CrossRef] [PubMed] 45. Skinner, J.S.; McLellan, T.M.; McLellan, T.H. The transition from aerobic to anaerobic metabolism. Res. Q. Exerc. Sport 1980, 51, 234–248. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2020, 17, 6196 14 of 14 46. Betts, J.A.; Smith, H.A.; Johnson-Bonson, D.A.; Ellis, T.I.; Dagnall, J.; Hengist, A.; Carroll, H.; Thompson, D.; Gonzalez, J.T.; Afman, G.H. The Energy Cost of Sitting versus Standing Naturally in Man. Med. Sci. Sports Exerc. 2018. [CrossRef] [PubMed] 47. Lazzer, S.; Salvadego, D.; Rejc, E.; Buglione, A.; Antonutto, G.; di Prampero, P.E. The energetics of ultra-endurance running. Eur. J. Appl. Physiol. 2012, 112, 1709–1715. [CrossRef] 48. Nikolaidis, P.T.; ´Cuk, I.; Knechtle, B. Pacing of Women and Men in Half-Marathon and Marathon Races. Medicina (Kaunas Lith.) 2019, 55, 14. [CrossRef] 49. de Blas, X.; Padullés, J.M.; López del Amo, J.L.; Guerra-Balic, M. Creation and Validation of Chronojump- Boscosystem: A Free Tool to Measure Vertical Jumps. Int. J. Sport Sci. 2012, 8, 334–356. [CrossRef] 50. Fritz, C.O.; Morris, P.E.; Richler, J.J. Effect size estimates: Current use, calculations, and interpretation. J. Exp. Psychol. Gen. 2012, 141, 2–18. [CrossRef] 51. Lenhard, W.; Lenhard, A. Calculation of Effect Sizes. Psychometrica 2016. [CrossRef] 52. Thomas, J.R.; Nelson, J.K.; Silverman, S.J. Research Methods in Physical Activity; Human Kinetics: Champaign, IL, USA, 2015; ISBN 978-1-4925-8528-2. 53. Nikolaidis, P.T.; Rosemann, T.; Knechtle, B. Force-Velocity Characteristics, Muscle Strength, and Flexibility in Female Recreational Marathon Runners. Front. Physiol. 2018, 9, 1563. [CrossRef] [PubMed] 54. Deaner, R.O.; Carter, R.E.; Joyner, M.J.; Hunter, S.K. Men are more likely than women to slow in the marathon. Med. Sci. Sports Exerc. 2015, 47, 607–616. [CrossRef] [PubMed] 55. Knechtle, B.; Nikolaidis, P.T. Physiology and Pathophysiology in Ultra-Marathon Running. Front. Physiol. 2018, 9, 634. [CrossRef] 56. Eichenberger, E.; Knechtle, B.; Knechtle, P.; Rüst, C.A.; Rosemann, T.; Lepers, R. Best performances by men and women open-water swimmers during the “English Channel Swim” from 1900 to 2010. J. Sports Sci. 2012, 30, 1295–1301. [CrossRef] 57. Senefeld, J.; Joyner, M.J.; Stevens, A.; Hunter, S.K. Sex differences in elite swimming with advanced age are less than marathon running. Scand. J. Med. Sci. Sports 2016, 26, 17–28. [CrossRef] 58. Frick, B. Gender Differences in Competitive Orientations: Empirical Evidence from Ultramarathon Running. J. Sports Econ. 2011, 12, 317–340. [CrossRef] 59. Tarnopolsky, L.J.; MacDougall, J.D.; Atkinson, S.A.; Tarnopolsky, M.A.; Sutton, J.R. Gender differences in substrate for endurance exercise. J. Appl. Physiol. (Bethesda Md. 1985) 1990, 68, 302–308. [CrossRef] [PubMed] 60. Tarnopolsky, M.A. Sex differences in exercise metabolism and the role of 17-beta estradiol. Med. Sci. Sports Exerc. 2008, 40, 648–654. [CrossRef] [PubMed] 61. Temesi, J.; Arnal, P.J.; Rupp, T.; Féasson, L.; Cartier, R.; Gergelé, L.; Verges, S.; Martin, V.; Millet, G.Y. Are Females More Resistant to Extreme Neuromuscular Fatigue? Med. Sci. Sports Exerc. 2015, 47, 1372–1382. [CrossRef] [PubMed] © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Using Accelerometry for Evaluating Energy Consumption and Running Intensity Distribution Throughout a Marathon According to Sex.

08-26-2020

Hernando, Carlos,Hernando, Carla,Martinez-Navarro, Ignacio,Collado-Boira, Eladio,Panizo, Nayara,Hernando, Barbara

eng

PMC6066218

RESEARCH ARTICLE Manipulating graded exercise test variables affects the validity of the lactate threshold and _VO2peak Nicholas A. Jamnick1☯*, Javier Botella1☯, David B. Pyne2,3☯, David J. Bishop1,4☯ 1 Institute for Health and Sport, College of Sport and Exercise Science, Victoria University, Melbourne, Australia, 2 Australian Institute of Sport, Canberra, Australia, 3 Research Institute for Sport and Exercise (UCRISE), University of Canberra, Canberra, Australia, 4 School of Medical and Health Sciences, Edith Cowan University, Joondalup, Australia ☯ These authors contributed equally to this work. * [email protected] Abstract Background To determine the validity of the lactate threshold (LT) and maximal oxygen uptake ( _VO2max) determined during graded exercise test (GXT) of different durations and using different LT calculations. Trained male cyclists (n = 17) completed five GXTs of varying stage length (1, 3, 4, 7 and 10 min) to establish the LT, and a series of 30-min constant power bouts to estab- lish the maximal lactate steady state (MLSS). _VO2 was assessed during each GXT and a subsequent verification exhaustive bout (VEB), and 14 different LTs were calculated from four of the GXTs (3, 4, 7 and 10 min)—yielding a total 56 LTs. Agreement was assessed between the highest _VO2 measured during each GXT ( _VO2peak) as well as between each LT and MLSS. _VO2peak and LT data were analysed using mean difference (MD) and intraclass correlation (ICC). Results The _VO2peak value from GXT1 was 61.0 ± 5.3 mL.kg-1.min-1 and the peak power 420 ± 55 W (mean ± SD). The power at the MLSS was 264 ± 39 W. _VO2peak from GXT3, 4, 7, 10 underesti- mated _VO2peak by ~1–5 mL.kg-1.min-1. Many of the traditional LT methods were not valid and a newly developed Modified Dmax method derived from GXT4 provided the most valid esti- mate of the MLSS (MD = 1.1 W; ICC = 0.96). Conclusion The data highlight how GXT protocol design and data analysis influence the determination of both _VO2peak and LT. It is also apparent that _VO2max and LT cannot be determined in a sin- gle GXT, even with the inclusion of a VEB. PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 1 / 21 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Jamnick NA, Botella J, Pyne DB, Bishop DJ (2018) Manipulating graded exercise test variables affects the validity of the lactate threshold and _VO2peak. PLoS ONE 13(7): e0199794. https:// doi.org/10.1371/journal.pone.0199794 Editor: Øyvind Sandbakk, Norwegian University of Science and Technology, NORWAY Received: April 30, 2018 Accepted: June 13, 2018 Published: July 30, 2018 Copyright: © 2018 Jamnick et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: The data underlying this study have been uploaded to the Open Science Framework and are accessible using the following link: https://osf.io/293ns/. Funding: Funding was provided by the Graduate Research Office (PhD Student Budget) at Victoria University. The funder had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. Introduction Sampling of expired gas and blood data during a graded exercise test (GXT) to exhaustion per- mits identification of the gas exchange threshold (GET), the respiratory compensation point (RCP), the lactate threshold (LT), and maximal oxygen uptake ( _VO2max). These indices can dis- tinguish cardiorespiratory fitness, and demarcate the domains of exercise [1, 2] that can be used to prescribe exercise and to optimize training stimuli [3–6]. However, despite the popu- larity of these indices, the methods used to determine them can differ substantially and there has been little systematic investigation of their validity [7–9]. The recommended duration of a GXT to assess _VO2max is 8 to 12 minutes [10–13]. How- ever, there is little consensus on an appropriate GXT protocol design, including duration, stage length, or number of stages, needed to establish the LT. A stage length of at least 3 min- utes has been recommended [13], although an 8-minute stage length has also been suggested for blood lactate concentrations to stabilize [14]. The number of stages and GXT duration will depend on the starting intensity and power increments. Power is typically increased identically [15], regardless of sex or fitness, leading to a heterogenous GXT duration and number of stages completed [16]. A customized approach to LT testing has been recommended to ensure a more homogenous GXT duration [17]. More than 25 methods have been proposed to calculate the LT [18]; these include the power preceding a rise in blood lactate concentration of more than 0.5, 1.0 or 1.5 mmol.L-1 from baseline [19], the onset of a fixed blood lactate accumulation (OBLA) ranging from 2.0 to 4.0 mmol.L-1 [20, 21], or the use of curve fitting procedures such as the Dmax or modi- fied Dmax methods (ModDmax) [22, 23]. However, many of these ‘accepted’ methods are influenced by GXT protocol design [8, 24] and their underlying validity has not been reported. Assessing the validity of a measurement requires comparison with a criterion measure. The maximal lactate steady state (MLSS) represents the highest intensity where blood lac- tate appearance and disappearance is in equilibrium and where energy demand is ade- quately met by oxidative phosphorylation [25]. Exercise performed above the MLSS results in accelerated blood lactate appearance and it has therefore been suggested as an appropriate criterion measure for the LT [25, 26]. The primary advantages of the MLSS test include its independence of participant effort, it’s submaximal and is reliable [27]. However, the disadvantage is the necessity of multiple laboratory visits and that it yields only one index of performance. _VO2max is considered the “gold standard” for assessing cardiorespiratory fitness [28] and the highest recorded _VO2 from a GXT is often accepted as the _VO2max [10]. Establish- ing the LT requires a GXT that typically exceeds 20 minutes [13]; however, in these instances the highest _VO2 may underestimate the _VO2max [12] and is termed _VO2peak. Recently, the use of a verification exhaustive bout (VEB) has been recommended to con- firm the _VO2max. However, it is unknown if a VEB performed after a longer duration GXT provides a valid estimate of _VO2max. The aim of this study was to determine the validity of the LT and _VO2max derived from a sin- gle visit GXT. We hypothesized that our results would yield one or more GXT stage length and LT calculation method combination that provides a valid estimation of the criterion measure of the LT (i.e., MLSS). We also hypothesized the highest _VO2 measured during longer duration GXTs would underestimate _VO2max and that the highest _VO2 value measured during each VEB would be similar to the _VO2peak measured during the 8- to 12-minute GXT. Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 2 / 21 Abbreviations: LT, lactate threshold; GXT, graded exercise testing; _VO2max, maximal oxygen uptake; OBLA, onset of fixed blood lactate accumulation; MLSS, maximal lactate steady state; VEB, verification exhaustive bout; BMI, body mass index; _Wmax, maximum power output; _VO2pea, highest measured oxygen uptake value; RCP, respiratory compensation point; RCPMLSS, estimate of the maximal lactate steady state via respiratory compensation point; B + mmol.L-1, blood lactate concentration increases above baseline value(s); ModDmax, modified Dmax; Exp-Dmax, exponential Dmax; Log-Poly-ModDmax, Log-log Modified Dmax; Log-Exp-ModDmax, Log-log Exponential Modified Dmax; CV, coefficient of the variation; ES, effect size; r, Pearson product moment correlation; ICC, intraclass correlation; SEM, standard error of the measurement. Materials and methods Ethical approval All procedures were performed in accordance with the ethical standards of the institutional and/or national research committee, and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. Participants/Experimental design Seventeen trained male cyclists ( _VO2max 62.1 ± 5.8 mL.kg-1.min-1, age 36.2 ± 7.4 years, body mass index (BMI): 24.1 ± 2.0 kg.m-2) volunteered for this study which required 7 to 10 visits to the laboratory. Informed consent was obtained from all individual participants included in the study. Visit one included risk stratification using the American College of Sports Medicine Risk Stratification guidelines [29], written informed consent, self-reported physical activity rating (PA-R) [30], measurement of height and body mass, and completion of a cycling GXT with 1-minute stages (GXT1) followed by a VEB. The remaining visits consisted of four cycling GXTs with varying stage length (3-, 4-, 7- and 10-min stages) and a series of 30-min constant power bouts to establish the MLSS. The GXTs and constant power bouts were performed in an alternating order and the order of the GXTs was randomised. Prior to each GXT and the con- stant power bouts a 5-min warm up was administered at a self-selected power followed by 5 min of passive rest. Participants performed each test at their preferred cadence determined during the initial visit. Antecubital venous blood (1.0 mL) was sampled during all visits (excluding GXT1) at rest, and at the end of every stage during the GXTs or every 5 min during the constant power exercise bouts. All participants self-reported abstaining from the consump- tion of alcohol and caffeine or engaging in heavy exercise 24 h prior to each visit. Participants were given at least 48 h between visits and all tests were completed within 6 weeks. The Victo- ria University Human Research Ethics Committee approved all procedures (HRE 017–035). Equipment/Instruments All exercise testing was conducted using an electronically-braked cycle ergometer (Lode Excal- ibur v2.0, The Netherlands). A metabolic analyser (Quark Cardiopulmonary Exercise Testing, Cosmed, Italy) was used to assess oxygen uptake ( _VO2) on a breath-by-breath basis, and heart rate was measured throughout all tests. Antecubital venous blood was analysed using a blood lactate analyser (YSI 2300 STAT Plus, YSI, USA). GXTs with verification exhaustive bout Demographic data, PA-R, and measurements of height and body mass were used to estimate _VO2max [31] and maximum power output _Wmax [30, 32]. Est:VO2max ¼ 56:363 þ ð1:921 x PA were designed for each of the remaining GXTs based on a percentage of the measured _Wmax from GXT1. The predicted _Wmax was 80%, 77%, 72% and 70% for GXT3, GXT4, GXT7, and GXT10, respectively. The target number of stages for each participant was nine; the initial stage and subsequent stages of the remaining GXTs were determined using the following equations: Stage 1 Power ¼ Predicted _Wmax  0:25 Eq 3 Subsequent power increments ¼ ðPredicted _Wmax 6. Exponential Dmax (Exp-Dmax): The point on the exponential plus-constant regression curve that yielded the maximum perpendicular distance to the straight line formed by the two end points of the curve [41, 42]. 7. Log-log Modified Dmax (Log-Poly-ModDmax): The intensity at the point on the third order polynomial regression curve that yielded the maximal perpendicular distance to the straight line formed by the intensity associated with the log-log LT and the final lactate point. 8. Log-log Exponential Modified Dmax method (Log-Exp-ModDmax): The intensity at the point on the exponential plus-constant regression curve that yielded the maximal perpen- dicular distance to the straight line formed by the intensity associated with the log-log LT and the final lactate point. 9. RCP: refer to Constant Power Exercise Bouts to Establish the Maximal Lactate Steady State method section. 10. The estimated MLSS was based on a regression equation based on the RCP from GXT1 (RCPMLSS) (Eq 5). Data analysis Breath-by-breath data were edited individually with values greater than three standard devia- tions from the mean excluded [43]. The data was interpolated on a second-by-second basis and averaged into 5- and 30-s bins [44, 45]. The highest measured _VO2 value from every GXT and VEB was determined as the highest 20-s rolling average. The _VO2max was computed as the Fig 1. Representative blood lactate curve with 14 LTs calculated from GXT4 (participant #9). The power of the MLSS was 302 W and the blood lactate concentration was 2.85 mmol.L-1. Log-log = power at the intersection of two linear lines with the lowest residual sum of squares; log = using the log-log method as the point of the initial data point when calculating the Dmax or Modified Dmax; poly = Modified Dmax method calculated using a third order polynomial regression equation; exp = Modified Dmax method calculated using a constant plus exponential regression equation; OBLA = onset of blood lactate accumulation; B + absolute value = the intensity where blood lactate increases above baseline. https://doi.org/10.1371/journal.pone.0199794.g001 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 5 / 21 highest _VO2 measured from any GXT or VEB. The _VO2peak for each GXT was defined as the highest measured _VO2 from either the GXT or the subsequent VEB. The _Wmax for every GXT was determined as the power from the last completed stage plus the time completed in the subsequent stage multiplied by the slope (Eq 6). The _VO2 response at the MLSS was determined by the average _VO2 value during the last two minutes of the 30-minute constant power bout. _Wmax ¼ Power of Last Stage ðWÞ þ ½slope ðW:s Table 1. The mean ± standard deviation (SD) of the 14 lactate thresholds calculated from the 4 prolonged graded exercise tests (i.e., GXT3, GXT4, GXT7 and GXT10), and the respiratory compensation point (RCP) and the maximal lactate steady state (MLSS) estimated from the RCP (RCPMLSS) calculated from GXT1. GXT3 GXT4 GXT7 GXT10 Log-log LT Mean SD (W) 211 ± 43 202 ± 38 200 ± 40 196 ± 41 MD (W) 53.1 62.8 64.8 68.3 r 0.84 0.89 0.87 0.78 ES 1.28 1.63 1.62 1.70 OBLA 2.0 Mean SD (W) 262 ± 40 249 ± 39 247 ± 39 245 ± 37 MD (W) 2.1 15.1 17.3 19.6 r 0.86 0.94 0.94 0.93 ES -0.05 -0.38 -0.44 -0.50 OBLA 2.5 Mean SD (W) 276 ± 42 262 ± 40 258 ± 40 255 ± 38 MD (W) -11.9 2.0 6.7 9.2 r 0.89 0.95 0.94 0.93 ES 0.30 -0.05 -0.17 -0.23 OBLA 3.0 Mean SD (W) 288 ± 43 273 ± 41 267 ± 41 264 ± 39 MD (W) -23.2 -8.8 -2.2 0.4 r 0.90 0.96 0.95 0.93 ES 0.59 0.22 0.06 -0.01 OBLA 3.5 Mean SD (W) 297 ± 45 282 ± 41 274 ± 41 272 ± 40 MD (W) -32.8 -18.1 -10.0 -7.3 r 0.91 0.96 0.95 0.93 ES 0.83 0.46 0.25 0.19 OBLA 4.0 Mean SD (W) 306 ± 46 291 ± 42 281 ± 42 279 ± 41 MD (W) -41.3 -26.3 -16.8 -14.2 r 0.91 0.97 0.95 0.93 ES 1.05 0.67 0.43 0.36 Baseline + 0.5 Mean SD (W) 235 ± 38 229 ± 40 228 ± 41 225 ± 37 MD (W) 29.4 35.6 36.6 39.5 r 0.74 0.81 0.83 0.82 ES -0.75 -0.90 -0.93 -1.00 Baseline + 1.0 Mean SD (W) 255 ± 39 239 ± 40 236 ± 39 235 ± 39 MD (W) 9.5 25.3 27.9 29.1 r 0.88 0.92 0.93 0.91 ES -0.24 -0.64 -0.71 -0.74 Baseline + 1.5 Mean SD (W) 270 ± 41 254 ± 41 250 ± 39 248 ± 39 MD (W) -6.0 10.1 14.7 16.8 r 0.90 0.94 0.94 0.92 ES 0.15 -0.26 -0.37 -0.43 Dmax Mean SD (W) 246 ± 34 232 ± 36 223 ± 31 216 ± 33 MD (W) 18.6 31.9 41.6 48.8 r 0.94 0.97 0.96 0.95 ES -0.47 -0.81 -1.06 -1.24 Modified Dmax Mean SD (W) 278 ± 37 267 ± 39 255 ± 40 248 ± 37 MD (W) -13.2 -2.9 9.7 15.9 r 0.90 0.91 0.93 0.92 ES 0.33 0.07 -0.25 -0.40 Log-Poly-MDmax Mean SD (W) 280 ± 42 265 ± 42 255 ± 39 248 ± 40 MD (W) -15.5 -1.1 9.5 16.5 (Continued) Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 7 / 21 total) and the MLSS (all log-log methods were excluded given an ES > 1.0). Ten of the calcu- lated LTs and the RCPMLSS met our inclusion criteria for final analysis—detailed comparisons with the MLSS are provided in Table 3 and Fig 3. Figs 3–7 shows Bland-Altman plots of the 11 estimations included in our analysis; the newly developed ModDmax LT calculations (Fig 5 Panel C and D; Fig 6 Panel C) had the lowest limits of agreement with the MLSS. The log-log polynomial modified Dmax (Log-Poly-ModDmax) method derived from GXT4 provided the best estimation of the MLSS (Fig 5 Panel C). There was an inverse relationship between the power calculated for each of the 14 LTs and stage length (Tables 1 and 4). _Wmax and _VO2max There was an inverse relationship between GXT duration and both _Wmax and _VO2peak (Table 5). The _VO2peak values derived from GXT3 and GXT4 were similar to the _VO2peak mea- sured during GXT1 (Table 6); however, the values were outside the variability of the measure- ment (CV > 3%) [27]. _VO2peak values from GXT1 and the corresponding VEB had the highest agreement (MD = 0.5 mL.kg-1.min-1, ICC = 0.96, SEM = 1.1 mL.kg-1.min-1 and CV = 2.0%) compared with any GXT and corresponding VEB. The remaining GXTs and corresponding Table 1. (Continued) GXT3 GXT4 GXT7 GXT10 r 0.94 0.96 0.96 0.92 ES 0.39 0.03 -0.24 -0.42 Exp-Dmax Mean SD (W) 256 ± 35 243 ± 36 234 ± 34 228 ± 35 MD (W) 8.0 21.8 30.8 36.8 r 0.92 0.97 0.96 0.94 ES -0.20 -0.55 -0.78 -0.93 Log-Exp-MDmax Mean SD (W) 286 ± 42 271 ± 42 260 ± 39 253 ± 40 MD (W) -21.7 -7.0 4.3 11.1 r 0.94 0.97 0.96 0.93 ES 0.55 0.18 -0.11 -0.28 GXT1 RCPMLSS Mean SD (W) 271 ± 39 MD (W) -6.71 r 0.92 ES -0.17 RCP Mean SD (W) 315 ± 40 MD (W) -50.4 r 0.91 ES 1.27 https://doi.org/10.1371/journal.pone.0199794.t001 Table 2. Mean, standard deviation, and range of the _VO2 and power associated with the maximal lactate steady state (MLSS) expressed as a percentage of the maxi- mal power ( _Wmax) and _VO2peak measured during each GXT. Note: The _VO2 at the MLSS was 81.4 ± 4.7% of the _VO2max. (Defined as the highest measured _VO2 during any GXT). GXT1 GXT3 GXT4 GXT7 GXT10 _VO2 at MLSS (% of _VO2peak) 83.0 ± 4.5 [75.5–90.7] 84.7 ± 4.7 [76.6–91.9] 86.1 ± 5.9 [73.9–94.2] 88.4 ± 6.0 [77.4–103.2] 90.2 ± 5.3 [78.7–99.9] Power at MLSS (% of _Wmax) 62.9 ± 3.9 [56.8–71.7] 78.4 ± 4.3 [69.8–84.4] 82.4 ± 3.6 [73.7–88.8] 87.3 ± 4.4 [79.8–96.0] 89.6 ± 4.7 [81.6–98.1] https://doi.org/10.1371/journal.pone.0199794.t002 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 8 / 21 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 9 / 21 VEB had a CV of 3.3, 2.0, 3.5 and 5.2%, for GXT3, GXT4, GXT7 and GXT10, respectively. The VEB performed following the longer duration GXTs (GXT3-10) underestimated the _VO2peak from GXT1 (Table 6). Discussion The main findings of the present study are as follows. Only 11 of the 58 threshold values met our inclusion criteria as valid estimates of the MLSS. Of the 11 methods included in our analy- sis, three of the ModDmax methods yielded the most favourable estimations of the MLSS, and the Log-Poly-ModDmax derived from GXT4 provided the best estimation of the MLSS. There was an inverse relationship between stage length and LT, and this effect was larger in all Dmax methods compared with the OBLA and baseline plus absolute lactate value methods. The _VO2peak values measured during the longer duration GXTs (GXT3-10) underestimated the _VO2max and the _VO2peak values obtained from GXT1 (MD = 1.2 to 4.8 mL.kg-1.min-1). Finally, contrary to our hypothesis, the VEB after the longer duration GXTs did not yield _VO2peak val- ues comparable to the _VO2peak derived from GXT1. The use of five GXT protocols, 14 common LT methods, the RCP and RCPMLSS resulted in 58 unique thresholds. However, despite their common use, we observed that only 11 of these values met our criteria for inclusion (MD < 7.9 W; ES < 0.2; r > 0.90). Of the four Dmax meth- ods included in our analysis, one consisted of the traditional ModDmax method [22]. This had the poorest agreement relative to the other ModDmax methods included in our analysis. The remaining three Dmax methods are new variations of the ModDmax method, and the Log-Poly- Fig 2. (A-D) Forrest Plots of the difference (ES ± 95% CI) between the MLSS and the power calculated from the 13 lactate thresholds derived from (A) GXT3, (B) GXT4, (C) GXT7 and (D) GXT10 (52 in total and excluding log-log). The solid vertical bar represents no difference from the MLSS and the dashed vertical bars represents the threshold between a trivial and small difference (ES = 0.2) established by Cohen (50) and Hopkins (49). log = using the log-log method as the initial data point when calculating the Dmax or Modified Dmax; poly = Modified Dmax method calculated using a third order polynomial regression equation; exp = Modified Dmax method calculated using a constant plus exponential regression equation; OBLA = onset of blood lactate accumulation. https://doi.org/10.1371/journal.pone.0199794.g002 Table 3. Mean ± standard deviation, mean difference (MD), intraclass correlation coefficient (ICC), Lin’s concordance correlation coefficient (ρc), standard error of the measurement (SEM), effect size (ES) with 95% confidence limits, and coefficient of the variation (%CV) between the maximal lactate steady state (MLSS) and the eleven thresholds included in our analysis. (RCPMLSS = MLSS estimate based on the respiratory compensation point; log = Modified Dmax method using the log-log method as the point of the initial lactate point; poly = Modified Dmax method calculated using a third order polynomial regression equation; exp = Modified Dmax method calculated using a constant plus exponential regression equation; OBLA = onset of blood lactate accumulation). Mean ± SD (W) MD (W) ICC [95% CI] ρc SEM [95% CI] (W) ES [95% CI] CV [95% CI] (%) MLSS 264 ± 39 GXT1 RCPMLSS 271 ± 39 6.7 0.92 [0.78–0.97] 0.90 11.2 [8.3–17.0] 0.17 [-0.04–0.38] 6.0 [4.4–9.4] GXT3 Baseline + 1.5 mmol.L-1 270 ± 41 6.0 0.90 [0.75–0.97] 0.90 12.5 [9.3–19.0] 0.15 [-0.08–0.38] 6.6 [4.9–10.4] GXT4 OBLA 2.5 mmol.L-1 262 ± 40 -2.0 0.95 [0.87–0.98] 0.95 8.7 [6.5–13.2] -0.05 [-0.21–0.11] 5.3 [3.9–8.4] Modified Dmax 267 ± 39 2.9 0.91 [0.76–0.98] 0.90 11.7 [8.7–17.9] 0.07 [-0.15–0.29] 7.0 [5.1–11.0] Log-Poly-MDmax 265 ± 42 1.1 0.96 [0.90–0.99] 0.96 7.9 [5.8–12.0] 0.03 [-0.11–0.17] 4.4 [3.2–6.9] Log-Exp-MDmax 271 ± 42 7.0 0.97 [0.91–0.99] 0.95 7.5 [5.6–11.4] 0.18 [0.04–0.32] 4.1 [3.0–6.3] GXT7 OBLA 2.5 mmol.L-1 258 ± 41 -6.7 0.94 [0.85–0.98] 0.93 9.4 [7.0–14.3] -0.17 [-0.34–0.00] 4.9 [3.6–7.7] OBLA 3.0 mmol.L-1 267 ± 41 2.2 0.95 [0.86–0.98[ 0.95 9.2 [6.9–14.1] 0.06 [-0.11–0.23] 5.1 [3.7–8.0] Log-Exp-MDmax 260 ± 39 -4.3 0.96 [0.89–0.99] 0.95 7.8 [5.8–11.9] -0.11 [-0.25–0.03] 4.1 [3.0–6.4] GXT10 OBLA 3.0 mmol.L-1 264 ± 39 -0.4 0.93 [0.82–0.98] 0.93 10.2 [7.6–15.5] -0.01 [-0.20–0.18] 5.5 [4.0–8.6] OBLA 3.5 mmol.L-1 (n = 16) 275 ± 39 6.9 0.93 [0.82–0.98] 0.91 10.3 [7.7–15.7] 0.19 [0.00–0.38] 5.5 [4.0–8.7] https://doi.org/10.1371/journal.pone.0199794.t003 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 10 / 21 ModDmax derived from GXT4 had the highest correlation and lowest mean difference with the MLSS. These variations of the ModDmax method use the power at the log-log LT as the initial intensity to calculate the ModDmax and then either the traditional third-order polynomial or exponential plus-constant regression curve to fit the lactate curve [23, 41]. Although the valid- ity of these three methods has not previously been assessed, the favourable estimations of the MLSS may be related to the greater objectivity with which they determine the intensity that corresponds with the initial rise in blood lactate concentration [37]. Fig 3. Bland-Altman plots displaying agreement between measures of the power associated with the RCP regression equation (RCPMLSS) calculated from GXT1 and the MLSS. The differences between measures (y-axis) are plotted as a function of the mean of the two measures (x-axis) in power (Watts). The horizontal solid line represents the mean difference between the two measures (i.e., bias). The two horizontal dashed lines represent the limits of agreement (1.96 x standard deviation of the mean difference between the estimated lactate threshold via the RCPMLSS and the maximal lactate steady state). The dotted diagonal lines represent the boundaries of the 95% CI for MLSS reliability (CV = 3.0%; 95%; CI = 3.8%) calculated from Hauser et al., 2014) (RCP = respiratory compensation point). https://doi.org/10.1371/journal.pone.0199794.g003 Fig 4. Bland-Altman plots displaying agreement between measures of the power associated with the baseline plus 1.5 mmol.L-1 calculated from GXT3 and the MLSS. The differences between measures (y-axis) are plotted as a function of the mean of the two measures (x-axis) in power (Watts). The horizontal solid line represents the mean difference between the two measures (i.e., bias). The two horizontal dashed lines represent the limits of agreement (1.96 x standard deviation of the mean difference between the lactate threshold and the maximal lactate steady state). The dotted diagonal lines represent the boundaries of the 95% CI for MLSS reliability (CV = 3.0%; 95%; CI = 3.8%) calculated from Hauser et al., 2014). https://doi.org/10.1371/journal.pone.0199794.g004 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 11 / 21 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 12 / 21 Although the original Dmax method is a commonly cited method for determining the LT [23], we observed large mean differences (19 to 49 W) between the Dmax and MLSS. Three pre- vious studies have purported to investigate the validity of this method to estimate the MLSS in trained male cyclists [15, 52, 53]. One concluded that the Dmax method derived from GXT3 was a valid estimation of the MLSS (r = 0.97) [54]. We also observed a high correlation between Dmax and the MLSS (r = 0.94 to 0.97) (Table 1), but, as indicated by the MD and other measures, a high correlation is not sufficient to establish validity [55]. Another study examined Dmax derived from two GXTs with similar durations (36 vs. 39 min), but with different stage lengths (30-s vs. 6-min) [15]. The Dmax derived from GXT30s was not correlated (r = 0.51) with the MLSS, even though the MD was 5 W, whilst the Dmax derived from GXT6 was correlated (r = 0.85); however, it underestimated the MLSS (MD = 22 W). The third study concluded the Dmax derived from GXT1 yielded poor estimates of the MLSS (r = 0.56; bias = -1.8 ± 38.1 W) [53]. Thus, although some studies [15, 54] have used correlation analysis to suggest the Dmax provides a valid estimate of the MLSS, this is not supported by the more comprehensive assess- ment of validity performed in the present and other studies [53]. There were five fixed blood LT methods and one baseline plus an absolute value that met our inclusion criteria, and, as previously reported [15, 24], these varied with the GXT protocol used. The baseline + 1.5 mmol.L-1 was the only LT derived from GXT3 included in our analysis (bias = -6 ± 35 W). This is consistent with the results of one previous study (bias = 0.5 ± 24 W), which also recruited trained male cyclists and had a similar GXT protocol design [56]. Consistent with our findings, this study also reported that an OBLA of 3.5 mmol.L-1 derived from GXT3 did not provide a valid estimation of the MLSS. In contrast, another study con- firmed the validity of the OBLA of 3.5 mmol.L-1 [52], despite recruiting trained cyclists and using an identical GXT protocol. These conflicting results are likely attributable to the low reproducibility of the OBLA methods [16]. While none of the OBLAs from GXT3 met our inclusion criteria, the OBLA methods of 2.5 mmol.L-1 derived from GXT4 and GXT7 provided valid estimations of the MLSS, as did the OBLA of 3.0 mmol.L-1 derived from GXT7 and GXT10. The OBLA of 3.5 mmol.L-1 from GXT10 was the highest fixed blood LT that identified the MLSS. There is no previous data investigating the validity of these OBLA methods. However, it is worth noting that these five methods provided superior estimations of the MLSS compared with the original ModDmax, but were less favourable than the newly-developed ModDmax methods. An OBLA of 4.0 mmol.L-1 is the most commonly-accepted fixed blood lactate value for esti- mating the LT or MLSS. Three previous studies have attempted to validate use of an OBLA of 4.0 mmol.L-1 with cycle ergometry [15, 53, 57]. One study found that it overestimated the MLSS (MD = 49 W) when derived from GXT1 [53]. The other study reported poor agreement (bias 7 ± 49 W) when OBLA of 4.0 mmol.L-1 was derived from GXT4 [57]. The final study observed a poor correlation between an OBLA of 4.0 mmol.L-1 and the MLSS (r = 0.71) [15]. Our results indicated the OBLA of 4.0 mmol.L-1 overestimated the MLSS across all GXTs. Fig 5. (A-D) Bland-Altman plots displaying agreement between measures of the power associated with the (A) OBLA 2.5 mmol.L-1, (B) Modified Dmax, (C) Log-Poly-Modified Dmax, (D) Log-Exp-Modified Dmax calculated from GXT4 and the MLSS. The differences between measures (y-axis) are plotted as a function of the mean of the two measures (x- axis) in power (Watts). The horizontal solid line represents the mean difference between the two measures (i.e., bias). The two horizontal dashed lines represent the limits of agreement (1.96 x standard deviation of the mean difference between the lactate threshold and the maximal lactate steady state). The dotted diagonal lines represent the boundaries of the 95% CI for MLSS reliability (CV = 3.0%; 95%; CI = 3.8%) calculated from Hauser et al., 2014) (log = Modified Dmax method using the log-log method as the point of the initial lactate point; poly = Modified Dmax method calculated using a third order polynomial regression equation; exp = Modified Dmax method calculated using a constant plus exponential regression equation; OBLA = onset of blood lactate accumulation.). https://doi.org/10.1371/journal.pone.0199794.g005 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 13 / 21 Fig 6. (A-C) Bland-Altman plots displaying agreement between measures of the power associated with the (A) OBLA 2.5 mmol.L-1 (GXT7), (B) OBLA 3.0 mmol.L-1 (GXT7), (C) Log-Exp-Modified Dmax calculated from GXT7 and the MLSS. The differences between measures (y-axis) are plotted as a function of the mean of the two measures (x-axis) in power (Watts). The horizontal solid line represents the mean difference between the two measures (i.e., bias). The two horizontal dashed lines represent the limits of agreement (1.96 x standard deviation of the mean difference between the lactate threshold and the maximal lactate steady state). The dotted diagonal lines represent the boundaries of the 95% Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 14 / 21 Thus, in agreement with previous research, our results indicate; the OBLA of 4.0 mmol.L-1 does not accurately estimate the MLSS. It is also worth noting that the original authors cau- tioned the use of this OBLA method, given the lack of a significant correlation when compar- ing OBLA methods from a GXT and the MLSS [24]. The RCP derived from an 8- to 12-minute GXT consistently overestimates the MLSS [44, 53], and this was confirmed in our study (Table 1). Therefore, we used a regression equation based on the RCP (RCPMLSS) (Eq 5) to estimate the starting intensity for establishing the MLSS [33]. Our results indicate there was good agreement between the MLSS and RCPMLSS CI for MLSS reliability (CV = 3.0%; 95%; CI = 3.8%) calculated from Hauser et al., 2014) (log = Modified Dmax method using the log-log method as the point of the initial lactate point; exp = Modified Dmax method calculated using a constant plus exponential regression equation; OBLA = onset of blood lactate accumulation.). https://doi.org/10.1371/journal.pone.0199794.g006 Fig 7. (A-B) Bland-Altman plots displaying agreement between measures of the power associated with the (A) OBLA 3.0 mmol.L-1, (B) OBLA 3.5 mmol.L-1 calculated from GXT10 and the MLSS. The differences between measures (y- axis) are plotted as a function of the mean of the two measures (x-axis) in power (Watts). The horizontal solid line represents the mean difference between the two measures (i.e., bias). The two horizontal dashed lines represent the limits of agreement (1.96 x standard deviation of the mean difference between the lactate threshold and the maximal lactate steady state). The dotted diagonal lines represent the boundaries of the 95% CI for MLSS reliability (CV = 3.0%; 95%; CI = 3.8%) calculated from Hauser et al., 2014) (OBLA = onset of blood lactate accumulation.). https://doi.org/10.1371/journal.pone.0199794.g007 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 15 / 21 (Table 3). Nonetheless, for many participants the difference between MLSS and RCPMLSS exceeded the CV% for the MLSS (Fig 3). Therefore, although the RCPMLSS can be used as a convenient ‘starting point’ when establishing the MLSS, we recommend methods based on blood sampling from the current study and assessing blood lactate kinetics in real time as rec- ommended by Hering et al. [58] for a more accurate estimation of the MLSS. Table 4. Mean difference (MD), effect size (ES), and p-value comparing the influence of graded exercise test stage length on all 14 lactate threshold methods. 3 vs. 4 3 vs. 7 3 vs. 10 4 vs. 7 4 vs. 10 7 vs. 10 Log-log LT MD (W) 10 12 15 2 6 3 ES 0.24 0.28 0.36 0.05 0.14 0.08 p-value 0.09 0.02 0.02 0.63 0.15 0.47 OBLA 4.0 mmol.L-1 MD (W) 15 24 27 9 12 3 ES 0.34 0.56 0.63 0.22 0.29 0.06 p-value 0.00 0.00 0.00 0.05 0.01 0.35 OBLA 3.5 mmol.L-1 MD (W) 15 23 25 8 11 3 ES 0.34 0.53 0.60 0.20 0.26 0.06 p-value 0.00 0.00 0.00 0.09 0.02 0.35 OBLA 3.0 mmol.L-1 MD (W) 14 21 24 7 9 3 ES 0.34 0.50 0.57 0.16 0.23 0.06 p-value 0.00 0.00 0.00 0.16 0.05 0.36 OBLA 2.5 mmol.L-1 MD (W) 14 19 21 5 7 2 ES 0.34 0.46 0.53 0.12 0.18 0.06 p-value 0.00 0.00 0.00 0.30 0.13 0.39 OBLA 2.0 mmol.L-1 MD (W) 13 15 18 2 4 2 ES 0.33 0.38 0.45 0.06 0.12 0.06 p-value 0.01 0.01 0.00 0.63 0.36 0.45 Baseline + 0.5 mmol.L-1 MD (W) 6 7 10 1 4 3 ES 0.16 0.18 0.27 0.03 0.10 0.07 p-value 0.25 0.27 0.10 0.85 0.46 0.50 Baseline + 1.0 mmol.L-1 MD (W) 16 18 20 3 4 1 ES 0.40 0.47 0.51 0.07 0.10 0.03 p-value 0.01 0.00 0.00 0.53 0.41 0.71 Baseline + 1.5 mmol.L-1 MD (W) 16 21 23 5 7 2 ES 0.39 0.52 0.57 0.12 0.17 0.05 p-value 0.00 0.00 0.00 0.27 0.14 0.49 Dmax MD (W) 13 23 30 10 17 7 ES 0.38 0.71 0.90 0.29 0.49 0.22 p-value 0.00 0.00 0.00 0.00 0.00 0.00 Modified Dmax MD (W) 10 23 29 13 19 6 ES 0.27 0.59 0.79 0.32 0.50 0.16 p-value 0.01 0.00 0.00 0.01 0.00 0.06 Log-Poly-ModDmax MD (W) 14 25 32 11 18 7 ES 0.35 0.62 0.78 0.26 0.43 0.18 p-value 0.00 0.00 0.00 0.00 0.00 0.02 Exp-Dmax MD (W) 14 23 29 9 15 6 ES 0.38 0.66 0.82 0.26 0.42 0.17 p-value 0.00 0.00 0.00 0.00 0.00 0.02 Log-Exp-ModDmax MD (W) 15 26 33 11 18 7 ES 0.35 0.64 0.80 0.28 0.44 0.17 p-value 0.00 0.00 0.00 0.00 0.00 0.01 https://doi.org/10.1371/journal.pone.0199794.t004 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 16 / 21 Although a single GXT can be used to estimate both _VO2max and LT, the optimal test dura- tion for each measure is different [11, 13]. To address this challenge, we added a supramaximal VEB after each GXT, equivalent to that performed following GXT1, expecting all VEBs would yield similar _VO2 values. However, the _VO2peak values from the VEB after the longer duration GXTs underestimated the _VO2peak from GXT1. Although the _VO2peak values from GXT3 and GXT4 were similar to GXT1, the differences were larger than the typical coefficient of variabil- ity for _VO2peak (CV < 3%) [59]. Our results are consistent with previous recommendations that longer duration GXTs are not optimal for establishing _VO2peak [10, 60]. Furthermore, while a VEB can be used to verify that _VO2peak was achieved, it appears that a VEB following a prolonged GXT cannot be used to establish _VO2max. Extending the duration of the GXT stages results in a lower _Wmax [61]. This has implica- tions for exercise prescription, as it is common in sport and exercise science research to pre- scribe exercise intensity as a percentage of _Wmax. For example, in the present study the MLSS ranged from 63 ± 4% (range = 52 to 72%) of _Wmax from GXT1 to 82 ± 4% (range = 74 to 88%) Table 5. Mean and standard deviation of _VO2max—highest measured _VO2 during any graded exercise test (GXT); GXT _VO2 -highest measured _VO2 during each GXT; VEB _VO2 highest measured _VO2 during each verification exhaustive bout (VEB); _VO2peak, highest measured _VO2 during either the GXT or corresponding VEB. Mean and standard deviation of GXT duration, max power (Watts) from each GXT, percentage of maximum power from the prolonged GXT expressed as a per- centage of W maximum power from GXT1 and power of each VEB (Watts) from the GXTs. Relative power of the verification exhaustive bout expressed as a percentage (%) of the maximal power measured during the GXT. The subscript (i.e., 1, 3, 4, 7 or 10) refers to the stage duration (minutes) for each test. GXT1 GXT3 GXT4 GXT7 GXT10 _VO2max (mL.kg-1.min-1) 62.1 ± 5.8 GXT _VO2 (mL.kg-1.min-1) 60.6 ± 5.4 58.2 ± 5.3 57.3 ± 5.7 56.4 ± 5.2 54.9 ± 4.9 VEB _VO2 (mL.kg-1.min-1) 60.1 ± 5.8 58.9 ± 5.9 58.8 ± 6.1 56.4 ± 5.9 54.7 ± 6.6 _VO2peak (mL.kg-1.min-1) 61.0 ± 5.3 59.7 ± 5.4 58.9 ± 6.0 57.3 ± 5.4 56.2 ± 5.5 GXT Duration (min) 11.3 ± 0.9 26.8 ± 1.4 34.9 ± 1.9 59.2 ± 3.3 81.6 ± 4.6 Maximum Power (Watts) 420 ± 55 337 ± 46 321 ± 47 303 ± 43 295 ± 43 Percent _Wmax of GXT1 (%) 100 80.3 ± 2.9 76.4 ± 3.1 72.1 ± 3.6 70.3 ± 4.0 VEB (Watts) 378 ± 50 VEB (% of GXT _Wmax) 90 109.7 ± 3.8 118.4 ± 18.7 125.4 ± 19.3 128.8 ± 20.4 https://doi.org/10.1371/journal.pone.0199794.t005 Table 6. Mean difference (MD) and standard deviation, effect size (ES), coefficient of the variation (CV) and p-value (p) for the measured _VO2peak values from GXT1 compared with the _VO2peak values from GXT3, GXT4, GXT7, and GXT10 and for the _VO2peak values from GXT1 compared with the _VO2peak values from the VEB following GXT3, GXT4, GXT7, and GXT10. The subscript (i.e., 1, 3, 4, 7 or 10) refers to the stage duration (minutes) for each test. GXT1 vs. GXT3 GXT1 vs. GXT4 GXT1 vs. GXT7 GXT1 vs. GXT10 MD (mL.kg-1.min-1) -1.2 ± 3.3 -2.1 ± 4.2 -3.7 ± 4.7 -4.8 ± 3.7 ES 0.23 0.36 0.69 0.88 CV (%) 3.8 4.9 5.6 4.6 p 0.13 0.06 < 0.01 < 0.01 GXT1 vs. VEB GXT3 GXT1 vs. VEB GXT4 GXT1 vs. VEB GXT7 GXT1 vs. VEB GXT10 MD (mL.kg-1.min-1) -2.1 ± 5.9 -2.1 ± 6.1 -4.6 ± 5.9 -6.2 ± 6.6 ES 0.37 0.37 0.81 1.04 CV (%) 4.2 4.9 6.1 5.9 p 0.02 0.98 0.03 0.03 https://doi.org/10.1371/journal.pone.0199794.t006 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 17 / 21 of _Wmax from GXT4. Prescribing exercise in the current study cohort at a fixed percentage of _Wmax (e.g., 73% of _Wmax), would result in all participants exercising above or below the MLSS, GXT1 and GXT4, respectively. This is important as it has previously been reported that pre- scribing exercise relative to LT results in a more homogenous physiological response than when exercise performed relative to _Wmax [62]. This also highlights why it is important to con- sider the GXT protocol and the method used to determine relative exercise intensity when comparing results between studies. The wide range of _Wmax for each GXT is also note-worthy, the _Wmax range for GXT1 was 320 to 517 W and the duration ranged from 9 to 12 minutes. Had we employed a standardized GXT (e.g., 35 W increments), and assuming _Wmax stayed constant, the range would have been 9- to 15 min. Applying this to our longer duration GXTs resulted in a homogenous duration (GXT4: 32- to 39 min), whereas a standardised approach (e.g., 35 W increments) would have resulted in a range of 27- to 46 min [57]. Thus, individualizing GXT protocol design is a useful approach to ensure homogenous test duration [17]. Conclusion In conclusion, the traditional Dmax and OBLA of 4.0 mmol.L-1 did not provide valid estimates of the MLSS. The best estimation of the MLSS was the Log-Poly-ModDmax derived from GXT4. The validity of our newly-developed ModDmax model may relate to the objectivity for determining the initial rise in blood lactate concentration. However, we must advise caution with the use of our newly-developed method until future research investigates the reliability and reproducibility. It is apparent that both _VO2max and LT cannot be determined in a single GXT, even if the GXT is followed by a VEB. Therefore, to appropriately determine _VO2max the optimum duration of a GXT is 8–12 minutes and the _VO2 values measured during the GXT and VEB be within 3% = CV [63]. Our data also highlight how differences in GXT protocol design and methods used to calculate the relative exercise intensity may contribute to the con- flicting findings reported in the literature. Author Contributions Conceptualization: Nicholas A. Jamnick, Javier Botella, David B. Pyne, David J. Bishop. Data curation: Nicholas A. Jamnick, Javier Botella, David B. Pyne, David J. Bishop. Formal analysis: Nicholas A. Jamnick, Javier Botella, David B. Pyne, David J. Bishop. Investigation: Nicholas A. Jamnick, Javier Botella, David J. Bishop. Methodology: Nicholas A. Jamnick, Javier Botella, David B. Pyne, David J. Bishop. Project administration: Nicholas A. Jamnick, David J. Bishop. Resources: Nicholas A. Jamnick, Javier Botella, David J. Bishop. Software: Nicholas A. Jamnick. Supervision: David B. Pyne. Validation: Nicholas A. Jamnick, Javier Botella, David J. Bishop. Visualization: Javier Botella. Writing – original draft: Nicholas A. Jamnick, Javier Botella, David B. Pyne, David J. Bishop. Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 18 / 21 Writing – review & editing: Nicholas A. Jamnick, Javier Botella, David B. Pyne, David J. Bishop. References 1. Skinner JS, McLellan TH. The transition from aerobic to anaerobic metabolism. Research Quarterly for Exercise and Sport. 1980; 51(1):234–48. https://doi.org/10.1080/02701367.1980.10609285 PMID: 7394286 2. Londeree BR. Effect of training on lactate/ventilatory thresholds: a meta-analysis. Medicine and Sci- ence in Sport and Exercise. 1997; 29(6):837–43. PMID: 9219214 3. Wenger HA, Bell GJ. The interactions of intensity, frequency and duration of exercise training in altering cardiorespiratory fitness. Sports Medicine. 1986; 3(5):346–56. PMID: 3529283 4. Mann T, Lamberts RP, Lambert MI. Methods of prescribing relative exercise intensity: physiological and practical considerations. Sports Medicine. 2013; 43(7):613–25. https://doi.org/10.1007/s40279-013- 0045-x PMID: 23620244 5. Coen B, Schwarz L, Urhausen A, Kindermann W. Control of training in middle-and long-distance run- ning by means of the individual anaerobic threshold. International Journal of Sports Medicine. 1991; 12 (06):519–24. 6. Granata C, Jamnick NA, Bishop DJ. Principles of Exercise Prescription, and How They Influence Exer- cise-Induced Changes of Transcription Factors and Other Regulators of Mitochondrial Biogenesis. Sports Medicine. 2018:1–19. 7. Gonza´lez-Haro C. Differences in Physiological Responses Between Short-vs. Long-Graded Laboratory Tests in Road Cyclists. Journal of Strength and Conditioning Research. 2015; 29(4):1040–8. https://doi. org/10.1519/JSC.0000000000000741 PMID: 25330085 8. McNaughton LR, Roberts S, Bentley DJ. The relationship among peak power output, lactate threshold, and short-distance cycling performance: effects of incremental exercise test design. Journal of Strength and Conditioning Research. 2006; 20(1):157. https://doi.org/10.1519/R-15914.1 PMID: 16506862 9. Bentley D, McNaughton L, Batterham A. Prolonged stage duration during incremental cycle exercise: effects on the lactate threshold and onset of blood lactate accumulation. European Journal of Applied Physiology. 2001; 85(3–4):351–7. https://doi.org/10.1007/s004210100452 PMID: 11560091 10. Poole DC, Jones AM. Measurement of the maximum oxygen uptake _Vo2max : _Vo2peak is no longer acceptable. Journal of Applied Physiology. 2017; 122(4):997–1002. https://doi.org/10.1152/ japplphysiol.01063.2016 PMID: 28153947 11. Buchfuhrer MJ, Hansen JE, Robinson TE, Sue DY, Wasserman K, Whipp BJ. Optimizing the exercise protocol for cardiopulmonary assessment. Journal of Applied Physiology. 1983; 55(5):1558–64. https:// doi.org/10.1152/jappl.1983.55.5.1558 PMID: 6643191 12. Yoon B-K, Kravitz L, Robergs R. VO2max, Protocol Duration, and the VO2 Plateau. Medicine and Sci- ence in Sport and Exercise. 2007; 39(7):1186–92. https://doi.org/10.1249/mss.0b13e318054e304 PMID: 17596788 13. Bentley DJ, Newell J, Bishop D. Incremental exercise test design and analysis. Sports Medicine. 2007; 37(7):575–86. PMID: 17595153 14. Foxdal P, Sjo¨din A, Sjo¨din B. Comparison of blood lactate concentrations obtained during incremental and constant intensity exercise. International Journal of Sports Medicine. 1996; 17(05):360–5. 15. Van Schuylenbergh R, Vanden EB, Hespel P. Correlations between lactate and ventilatory thresholds and the maximal lactate steady state in elite cyclists. International Journal of Sports Medicine. 2004; 25 (6):403–8. https://doi.org/10.1055/s-2004-819942 PMID: 15346226 16. Morton RH, Stannard SR, Kay B. Low reproducibility of many lactate markers during incremental cycle exercise. British Journal of Sports Medicine. 2012; 46(1):64–9. https://doi.org/10.1136/bjsm.2010. 076380 PMID: 21343140 17. Pettitt R, Clark I, Ebner S, Sedgeman D, Murray S. Gas Exchange Threshold and VO2max Testing for Athletes: An Update. Journal of Strength and Conditioning Research. 2013; 27(2):549–55. https://doi. org/10.1519/JSC.0b013e31825770d7 PMID: 22531615 18. Faude O, Kindermann W, Meyer T. Lactate threshold concepts. Sports Medicine. 2009; 39(6):469–90. https://doi.org/10.2165/00007256-200939060-00003 PMID: 19453206 19. Ivy JL, Withers RT, Vanhandel PJ, Elger DH, Costill DL. Muscle respiratory capacity and fiber type as determinants of the lactate threshold. Journal of Applied Physiology. 1980; 48(3):523–7. PubMed PMID: WOS:A1980JJ95100018. https://doi.org/10.1152/jappl.1980.48.3.523 PMID: 7372524 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 19 / 21 20. Kindermann W, Simon G, Keul J. Significance of the aerobic-anaerobic transition for the determination of work load intensities during endurance training. European Journal of Applied Physiology and Occu- pational Physiology. 1979; 42(1):25–34. https://doi.org/10.1007/bf00421101 PubMed PMID: WOS: A1979HN15900003. PMID: 499194 21. Heck H, Mader A, Hess G, Mucke S, Muller R, Hollmann W. Justification of the 4-mmol/L lactate thresh- old. International Journal of Sports Medicine. 1985; 6(3):117–30. https://doi.org/10.1055/s-2008- 1025824 PubMed PMID: WOS:A1985ALS1800002. PMID: 4030186 22. Bishop D, Jenkins DG, Mackinnon LT. The relationship between plasma lactate parameters, Wpeak and 1-h cycling performance in women. Medicine and Science in Sports and Exercise. 1998; 30 (8):1270–5. PMID: 9710868 23. Cheng B, Kuipers H, Snyder A, Keizer H, Jeukendrup A, Hesselink M. A new approach for the determi- nation of ventilatory and lactate thresholds. International Journal of Sports Medicine. 1992; 13(7):518– 22. https://doi.org/10.1055/s-2007-1021309 PMID: 1459746 24. Heck H, Mader A, Hess G, Mu¨cke S, Mu¨ller R, Hollmann W. Justification of the 4-mmol/l lactate thresh- old. International Journal of Sports Medicine. 1985;(6):117–30. 25. Billat VL, Sirvent P, Py G, Koralsztein J-P, Mercier J. The concept of maximal lactate steady state. Sports Medicine. 2003; 33(6):407–26. PMID: 12744715 26. Beneke R. Methodological aspects of maximal lactate steady state—implications for performance test- ing. European Journal of Applied Physiology. 2003; 89(1):95–9. https://doi.org/10.1007/s00421-002- 0783-1 PMID: 12627312 27. Hauser T, Bartsch D, Baumga¨rtel L, Schulz H. Reliability of maximal lactate-steady-state. International Journal of Sports Medicine. 2013; 34(3):196–9. https://doi.org/10.1055/s-0032-1321719 PMID: 22972242 28. Hill AV, Long C, Lupton H. Muscular exercise, lactic acid, and the supply and utilisation of oxygen. Pro- ceedings of the Royal Society of London Series B, Containing Papers of a Biological Character. 1924; 97(681):84–138. 29. Riebe D, Franklin BA, Thompson PD, Garber CE, Whitfield GP, Magal M, et al. Updating ACSM’s rec- ommendations for exercise preparticipation health screening. Medicine and Science in Sport and Exer- cise. 2015; 47(11):2473–9. https://doi.org/10.1249/MSS.0000000000000664 PMID: 26473759 30. Jamnick NA, By S, Pettitt CD, Pettitt RW. Comparison of the YMCA and a Custom Submaximal Exer- cise Test for Determining VO2max. Medicine and Science in Sport and Exercise. 2016; 48(2):254–9. https://doi.org/10.1249/MSS.0000000000000763 PMID: 26339726 31. Jackson AS, Blair SN, Mahar MT, Wier LT, Ross RM, Stuteville JE. Prediction of functional aerobic capacity without exercise testing. Medicine and Science in Sports and Exercise. 1990; 22(6):863–70. Epub 1990/12/01. PMID: 2287267. 32. Medicine ACoS. ACSM’s guidelines for exercise testing and prescription: Lippincott Williams & Wilkins; 2013. 33. Smekal G, von Duvillard SP, Pokan R, Hofmann P, Braun WA, Arciero PJ, et al. Blood lactate concen- tration at the maximal lactate steady state is not dependent on endurance capacity in healthy recreation- ally trained individuals. European Journal of Applied Physiology. 2012; 112(8):3079–86. https://doi.org/ 10.1007/s00421-011-2283-7 PMID: 22194004 34. Beaver W, Wasserman K, Whipp B. A new method for detecting anaerobic threshold by gas exchange. Journal of Applied Physiology. 1986; 60(6):2020–7. https://doi.org/10.1152/jappl.1986.60.6.2020 PMID: 3087938 35. Whipp BJ, Davis JA, Wasserman K. Ventilatory control of the ‘isocapnic buffering’region in rapidly-incre- mental exercise. Respiration Physiology. 1989; 76(3):357–67. PMID: 2501844 36. Caiozzo VJ, Davis JA, Ellis JF, Azus JL, Vandagriff R, Prietto C, et al. A comparison of gas exchange indices used to detect the anaerobic threshold. Journal of Applied Physiology. 1982; 53(5):1184–9. https://doi.org/10.1152/jappl.1982.53.5.1184 PMID: 7174412 37. Beaver W, Wasserman K, Whipp B. Improved detection of lactate threshold during exercise using a log- log transformation. Journal of Applied Physiology. 1985; 59(6):1936–40. https://doi.org/10.1152/jappl. 1985.59.6.1936 PMID: 4077801 38. Kindermann W, Simon G, Keul J. The significance of the aerobic-anaerobic transition for the determina- tion of work load intensities during endurance training. European Journal of Applied Physiology and Occupational Physiology. 1979; 42(1):25–34. PMID: 499194 39. Zoladz JA, Rademaker A, Sargeant AJ. Non-linear relationship between O2 uptake and power output at high intensities of exercise in humans. Journal of Physiology. 1995; 488(Pt 1):211. 40. Berg A, Jakob E, Lehmann M, Dickhuth H, Huber G, Keul J. Current aspects of modern ergometry. Pneumologie (Stuttgart, Germany). 1990; 44(1):2. Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 20 / 21 41. Machado FA, Nakamura FY, Moraes SMFD. Influence of regression model and incremental test proto- col on the relationship between lactate threshold using the maximal-deviation method and performance in female runners. Journal of Sports Sciences. 2012; 30(12):1267–74. https://doi.org/10.1080/ 02640414.2012.702424 PMID: 22775431 42. Hughson RL, Weisiger KH, Swanson GD. Blood lactate concentration increases as a continuous func- tion in progressive exercise. Journal of Applied Physiology. 1987; 62(5):1975–81. https://doi.org/10. 1152/jappl.1987.62.5.1975 PMID: 3597269 43. Lamarra N, Whipp BJ, Ward SA, Wasserman K. Effect of interbreath fluctuations on characterizing exercise gas exchange kinetics. Journal of Applied Physiology. 1987; 62(5):2003–12. https://doi.org/10. 1152/jappl.1987.62.5.2003 PMID: 3110126 44. Keir DA, Fontana FY, Robertson TC, Murias JM, Paterson DH, Kowalchuk JM, et al. Exercise Intensity Thresholds: Identifying the Boundaries of Sustainable Performance. Medicine and Science in Sport and Exercise. 2015; 47(9):1932–40. https://doi.org/10.1249/MSS.0000000000000613 PMID: 25606817 45. Keir DA, Murias JM, Paterson DH, Kowalchuk JM. Breath-by-breath pulmonary O2 uptake kinetics: effect of data processing on confidence in estimating model parameters. Experimental Physiology. 2014; 99(11):1511–22. https://doi.org/10.1113/expphysiol.2014.080812 PMID: 25063837 46. Lawrence I, Lin K. A concordance correlation coefficient to evaluate reproducibility. Biometrics. 1989:255–68. PMID: 2720055 47. Bland JM, Altman D. Statistical methods for assessing agreement between two methods of clinical mea- surement. The Lancet. 1986; 327(8476):307–10. 48. Hopkins W. Analysis of validity by linear regression (Excel spreadsheet). 2015; 19:36–44. 49. Hopkins W. Measures of reliability in sports medicine and science. Sports Medicine. 2000; 30(1):1–15. PMID: 10907753 50. Hopkins W, Marshall S, Batterham A, Hanin J. Progressive statistics for studies in sports medicine and exercise science. Medicine and Science in Sports and Exercise. 2009; 41(1):3. https://doi.org/10.1249/ MSS.0b013e31818cb278 PMID: 19092709 51. Cohen J. A power primer. Psychological Bulletin. 1992; 112(1):155. PMID: 19565683 52. Denadai B, Figueira T, Favaro O, Gonc¸alves M. Effect of the aerobic capacity on the validity of the anaerobic threshold for determination of the maximal lactate steady state in cycling. Brazilian Journal of Medical and Biological Research. 2004; 37(10):1551–6. https://doi.org//S0100-879X2004001000015 PMID: 15448877 53. Pallare´s JG, Mora´n-Navarro R, Ortega JF, Ferna´ndez-Elı´as VE, Mora-Rodriguez R. Validity and Reli- ability of Ventilatory and Blood Lactate Thresholds in Well-Trained Cyclists. PloS one. 2016; 11(9): e0163389. https://doi.org/10.1371/journal.pone.0163389 PMID: 27657502 54. Czuba M, Zając A, Cholewa J, Poprzęcki S, Waśkiewicz Z, Mikołajec K. Lactate threshold (D-max method) and maximal lactate steady state in cyclists. Journal of Human Kinetics. 2009; 21:49–56. 55. Aldrich J. Correlations genuine and spurious in Pearson and Yule. Statistical Science. 1995:364–76. 56. Grossl T, De Lucas RD, De Souza KM, Antonacci Guglielmo LG. Maximal lactate steady-state and anaerobic thresholds from different methods in cyclists. European Journal of Sport Science. 2012; 12 (2):161–7. 57. Hauser T, Adam J, Schulz H. Comparison of selected lactate threshold parameters with maximal lactate steady state in cycling. International Journal of Sports Medicine. 2014; 35(6):517–21. https://doi.org/10. 1055/s-0033-1353176 PMID: 24227122 58. Hering GO, Hennig EM, Riehle HJ, Stepan J. A Lactate Kinetics Method for Assessing the Maximal Lac- tate Steady State Workload. Frontiers in Physiology. 2018; 9(310). https://doi.org/10.3389/fphys.2018. 00310 PMID: 29651253 59. Kirkeberg J, Dalleck L, Kamphoff C, Pettitt R. Validity of 3 protocols for verifying VO2max. International Journal of Sports Medicine. 2011; 32(04):266–70. 60. Bishop D, Jenkins DG, Mackinnon LT. The effect of stage duration on the calculation of peak VO2 dur- ing cycle ergometry. Journal of Science and Medicine in Sport. 1998; 1(3):171–8. PMID: 9783518 61. Adami A, Sivieri A, Moia C, Perini R, Ferretti G. Effects of step duration in incremental ramp protocols on peak power and maximal oxygen consumption. European Journal of Applied Physiology. 2013; 113 (10):2647–53. https://doi.org/10.1007/s00421-013-2705-9 PMID: 23949790 62. Baldwin J, Snow RJ, Febbraio MA. Effect of training status and relative exercise intensity on physiological responses in men. Medicine and Science in Sport and Exercise. 2000; 32(9):1648–54. PMID: 10994919 63. Pettitt RW, Jamnick NA. Commentary on “Measurement of the maximum oxygen uptake _Vo2max : _Vo2peak is no longer acceptable”. Journal of Applied Physiology. 2017; 123(3):696–. https://doi.org/10. 1152/japplphysiol.00338.2017 PMID: 28947630 Validation of a single visit graded exercise test PLOS ONE | https://doi.org/10.1371/journal.pone.0199794 July 30, 2018 21 / 21

Manipulating graded exercise test variables affects the validity of the lactate threshold and [Formula: see text].

07-30-2018

Jamnick, Nicholas A,Botella, Javier,Pyne, David B,Bishop, David J

eng

PMC7883807

Physiological Reports. 2021;9:e14739. | 1 of 9 https://doi.org/10.14814/phy2.14739 wileyonlinelibrary.com/journal/phy2 Received: 22 November 2020 | Revised: 7 January 2021 | Accepted: 11 January 2021 DOI: 10.14814/phy2.14739 R E V I E W A R T I C L E The effect of L- arginine supplementation on maximal oxygen uptake: A systematic review and meta- analysis Shahla Rezaei1,2 | Maryam Gholamalizadeh3 | Reza Tabrizi4 | Peyman Nowrouzi- Sohrabi5 | Samira Rastgoo6 | Saeid Doaei3 This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2021 The Authors. Physiological Reports published by Wiley Periodicals LLC on behalf of The Physiological Society and the American Physiological Society 1Student Research Committee, Department of Clinical Nutrition, School of Nutrition and Food Sciences, Shiraz University of Medical Sciences, Shiraz, Iran 2Nutrition Research Center, School of Nutrition and Food Sciences, Shiraz University of Medical Sciences, Shiraz, Iran 3Student Research Committee, Cancer Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran 4Non- Communicable Diseases Research Center, Fasa University of Medical Sciences, Fasa, Iran 5Department of Biochemistry, School of Medicine, Shiraz University of Medical Sciences, Shiraz, Iran 6Department of Clinical Nutrition and Dietetics, National Nutrition and Food Technology Research Institute Shahid Beheshti University of Medical Science, Tehran, Iran Correspondence Saeid Doaei, Student Research Committee, Cancer Research Center, Shahid Beheshti University of Medical Sciences, Tehran, Iran. Email: [email protected] Funding information Funding for this study was provided by the student research committee of Shahid Beheshti University of Medical Sciences, Tehran, Iran (code: 13172). Abstract Background: The efficacy and safety of L- arginine supplements and their effect on maximal oxygen uptake (VO2 max) remained unclear. This systematic review aimed to investigate the effect of L- arginine supplementation (LAS) on VO2 max in healthy people. Methods: We searched PubMed, Scopus, Web of Science, Cochrane, Embase, ProQuest, and Ovid to identify all relevant literature investigating the effect of LAS on VO2 max. This meta- analysis was conducted via a random- effects model for the best estimation of desired outcomes and studies that meet the inclusion criteria were considered for the final analysis. Results: The results of 11 randomized clinical trials indicated that LAS increased VO2 max compared to the control group. There was no significant heterogeneity in this meta- analysis. Subgroup analysis detected that arginine in the form of LAS significantly increased VO2 max compared to the other forms (weighted mean differ- ence = 0.11 L min−1, I2 = 0.0%, p for heterogeneity = 0.485). Conclusions: This meta- analysis indicated that supplementation with L- arginine could increase VO2 max in healthy people. Further studies are warranted to confirm this finding and to identify the underlying mechanisms. K E Y W O R D S L- arginine, maximal oxygen uptake, meta- analysis, VO2 max 2 of 9 | REZAEI Et Al. 1 | INTRODUCTION Cardiovascular endurance is one of the most important mea- sures of overall health (Ruiz et al., 2006). A person's level of cardiovascular endurance helps predict ability to react to acute physical and mental stress (Gutin et al., 2005). For healthy individuals, higher cardiovascular endurance also can indicate an elevated level of physical fitness (Haghshenas et al., 2019). One of the best indicators for the athlete's car- diovascular performance is the maximal oxygen uptake (VO2 max) assessment (Campbell et al., 2006). A greater amount of oxygen consumed by the body is related to higher cardio- vascular efficiency (Adams et al., 1995). Higher cardiovascu- lar efficiency allows muscle to work at a higher intensity for a longer time period. The body can only exercise as long as oxygen is delivered to the muscle and waste products such as carbon dioxide are removed (Haghshenas et al., 2019). Many factors such as proteins could be associated with cardiovascu- lar risk factors and other diseases (Doaei et al., 2018; Shidfar et al., 2018). Amino acids are among the most common nutritional sup- plements which are used by athletes to improve athletic per- formance under aerobic and anaerobic conditions (Mashiko et al., 2004). L- arginine is one of the semi- essential amino acids that has positive effects on muscle metabolism (Preli et al., 2002). L- arginine may also have a key role in the car- diac function of athletes. Arginine is a precursor of nitric oxide (NO) and NO causes vasodilatory effects, increased blood flow to the muscles, and increased release of cer- tain hormones such as insulin and human growth hormone (Adams et al., 1995; Moazami et al., 2015). Oral L- arginine supplements improved coronary endothelium- dependent di- lation (Melik et al., 2017). L- arginine may have led to delayed fatigue by altering blood lactate concentration and metabolic indices of respi- ration. It is frequently reported that using L- arginine supple- ment may improve athletic performance in sports activities. (Yaman et al., 2010). Yaman et al. found that L- arginine sup- plementation (LAS) significantly reduced blood pressure and increased VO2 max and may influence athletic performance capacity (Kalman et al., 2016). However, the studies on the association between LAS and VO2 max reported contradictory results. Therefore, this systematic review and meta- analysis aimed to investigate the effect of LAS on VO2 max. 2 | METHODS AND MATHERIALS 2.1 | Literature search strategy This systematic review and meta- analysis was performed in accordance with PRISMA (Preferred Reporting Items for Systematic Reviews and Meta- Analyses) guidelines (Liberati et al., 2009). The scientific databases includ- ing PubMed, Scopus, Web of Science, Cochrane, Embase, ProQuest, and Ovid were reviewed to identify all relevant literature on the effects of LAS on VO2 max that were published by August 2020. The following search strat- egy was used to finalize the first step of data gathering: (Arginine[Mesh] OR Arginine[tiab]) AND (VO2[tiab] OR "maximal aerobic"[tiab] OR "aerobic capacity"[tiab] OR "maximal O2"[tiab] OR "maximal O2 consumption"[tiab] OR "maximal O2 uptake"[tiab] OR "peak O2"[tiab] OR "maximal oxygen consumption"[tiab] OR "maximal oxygen uptake"[tiab] OR "peak oxygen uptake"[tiab] OR "maximal aerobic capacity"[tiab]). To enhance the quality of the searches, hand search- ing was performed to find all relevant articles using the references of the collected articles. The searches were limited to human studies and no language restriction was used in the search process. Two authors (Sh. R and P. N) independently screened the title and the abstracts of the included papers, performed data extraction, and carried out the quality assessments of the eligible studies. All dis- agreements were resolved by consulting with a third author (R. T). 2.2 | Study selection The following strategy was used to select the eligible pa- pers for performing the meta- analysis: Randomized clini- cal trials (parallel or cross- over) experiments, investigated the effect of LAS on VO2 max in healthy human subjects, individuals supplemented with arginine were compared to placebo- control individuals, arginine supplementation administered for at least 1 week, papers with enough in- formation to measure the VO2 max, papers contained data for SD, SE, and CI parameters in the beginning and the end of the study for both of the intervention and control groups. 2.3 | Data extraction All eligible randomized controlled trials were separately re- checked and the following data were extracted: the name of the first author, country, the number of individu- als in the intervention and control groups, the form of supplemented arginine, arginine doses, duration of the study, type of the study, and the related data for further steps (Table 1). For each study, the value of mean and SD for VO2 max in the beginning and at the end of the study was extracted. The following formula was used to calculate the mean difference of SDs:  | 3 of 9 REZAEI Et Al. TABLE 1 Participant and intervention characteristics of the studies included in the systematic review and meta- analysis ID Authors Year Country Population Age Number of subjects in intervention/ control groups Type of study Type of Intervention Control group Duration of study 1 Camic et al. 2010 USA College- aged male 22.1 ± 2.4 21/20 Randomized, double- blind, placebo, controlled, parallel design 3 g/day, LAS+300 mg of grape seed extract + 300 mg of polyethylene glycol Placebo 28 days 2 Chen et al. 2010 USA Male cyclists ARG:57.6 ± 4.6 PLA: 60.6 ± 8.7 8/8 Two- arm prospectively randomized double- blinded and placebo- controlled trial 5.2 g/day, LAS + L- citrulline + 500 mg ascorbic acid, 400 IU vitamin E, 400 μg folic acid, 300 mg L- taurine, and 10 mg alpha lipoic acid Placebo 21 days 3 Muazzezzaneh et al. 2010 Iran Healthy athletes 22.66 ± 1.46 14/13 Based on a single- blind placebo- controlled trial 5 g/day, LAS Placebo 21 days 4 Sunderland et al. 2011 USA Endurance- trained male cyclists 36.3 ± 7.9 9/9 Randomized, conducted in a double- blind manner 12 g/day, LAS Placebo 28 days 5 Moazami et al. 2014 Iran Female handballists 2.49 ± 22.15 8/8 Randomized clinical trial 3 g/day, LAS Placebo 7 days 6 Zak et al. 2015 USA Untrained men 22.0 ± 1.7 19/19 Double blinded, placebo- controlled, within subjects’ crossover design 3 g/day, LAS + 300 mg of grape seed extract (95% procyanidins), and 300 mg of polyethylene glycol Placebo 7 days 7 Hosseini et al. 2015 Iran Healthy futsal players 22.5 ± 1.39 10/10 Randomized control trial 4 g/day, LAS Placebo 28 days 8 Pahlavani et al. 2017 Iran Soccer players 20.85 ± 4.29 25/27 Double- blinded, randomized, placebo- controlled trial 2 g/day, LAS Placebo 45 days 9 Dennis et al. 1991 France Medical students, active in recreational activities 19– 26 15/15 Double- blind, cross- over study 5 g/day, AA Placebo 10 days 10 Burtscher et al. 2005 Austria Healthy athletes 22 ± 3 8/8 Double blind placebo- controlled trial 3 g/day, AA Placebo 21 days 11 Campbel et al. 2005 USA Resistance- trained healthy adult men 38.9 ± 5.8 20/15 Randomized, double- blind, controlled design 12 g/day, AAKG Placebo 56 days 4 of 9 | REZAEI Et Al. A correlation coefficient of 0.5 was used for R, esti- mated between 0 and 1 values, respectively. Also, the for- mula SD = SE × √ n (n = the number of individuals in each group) was used to measure SD in each article that reported SE instead of SD. 2.4 | Quality assessment The quality assessment of the included papers in this meta- analysis was performed according to Cochrane criteria (Higgins, 2011). According to this guideline, any source of bias including selection bias, performance bias, detection bias, attrition bias, and reporting bias were judged for all in- cluded studies (Figure 1). 2.5 | Statistical analysis This meta- analysis was conducted using Stata version 11. Due to the population selection from different countries, a random- effects model was employed with a 95% confidence interval (CI) for the calculation of the pooled weighted mean difference (WMD). Analysis endpoints were calcu- lated as the difference in mean between baseline and post- treatment (measure at the end of follow- up − measure at baseline); also, the SD of mean change was calculated by the pooled SD. The statistical heterogeneity between trials was calculated by p- value and using I2 statistic (p < 0.05 and I2 > 50%). Publication bias was checked by the funnel plot, Begg's test (p = 0.815), and Egger's tests (p = 0.218; Figure 2). SD =square root [(SD at baseline)2 +(SD at the end of study)2 −(2R×SD at baseline×SD at the end of study)] . FIGURE 1 Summary of risk of bias: According to Cochrane criteria, any source of bias including selection bias, performance bias, detection bias, attrition bias, and reporting bias were judged for all included studies FIGURE 2 Publication bias was checked by the funnel plot, Begg's (p = 0.815) test, and Egger's (p = 0.218) tests. SE, standard error; WMD, weighted mean difference  | 5 of 9 REZAEI Et Al. 3 | RESULTS 3.1 | Search results and study selection The flow chart presented in Figure 3 describes the process of selection and the references retrieved in the database. A total number of 945 articles was identified in the first step of the literature search of electronic databases. After excluding duplicated studies (n = 617), irrelevant studies based on title and abstracts (n = 295), type of intervention (n = 1), type of outcomes (n = 5), and the required data (n = 4), 23 poten- tially relevant articles were considered for full text review. After screening, 12 articles were excluded for the follow- ing reasons: studies population, insufficient data reporting of outcome, and type of LAS. Finally, 11 studies were in- cluded in the present meta- analysis (Burtscher et al., 2005; Camic et al., 2010; Campbell et al., 2006; Chen et al., 2010; Denis et al., 1991; Hosseini et al., 2015; Moazami et al., 2015; Muazzezzaneh et al., 2010; Pahlavani et al., 2017; Sunderland et al., 2011; Zak et al., 2015). 3.2 | Quantitative data synthesis Marginal significant increase in VO2 max (WMD  =  0.07  L  min−1; 95% CI, 0.00– 0.13, p  =  0.047; I2 = 23.2%) was found after L- arginine supplementation in comparison with the control group (Figure 4). 3.3 | Subgroup analysis Subgroup analysis was performed based on the study dura- tion (≥14 days), dosage of L- arginine (<5 g/day), and the type of arginine supplementation including LAS, arginine aspartate, arginine alpha- Ketoglutarate, and arginine in com- bination with antioxidants to detect the source of heteroge- neity. There was a significant increase in VO2 max in the subgroup analysis of trials with LAS (WMD = 0.11 L min−1, I2 = 0.0%, p for heterogeneity = 0.485; Table 2). 3.4 | Sensitivity analysis The sensitivity analysis was performed using “one- study- removed” method to survey the impact of each study on the effect size. The results of sensitivity analysis identified the higher and lower pooled weight mean difference for VO2 max (WMD  =  0.1  L  min−1; 95% CI 0.08, 0.13) after ex- cluding the Burtscher et al. (2005) study and (WMD = 0.03 L.min- 1; 95% CI 0.04, 0.1) after excluding Hosseini et al. (2015) study, respectively (Figure 5). FIGURE 3 Preferred Reporting Items for Systematic Review and Meta- Analyses flow diagram Records idenfied through database searching (n =945) Eligibility Addional records idenfied by hand searching (n = 0) Records aer duplicates removed (n = 617) Records screened (n = 328) Records excluded by tle and abstracts (n =295) Full-text arcles assessed for eligibility (n = 33) Full-text arcles excluded, with reasons (n = 10) Intervenon (1) ( ) Studies included in qualitave synthesis (n = 23) Studies included in quantave synthesis (meta-analysis) (n = 11) Records excluded (n = 12 ) Populaon (2) Type of intervenon (2) Screening Idenficaon Included 6 of 9 | REZAEI Et Al. 4 | DISCUSSION This is the first meta- analysis that evaluates the effect of LAS on VO2 max in healthy human subjects. The results indicated that LAS resulted in a mean increase of 0.07 L min−1 for VO2 max compared with placebo (95% CI, 0.00– 0.13). No significant heterogeneity was detected in this meta- analysis. The subgroup analysis indicated that supplementation with L- arginine alone significantly increased VO2 max compared to the other types of arginine or combined with other metabolites or supplements. Evidence suggest the relationship between LAS and im- proved exercise performance. L- arginine is reported to have a key role in creatine synthesis as well as in increase endogenous growth hormone (Campbell et al., 2004). L- arginine is also the substrate for nitric oxide synthesis that plays a role in the autoregulation of blood flow, myocyte differentiation, respira- tion, and glucose homeostasis in muscle (Stamler & Meissner, 2001). Although some studies have shown a positive effect of LAS on exercise performance, the results of the trials which assessed the effect of LAS on VO2 max were inconsistent. A positive effect of LAS on VO2 max was identified in the present meta- analysis. This finding is generally in line with some of the individual studies selected for this re- view. Hosseini et al. (2015) reported that 4 g/day arginine FIGURE 4 Forest plot comparing the effects of L- Arg supplementation on VO2 max Subgroup analysis for VO2 max Subgroup No. of trials WMD (95% CI) Test for the overall effect Test for heterogeneity I2 (%) Duration of study (days) >14 days 8 0.05 (−0.04, −0.13) p = 0.261 p = 0.102 41.5 ≤14 days 3 0.11 (−0.06, 0.28) p = 0.188 p = 0.596 0.0 L- arginine dose (g/day) <5 6 0.07 (−0.03, 0.17) p = 0.186 p = 0.047 55.4 ≥5 5 0.03 (−0.11, −0.16) p = 0.704 p = 0.969 0.0 Type of L- arginine LAS 5 0.11 (0.08, 0.14) p = 0.000 p = 0.485 0.0 AA 2 −0.06 (−0.25, 0.12) p = 0.506 p = 0.250 24.5 Other 4 0.01 (−0.14, 0.17) p = 0.852 p = 0.956 0.0 Abbreviations: AA, arginine aspartate; AAKG, arginine alpha- Ketoglutarate; LAS, L- arginine supplementation. TABLE 2 Subgroup analysis was performed based on the study duration, dosage of L- arginine, and the form of arginine supplementation  | 7 of 9 REZAEI Et Al. supplementation for 4  weeks could significantly increase VO2 max and subsequently improved sports performance in athletes. Another study conducted by Moazami et al. (2015) reported that VO2 max was significantly increased after a 7- day supplementation of L- arginine at the dose of 21 g/day in female athletes. In addition, a placebo- controlled trial (Pahlavani et al., 2017) indicated that the oral supple- mentation of L- arginine at the dose of 2 g/day for 45 days could increase VO2 max. Conversely, Burtscher et al. (2005) found that 3 weeks of L- arginine- L- aspartate supplementa- tion at the dose of 3 g/day resulted in lower oxygen con- sumption and reduced ventilation during submaximal cycle exercise. This may be explained by the difference in phys- iological functions at a maximum level of effort compared with a submaximal (Larsen et al., 2007). It seems that nitric oxide derived from L- arginine through competitive inhi- bition of oxygen use in the electron transport chain result in lower whole- body oxygen consumption at submaximal work (Burtscher et al., 2005; Larsen et al., 2007; Schweizer & Richter, 1994). However, some studies did not observe any significant asso- ciation between the intake of LAS and VO2 max (Abel et al., 2005; Zak et al., 2015). These inconsistent results may be ex- plained by the different test protocols applied, study duration, dosages of L- arginine, form of L- arginine supplement, and also the level of physical fitness. For example, oral supplementation of L- arginine was used in combination with various other me- tabolites/salts in several studies that may cause synergistic or antagonistic effects (McConell, 2007). Furthermore, the train- ing status of the subjects seems to be an important factor related to the positive effect of LAS. LAS could have lower positive effects in well- trained participants comparing with untrained people (Besco et al., 2012). Moreover, different L- arginine dos- ages used in chronic and acute supplementation protocols could have different physiological mechanisms of action. A recent meta- analysis reported that the effective dose of LAS should be adjusted to 0.15 g/kg body weight taken 60– 90 min before exercise in the acute protocol or 10– 12 g LAS for 8 weeks in chronic protocol for improving both aerobic and anaerobic per- formances (Viribay et al., 2020). L- arginine can improve exercise performance by en- hancing protein synthesis and reducing muscle fiber damage (Lomonosova et al., 2014). It is also the precursor of nitric oxide that is used to increase endurance and improvement in blood flow (Alvares et al., 2011; Moncada & Higgs, 1993).One of the possible mechanisms to describe the increase in VO2 max is the nitric oxide derived from L- arginine that results in vessel vasodilatation and flow, which, in turn, may positively influence coronary perfusion. An increase in NO production may enhance oxygen and nutrients delivery to the active muscles. Therefore, oxygen consumption increases dramatically in the active mus- cles with a parallel increase in muscle blood flow. (Burgomaster et al., 2006; Nagaya et al., 2001; Stamler & Meissner, 2001). The oral LAS also facilitates the phase II pulmonary VO2 response. The proposed mechanism to explain this effect is an increase in L- arginine availability, with subsequent increases in certain tricarboxylic acid cycle intermediates which fi- nally lead to enhance the oxidative metabolism (Koppo et al., 2009). However, further studies are needed to understand the exact mechanisms of how L- arginine affects VO2 max in healthy human subjects. Although this is the first meta- analysis that evaluates the effect of LAS on VO2 max in healthy human subjects, it has some limitations. There were some differences in the sup- plementation protocols, doses, timing, and also form of L- arginine in the included trials which limited the extraction of strong conclusions. 5 | CONCLUSIONS This meta- analysis indicated that LAS had a positive effect on increasing VO2 max. Future homogeneous and well- designed randomized clinical trials are needed to a deep un- derstand of the effects of L- arginine on VO2 max in healthy human subjects. FIGURE 5 The sensitivity analysis was performed using the “one- study- removed” method to survey the impact of each study on the effect size 8 of 9 | REZAEI Et Al. ETHICS APPROVAL AND CONSENT TO PARTICIPATE This study has been approved by Local ethics review boards at Shahid Beheshti University of Medical Sciences. CONSENT FOR PUBLICATION Institutional consent forms were used in this study. ACKNOWLEDGMENTS This study was conducted at the Student research center of Shahid Beheshti University of Medical Sciences, Tehran, Iran (code 13172). CONFLICT OF INTEREST The authors declare that they have no conflict of interest. AUTHORS' CONTRIBUTIONS Maryam Gholamalizadeh, Saeid Doaei, and Shahla Rezaei designed the study, and were involved in the data collec- tion, analysis, and drafting of the manuscript. Reza Tabrizi, Peyman Nowrouzi- Sohrabi, and Samira Rastgoo were in- volved in the design of the study, analysis of the data, and critically reviewed the manuscript. All authors read and ap- proved the final manuscript. DATA AVAILABILITY STATEMENT Not applicable. ORCID Saeid Doaei  https://orcid.org/0000-0002-2532-7478 REFERENCES Abel, T., Knechtle, B., Perret, C., Eser, P., Von Arx, P., & Knecht, H. (2005). Influence of chronic supplementation of arginine aspar- tate in endurance athletes on performance and substrate metabo- lism. International Journal of Sports Medicine, 26(05), 344– 349. https://doi.org/10.1055/s- 2004- 821111 Adams, M. R., Forsyth, C. J., Jessup, W., Robinson, J., & Celermajer, D. S. (1995). Oral L- arginine inhibits platelet aggregation but does not enhance endothelium- dependent dilation in healthy young men. Journal of the American College of Cardiology, 26(4), 1054– 1061. https://doi.org/10.1016/0735- 1097(95)00257 - 9 Alvares, T. S., Meirelles, C. M., Bhambhani, Y. N., Paschoalin, V. M., & Gomes, P. S. (2011). L- Arginine as a potential ergogenic aidin healthy subjects. Sports Medicine, 41(3), 233– 248. https://doi. org/10.2165/11538 590- 00000 0000- 00000 Besco, R., Sureda, A., Tur, J. A., & Pons, A. (2012). The effect of nitric- oxide- related supplements on human performance. Sports Medicine, 42(2), 99– 117. https://doi.org/10.2165/11596 860- 00000 0000- 00000 Burgomaster, K. A., Heigenhauser, G. J., & Gibala, M. J. (2006). Effect of short- term sprint interval training on human skeletal muscle carbohydrate metabolism during exercise and time- trial perfor- mance. Journal of Applied Physiology, 100(6), 2041– 2047. https:// doi.org/10.1152/jappl physi ol.01220.2005 Burtscher, M., Brunner, F., Faulhaber, M., Hotter, B., & Likar, R. (2005). The prolonged intake of L- arginine- L- aspartate reduces blood lactate accumulation and oxygen consumption during sub- maximal exercise. Journal of Sports Science & Medicine, 4(3), 314. Camic, C. L., Housh, T. J., Mielke, M., Zuniga, J. M., Hendrix, C. R., Johnson, G. O., Schmidt, R. J., & Housh, D. J. (2010). The effects of 4 weeks of an arginine- based supplement on the gas exchange threshold and peak oxygen uptake. Applied Physiology, Nutrition, and Metabolism, 35(3), 286– 293. https:// doi.org/10.1139/H10- 019 Campbell, B. I., La Bounty, P. M., & Roberts, M. (2004). The ergogenic potential of arginine. Journal of the International Society of Sports Nutrition, 1(2), 1– 4. https://doi.org/10.1186/1550- 2783- 1- 2- 35 Campbell, B., Roberts, M., Kerksick, C., Wilborn, C., Marcello, B., Taylor, L., Nassar, E., Leutholtz, B., Bowden, R., Rasmussen, C., Greenwood, M., & Kreider, R. (2006). Pharmacokinetics, safety, and effects on exercise performance of L- arginine α- ketoglutarate in trained adult men. Nutrition, 22(9), 872– 881. https://doi. org/10.1016/j.nut.2006.06.003 Chen, S., Kim, W., Henning, S. M., Carpenter, C. L., & Li, Z. (2010). Arginine and antioxidant supplement on performance in elderly male cyclists: A randomized controlled trial. Journal of the International Society of Sports Nutrition, 7(1), 13. https://doi. org/10.1186/1550- 2783- 7- 13 Denis, C., Dormois, D., Linossier, M., Eychenne, J., Hauseux, P., & Lacour, J. (1991). Effect of arginine aspartate on the exercise- induced hyperammoniemia in humans: a two periods cross- over trial. Archives Internationales de Physiologie, de Biochimie et de Biophysique, 99(1), 123– 127. https://doi.org/10.3109/13813 45910 9145914 Doaei, S., Hajiesmaeil, M., Aminifard, A., Mosavi- Jarrahi, S., Akbari, M., & Gholamalizadeh, M. (2018). Effects of gene polymorphisms of metabolic enzymes on the association between red and pro- cessed meat consumption and the development of colon cancer; a literature review. Journal of Nutritional Science, 7. https://doi. org/10.1017/jns.2018.17 Epstein, F. H., Moncada, S., & Higgs, A. (1993). The L- arginine- nitric oxide pathway. New England Journal of Medicine, 329(27), 2002– 2012. https://doi.org/10.1056/NEJM1 99312 30329 2706 Gutin, B., Yin, Z., Humphries, M. C., & Barbeau, P. (2005). Relations of moderate and vigorous physical activity to fitness and fatness in adolescents. The American Journal of Clinical Nutrition, 81(4), 746– 750. https://doi.org/10.1093/ajcn/81.4.746 Haghshenas, R., Jamshidi, Z., Doaei, S., & Gholamalizadeh, M. (2019). The effect of a high- intensity interval training on plasma vitamin D level in obese male adolescents. Indian Journal of Endocrinology and Metabolism, 23(1), 72– 75. https://doi.org/10.4103/ijem. IJEM_267_18 Higgins, J. (2011). Cochrane handbook for systematic reviews of in- terventions. Version 5.1. 0 [updated March 2011]. The Cochrane Collaboration. Retrieved from www.cochr ane- handb ook.org Hosseini, A., & Valipour Dehnou, V., Azizi, M., & Khanjari Alam, M. (2015). Effect of high- intensity interval training (HIT) for 4 weeks with and without L- arginine supplementation on the per- formance of women's futsal players. Quarterly of Horizon of Medical Sciences, 21(2), 113– 119. https://doi.org/10.18869/ acadp ub.hms.21.2.113 Kalman, D., Harvey, P. D., Perez Ojalvo, S., & Komorowski, J. (2016). Randomized prospective double- blind studies to evaluate the  | 9 of 9 REZAEI Et Al. cognitive effects of inositol- stabilized arginine silicate in healthy physically active adults. Nutrients, 8(11), 736. https://doi. org/10.3390/nu811 0736 Koppo, K., Taes, Y. E., Pottier, A., Boone, J., Bouckaert, J., & Derave, W. (2009). Dietary arginine supplementation speeds pulmo- nary VO2 kinetics during cycle exercise. Medicine & Science in Sports & Exercise, 41(8), 1626– 1632. https://doi.org/10.1249/ MSS.0b013 e3181 9d81b6 Larsen, F., Weitzberg, E., Lundberg, J., & Ekblom, B. (2007). Effects of dietary nitrate on oxygen cost during exercise. Acta Physiologica, 191(1), 59– 66. https://doi.org/10.1111/j.1748- 1716.2007.01713.x Liberati, A., Altman, D. G., Tetzlaff, J., Mulrow, C., Gøtzsche, P. C., Ioannidis, J. P., Clarke, M., Devereaux, P. J., Kleijnen, J., & Moher, D. (2009). The PRISMA statement for reporting system- atic reviews and meta- analyses of studies that evaluate health care interventions: Explanation and elaboration. Journal of Clinical Epidemiology, 62(10), e1– e34. https://doi.org/10.1016/j.jclin epi.2009.06.006 Lomonosova, Y. N., Shenkman, B. S., Kalamkarov, G. R., Kostrominova, T. Y., & Nemirovskaya, T. L. (2014). L- arginine supplementation protects exercise performance and structural integrity of muscle fibers after a single bout of eccentric exercise in rats. PLoS One, 9(4), e94448. https://doi.org/10.1371/journ al.pone.0094448 Mashiko, T., Umeda, T., Nakaji, S., & Sugawara, K. (2004). Position related analysis of the appearance of and relationship between post- match physical and mental fatigue in university rugby foot- ball players. British Journal of Sports Medicine, 38(5), 617– 621. https://doi.org/10.1136/bjsm.2003.007690 McConell, G. K. (2007). Effects of L- arginine supplementation on exercise metabolism. Current Opinion in Clinical Nutrition and Metabolic Care, 10(1), 46– 51. https://doi.org/10.1097/ MCO.0b013 e3280 1162fa Melik, Z., Zaletel, P., Virtic, T., & Cankar, K. (2017). L- arginine as dietary supplement for improving microvascular function. Clinical Hemorheology and Microcirculation, 65(3), 205– 217. Moazami, M., Taghizadeh, V., Ketabdar, A., Dehbashi, M., & Jalilpour, R. (2015). Effects of oral L- arginine supplementation for a week, on changes in respiratory gases and blood lactate in female hand- ballists. Iranian Journal of Nutrition Sciences & Food Technology, 9(4), 45– 52. Muazzezzaneh, A., Keshavarz, S. A., Sabour Yaraghi, A. A., Djalali, M., & Rahimi, A. (2010). Effect of L- arginine supplementation on blood lactate level and VO2 max at anaerobic threshold perfor- mance. KAUMS Journal (FEYZ), 14(3), 200– 208. Nagaya, N., Uematsu, M., Oya, H., Sato, N., Sakamaki, F., Kyotani, S., Ueno, K., Nakanishi, N., Yamagishi, M., & Miyatake, K. (2001). Short- term oral administration of L- arginine improves hemodynamics and exercise capacity in patients with precapillary pulmonary hypertension. American Journal of Respiratory and Critical Care Medicine, 163(4), 887– 891. https://doi.org/10.1164/ ajrccm.163.4.2007116 Pahlavani, N., Entezari, M., Nasiri, M., Miri, A., Rezaie, M., Bagheri- Bidakhavidi, M., & Sadeghi, O. (2017). The effect of l- arginine supplementation on body composition and performance in male athletes: A double- blinded randomized clinical trial. European Journal of Clinical Nutrition, 71(4), 544– 548. https://doi. org/10.1038/ejcn.2016.266 Preli, R. B., Klein, K. P., & Herrington, D. M. (2002). Vascular effects of dietary L- arginine supplementation. Atherosclerosis, 162(1), 1– 15. https://doi.org/10.1016/S0021 - 9150(01)00717 - 1 Ruiz, J. R., Rizzo, N. S., Hurtig- Wennlöf, A., Ortega, F. B., W àrnberg, J., & Sjöström, M. (2006). Relations of total physical activity and intensity to fitness and fatness in children: The European Youth Heart Study. The American Journal of Clinical Nutrition, 84(2), 299– 303. https://doi.org/10.1093/ajcn/84.2.299 Schweizer, M., & Richter, C. (1994). Nitric oxide potently and revers- ibly deenergizes mitochondria at low oxygen tension. Biochemical and Biophysical Research Communications, 204(1), 169– 175. https://doi.org/10.1006/bbrc.1994.2441 Shidfar, F., Bahrololumi, S. S., Doaei, S., Mohammadzadeh, A., Gholamalizadeh, M., & Mohammadimanesh, A. (2018). The ef- fects of extra virgin olive oil on alanine aminotransferase, aspartate aminotransferase, and ultrasonographic indices of hepatic steatosis in nonalcoholic fatty liver disease patients undergoing low calo- rie diet. Canadian Journal of Gastroenterology and Hepatology, 2018, 1– 7. https://doi.org/10.1155/2018/1053710 Stamler, J. S., & Meissner, G. (2001). Physiology of nitric oxide in skel- etal muscle. Physiological Reviews, 81(1), 209– 237. https://doi. org/10.1152/physr ev.2001.81.1.209 Sunderland, K. L., Greer, F., & Morales, J. (2011). JOURNAL/ jscr/04.02/00124278- 201103000- 00034/ENTITY_OV0312/ v/2017- 07- 20T235437Z/r/image- pngo2max and ventilatory threshold of trained cyclists are not affected by 28- day L- arginine supplementation. The Journal of Strength and Conditioning Research, 25(3), 833– 837. https://doi.org/10.1519/jsc.0b013 e3181 c6a14d Viribay, A., Burgos, J., Fernández- Landa, J., Seco- Calvo, J., & Mielgo- Ayuso, J. (2020). Effects of arginine supplementation on athletic performance based on energy metabolism: A systematic review and meta- analysis. Nutrients, 12(5), 1300. Yaman, H., Tiryaki- Sönmez, G., & Gürel, K. (2010). Effects of oral L- arginine supplementation on vasodilation and VO2 max in male soccer players. Biomedical Human Kinetics, 2(1), 25– 29. https:// doi.org/10.2478/v1010 1- 010- 0006- x Zak, R. B., Camic, C. L., Hill, E. C., Monaghan, M. M., Kovacs, A. J., & Wright, G. A. (2015). Acute effects of an arginine- based supplement on neuromuscular, ventilatory, and metabolic fatigue thresholds during cycle ergometry. Applied Physiology, Nutrition, and Metabolism, 40(4), 379– 385. https://doi.org/10.1139/ apnm- 2014- 0379 How to cite this article: Rezaei S, Gholamalizadeh M, Tabrizi R, Nowrouzi- Sohrabi P, Rastgoo S, Doaei S. The effect of L- arginine supplementation on maximal oxygen uptake: A systematic review and meta- analysis. Physiol Rep. 2021;9:e14739. https:// doi.org/10.14814/phy2.14739

The effect of L-arginine supplementation on maximal oxygen uptake: A systematic review and meta-analysis.

[]

Rezaei, Shahla,Gholamalizadeh, Maryam,Tabrizi, Reza,Nowrouzi-Sohrabi, Peyman,Rastgoo, Samira,Doaei, Saeid

eng

PMC9864675

Citation: Helwig, J.; Diels, J.; Röll, M.; Mahler, H.; Gollhofer, A.; Roecker, K.; Willwacher, S. Relationships between External, Wearable Sensor-Based, and Internal Parameters: A Systematic Review. Sensors 2023, 23, 827. https://doi.org/10.3390/s23020827 Academic Editor: George Grouios Received: 26 October 2022 Revised: 5 January 2023 Accepted: 9 January 2023 Published: 11 January 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). sensors Systematic Review Relationships between External, Wearable Sensor-Based, and Internal Parameters: A Systematic Review Janina Helwig 1,2,*, Janik Diels 1, Mareike Röll 1, Hubert Mahler 1,3, Albert Gollhofer 1, Kai Roecker 1,4 and Steffen Willwacher 2 1 Institute of Sport and Sport Science, Albert-Ludwigs University Freiburg, 79117 Freiburg, Germany 2 Institute for Advanced Biomechanics and Motion Studies, Offenburg University, Max-Planck Straße 1, 77656 Offenburg, Germany 3 Sport-Club Freiburg e.V., Achim-Stocker-Str. 1, 79108 Freiburg, Germany 4 Institute for Applied Health Promotion and Exercise Medicine, Furtwangen University, 78120 Furtwangen, Germany * Correspondence: [email protected] Abstract: Micro electro-mechanical systems (MEMS) are used to record training and match play of intermittent team sport athletes. Paired with estimates of internal responses or adaptations to exercise, practitioners gain insight into players’ dose–response relationship which facilitates the prescription of the training stimuli to optimize performance, prevent injuries, and to guide rehabilitation processes. A systematic review on the relationship between external, wearable-based, and internal parameters in team sport athletes, compliant with the PRISMA guidelines, was conducted. The literature research was performed from earliest record to 1 September 2020 using the databases PubMed, Web of Science, CINAHL, and SportDISCUS. A total of 66 full-text articles were reviewed encompassing 1541 athletes. About 109 different relationships between variables have been reviewed. The most investigated relationship across sports was found between (session) rating of perceived exertion ((session-)RPE) and PlayerLoad™ (PL) with, predominantly, moderate to strong associations (r = 0.49–0.84). Relation- ships between internal parameters and highly dynamic, anaerobic movements were heterogenous. Relationships between average heart rate (HR), Edward’s and Banister’s training impulse (TRIMP) seem to be reflected in parameters of overall activity such as PL and TD for running-intensive team sports. PL may further be suitable to estimate the overall subjective perception. To identify high fine-structured loading—relative to a certain type of sport—more specific measures and devices are needed. Individualization of parameters could be helpful to enhance practicality. Keywords: player monitoring; external load; internal load; MEMS; wearable sensors 1. Introduction Player monitoring in sports aims at optimizing training adaptations to improve per- formance and reduce injury risk [1]. Adaptations occur based on psycho-physiological responses to exercise. These internal responses are stimulated by the internal load expe- rienced during exercise; they are difficult to measure directly in a non-invasive way and can only be estimated in typical sports settings. Estimates of internal load and an athlete’s response to exercise are commonly provided by markers of cardiovascular, neuromuscular, or metabolic functioning, e.g., measurements of heart rate (HR) or ratings of perceived exertion (RPE) [2]. Adaptations to training and match demands may be estimated by detecting a change in fitness or fatigue state using, e.g., spiroergometry, cardiopulmonary fitness tests, immunological or hormonal blood markers. Adaptations may be positive or negative, or the fitness state may be maintained. Negative adaptations occur during detraining phases (i.e., off-season) and overtraining, whereas positive adaptations occur after optimal loading and adequate recovery periods. Sensors 2023, 23, 827. https://doi.org/10.3390/s23020827 https://www.mdpi.com/journal/sensors Sensors 2023, 23, 827 2 of 22 Internal loading due to sports activities primarily results from movement-related force application demands. Forces need to be applied to the environment to cover running distances, perform changes in movement direction or accelerate or decelerate the body, e.g., during acceleration, stopping, or jumping tasks. Applying forces to the environment results in reaction forces acting on athletes’ bodies, determining the external load stimulus applied to the biological system. Physical external loads applied over time result in different types of internal loads (e.g., mechanical or physiological), which determine the body’s adaptations. Knowledge of the internal response and adaptation to a given dose of external load is crucial for optimal, injury-free training progress. The internal load is influenced by individual factors such as age, gender, training experience, health status, and nutrition [3]. The link between individual characteristics, external load, internal load, exercise-induced responses, and performance adaptations is depicted in Figure 1. In the context of this paper, we refer to internal load, exercise-induced response and adaptations, and individual characteristics as internal parameters. In the same figure, the possibilities to assess these categories are displayed, as they are included in this systematic review. Sensors 2023, 23, x FOR PEER REVIEW 2 of 24 may be positive or negative, or the fitness state may be maintained. Negative adaptations occur during detraining phases (i.e., off-season) and overtraining, whereas positive adaptations occur after optimal loading and adequate recovery periods. Internal loading due to sports activities primarily results from movement-related force application demands. Forces need to be applied to the environment to cover running distances, perform changes in movement direction or accelerate or decelerate the body, e.g., during acceleration, stopping, or jumping tasks. Applying forces to the environment results in reaction forces acting on athletes’ bodies, determining the external load stimulus applied to the biological system. Physical external loads applied over time result in different types of internal loads (e.g., mechanical or physiological), which determine the body’s adaptations. Knowledge of the internal response and adaptation to a given dose of external load is crucial for optimal, injury-free training progress. The internal load is influenced by individual factors such as age, gender, training experience, health status, and nutrition [3]. The link between individual characteristics, external load, internal load, exercise-induced responses, and performance adaptations is depicted in Figure 1. In the context of this paper, we refer to internal load, exercise-induced response and adaptations, and individual characteristics as internal parameters. In the same figure, the possibilities to assess these categories are displayed, as they are included in this systematic review. Figure 1. Interaction of external and internal parameters and possibilities to assess those parameters. In this framework, it can be distinguished between four different categories of internal parameter assessments: First, the internal load estimates collected during exercise, primarily made up of HR-based indices and RPE or session-RPE. Second, the exercise-induced responses measured post-exercise due to the delayed response of specific systems to activity, such as creatine kinase (CK), an indicator of muscle damage. Third, the body’s adaptations may be assessed over time (usually tested under standardized conditions, e.g., maximal oxygen uptake (VO2max) tests using ergometry). Fourth, the assessment of the current health and fitness status, which, among other parameters such as genetics, age, and gender, make up the individual characteristics. Parameters of each category are included in this systematic review if a relationship to an external load parameter, measured during training or match play using a MEMS device, was assessed. Most sports science research groups term the responses as exercise and the training or match stimuli as internal and external load, workload, or training load, respectively [1,4–7]. We acknowledged that this terminology might be misleading considering the Figure 1. Interaction of external and internal parameters and possibilities to assess those parameters. In this framework, it can be distinguished between four different categories of internal parameter assessments: First, the internal load estimates collected during exercise, primarily made up of HR-based indices and RPE or session-RPE. Second, the exercise-induced responses measured post-exercise due to the delayed response of specific systems to activity, such as creatine kinase (CK), an indicator of muscle damage. Third, the body’s adaptations may be assessed over time (usually tested under standardized conditions, e.g., maximal oxygen uptake (VO2max) tests using ergometry). Fourth, the assessment of the current health and fitness status, which, among other parameters such as genetics, age, and gender, make up the individual characteristics. Parameters of each category are included in this systematic review if a relationship to an external load parameter, measured during training or match play using a MEMS device, was assessed. Most sports science research groups term the responses as exercise and the training or match stimuli as internal and external load, workload, or training load, respectively [1,4–7]. We acknowledged that this terminology might be misleading considering the mechanical concepts where the load is weight or resistance, which is expressed in Newtons (N), as defined by the Système International d’Unites (SI), as various other research groups have indicated [8–11]. In order to cover the literature comprehensively, the terms external and internal load were included during the search process and are further used throughout this systematic review, but with their meaning as outlined in Figure 1. Sensors 2023, 23, 827 3 of 22 Internal parameters, such as biochemical, hormonal and immunological parameters, are often impractical to collect during training sessions and competitions; doing so might be time and cost-intensive [1]. External load variables can be measured more efficiently and time-effectively. Thus, knowing the relationship between external and internal parameters would be practical to learn about potential dose-response relationships [12,13]. External load variables can be tested in laboratory or field settings. While labora- tory settings offer access to accurate gold-standard approaches to quantify external load (e.g., through direct measurements of ground reaction forces (GRFs) using force platforms or the inverse dynamics-based calculation of external joint moments), field settings offer greater ecological validity and the potential to reach larger numbers of athletes. In the field setting, external load variables can be measured using lightweight, body- worn sensors. With the introduction of global navigation satellite systems (GNSS) devices into the player monitoring market, the research around workload quantification and load monitoring has increased exponentially in the last 15–20 years [14,15]. Next to GNSS, wearable sensor-based load monitoring systems may consist of local positioning systems (LPS), offering higher accuracy in the location of, e.g., team sports athletes in the field of play. Another promising combination of sensor technologies are inertial measurement units (IMUs), commonly combining 3D accelerometers, 3D gyroscope, and 3D magnetometers. Combined in one unit, these systems belong to the group of micro-electro-mechanical systems (MEMS). Commercially available physical activity trackers have gained tremen- dous interest in the recent decade. “Wearable technology was the top worldwide fitness trend in 2016 and 2017” [16]. Besides physical parameters (i.e., step count), wearables aim to estimate the internal load a person is experiencing. Some smartwatches provide an estimate of, e.g., the metabolic work and power [17]. However, the validity and reliability of these parameters may be questionable and highly dependent on the hardware used, and algorithms applied [17]. Thus, the factors mentioned above are not always clearly defined or explained; yet, the parameters are still widely used to quantify the general population’s activity and the external load and internal parameters of team sport athletes. However, keeping track of loading in team sports is a complex task: Running-based team sports are intermittent sports, consisting of hundreds of brief and very intense actions, such as jumps, tackles, changes of directions, accelerations, and decelerations [18]. These movements are metabolically and physically demanding, more than the same distance covered at a constant speed [19]; thus, specific approaches to quantifying loads for team sport athletes are needed. Consequently, sports scientists and tracking device manufacturers have created several parameters such as “PlayerLoad™” (PL), “impact load”, or “leg stiffness”, intending to capture load characteristics and their changes with, e.g., fatigue or training status. One of the main challenges in developing load parameters is to capture the demands of accelerating and decelerating, as well as turns and tackles. “Metabolic power”, for example, is one more recently developed approach that attempts to capture the demands of accelerating based on the assumption that this is comparable to the metabolic demands of running uphill [20]. Nevertheless, it does not capture the lateral movements, turns, and tackles. New possibilities have been created using MEMS to quantify loading in team sports athletes. Nevertheless, a consensus on quantifying the “internal” load of team sport athletes by “external” locomotor measurements is still missing [1,21–23]. Consequently, common ground for best practice in load monitoring of team sport athletes has not been established so far [1,22]. In particular, detailed knowledge about the relationship between a recorded external load and internal parameters is rare. A recent meta-analysis has analyzed the relationship between external and internal load parameters in team sport athletes [24]. This work focused on the relationship between HR indices, RPE, and various external load parameters. However, as outlined above, beyond internal load, a multitude of internal processes are stimulated, which are relevant for the psycho-physiological response and adaptation to exercise, as well as the risk of injury. Sensors 2023, 23, 827 4 of 22 Consequently, two main challenges regarding load monitoring in team sport athletes have been identified: First, the complex task of quantifying the complex loading situation of intermittent team sports, and second, the difficulty of knowing the relationship between the given MEMS-based external load and the athlete’s individual internal loading and consequently exercise-responses and adaptations within different domains. Identifying these relationships offers great potential to improve the understanding of individual load- response profiles. Therefore, this systematic review addresses MEMS-based external load parameters and their relationship to various internal parameters, encompassing biochemical, neuromus- cular, subjective, cardiovascular, and further domains. This work could aid practitioners in choosing and interpreting appropriate parameters to monitor load in a time- and cost- effective manner to provide the appropriate stimuli to induce adaptations to improve sports performance and decrease the risk of injury. 2. Materials and Methods 2.1. Article Search, Inclusion, Exclusion A systematic literature review was conducted based on the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines [25]. The following elec- tronic databases were searched: PubMed, Web of Science, SPORTDiscus, and CINAHL. The search term was created by linking four sections with the Boolean operator “AND”, en- suring that at least one word from each section will appear in the results. Keywords within one section were connected with the operator “OR”. The first section contained various team sports. The second section contained methods and systems used to monitor athletes. The third and fourth sections contained numerous external and internal or performance measurement parameters. Truncation searching was employed to find variations of certain words (see Table 1 for the complete search term). The databases were searched with no restrictions from the earliest records available up to September 1, 2020. Results were stored in a citation manager, and all duplicates were removed (search process see Figure 2). All abstracts were then screened for eligibility regarding the inclusion and exclusion criteria assessed. Any studies including athletes younger than 18 years were excluded as cognitive development influences the accuracy of the RPE [26]. Articles were considered if they showed a relationship measure between one external and one internal or performance parameter obtained from able-bodied team sport athletes during regular training or match play which did not include additional interventions, such as nutritional interventions or manipulated play. For the complete list of inclusion and exclusion criteria, please refer to Table 2. All data were independently extracted by two researchers (JH, JD). In case of disagreement, a consensus was found by a third reviewer (KR). The study further adheres to the ethical standards in sports and exercise science research [27]. Table 1. Search Term: Categories are connected with the Boolean operator “AND”; key words within a category are connected with “OR”. Category Keywords Team Sport “Team Sport*” OR soccer OR football OR handball OR basketball OR rugby OR volleyball OR futsal OR netball Monitoring system monitoring OR tracking OR GPS OR “Global Positioning System”[MeSH] OR LPS OR “Local Positioning System”[MeSH] OR IMU OR “inertial measurement unit” OR acceleromet* OR MEMS OR microsensor OR “time motion” OR TMA OR “motion analysis”[MeSH] OR “wearable technologies”[MeSH] External load workload OR load OR speed OR ACWR OR “acute to chronic work ratio” OR “work:rest” OR distance OR acceleration OR “metabolic power” OR “metabolic load” OR PlayerLoad OR intensit* OR “energy expenditure” OR “high intensity burst*” OR “work ratio” OR “fatigue index” OR “physical” OR “repeated sprintability Internal load “internal load” OR RPE OR “rate of perceived exertion” OR RPE OR sRPE OR “heart rate” OR HR OR TRIMP OR questionnaire OR biochemical OR physiological OR neurological OR fatigue OR blood OR lactate OR SPX OR Spiroergometry OR “breath gas analysis” OR CK OR “creatine kinase” OR VO2 OR “anaerobic threshold” Sensors 2023, 23, 827 5 of 22 Figure 2. Flow diagram of the study selection process adopted from the PRISMA guidelines. Table 1. Search Term: Categories are connected with the Boolean operator “AND”; key words within a category are connected with “OR”. Category Keywords Figure 2. Flow diagram of the study selection process adopted from the PRISMA guidelines. Table 2. Inclusion and exclusion criteria. Inclusion Exclusion Topic of the article is human physical performance Topic not related to physical performance or non-human subjects Original research Surveys, opinions, books, case studies, non-academic text, reviews, conference abstracts Competitive field- or court-based team sport athletes Individual sports, ice-, sand-, or water-based team sports, referees Adult athletes Athletes under 18 years of age Able-bodied, non-injured athletes Special populations (i.e., clinical), mentally or physically impaired athletes, injured athletes Training or match play Laboratory settings, and field-based settings coupled with an intervention (i.e., nutritional intervention). Report of at least one external and one internal load measure or physiological fitness assessment Report of only internal or only external measures Report of a relationship between internal and external measures No relationship between internal and external measures reported Use of GNSS, MEMS, IMU, LPS Use of timing gates, measuring tapes, video-based tracking Good, very good, or excellent methodological quality based on the checklist used for this review Poor methodological quality based on the checklist used for this review 2.2. Study Quality Assessment After the final selection was made, the quality of the selected studies was assessed using a 16-item checklist developed by Law et al. [28] and modified by Sarmento et al. [29], which has also been used in previous reviews [29,30]. The authors are aware that a risk of bias assessment may be superior to a checklist summarizing components into a single number, especially when concerned about randomized controlled trials. This systematic review, however, is concerned with observational studies, and thus, the authors decided Sensors 2023, 23, 827 6 of 22 on a quality checklist that is applicable to the topic at hand. The items on the checklist were scored on a binary scale (0 = no, 1 = yes). Items 6 and 13 included the option “not applicable”. The sum of the scores for each study divided by the maximum value possible for that study represented the quality score. Expressed as a percentage, the score then indicated the methodological quality of the studies. The following meaning was associated with the final percentage: low methodological quality ≤ 50%; good methodological quality 51–75%, excellent methodological quality > 75%. The quality assessment was carried out independently by two reviewers (JH and KR). Disagreement was solved by discussion. 2.3. Data Extraction Data extraction was done using a custom-made sheet pilot tested on five randomly chosen articles. The sheet was redefined, and its final version was used by one reviewer (JH) who performed the data extraction. In case of unclear or missing data, corresponding authors were contacted. The following data were extracted from the studies: (1.) The type of team sport; (2.) the study sample (along with the number of participants, gender, and level/league athletes competed in); (3.) the external parameters recorded or calcu- lated; (4.) the internal parameters measured or calculated and/or the fitness assessment; (5.) the relationship between the external and internal parameters as indicated by statistical association or predictive measures. 2.4. Data Synthesis Data were categorized into groups consisting of the different team sports. Then, subgroups according to the parameters analyzed were created. The subgroups are based on Figure 1 and consist of: the assessment of the relationship between external load parameters and internal load collected during exercise, exercise-induced responses, adaptations, and individual characteristics. A descriptive synthesis was undertaken with the data structured in a table contain- ing the team sport, the studies included, the load parameters collected, including their frequency of use per sport, and the statistical relational measures between the external and internal parameters. The overall frequency of use of each external and each internal parameter was visualized using pie charts. 3. Results 3.1. Search Results The initial search returned 3573 articles. A total of 2234 records remained after re- moving duplicates; these articles were screened by title and abstracts against the eligibility criteria. After further exclusion of studies (n = 2178) that did not meet the criteria, 66 articles remained for the final analysis (Figure 2). The main reasons for exclusion were not using MEMS-based parameters, not analyzing regular training or match play, and analyzing only internal or only external parameters. The references of the included articles were screened, but no further study met the inclusion criteria. The mean methodological quality score of the included studies was 84.6% (+/−8.4%). No article was excluded due to low quality. Ten studies scored between 51 and 75% as good methodological quality. The remainder (n = 56) qualified as excellent regarding methodological quality. The most common item to lose quality points on was item 5: justification of the study sample size. 3.2. Basic Characteristics of Included Studies The articles included in this systematic review ranged from 2011 to 2019. The sports analyzed were: American football (n = 6), Australian football (n = 11), basketball (n = 4), field hockey (n = 1), rugby union and rugby league (n = 8), soccer (n = 35), and tag football (n = 1). The participants were professional (n = 606), elite (n = 413), college/university (n = 402), and semi-professional (n = 120) athletes. n = 62 studies included male participants, totaling 1479 male athletes. Four studies studied female participants, accounting for n = 62 female athletes. The three most commonly external parameters recorded were distances in Sensors 2023, 23, 827 7 of 22 speed zones (n = 55), total distance (n = 46), and PL (n = 34), as depicted in Figure 3. The most frequently recorded internal parameters were (session) RPE (this includes RPE as well as session RPE and thus termed “(session-)RPE” going forward) (n = 29), HR-based indices (n = 19), and well-being questionnaires (n = 17), as depicted in Figure 3. The articles included in this systematic review ranged from 2011 to 2019. The sports analyzed were: American football (n = 6), Australian football (n = 11), basketball (n = 4), field hockey (n = 1), rugby union and rugby league (n = 8), soccer (n = 35), and tag football (n = 1). The participants were professional (n = 606), elite (n = 413), college/university (n = 402), and semi-professional (n = 120) athletes. n = 62 studies included male participants, totaling 1479 male athletes. Four studies studied female participants, accounting for n = 62 female athletes. The three most commonly external parameters recorded were distances in speed zones (n = 55), total distance (n = 46), and PL (n = 34), as depicted in Figure 3. The most frequently recorded internal parameters were (session) RPE (this includes RPE as well as session RPE and thus termed “(session-)RPE” going forward) (n = 29), HR-based indices (n = 19), and well-being questionnaires (n = 17), as depicted in Figure 3. Figure 3. External and internal parameters with the number of studies they are appearing in. Parameters occurring in one or two studies only are pooled under “Further”. Acc/Dec, acceleration and deceleration parameters; Avg., average; CK, creatine kinase; CMJ, countermovement jump; Dyn., dynamic; exp., expenditure; HR, heart rate; Im.gl., immunoglobulin parameters; Max., Figure 3. External and internal parameters with the number of studies they are appearing in. Parameters occurring in one or two studies only are pooled under “Further”. Acc/Dec, acceleration and deceleration parameters; Avg., average; CK, creatine kinase; CMJ, countermovement jump; Dyn., dynamic; exp., expenditure; HR, heart rate; Im.gl., immunoglobulin parameters; Max., maximal; Met., metabolic; RHIE, repeated high-intensity efforts; (s)RPE (session) rating of perceived exertion; Well-being, well-being questionnaires; YYIR, Yo-Yo intermittent recovery test. 3.3. External and Internal Parameters About 34 external and 32 internal parameters and parameter groups were included across all studies. Different HR-based and various (session-)RPE parameters were grouped and displayed in Figures 3 and 4. The most often investigated external parameter was distance covered in specific speed zones (n = 55) which was investigated in 82% of studies included in this review, followed by total distance (n = 46), analyzed in 67% of the re- search articles in this review, and PL (n = 34), occurring in 51%. (session-)RPE (n = 29) was most often investigated amongst the internal parameters, followed by HR-based indices (n = 19) and well-being questionnaires (n = 17), occurring in 45, 28, and 27% of research articles included in this systematic review, respectively. From the 66 articles included, 109 different relationships between external and internal parameters have been extracted. The most frequently analyzed relationship was between (session-)RPE and PL with pre- dominantly moderate to strong associations (r = 0.49–0.84). The second most frequently analyzed relationship was between (session-)RPE and distances in speed zones with het- erogeneous results. All results for the 109 relationships can be found in Table S1 in the supplementary material. 3.4. Summary of Individual Studies Table 3 provides an overview of all studies included in this systematic review, grouped by sport. It includes the number of participants, their playing level, and the collected external and internal parameters. Study designs, participants, hard- and software used, and outcome measures varied noticeably such that the authors focused on describing the Sensors 2023, 23, 827 8 of 22 results of the studies rather than performing a meta-analysis. Table S1 in the supplementary material further shows the relationship measures between parameters. Sensors 2023, 23, 827 9 of 22 Table 3. Studies included in this systematic review sorted by type of team sport. The table includes information about the player level and the parameters collected. Sport Study Player Level (n = Number of Athletes) External Parameters (n = Number of Studies) Internal Parameters (n = Number of Studies) American football [31–36] University Divison I (n = 225, male) PL (AU) (n = 4) Acceleration/Deceleration (m·s−2) (n = 4) Distance in speed zones (m) (n = 2) Impacts (n) (n = 2) Stride variability (n = 1) INTERNAL LOAD PARAMETERS (session-)RPE (AU) (n = 1) EXERCISE-INDUCED RESPONSES Well-being questionnaire (5-point scale) (n = 4) S100beta (pg/mL) (n = 1) Tau concentration (pg/mL) (n = 1) Australian football [13,37–45] Professional (n = 202, male) Elite (n = 118, male) Distance in speed zones (m) (n = 13) PL (au) (n = 9) Total/Relative distance (m, m/min) (n = 9) Duration (min) (n = 5) Average speed (m/s) (n = 4) Acceleration/Deceleration (m·s−2) (n = 3) Energy expenditure (kJ/kg) (n = 2) Metabolic power concept (W/kg) (n = 2) Distance load (m2/s) (distance x mean speed) (n = 1) Effort zones (n) (n = 1) Equivalent distance (m) (n = 1) Explosive efforts (n) (n = 1) Impacts (n) (n = 1) Match exercise intensity (AU) (n = 1) INTERNAL LOAD PARAMETERS (session-)RPE (AU) (n = 7) Core temperature (C) (n = 1) EXERCISE-INDUCED RESPONSES Well-being questionnaire (5-point scale) (n = 3) CMJ (cm) (n = 1) CK (U/L) (n = 1) INDIVIDUAL CHARACTERISTICS Maximal aerobic speed (m/s) (n = 1) YYIR (m) (n = 1) Basketball [46–49] Elite (n = 12, male) Professional (n = 26, male) Semiprofessional (n = 8, male) University (n = 5, female) PL (AU) (n = 4) Acceleration/Deceleration (m·s−2) (n = 4) Jumps (n) (n = 2) IMA™ (AU) (n = 1) INTERNAL LOAD PARAMETERS (session-)RPE (AU) (n = 3) HR-based indices (n = 1) EXERCISE-INDUCED RESPONSES Tensiomyography (ms, mm) (n = 1) Field Hockey [50] Elite (n = 12, male) Acceleration/Deceleration (m·s−2) (n = 1) Distances in speed zones (m) (n = 1) Total/relative distance (m, m/min) (n = 1) EXERCISE-INDUCED RESPONSES Well-being questionnaire (5-point scale) (n = 1) Rugby Sevens [51,52] Elite (n = 24, 12 female, 12 male) Amateur (n = 10, female) Total/relative distance (m, m/min) (n = 2) Distance in speed zones (m) (n = 2) Impacts (n) (n = 1) EXERCISE-INDUCED RESPONSES CK (U/L) (n =1) Bicarbonate concentration (mmol/L) (n = 1) Lactate concentration (mmol/L) (n = 1) pH (n = 1) Sensors 2023, 23, 827 10 of 22 Table 3. Cont. Sport Study Player Level (n = Number of Athletes) External Parameters (n = Number of Studies) Internal Parameters (n = Number of Studies) Rugby League [53–56] Professional (n = 46, male) Elite (n = 45, male) Distance in speed zones (m) (n = 3) Impacts (n) (n = 3) Acceleration/Deceleration (m·s−2) (n = 2) Total/Relative distance (m, m/min) (n = 2) Duration (min) (n = 1) PL (AU) (n = 1) RHIE (n) (n = 1) INTERNAL LOAD PARAMETERS (session-)RPE (AU) (n = 2) EXERCISE-INDUCED RESPONSES Well-being questionnaire (5-point scale) (n = 1) CK (U/L) (n = 2) Salivary cortisol (nmol/L) (n = 1) Repeated plyometric push-ups (n) (n = 1) Sleep (h) (n = 1) ADAPTATION PARAMETERS Sleep (h) (n = 1) Rugby Union [57,58] Professional (n = 51, male) Distance in speed zones (m) (n = 2) Impacts (n) (n = 2) PL (au) (n = 1) Total/Relative distance (m, m/min) (n = 1) EXERCISE-INDUCED RESPONSES CK (U/L) (n = 1) Urinary n-terminal prohormone of brain natriuretic peptide (pg/mL) (n = 1) Soccer [59–93] Professional (n = 311, male) Elite (n = 236, male) Semi-professional (n = 61, male) University (n = 114, 79 male, 35 female) Distance in speed zones (m) (n = 31) Total/Relative distance (m, m/min) (n = 30) PL (AU) (n = 15) Acceleration/Deceleration (m·s−2) (n = 13) Duration (min) (n = 12) Impacts (n) (n = 5) Average Speed (m/s) (n = 4) Dynamic stress load (AU) (n = 4) Metabolic power concept (W/kg) (n = 4) Maximal velocity (m/s) (n = 3) Effindex (AU) (n = 2) RHIE (n) (n = 2) Body load (AU) (n = 1) Energy expenditure (kJ/kg) (n = 2) Equivalent distance (m) (n = 1) Explosive distance (m) (n = 1) Impulse Load (Ns) (n = 1) Force load (AU) (n = 1) Mechanical work (AU) (n = 1) Training load score by Polar (AU) (n = 1) Total accelerometer load (AU) (n = 1) Total forces (AU) (n = 1) Velocity load (AU) (n = 1) Work:rest ratio (n = 1) INTERNAL LOAD PARAMETERS HR-based indices (n = 17) (session-)RPE (AU) (n = 16) Effindex (AU) (n = 2) EXERCISE-INDUCED RESPONSES Well-being questionnaire (5-point scale) (n = 8) CMJ (cm) (n = 6) CK (U/L) (n = 5) Immunoglobulin (µg/mL) (n = 3) C-reactive protein (mg/L) (n = 1) HR-based indices (n = 1) Myoglobin concentration (ng/mL) (n = 1) Plasma lactate dehydrogenase (U/L) (n = 1) Body mass measures (kg) (n = 1) ADAPTATION PARAMETERS HR-based indices (n = 2) Body mass measures (kg) (n = 2) Strength test (Nm) (n = 1) VO2max (ml/kg/min) (n = 1) 30-15 intermittent fitness test (m) (n = 1) INDIVIDUAL CHARACTERISTICS VO2max (ml/kg/min) (n = 1) YYIR (m) (n = 1) Repeated sprint ability (m) (n = 1) Body mass measures (kg) (n = 1) Muscle characteristics (cm) (n = 1) Sprint test (s) (n = 1) Sensors 2023, 23, 827 11 of 22 Table 3. Cont. Sport Study Player Level (n = Number of Athletes) External Parameters (n = Number of Studies) Internal Parameters (n = Number of Studies) Tag football [94] Regional (n = 16, male) Acceleration/Deceleration (m·s−2) (n = 1) Distance in speed zones (m) (n = 1) RHIE (n) (n = 1) Total/relative distance (m, m/min) (n = 1) INDIVIDUAL CHARACTERISTICS CMJ (cm) (n = 1) Sprint test (m/s) (n = 1) YYIR (m) (n = 1) AU arbitrary unit, CK creatine kinase, CMJ countermovement jump, HRmax maximal heart rate, HR heart rate, IMA™ inertial movement analysis, PL player load, RHIE repeated high-intensity events, RPE rating of perceived exertion, TRIMP training impulse, VO2max maximal oxygen uptake, YYIR Yo-Yo intermittent recovery test. Sensors 2023, 23, 827 12 of 22 4. Discussion This systematic review aimed to enhance the knowledge around relationships between external, wearable-based load parameters and internal load, exercise-induced responses, adaptation parameters, and parameters of individual characteristics in running-based team sports. Knowledge about these relationships may reduce time- and possibly cost-intensive testing outside regular training. Acute fitness and fatigue states may be drawn only based on external load parameters. Additionally, the amount of data to be collected and analyzed could be reduced by collecting fewer internal parameters. Our systematic review is the first to include a myriad of external and internal param- eters, focusing on external parameters collected from wearables only. This is crucial to enhance practicality and usability of parameters collected on-field. As the amount of data from wearable sensors and their use increase, it is inevitable to enhance the knowledge around these parameters and understand the dose–response relationship of team sport athletes. The findings are discussed in the following sections. As some relationships have been examined by a minimal number of studies, results are discussed only when a systematic synthesis of results is feasible. In the following, results are discussed in categories of internal load, exercise-induced response, adaptation parameters, and individual characteristics (Figure 1). Figure 4 additionally highlights the findings of moderate to large relationships which are explored in the following sections, separated by sports. Then, general aspects and future directions are discussed and outlined. Sensors 2023, 23, x FOR PEER REVIEW 12 of 24 4. Discussion This systematic review aimed to enhance the knowledge around relationships be- tween external, wearable-based load parameters and internal load, exercise-induced re- sponses, adaptation parameters, and parameters of individual characteristics in running- based team sports. Knowledge about these relationships may reduce time- and possibly cost-intensive testing outside regular training. Acute fitness and fatigue states may be drawn only based on external load parameters. Additionally, the amount of data to be collected and analyzed could be reduced by collecting fewer internal parameters. Our systematic review is the first to include a myriad of external and internal param- eters, focusing on external parameters collected from wearables only. This is crucial to enhance practicality and usability of parameters collected on-field. As the amount of data from wearable sensors and their use increase, it is inevitable to enhance the knowledge around these parameters and understand the dose–response relationship of team sport athletes. The findings are discussed in the following sections. As some relationships have been examined by a minimal number of studies, results are discussed only when a systematic synthesis of results is feasible. In the following, re- sults are discussed in categories of internal load, exercise-induced response, adaptation parameters, and individual characteristics (Figure 1). Figure 4 additionally highlights the findings of moderate to large relationships which are explored in the following sections, separated by sports. Then, general aspects and future directions are discussed and out- lined. Figure 4. Displayed are parameters for which a systematic synthesis was feasible and that exhibited a moderate to strong relationship. Internal load parameters are sorted by sport and colored as indi- cated in the legend. The relationship to an external load parameter is marked by an arrow. External parameters are displayed with the number of studies they are appearing in Figure 4. Displayed are parameters for which a systematic synthesis was feasible and that ex- hibited a moderate to strong relationship. Internal load parameters are sorted by sport and col- ored as indicated in the legend. The relationship to an external load parameter is marked by an arrow. External parameters are displayed with the number of studies they are appearing in [13,31,32,38,39,43,47–49,51,56,57,59,60,69,80,91,93,94]. Acc/Dec acceleration/deceleration parame- ters, CK creatine kinase, CMJ countermovement jump, HR heart rate, RPE rating of perceived exertion, sIgA secretory immunoglobulin A, TRIMP training impulse, LI low intensity, MI medium intensity. Sensors 2023, 23, 827 13 of 22 4.1. Internal Load 4.1.1. (session-)RPE Internal load parameters predominantly encompassed subjective ratings of exertion and HR-based indices. (session-)RPE had moderate to strong associations with total and relative distance in Australian football [13,43] and soccer players [59,60,93]. Moderate to strong asso- ciations were also present between (session-)RPE and PL in Australian football [43], basketball [47,48], and soccer [59,60]. Similar strength of associations was detected be- tween (session-)RPE and distances covered in different speed zones in Australian football players [43], and acceleration and deceleration parameters in basketball players [48,49]. A weak relationship between (session-)RPE and PL was only found in American football play- ers [31], and heterogenous results were present for the relationship between (session-)RPE and distances covered in speed zones and (session-)RPE and acceleration and deceleration parameters in soccer players. The weak relationship between (session-)RPE and PL in American football may be due to different match demands compared to other team sports analyzed. American football players generally cover lower overall distances in games (3000 to 5500 m in NCAA I football [59]) than soccer players (male elite outfield players 9000–14,000 m [95]), Australian football players (elite level 12,939 ± 1145 m [96]), and even basketball players (4404 to 7558 m [97]). Bartlett et al. (2017) found collinearity between session duration, PL, and total distance. This may suggest that the lower the overall distance, the lower the association between PL and RPE or session-RPE. Heterogenous results between (session-)RPE and distances in speed zones for soc- cer players may be due to the larger volume of studies compared to other team sports included and due to the different methods used: partial correlations [92], within-individual correlations [73,92], Pearson product-moment correlations [59,60], and machine learning techniques [61,64]. Furthermore, speed thresholds to define speed zones were either fixed or individualized, and distances were expressed in various ways (as absolute, percentage of total distance, frequency, number of efforts, or distance per minute). Generally, indicators of total volume seem to result in higher associations than those expressed per minute or as a percentage of total volume. Rago et al. (2019) found a tendency of increasing correlations when speed thresholds were individualized rather than identical for all players. Heterogenous results between (session-)RPE and parameters describing accelerations and decelerations in soccer may be due to varying methods. Using partial and within- individual correlations [92], small to moderate correlations were detected between those parameters. Furthermore, correlations describing total acceleration were higher than those describing accelerations per minute [92]. This supports the above findings that (session-)RPE seems to have stronger associations with parameters describing total volume. Machine learning techniques identified the number of acceleration efforts and decelerating distances as the main contributors to RPE in soccer [64]. Correlations between (session-)RPE and acceleration and deceleration parameters may be higher in basketball due to the high frequency of accelerations of 29.6 ± 3.9–32.7 ± 11.0 per minute in professional male players [98] compared to 90 ± 21 total accelerations per match in soccer players [99]. This places a greater total amount of accelerations and decelerations on basketball players compared to i.e., soccer. Thus, accelerations and deceleration may have a greater impact on perceived exertion compared to team sports in which they occur less frequently. For practitioners, this means that estimation of (session-)RPE may be done most adequately with indicators of total volume such as total distance or PL in Australian football, soccer, and basketball players. In indoor team sports such as basketball, where total distance is not available from wearable sensors due to a lack of GNSS signal, parameters describing acceleration and deceleration may be used instead of total distance. Omitting Sensors 2023, 23, 827 14 of 22 (session-)RPE scales would save practitioners and players the time to analyze and fill out the scales. 4.1.2. HR-Based Indices The relationships of HR-based parameters of internal load to external load parameters were analyzed in soccer players only. HR was divided in zones [79,80], predicted from external load parameters [65], expressed as percentage of HRmax [79,84], and used for calculations of Edward’s and Banister’s TRIMP [59,60,72,79]. TRIMP is a method, originally proposed by Banister et al. [100,101] that integrates training duration, maximal, resting, and average exercise HR, and a weighting factor to address high intensities [102]. Banister’s TRMIP has further been modified, including a summated HR zone method proposed by Edwards [103], here Edwards’ TRIMP. This method takes the time spent in predefined HR zones into account. The result of both methods is a training score per session indicating the cardio-vascular demands experienced by the athlete. Moderate to large correlations existed between time spent in the low- and medium-intensity velocity and the low- and medium- intensity HR zones, respectively. Time spent in the low- and medium- intensity HR zones also showed moderate to large correlations with PL [80]. Similar strength in associations was found between total distance and Edwards’ TRIMP [60] and between PL and Edwards’ and Banister’s TRIMP [59,60]. Correlations were not significant or weak between time spent in the high-intensity velocity and the high-intensity HR zone [80], between high-speed distance and number of efforts at sprinting speed and Edwards’ TRIMP [60], between PL and the high-intensity HR zone [80], between the number of rapid accelerations per minute and the time spent above 80% of HRmax [79], and between repeated high-intensity events and Banister’s TRIMP [79] and percent time spent above 80% of HRmax [79]. With increasing speed, correlations between HR-based parameters and distances in speed zones seem to weaken. This finding is similar to (session-)RPE which may be due to the high relationship between HR-based parameters and session-RPE [104,105]. Previous research has highlighted that HR measures may not be appropriate for high-intensity interval training or intermittent exercise due to the increase in anaerobic contribution [102,106]. For practitioners, this means that high-intensity running parameters may not be an appropriate indicator of HR. Low-intensity running and indicators of total volume such as total distance or PL may be more suited for low- and medium-intensity HR parameter estimation. Overall, among the internal load parameters, (session-)RPE, time spent in low- and medium-intensity HR zones, and TRIMP are best estimated using parameters of total volume such as PL and total distance. Time spent in high-intensity HR zones are not represented adequately by the external load parameters examined and may be recorded separately if of interest. Noteworthy is that HR-based indices do not adequately represent anaerobic training [102,106]. The latter two points support the idea of collecting both HR-based indices and external load parameters. 4.2. Exercise-Induced Responses Exercise-induced response are short-term changes in parameters. To detect changes, parameters are collected at two or more time points several hours apart. The first time point serves as a baseline measure, usually in a non-fatigued state. The next time point(s) occur(s) following exercise when athletes may be fatigued. Exercise-induced responses extracted from the studies included consist of well-being indicators, CK concentrations, HR-based indices, neuromuscular functioning, and biomarkers, among others. The relationships between well-being parameters and external load indicators were heterogeneous, possibly due to studies using different questionnaires analyzing either overall wellness [35,36,38,50,68,76] or single parameters of wellness (e.g., sleep, stress, recovery, muscles soreness, or fatigue) [56,83,84,87,88] and related it to external load pa- rameters of the same day [38,50,56,76], the previous day [35,36,63,83], the previous two to four days [36,84], or the previous weeks [88]. Thus, practitioners may need to be careful Sensors 2023, 23, 827 15 of 22 when choosing the specific parameter and the time period being analyzed as this provides different information about the athlete. Results in Table S1 in the supplementary material indicate that especially high-intensity distance parameters such as distance covered, time spent, or number of efforts at high speeds may inform about potential muscle damage as indicated by CK levels. In col- lision team sports such as Australian football, rugby league, rugby union, and rugby sevens, moderate to strong relationships were found between CK levels and impact parameters [39,51,56,57]. In these sports, impact parameters may additionally indicate muscle damage. No study included addressed those parameters in American football players. However, since American football belongs to the collision team sports, similar associations to CK levels can be expected. This, however, needs further verification. Exercise-induced HR-based indices were found only in studies observing soccer play- ers. Here, mostly negligible to small associations between several external load indicators and heart rate variability (HRV) were found [71,83,84]. For practitioners, this means that common external load indicators, as included in this review, may not serve as an estimation of HRV. Thus, this metric should be recorded separately if it is of interest. The relationship between change in countermovement jump (CMJ) parameters as an indicator of neuromuscular fatigue and high-speed running parameters varied from negligible and nonsignificant to large and was assessed in soccer and tag football players. Three studies found negligible to small correlations [83,84,86], whereas another three studies found moderate to large correlations [74,75,94]. This may be due to the different CMJ parameters collected (jump height, GRFs, or power output) and different time points when fatigued jumps were executed (24 to 72 h post-exercise). CMJ parameters’ relationship to acceleration and deceleration parameters exhibited varying results. One study assessed relationships to parameters of overall volume, such as total distance, duration, and PL, only finding trivial or unclear effects in soccer players [86]. For practitioners, this means that a high volume of acceleration, deceleration, and high- speed running efforts possibly negatively influence neuromuscular performances for up to 24 h. Further research is needed to assess CMJ parameters relationships. In American football, the number of impacts and peak head accelerations may indi- cate S100beta levels but not tau concentrations [32]. Here, data were collected from an instrumented mouthguard. S100beta is a blood biomarker that may be useful in detecting mechanical stress in the brain [32]. In the field, symptom scores offer a quick and easy-to- use method to detect symptoms of concussions [107]. Kawata et al. (2017) did not find higher symptom scores in players who sustained more impacts. This finding is particu- larly important to recognize for practitioners as the easy-to-use method (symptom scores) may fail to detect exposure to repeated sub-concussive (head) impacts. Accelerometer- embedded mouth guards, however, seem to pose a reasonable method to detect elevated S100beta levels. Thus, it may be beneficial to implement external monitoring systems such as accelerometers. Further studies are needed to analyze these relationships. Concentrations of secretory immunoglobulin A (sIgA), a marker of immune function, have been linked to high-intensity distance, total distance, and acceleration and deceleration parameters in soccer [69,91]. Small to large negative relationships were observed. If the overall volume is high, sIgA is reduced; thus, immune function and the risk of contracting an upper respiratory tract infection (URTI) may be increased. This finding is in line with the previous findings that reported reduced sIgA levels after interval runs [108] and in athletes with a high workload [109]. For practitioners, this means that players are at a higher risk of falling sick following a period of congested schedule or high volume in general. Players may be at greatest risk 3 to 72 h post intense exercise according to the “open window theory” of altered immunity [109]. Overall, numerous different exercise-induced responses were analyzed, as previously depicted in Figure 3. Most of them, however, appeared in few studies such that a systematic synthesis was not feasible. For the parameters discussed above, practitioners need to carefully consider time point of collection and the specific parameter collected, as changes Sensors 2023, 23, 827 16 of 22 may result in varying outcomes. Consistent findings were present regarding muscle damage in collision-based team sports which may be estimated using impact parameters. Practitioners shall further consider that sIgA levels may be low following high total activity volume and players may be at risk of attracting an URTI. 4.3. Adaptation Parameters Adaptation parameters are assessed at a minimum of two time points several days, weeks, or months apart and they are commonly carried out in a non-fatigued state. Changes between measurements can be analyzed and viewed as adaptation. Adaptation parameters extracted from the studies included were related to changes in body mass, HR-based indices, intermittent and aerobic endurance capacities, sleep efficiency, and strength parameters. Body mass seems to change in relation to 10-week sprinting distance, and total distance, but not session duration or average speed [78,82]. Change in HRmax was positively related to 10-week sprinting distance [82], and HRV negatively to acute-to-chronic workload ratio (ACWR)-based session time during a season [77]. ACWR is commonly used for injury prevention purposes. The model says that a greatly increased or decreased acute workload, compared to the chronic workload, increases the risk of sustaining an injury [110]. Change in hamstring peak torque, quadriceps to hamstring ratio, percent change in peak torque, and quadriceps to hamstring ratio was positively related to 10-week accumulated PL and acceleration sum (accumulated acceleration data in all three axes), sprinting distance, dura- tion, and total distance, respectively [82]. Meaning volume and intensity improve strength test performances and may thus reduce hamstring injuries and the risk of suffering an anterior cruciate ligament (ACL) injury as imbalances between quadriceps and hamstring constitute an ACL risk factor [111]. Monitoring loads long term to ensure they are suffi- ciently high to cause adaptations may be beneficial. Further, team sports practitioners can gain a better understanding of the individual dose-response patterns. As indicated by a positive change in VO2max, improvements in aerobic endurance were strongly correlated to 10-week accumulated PL and acceleration [82]. Session duration, however, showed an inverse relationship to changes in VO2max [82]. The duration in this study, however, decreased throughout the season. Meaning VO2max increased despite decreasing duration. In this case, other factors, such as high mechanical loading, seem to represent improvements in aerobic capacity better than training duration. Changes in intermittent fitness, as indicated by the 30-15 intermittent fitness test, were observed by one study only, which found unclear and large relationships to high-intensity running, total distance, and PL, respectively [72]. More research is needed regarding these relationships to synthesize results systematically. Overall, findings suggest that intensity seems particularly important to improve certain physiological capacities related to intermittent team sports. Volume and intensity need to be well-balanced in training programs to cause optimal adaptations. 4.4. Individual Characteristics Individual characteristics analyzed in relation to external parameters collected using wearables include intermittent and aerobic endurance capacity, neuromuscular perfor- mance parameters, and muscle architecture. As indicated by the Yo-Yo intermittent recov- ery test (YYIR), players with larger intermittent endurance capacities covered greater total distances in soccer and tag football [66,94]. Greater intensities, as indicated by high-speed running meters per minute and the number of repeated high-intensity events, were covered by tag football players with better YYIR performance [94]. Similar findings were present for players with greater VO2max regarding total distance and intensity parameters in soccer players [67]. For practitioners, this means that players who generally cover more distances likely have greater endurance capacities. As such, rigorous and time-intensive testing in the laboratory may not be necessary to find out endurance deficits and strengths in players. Sensors 2023, 23, 827 17 of 22 Parameters of neuromuscular performance and muscle architecture were analyzed each by one study only; thus, a systematic synthesis is not feasible, and more research is needed to draw conclusions regarding those parameter relationships. 4.5. General Aspects Generally, based on the intensity and volume of external load experienced during training and match play, internal bodily reactions take place, which, in the long term, lead to adaptations and influence individual characteristics. Those characteristics determine how well a player can handle the external load. However, no consensus exists on the parameters encompassing internal load yet. Some researchers have a broad understanding of internal load parameters, including biochemical, neuromuscular, and hormonal responses [1,2]. Others have a more narrow definition of internal load, including only measures that can prescribe exercise intensity, comprising mainly HR and session-RPE [3,4,110]. In this systematic review, parameters were categorized mainly based on the time point of measurement, as depicted in Figure 1. This, however, was not always straightforward: Couderc et al. (2017) collected blood from the fingertip 3 min post-exercise to determine lactate and bicarbonate concentration and pH levels. As those parameters are not monitored continuously during exercise, they fall into the category of exercise-induced responses. RPE, however, is commonly collected around 15–30 min after exercise [53,59,73,92] to reduce bias that may result from particular easy or challenging segments during the final exercise period [112]. However, RPE is deemed an internal load parameter by both parties of broad and narrow definitions of internal load. This might be due to the strong relationship of RPE with internal load parameters (HR and blood lactate) [113,114]. Some limitations to this systematic review are acknowledged in the following. These include the non-feasibility of synthesizing results for some parameter relationships. Given the wide variety of parameters, some relationships were analyzed by fewer studies to synthesize results systematically. Further, thresholds for speed zones differ across studies such that results may vary due to varying absolute or individualized thresholds used. Different hard- and software was implemented in the studies analyzed, which may cause a discrepancy in results. Manufacturers apply filters to the data during post-processing such that the same parameter could supposedly differ when obtained from another product. Even a software update could result in inconsistent results. Correlations found do not mean causality; parameters might correlate because of other circumstances. Clemente et al. (2019) found a large negative correlation between training duration and VO2max. This finding, however, likely does not mean shorter training durations cause an increase in aerobic endurance, but rather other circumstances were in place, such as high running volume and repeated high-intensity events, that may elicit improved aerobic endurance. Few studies (n = 4) included in this systematic review studied female athletes [46,51,67,93], totaling 62 female athletes compared to 1479 male athletes. Even though the parameter rela- tionships were comparable between males and females, this can only be said about the few parameters’ relationships analyzed. Thus, more research, including female athletes, is needed. 4.6. Outlook and Future Work Besides ever ongoing enhancements in hard- and software that will provide more accurate and reliable data in the future, research around accelerometer- or inertial-based GRF and moment estimation has shown promising results. Estimating GRFs and joint moments during training and match play would provide valuable biomechanical insight. Continuous monitoring of forces and moments acting on the player’s body could enhance the knowledge of the optimal individual dose-response relationship, injury mechanisms, and performance indicators from a biomechanical perspective. Research groups have placed MEMS on the shank [115], the sacrum [116], and the trunk [117] or used a full-body sensor suit [118]. So far, movements such as walking, jumping, running, and squatting, have been analyzed separately. This approach needs further development for more complex Sensors 2023, 23, 827 18 of 22 and compound movements to be transferrable to team sports. Players will either have to wear sensors in more locations (possibly embedded in clothing), or algorithms based on one trunk-mounted sensor only will have to be developed further to gain valuable insights into the said domains. Recently, advancements in continuous lactate and glucose monitoring have been made. This methodology will provide new insights into the external–internal load relationships of those parameters which need to be studied in the future. More possibilities to monitor internal load parameters continuously during training and match play as well as knowledge about the relationships to external parameters will move testing and identification of adaptations and health and fitness status away from the laboratory and more toward on- field assessments. Thus, separate time-consuming and fatiguing testing can be eliminated and replaced with data collected during training and match play. An ever-increasing amount of data will be collected, so the ability to analyze and select data appropriately according to the context becomes increasingly important. Even though some external parameters show strong relationships with, e.g., HR, it is essential to use parameters according to their context. If a highly anaerobic training session takes place, valid measures may differ from those of a more aerobic-based session. Despite strong correlations between parameters, and even if both external and internal parameters are considered, it is still essential to know what type of parameters to inspect depending on the demands placed on the athletes and the stressed biological systems. Having a sound understanding of those differences is inevitable for practitioners to harness the power of the collected data. Besides selecting parameters, verbal and visual transfer of information becomes increasingly relevant to create a common understanding between athletes, coaches, data analysts, and medical staff to enhance performance. Future work will need to validate novel methods of collecting internal parameters, analyze relationships of those to external parameters, and include more female athletes. With more reliable data, captured from highly dynamic movements, the impact of those movements on players can be explored in more depth, as the parameter relationships in this domain are currently ambiguous. 5. Conclusions Strong correlations have been detected, especially between parameters of total activity volume and the internal load parameters HR-indices and RPE or session-RPE. These pa- rameter relationships, were analyzed most often making the state of evidence clearer than less researched parameter relationships. Fitness tests assessing aerobic and intermittent endurance, or (session-) RPE, and in collision-based team sports, additionally markers of muscle damage may be omitted and replaced by external, on-field measurements, facili- tating the work of practitioners. Relationships between external load and the other three internal parameter categories, exercise-induced responses, adaptations, and individual characteristics, are mostly ambiguous and need further verification. Until then, a holistic picture of an athlete may best be obtained by collecting external and internal parameters for those parameter groups. Due to the ever-increasing amount of data collected in both areas, external and internal, a sound understanding of the data and their sport-specific context becomes increasingly important. Good communication is crucial for all stakeholders to attain a common understanding of the data. Studies including female athletes have been noticeably little in number and should be increased in the future. Future work will need to validate novel methods of collecting internal parameters and their relationships to external parameters in order to understand the individual dose-response patterns. Supplementary Materials: The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/s23020827/s1. Table S1: Studies included in this systematic review, the parameters collected and the relationship found between them. Results are listed by sport in alphabetical order. Sensors 2023, 23, 827 19 of 22 Author Contributions: Conceptualization, K.R., H.M., S.W. and M.R.; methodology, J.H.; formal analysis, J.H.; investigation, J.H., J.D.; data curation, J.H. and J.D.; writing—original draft preparation, J.H.; writing—review and editing, A.G., K.R. and S.W.; supervision and editing, A.G., K.R. and S.W. All authors have read and agreed to the published version of the manuscript. Funding: We acknowledge support by the Open Access Publication Fund of the University of Freiburg. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Conflicts of Interest: The authors declare no conflict of interest. References 1. Halson, S.L. Monitoring training load to understand fatigue in athletes. Sports Med. 2014, 44, 139–147. [CrossRef] [PubMed] 2. Bourdon, P.C.; Cardinale, M.; Murray, A.; Gastin, P.; Kellmann, M.; Varley, M.C.; Gabbett, T.J.; Coutts, A.J.; Burgess, D.J.; Gregson, W.; et al. Monitoring Athlete Training Loads: Consensus Statement. Int. J. Sports Physiol. Perform. 2017, 12, 161–170. [CrossRef] [PubMed] 3. Impellizzeri, F.M.; Marcora, S.M.; Coutts, A.J. Internal and External Training Load: 15 Years on. Int. J. Sports Physiol. Perform. 2019, 14, 270–273. [CrossRef] [PubMed] 4. Impellizzeri, F.; Rampinini, E.; Marcora, S. Physiological assessment of aerobic training in soccer. J. Sports Sci. 2005, 23, 583–592. [CrossRef] 5. Impellizzeri, F.M.; Rampinini, E.; Coutts, A.J.; Sassi, A.; Marcora, S.M. Use of RPE-Based Training Load in Soccer. Med. Sci. Sports Exerc. 2004, 36, 1042–1047. [CrossRef] [PubMed] 6. Wallace, L.K.; Slattery, K.M.; Coutts, A.J. The Ecological Validity and Application of the Session-RPE Method for Quantifying Training Loads in Swimming. J. Strength Cond. Res. 2009, 23, 33–38. [CrossRef] 7. Vanrenterghem, J.; Nedergaard, N.J.; Robinson, M.A.; Drust, B. Training Load Monitoring in Team Sports: A Novel Framework Separating Physiological and Biomechanical Load-Adaptation Pathways. Sports Med. 2017, 47, 2135–2142. [CrossRef] 8. Ide, B.; Silvatti, A.; Staunton, C.; Marocolo, M.; Mota, G.; Lara, J.; Oranchuk, D. External and Internal Loads in Sports Science: Time to Rethink? Preprints 2021, 2021110207. [CrossRef] 9. Staunton, C.A.; Abt, G.; Weaving, D.; Wundersitz, D.W. Misuse of the term ‘load’ in sport and exercise science. J. Sci. Med. Sport 2021, 25, 439–444. [CrossRef] 10. Winter, E. “Workload”– time to abandon? J. Sports Sci. 2006, 24, 1237–1238. [CrossRef] 11. Winter, E.M.; Abt, G.; Brookes, F.C.; Challis, J.H.; Fowler, N.E.; Knudson, D.V.; Knuttgen, H.G.; Kraemer, W.J.; Lane, A.M.; van Mechelen, W.; et al. Misuse of “Power” and Other Mechanical Terms in Sport and Exercise Science Research. J. Strength Cond. Res. 2016, 30, 292–300. [CrossRef] 12. Burgess, D.J. The Research Doesn’t Always Apply: Practical Solutions to Evidence-Based Training-Load Monitoring in Elite Team Sports. Int. J. Sport. Physiol. Perform. 2017, 12, S2136–S2141. [CrossRef] [PubMed] 13. Bartlett, J.D.; O’Connor, F.; Pitchford, N.; Torres-Ronda, L.; Robertson, S.J. Relationships Between Internal and External Training Load in Team-Sport Athletes: Evidence for an Individualized Approach. Int. J. Sports Physiol. Perform. 2017, 12, 230–234. [CrossRef] [PubMed] 14. Buchheit, M.; Simpson, B.M. Player-Tracking Technology: Half-Full or Half-Empty Glass? Int. J. Sports Physiol. Perform. 2017, 12 (Suppl. 2), S235–S241. [CrossRef] [PubMed] 15. Lacome, M.; Simpson, B.; Buchheit, M. Monitoring Training Status with Player-Tracking Technology. Still on the Road to Rome. Part 1 2018, 7, 55–63. 16. Bunn, J.A.; Navalta, J.W.; Fountaine, C.J.; Reece, J.D. Current State of Commercial Wearable Technology in Physical Activity Monitoring 2015–2017. Int. J. Exerc. Sci. 2018, 11, 503–515. 17. Henriksen, A.; Mikalsen, M.H.; Woldaregay, A.Z.; Muzny, M.; Hartvigsen, G.; Hopstock, L.A.; Grimsgaard, S. Using Fitness Trackers and Smartwatches to Measure Physical Activity in Research: Analysis of Consumer Wrist-Worn Wearables. J. Med. Internet Res. 2018, 20, e110. [CrossRef] 18. Taylor, J.B.; Wright, A.A.; Dischiavi, S.L.; Townsend, M.A.; Marmon, A.R. Activity Demands during Multi-Directional Team Sports: A Systematic Review. Sports Med. 2017, 47, 2533–2551. [CrossRef] 19. di Prampero, P.; Osgnach, C. Metabolic Power in Team Sports—Part 1: An Update. Int. J. Sports Med. 2018, 39, 581–587. [CrossRef] 20. di Prampero, P.E.; Fusi, S.; Sepulcri, L.; Morin, J.-B.; Belli, A.; Antonutto, G. Sprint running: A new energetic approach. J. Exp. Biol. 2005, 208, 2809–2816. [CrossRef] 21. Jaspers, A.; Brink, M.S.; Probst, S.G.M.; Frencken, W.G.P.; Helsen, W.F. Relationships between Training Load Indicators and Training Outcomes in Professional Soccer. Sports Med. 2017, 47, 533–544. [CrossRef] [PubMed] 22. Akenhead, R.; Nassis, G.P. Training Load and Player Monitoring in High-Level Football: Current Practice and Perceptions. Int. J. Sports Physiol. Perform. 2016, 11, 587–593. [CrossRef] Sensors 2023, 23, 827 20 of 22 23. Thornton, H.R.; Delaney, J.A.; Duthie, G.M.; Dascombe, B.J. Developing Athlete Monitoring Systems in Team Sports: Data Analysis and Visualization. Int. J. Sports Physiol. Perform. 2019, 14, 698–705. [CrossRef] 24. McLaren, S.J.; Macpherson, T.W.; Coutts, A.J.; Hurst, C.; Spears, I.R.; Weston, M. The Relationships between Internal and External Measures of Training Load and Intensity in Team Sports: A Meta-Analysis. Sports Med. 2018, 48, 641–658. [CrossRef] 25. Moher, D.; Liberati, A.; Tetzlaff, J.; Altman, D.G.; PRISMA Group. Preferred reporting items for systematic reviews and meta-analyses: The PRISMA statement. PLoS Med. 2009, 6, e1000097. [CrossRef] 26. Alain, G.; Mahon, A. Perceived Exertion: Influence of Age and Cognitive Development. Sports Med. 2006, 36, 911–928. 27. Harriss, D.; MacSween, A.; Atkinson, G. Ethical Standards in Sport and Exercise Science Research: 2020 Update. Int. J. Sports Med. 2019, 40, 813–817. [CrossRef] [PubMed] 28. Law, M.; Stewart, D.; Pollock, N.; Letts, L.; Bosch, J.; Westmorland, M. Criticial Review Form—Quantitative Studies; McMaster University: Hamilton, ON, Canada, 1998. 29. Sarmento, H.; Clemente, F.M.; Harper, L.D.; Da Costa, I.T.; Owen, A.; Figueiredo, A.J. Small sided games in soccer—-A systematic review. Int. J. Perform. Anal. Sport 2018, 18, 693–749. [CrossRef] 30. Low, B.; Coutinho, D.; Gonçalves, B.; Rein, R.; Memmert, D.; Sampaio, J. A Systematic Review of Collective Tactical Behaviours in Football Using Positional Data. Sports Med. 2019, 50, 343–385. [CrossRef] [PubMed] 31. Govus, A.D.; Coutts, A.; Duffield, R.; Murray, A.; Fullagar, H. Relationship Between Pretraining Subjective Wellness Measures, Player Load, and Rating-of-Perceived-Exertion Training Load in American College Football. Int. J. Sports Physiol. Perform. 2018, 13, 95–101. [CrossRef] 32. Kawata, K.; Rubin, L.H.; Takahagi, M.; Lee, J.H.; Sim, T.; Szwanki, V.; Bellamy, A.; Tierney, R.; Langford, D. Subconcussive Impact-Dependent Increase in Plasma S100β Levels in Collegiate Football Players. J. Neurotrauma 2017, 34, 2254–2260. [CrossRef] [PubMed] 33. Kawata, K.; Rubin, L.H.; Wesley, L.; Lee, J.H.; Sim, T.; Takahagi, M.; Bellamy, A.; Tierney, R.; Langford, D. Acute Changes in Plasma Total Tau Levels Are Independent of Subconcussive Head Impacts in College Football Players. J. Neurotrauma 2018, 35, 260–266. [CrossRef] [PubMed] 34. Murray, A.; Buttfield, A.; Simpkin, A.; Sproule, J.; Turner, A.P. Variability of within-step acceleration and daily wellness monitoring in Collegiate American Football. J. Sci. Med. Sport 2019, 22, 488–493. [CrossRef] [PubMed] 35. Wellman, A.D.; Coad, S.C.; Flynn, P.J.; Climstein, M.; McLellan, C.P. Movement Demands and Perceived Wellness Associated With Preseason Training Camp in NCAA Division I College Football Players. J. Strength Cond. Res. 2017, 31, 2704–2718. [CrossRef] [PubMed] 36. Wellman, A.D.; Coad, S.C.; Flynn, P.J.; Siam, T.K.; McLellan, C.P. Perceived Wellness Associated With Practice and Competition in National Collegiate Athletic Association Division I Football Players. J. Strength Cond. Res. 2019, 33, 112–124. [CrossRef] 37. Carey, D.L.; Ong, K.; Morris, M.; Crow, J.; Crossley, K.M. Predicting ratings of perceived exertion in Australian football players: Methods for live estimation. Int. J. Comput. Sci. Sport 2016, 15, 64–77. [CrossRef] 38. Gallo, T.F.; Cormack, S.; Gabbett, T.J.; Lorenzen, C.H. Pre-training perceived wellness impacts training output in Australian football players. J. Sports Sci. 2015, 34, 1445–1451. [CrossRef] 39. Gastin, P.B.; Hunkin, S.L.; Fahrner, B.; Robertson, S. Deceleration, Acceleration, and Impacts Are Strong Contributors to Muscle Damage in Professional Australian Football. J. Strength Cond. Res. 2019, 33, 3374–3383. [CrossRef] 40. Graham, S.R.; Cormack, S.; Parfitt, G.; Eston, R. Relationships Between Model Estimates and Actual Match-Performance Indices in Professional Australian Footballers During an In-Season Macrocycle. Int. J. Sports Physiol. Perform. 2018, 13, 339–346. [CrossRef] 41. Johnston, R.J.; Watsford, M.L.; Austin, D.J.; Pine, M.J.; Spurrs, R.W. An Examination of the Relationship Between Movement Demands and Rating of Perceived Exertion in Australian Footballers. J. Strength Cond. Res. 2015, 29, 2026–2033. [CrossRef] 42. Ryan, S.; Coutts, A.J.; Hocking, J.; Dillon, P.A.; Whitty, A.; Kempton, T. Physical Preparation Factors That Influence Technical and Physical Match Performance in Professional Australian Football. Int. J. Sports Physiol. Perform. 2018, 13, 1021–1027. [CrossRef] [PubMed] 43. Gallo, T.; Cormack, S.; Gabbett, T.; Williams, M.; Lorenzen, C. Characteristics impacting on session rating of perceived exertion training load in Australian footballers. J. Sports Sci. 2014, 33, 467–475. [CrossRef] 44. Weston, M.; Siegler, J.; Bahnert, A.; McBrien, J.; Lovell, R. The application of differential ratings of perceived exertion to Australian Football League matches. J. Sci. Med. Sport 2014, 18, 704–708. [CrossRef] [PubMed] 45. Cormack, S.J.; Mooney, M.G.; Morgan, W.; McGuigan, M.R. Influence of Neuromuscular Fatigue on Accelerometer Load in Elite Australian Football Players. Int. J. Sports Physiol. Perform. 2013, 8, 373–378. [CrossRef] [PubMed] 46. Peterson, K.; Quiggle, G.T. Tensiomyographical responses to accelerometer loads in female collegiate basketball players. J. Sports Sci. 2016, 35, 2334–2341. [CrossRef] [PubMed] 47. Scanlan, A.; Wen, N.; Tucker, P.S.; Dalbo, V. The Relationships Between Internal and External Training Load Models During Basketball Training. J. Strength Cond. Res. 2014, 28, 2397–2405. [CrossRef] 48. Svilar, L.; Castellano, J.; Jukic, I. Load Monitoring System in Top-Level Basketball Team: Relationship between External and Internal Training Load. Kinesiology 2018, 50, 25–33. [CrossRef] 49. Svilar, L.; Castellano, J.; Jukic, I.; Casamichana, D. Positional Differences in Elite Basketball: Selecting Appropriate Training-Load Measures. Int. J. Sports Physiol. Perform. 2018, 13, 947–952. [CrossRef] Sensors 2023, 23, 827 21 of 22 50. Ihsan, M.; Tan, F.; Sahrom, S.; Choo, H.C.; Chia, M.; Aziz, A.R. Pre-game perceived wellness highly associates with match running performances during an international field hockey tournament. Eur. J. Sport Sci. 2017, 17, 593–602. [CrossRef] 51. Clarke, A.; Anson, J.M.; Pyne, D. Neuromuscular Fatigue and Muscle Damage after a Women’s Rugby Sevens Tournament. Int. J. Sports Physiol. Perform. 2015, 10, 808–814. [CrossRef] 52. Couderc, A.; Thomas, C.; Lacome, M.; Piscione, J.; Robineau, J.; Delfour-Peyrethon, R.; Borne, R.; Hanon, C. Movement Patterns and Metabolic Responses During an International Rugby Sevens Tournament. Int. J. Sports Physiol. Perform. 2017, 12, 901–907. [CrossRef] [PubMed] 53. Lovell, T.W.; Sirotic, A.C.; Impellizzeri, F.M.; Coutts, A.J. Factors Affecting Perception of Effort (Session Rating of Perceived Exertion) during Rugby League Training. Int. J. Sports Physiol. Perform. 2013, 8, 62–69. [CrossRef] [PubMed] 54. Thornton, H.R.; Delaney, J.A.; Duthie, G.M.; Dascombe, B.J. Effects of Preseason Training on the Sleep Characteristics of Professional Rugby League Players. Int. J. Sports Physiol. Perform. 2018, 13, 176–182. [CrossRef] [PubMed] 55. McLellan, C.P.; I Lovell, D.; Gass, G.C. Biochemical and Endocrine Responses to Impact and Collision During Elite Rugby League Match Play. J. Strength Cond. Res. 2011, 25, 1553–1562. [CrossRef] 56. Oxendale, C.L.; Twist, C.; Daniels, M.; Highton, J. The Relationship Between Match-Play Characteristics of Elite Rugby League and Indirect Markers of Muscle Damage. Int. J. Sports Physiol. Perform. 2016, 11, 515–521. [CrossRef] 57. Jones, M.R.; West, D.J.; Harrington, B.J.; Cook, C.J.; Bracken, R.M.; A Shearer, D.; Kilduff, L.P. Match play performance characteristics that predict post-match creatine kinase responses in professional rugby union players. BMC Sports Sci. Med. Rehabilitation 2014, 6, 38. [CrossRef] 58. Lindsay, A.; Lewis, J.G.; Gill, N.; Draper, N.; Gieseg, S.P. No relationship exists between urinary NT-proBNP and GPS technology in professional rugby union. J. Sci. Med. Sport 2017, 20, 790–794. [CrossRef] 59. Scott, B.R.; Lockie, R.G.; Knight, T.J.; Clark, A.C.; de Jonge, X.A.K.J. A Comparison of Methods to Quantify the In-Season Training Load of Professional Soccer Players. Int. J. Sports Physiol. Perform. 2013, 8, 195–202. [CrossRef] 60. Casamichana, D.; Castellano, J.; Calleja-Gonzalez, J.; San Román, J.; Castagna, C. Relationship Between Indicators of Training Load in Soccer Players. J. Strength Cond. Res. 2013, 27, 369–374. [CrossRef] 61. Geurkink, Y.; Vandewiele, G.; Lievens, M.; de Turck, F.; Ongenae, F.; Matthys, S.P.; Boone, J.; Bourgois, J.G. Modeling the Prediction of the Session Rating of Perceived Exertion in Soccer: Unraveling the Puzzle of Predictive Indicators. Int. J. Sports Physiol. Perform. 2019, 14, 841–846. [CrossRef] 62. Gomez-Piriz, P.T.; Jiménez-Reyes, P.; Ruiz-Ruiz, C. Relation between Total Body Load and Session Rating of Perceived Exertion in Professional Soccer Players. J. Strength Cond. Res. 2011, 25, 2100–2103. [CrossRef] [PubMed] 63. Malone, J.J.; Jaspers, A.; Helsen, W.F.; Merks, B.; Frencken, W.G.; Brink, M.S. Seasonal Training Load and Wellness Monitoring in a Professional Soccer Goalkeeper. Int. J. Sports Physiol. Perform. 2018, 13, 672–675. [CrossRef] [PubMed] 64. Jaspers, A.; De Beéck, T.O.; Brink, M.S.; Frencken, W.G.; Staes, F.; Davis, J.J.; Helsen, W.F. Relationships between the External and Internal Training Load in Professional Soccer: What Can We Learn From Machine Learning? Int. J. Sports Physiol. Perform. 2018, 13, 625–630. [CrossRef] [PubMed] 65. Lacome, M.; Simpson, B.; Broad, N.; Buchheit, M. Monitoring Players’ Readiness Using Predicted Heart-Rate Responses to Soccer Drills. Int. J. Sports Physiol. Perform. 2018, 13, 1273–1280. [CrossRef] 66. Lee, M.; Mukherjee, S. Relationship of Training Load with High-intensity Running in Professional Soccer Players. Int. J. Sports Med. 2019, 40, 336–343. [CrossRef] 67. McCormack, W.P.; Stout, J.R.; Wells, A.J.; Gonzalez, A.M.; Mangine, G.T.; Fragala, M.S.; Hoffman, J.R. Predictors of High-Intensity Running Capacity in Collegiate Women during a Soccer Game. J. Strength Cond. Res. 2014, 28, 964–970. [CrossRef] 68. De Beéck, T.O.; Jaspers, A.; Brink, M.S.; Frencken, W.G.; Staes, F.; Davis, J.J.; Helsen, W.F. Predicting Future Perceived Wellness in Professional Soccer: The Role of Preceding Load and Wellness. Int. J. Sports Physiol. Perform. 2019, 14, 1074–1080. [CrossRef] 69. Owen, A.L.; Wong, D.P.; Dunlop, G.; Groussard, C.; Kebsi, W.; Dellal, A.; Morgans, R.; Zouhal, H. High-Intensity Training and Salivary Immunoglobulin A Responses in Professional Top-Level Soccer Players: Effect of Training Intensity. J. Strength Cond. Res. 2016, 30, 2460–2469. [CrossRef] 70. Thorpe, R.; Sunderland, C. Muscle Damage, Endocrine, and Immune Marker Response to a Soccer Match. J. Strength Cond. Res. 2012, 26, 2783–2790. [CrossRef] 71. Chrismas, B.C.R.; Taylor, L.; Thornton, H.R.; Murray, A.; Stark, G. External training loads and smartphone-derived heart rate variability indicate readiness to train in elite soccer. Int. J. Perform. Anal. Sport 2019, 19, 143–152. [CrossRef] 72. Rabbani, A.; Kargarfard, M.; Castagna, C.; Clemente, F.M.; Twist, C. Associations between Selected Training-Stress Measures and Fitness Changes in Male Soccer Players. Int. J. Sports Physiol. Perform. 2019, 14, 1050–1057. [CrossRef] [PubMed] 73. Rago, V.; Brito, J.; Figueiredo, P.; Krustrup, P.; Rebelo, A. Relationship between External Load and Perceptual Responses to Training in Professional Football: Effects of Quantification Method. Sports 2019, 7, 68. [CrossRef] [PubMed] 74. Russell, M.; Sparkes, W.; Northeast, J.; Cook, C.; Bracken, R.; Kilduff, L. Relationships between match activities and peak power output and Creatine Kinase responses to professional reserve team soccer match-play. Hum. Mov. Sci. 2015, 45, 96–101. [CrossRef] 75. Malone, S.; Mendes, B.; Hughes, B.; Roe, M.; Devenney, S.; Collins, K.; Owen, A. Decrements in Neuromuscular Perfromance and Increases in Creatine Kinase Impact Training Outputs in Elite Soccer Players. J. Strength Cond. Res. 2018, 32, 1342–1351. [CrossRef] [PubMed] Sensors 2023, 23, 827 22 of 22 76. Malone, S.; Owen, A.; Newton, M.; Mendes, B.; Tiernan, L.; Hughes, B.; Collins, K. Wellbeing perception and the impact on external training output among elite soccer players. J. Sci. Med. Sport 2018, 21, 29–34. [CrossRef] 77. Sekiguchi, Y.; Huggins, R.A.; Curtis, R.M.; Benjamin, C.L.; Adams, W.M.; Looney, D.P.; West, C.A.; Casa, D.J. Relationship Between Heart Rate Variability and Acute:Chronic Load Ratio Throughout a Season in NCAA D1 Men’s Soccer Players. J. Strength Cond. Res. 2021, 35, 1103–1109. [CrossRef] 78. Sekiguchi, Y.; Adams, W.M.; Curtis, R.M.; Benjamin, C.L.; Casa, D.J. Factors influencing hydration status during a National Collegiate Athletics Association division 1 soccer preseason. J. Sci. Med. Sport 2019, 22, 624–628. [CrossRef] 79. Silva, P.; Dos Santos, E.; Grishin, M.; Rocha, J.M. Validity of Heart Rate-Based Indices to Measure Training Load and Intensity in Elite Football Players. J. Strength Cond. Res. 2018, 32, 2340–2347. [CrossRef] 80. Sparks, M.; Coetzee, B.; Gabbett, T.J. Internal and External Match Loads of University-Level Soccer Players: A Comparison Between Methods. J. Strength Cond. Res. 2017, 31, 1072–1077. [CrossRef] 81. Suarez-Arrones, L.; Torreño, N.; Requena, B.; De Villarreal, E.S.; Casamichana, D.; Barbero-Alvarez, J.C.; Munguía-Izquierdo, D. Match-play activity profile in professional soccer players during official games and the relationship between external and internal load. J. Sports Med. Phys. Fit. 2015, 55, 1417–1422. 82. Clemente, F.M.; Nikolaidis, P.T.; Rosemann, T.; Knechtle, B. Dose-Response Relationship Between External Load Variables, Body Composition, and Fitness Variables in Professional Soccer Players. Front. Physiol. 2019, 10, 443. [CrossRef] [PubMed] 83. Thorpe, R.T.; Strudwick, A.J.; Buchheit, M.; Atkinson, G.; Drust, B.; Gregson, W. Monitoring Fatigue During the In-Season Competitive Phase in Elite Soccer Players. Int. J. Sports Physiol. Perform. 2015, 10, 958–964. [CrossRef] [PubMed] 84. Thorpe, R.T.; Strudwick, A.J.; Buchheit, M.; Atkinson, G.; Drust, B.; Gregson, W. The Influence of Changes in Acute Training Load on Daily Sensitivity of Morning-Measured Fatigue Variables in Elite Soccer Players. Int. J. Sports Physiol. Perform. 2017, 12, S2107–S2113. [CrossRef] 85. Torreño, N.; Izquierdo, D.M.; Coutts, A.; De Villarreal, E.S.; Asian-Clemente, J.; Suarez-Arrones, L. Relationship Between External and Internal Loads of Professional Soccer Players During Full Matches in Official Games Using Global Positioning Systems and Heart-Rate Technology. Int. J. Sports Physiol. Perform. 2016, 11, 940–946. [CrossRef] 86. Wiig, H.; Raastad, T.; Luteberget, L.S.; Ims, I.; Spencer, M. External Load Variables Affect Recovery Markers up to 72 h after Semiprofessional Football Matches. Front. Physiol. 2019, 10, 689. [CrossRef] 87. Zurutuza, U.; Castellano, J.; Echeazarra, I.; Casamichana, D. Absolute and Relative Training Load and Its Relation to Fatigue in Football. Front. Psychol. 2017, 8, 878. [CrossRef] [PubMed] 88. Coker, N.A.; Wells, A.J.; Ake, K.M.; Griffin, D.L.; Rossi, S.J.; McMillan, J.L. Relationship between Running Performance and Recovery-Stress State in Collegiate Soccer Players. J. Strength Cond. Res. 2017, 31, 2131–2140. [CrossRef] 89. Coppalle, S.; Rave, G.; Ben Abderrahman, A.; Ali, A.; Salhi, I.; Zouita, S.; Zouita, A.; Brughelli, M.; Granacher, U.; Zouhal, H. Relationship of Pre-season Training Load With In-Season Biochemical Markers, Injuries and Performance in Professional Soccer Players. Front. Physiol. 2019, 10, 409. [CrossRef] 90. Djaoui, L.; Pialoux, V.; Vallance, E.; Owen, A.; Dellal, A. Relationship between Fluid Loss Variation and Physical Activity during Official Games in Elite Soccer Players. RICYDE. Rev. Int. Ciencias del Deport. 2018, 14, 5–15. [CrossRef] 91. Figueiredo, P.; Nassis, G.P.; Brito, J. Within-Subject Correlation Between Salivary IgA and Measures of Training Load in Elite Football Players. Int. J. Sports Physiol. Perform. 2019, 14, 847–849. [CrossRef] 92. Gaudino, P.; Iaia, F.M.; Strudwick, A.J.; Hawkins, R.D.; Alberti, G.; Atkinson, G.; Gregson, W. Factors Influencing Perception of Effort (Session Rating of Perceived Exertion) during Elite Soccer Training. Int. J. Sports Physiol. Perform. 2015, 10, 860–864. [CrossRef] [PubMed] 93. Gentles, J.; Coniglio, C.; Besemer, M.; Morgan, J.; Mahnken, M. The Demands of a Women’s College Soccer Season. Sports 2018, 6, 16. [CrossRef] [PubMed] 94. Hogarth, L.W.; Burkett, B.J.; Mckean, M.R. Activity Profiles and Physiological Responses of Representative Tag Football Players in Relation to Playing Position and Physical Fitness. PLoS ONE 2015, 10, e0144554. [CrossRef] [PubMed] 95. Dolci, F.; Hart, N.H.; Kilding, A.E.; Chivers, P.; Piggott, B.; Spiteri, T. Physical and Energetic Demand of Soccer: A Brief Review. Strength Cond. J. 2020, 42, 70–77. [CrossRef] 96. Coutts, A.J.; Quinn, J.; Hocking, J.; Castagna, C.; Rampinini, E. Match running performance in elite Australian Rules Football. J. Sci. Med. Sport 2010, 13, 543–548. [CrossRef] [PubMed] 97. Stojanovi´c, E.; Stojiljkovi´c, N.; Scanlan, A.T.; Dalbo, V.J.; Berkelmans, D.M.; Milanovi´c, Z. The Activity Demands and Physiological Responses Encountered during Basketball Match-Play: A Systematic Review. Sports Med. 2018, 48, 111–135. [CrossRef] 98. Vázquez-Guerrero, J.; Suarez-Arrones, L.; Gómez, D.C.; Rodas, G. Comparing external total load, acceleration and deceleration outputs in elite basketball players across positions during match play. Kinesiology 2018, 50, 228–234. [CrossRef] 99. Ingebrigtsen, J.; Dalen, T.; Hjelde, G.H.; Drust, B.; Wisløff, U. Acceleration and sprint profiles of a professional elite football team in match play. Eur. J. Sport Sci. 2014, 15, 101–110. [CrossRef] 100. Banister, E.W. Modeling Elite Athletic Performance. Physiol. Test. Elit. Athletes 1991, 347, 403–422. 101. Banister, E.W.; Calvert, T.W.; Savage, M.V.; Bach, T.M. A Systems Model of Training for Athletic Performance. Aust. J. Sci. Med. Sport 1975, 7, 57–61. 102. Borresen, J.; Lambert, M.I. The Quantification of Training Load, the Training Response and the Effect on Performance. Sports Med. 2009, 39, 779–795. [CrossRef] Sensors 2023, 23, 827 23 of 22 103. Edwards, S. The Heart Rate Monitor Book; Feet Fleet Press: Sacramento, CA, USA, 1993. 104. Scott, T.J.; Black, C.R.; Quinn, J.; Coutts, A.J. Validity and Reliability of the Session-RPE Method for Quantifying Training in Australian Football: A Comparison of the CR10 and CR100 Scales. J. Strength Cond. Res. 2013, 27, 270–276. [CrossRef] 105. Herman, L.; Foster, C.; Maher, M.; Mikat, R.; Porcari, J. Validity and reliability of the session RPE method for monitoring exercise training intensity. S. Afr. J. Sports Med. 2006, 18, 14. [CrossRef] 106. Alexiou, H.; Coutts, A.J. A Comparison of Methods Used for Quantifying Internal Training Load in Women Soccer Players. Int. J. Sports Physiol. Perform. 2008, 3, 320–330. [CrossRef] 107. Merritt, V.C.; Meyer, J.E.; Arnett, P.A. A novel approach to classifying postconcussion symptoms: The application of a new framework to the Post-Concussion Symptom Scale. J. Clin. Exp. Neuropsychol. 2015, 37, 764–775. [CrossRef] 108. Mackinnon, L.T.; Jenkins, D.G. Decreased salivary immunoglobulins after intense interval exercise before and after training. Med. Sci. Sports Exerc. 1993, 25, 678–683. [CrossRef] 109. Nieman, D.C. Current Perspective on Exercise Immunology. Curr. Sports Med. Rep. 2003, 2, 239–242. [CrossRef] 110. Gabbett, T.J. The training—Injury prevention paradox: Should athletes be training smarter and harder? Br. J. Sports Med. 2016, 50, 273–280. [CrossRef] 111. Kim, D.; Hong, J. Hamstring to quadriceps strength ratio and noncontact leg injuries: A prospective study during one season. Isokinet. Exerc. Sci. 2011, 19, 1–6. [CrossRef] 112. Foster, C.; JA, F.; Franklin, J.; Gottschall, L.; LA, H.; Parker, S.; Doleshal, P.; Dodge, C. A New Approach to Monitoring Exercise Training. J. Strength Cond. Res. 2001, 15, 109–115. 113. Borg, G.A. Perceived Exertion. Exerc. Sport Sci. Rev. 1974, 2, 131–153. [CrossRef] 114. Borg, G.A. Psychophysical bases of perceived exertion. Med. Sci. Sports Exerc. 1982, 14, 377–381. [CrossRef] 115. Leporace, G.; Batista, L.A.; Metsavaht, L.; Nadal, J. Residual analysis of ground reaction forces simulation during gait using neural networks with different configurations. Annu. Int. Conf. IEEE Eng. Med. Biol. Soc. 2015, 2015, 2812–2815. [CrossRef] 116. Lee, M.; Park, S. Estimation of Three-Dimensional Lower Limb Kinetics Data during Walking Using Machine Learning from a Single IMU Attached to the Sacrum. Sensors 2020, 20, 6277. [CrossRef] 117. Pogson, M.; Verheul, J.; Robinson, M.A.; Vanrenterghem, J.; Lisboa, P. A neural network method to predict task- and step-specific ground reaction force magnitudes from trunk accelerations during running activities. Med. Eng. Phys. 2020, 78, 82–89. [CrossRef] 118. Karatsidis, A.; Jung, M.K.; Schepers, H.M.; Bellusci, G.; de Zee, M.; Veltink, P.H.; Andersen, M.S. Musculoskeletal model-based inverse dynamic analysis under ambulatory conditions using inertial motion capture. Med. Eng. Phys. 2019, 65, 68–77. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Relationships between External, Wearable Sensor-Based, and Internal Parameters: A Systematic Review.

01-11-2023

Helwig, Janina,Diels, Janik,Röll, Mareike,Mahler, Hubert,Gollhofer, Albert,Roecker, Kai,Willwacher, Steffen

eng

PMC3966785

Keeping Your Eyes Continuously on the Ball While Running for Catchable and Uncatchable Fly Balls Dees B. W. Postma, A. Rob den Otter, Frank T. J. M. Zaal* Center for Human Movement Sciences, University Medical Center Groningen, Sector F, University of Groningen, Groningen, the Netherlands Abstract When faced with a fly ball approaching along the sagittal plane, fielders need information for the control of their running to the interception location. This information could be available in the initial part of the ball trajectory, such that the interception location can be predicted from its initial conditions. Alternatively, such predictive information is not available, and running to the interception location involves continuous visual guidance. The latter type of control would predict that fielders keep looking at the approaching ball for most of its flight, whereas the former type of control would fit with looking at the ball during the early part of the ball’s flight; keeping the eyes on the ball during the remainder of its trajectory would not be necessary when the interception location can be inferred from the first part of the ball trajectory. The present contribution studied visual tracking of approaching fly balls. Participants were equipped with a mobile eye tracker. They were confronted with tennis balls approaching from about 20 m, and projected in such a way that some balls were catchable and others were not. In all situations, participants almost exclusively tracked the ball with their gaze until just before the catch or until they indicated that a ball was uncatchable. This continuous tracking of the ball, even when running close to their maximum speeds, suggests that participants employed continuous visual control rather than running to an interception location known from looking at the early part of the ball flight. Citation: Postma DBW, den Otter AR, Zaal FTJM (2014) Keeping Your Eyes Continuously on the Ball While Running for Catchable and Uncatchable Fly Balls. PLoS ONE 9(3): e92392. doi:10.1371/journal.pone.0092392 Editor: Todd W. Troyer, University of Texas at San Antonio, United States of America Received July 25, 2013; Accepted February 21, 2014; Published March 26, 2014 Copyright:  2014 Postma et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: These authors have no support or funding to report. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected] Introduction Catching a fly ball not only adds to a good result in a baseball game but also keeps fascinating spectators and scientists alike. A particularly famous catch was the one made by Willie Mays in the 1954 World Series (e.g., see http://en.wikipedia.org/wiki/ The_Catch_(baseball)). He managed to catch a seemingly uncatchable ball, after looking at the ball and running to the interception location about 475 feet (145 m) from the home plate [1]. Willie Mays’s catch made it to an illustration accompanying the contribution of Chodosh, Lifson, and Tabin in the 1995 volume of the journal Science [2]. These authors claimed that Willie Mays, and other adept outfielders, do not need to track the ball with their gaze because they are able to predict where and when to intercept the ball from the initial part of the ball trajectory. This will be the issue that we address in the present contribution: Does it suffice to view only the initial part of the ball’s flight to predict the interception location or do fielders continuously track the ball with their gaze while running to that interception location? Two types of strategy for the interception of moving targets have been distinguished in the literature (e.g., see [3–9]). On the one hand are the predictive strategies. In the context of fly-ball catching, this type of strategy amounts to looking at the ball’s trajectory and predicting the interception location from the initial conditions of the ball’s trajectory (i.e. its initial velocity and initial angle; cf. [10]). It should be noted that the use of such predictive strategy depends on a priori knowledge about gravity and air resistance. Because of drag and spin, fly balls do not follow parabolic trajectories (cf. [1,8,11]), which implies that a sophisti- cated internal model of ball-flight dynamics would have to be postulated. An alternative to a predictive strategy is to use continuous visual guidance of locomotion on the basis of prospective information. Rather than having to know the interception location and time from early conditions, prospective strategies (e.g., see [12–14]) involve continuously available information that can be used to know whether the current action (e.g., running speed) will lead to a successful interception. In the context of the interception of fly balls, one such model states that if a fielder keeps the ball moving on a linear optical trajectory (LOT), he or she will arrive at the interception location in time, without knowing when and where the interception will take place [15–18]. The LOT strategy boils down to making sure that the horizontal and vertical components of the gaze angle (the angle between the heading and the gaze direction) change proportionally. Locomotion patterns of fielders running to catch fly balls travelling to locations in front or behind, and to the side of the fielders’ initial positions have been reported to be in line with a LOT strategy (e.g., [16,19]). However, several authors have claimed that keeping a linear optical trajectory is not sufficient to guarantee interception because linear optical trajec- tories can also occur for unsuccessful interceptions [20–23]. Furthermore, for balls approaching a fielder along the sagittal plane, a LOT strategy cannot be applied because there is only a vertical gaze angle; because there is no horizontal component of the gaze angle, all ball trajectories, whether leading to catches or not, will result in linear optical trajectories. PLOS ONE | www.plosone.org 1 March 2014 | Volume 9 | Issue 3 | e92392 When a fly ball approaches along the sagittal plane, only running in the forward and backward direction needs to be controlled. In 1968, the physicist Neville Chapman [24] consid- ered the mathematics of the situation of a fly ball on a parabolic trajectory approaching a fielder head on. He showed that the rate of change of the tangent of the gaze angle (the angle between the line of gaze and the horizontal, assuming that the gaze is directed at the ball) would remain constant if the fielder runs to the interception location at a constant speed. Thus, for fielders to arrive at the right place in the right time, the Chapman strategy amounts to keeping this rate of change constant. Because the rate of change of the tangent of the gaze angle is equivalent to the speed of the projection of the ball onto an image plane, and because keeping speed constant is equivalent to keeping acceler- ation at zero, the Chapman strategy is also known as the Optical Acceleration Cancellation (OAC) strategy (cf. [25,26]; see also [6,21,27]). Empirical studies have shown that participants, running to catch fly balls, show locomotion patterns that are consistent with the use of the OAC strategy [6,26,28,29]. Because the OAC strategy is a strategy based on prospective information, it predicts that locomotion paths will differ for balls that land in the same spot but with different trajectories. This has been demonstrated in catching cricket balls [6], baseballs [23], and in heading virtual soccer balls [22]. The Chapman strategy specifically applies to fly balls that approach the fielder head on. As mentioned before, the textbook (e.g., [30,31]) candidate complementary strategy to deal with the lateral component of running is the LOT strategy. Recent research using virtual reality has shown that the LOT strategy might not be the final answer [22,23], and other strategies to complement the OAC strategy have been put forward (e.g., strategies of keeping constant the bearing angle—the CBA strategy, see [24], or its first temporal derivative—see [21]). If fielders control their locomotor trajectories on a moment-to- moment basis and use prospective information, they need to rely on a constant informational coupling with the approaching fly ball. However, according to Chodosh and colleagues [2] (cf. [1,10,11]), there is no need for such continuous visual coupling because fielders are capable of predicting the future landing location of the ball based on early available information of its trajectory: These authors argued that real fielders, like Willie Mays, simply look at the ball, predict the interception location, run there at maximal speed, and wait for the ball to arrive. Quite surprisingly, the issue of whether or not the catching of fly balls involves a constant visual coupling has not yet received much scientific attention. A notable exception is the study by Oudejans, Michaels, Bakker, and Davids [32], which examined gaze direction of fielders confronted with approaching fly balls. Oudejans and colleagues [32] were interested in the potential contribution of extraretinal systems for picking up the information to guide running to intercept fly balls (see also [29,33,34]). They argued that if the ball is tracked with gaze, not only the retinal system but also vestibular or proprioceptive systems might be used to pick up optical acceleration. Participants were equipped with a gaze-tracking system, and were allowed to make a few steps in the right direction for fly balls projected at them head on. Because the gaze tracker was connected with a cable to the recording unit, participants could only move about one or two steps forward or backward. Interesting in the present context is the finding that participants in the Oudejans et al. study, indeed, continuously kept their eyes on the ball, by moving both their heads and eyes. When using a predictive strategy, fielders obviously need to look at the ball during the initial part of its flight. Certainly, there is no need to keep the eyes on the ball during its entire flight. Although the use of a prospective strategy does not necessitate such continuous tracking of the ball (intermittent looking at the ball would suffice), the finding that participants do continuously track the ball would fit the use of a prospective strategy better than it would the use of a predictive strategy. The present paper reports an experiment in which we tracked the gaze of participants in a setting in which approaching fly balls either were within their locomotor reach (i.e., balls were catchable) or were not within their locomotor reach (i.e., balls were uncatchable). Importantly, the gaze tracker that we used was mobile, and allowed the participants to use their natural range of motion. That is to say, whereas Oudejans et al. [32] have shown that their participants continu- ously tracked the balls with their gaze for balls falling at or near the initial position of the participant or when simply watching balls that landed farther away than two steps, the present study allows to establish this behavior while participants are free to run much greater distances, even reaching their top speeds. Furthermore, we studied two situations. In line with the majority of previous studies on catching fly balls, we considered balls that would fly close enough to the participants that they would be able to arrive at the interception location in time. In addition, we also studied the situation in which balls were projected so far away from the participants’ starting position that they would not be able to reach the ball before it would hit the floor. We instructed participants to indicate when they knew that a ball would be uncatchable, and when this occurred we inspected the direction of gaze up until this point. In short, the present study considered running to catch fly balls under the demanding circumstances as seen in regular ball games. Tracking the ball might be regarded much more difficult when running close to full speed. When even under these strenuous conditions we would find pursuit tracking of the ball, we argue, this gaze behavior must have a functional origin, which most probably would be related with continuous visual control. Methods Participants Ten female volunteers (mean age 21.762.2 years) participated in the experiment. To be included, they needed to have at least two years of experience in ball sports. All participants reported normal, or corrected to normal (lenses) vision. Prior to the experiment, participants were informed about the procedure of the study and gave their written informed consent. The study was approved by the Ethics Board of the Center for Human Movement Sciences (University Medical Center Groningen, the Netherlands), and the protocol was in accordance with the Declaration of Helsinki. Apparatus To determine the point of gaze (PoG), participants were equipped with a monocular, mobile eye tracker (Mobile Eye, Applied Science Laboratories, Bedford, MA). The tracking system consists of a scene camera (recording the field of view of the participant), and an optics module that consists of a near infrared light source and an eye camera. All components are mounted on a pair of lightweight spectacles. Calculation of the point of gaze is based on ‘dark pupil tracking’ and involves detection of the center of the pupil and the reflection of a cluster of three infrared LEDs on the cornea. Eye rotations are calculated from the angle and length of the vector connecting the pupil center and the corneal reflection. After calibration (see below), eye rotations are mapped onto the scene view, establishing the PoG in the scene. Interleaved images of the eye camera and the scene camera were recorded on Keeping Your Eyes Continuously on the Ball PLOS ONE | www.plosone.org 2 March 2014 | Volume 9 | Issue 3 | e92392 tape using a portable video recorder (Sony GV-D1000E DVCR), at 30 Hz. The PoG was represented in the scene view by a crosshair with an approximate accuracy of 1u visual angle. The visual range of the eye tracker is 50 degrees horizontally, and 40 degrees vertically. The weight of the system, excluding the video recorder, is 76 grams. During testing, the video recorder was worn in a pouch around the waist, and allowed near-normal mobility for the participant. The eye tracker was calibrated using a 3.3 m high by 4.4 m wide grid with 20 equally spaced points (4 rows of 5 dots each), representing a visual angle of 16.9u in the horizontal and 12.6u in the vertical direction. During calibration, the participants were positioned 15 meters from the grid, while their head rested on a chinrest that fixated the head. The gaze tracker was calibrated prior to the start of the experiment, after each set of 18 trials, and also when the participant indicated that the tracker had changed its position on the head during testing. Setup and procedure The experiment was performed in a well-lit gymnasium (50630610 m). A ball-projecting machine (Louisville Slugger, type UPM45 Blue Flame) with adjustable force and projection angle was used to deliver tennis balls along the sagittal plane of the participant, at different projection distances. Because the projec- tion angle could be manipulated only within a limited range, we used wooden blocks that were placed underneath the ball projection machine to generate the desired trajectories. The projected distance of the fly balls was varied systematically by adjusting the projection force and angle, and ranged approxi- mately from 10 to 29 m. The apex of the trajectory was about 8.5 m for every trial, so that all fly balls had an approximate flight time of 2.5 s. The ball-projecting machine was occluded from sight to prevent visual anticipation of the ball’s trajectory. Participants completed 54 trials. The initial position of the participant was 20 m from ball projection, and was identical in all trials. At the start of each trial, the experimenter verbally cued the participant before ball delivery. Participants were instructed not to make a dive to perform a successful catch. Other than that, participants were free to move as they felt necessary to catch the ball. No instructions were given with regard to catching strategies (e.g. overhand or underhand catching). Not all projected balls were catchable. Participants were instructed to call ‘no’ at the instant that they realized that they were unable to catch the ball. Data analysis We used EyeVision software (Applied Science Laboratories, Bedford, MA) to convert the video data that were stored on tape into AVI files. We analyzed the data from the moment of ball projection until the moment the ball was either intercepted or until the moment that the participant indicated that the ball was uncatchable by calling ‘no’. We used the audio signal from the internal microphone of the eye tracking system to detect these moments. More specifically, we marked the first video frame in which the sound of ball projection was audible as the first video frame for further analysis. The final frame that was analyzed for each trial corresponded either with the first frame in which the sound of the ball hitting the hand of the participant was audible, or the first frame in which the sound of the participant calling ‘no’ was audible. Audio analysis of the video data was performed in Adobe Premiere CS6. To assess whether gaze tracked the ball, we considered the distance between the point of gaze (PoG) and the ball image, for each video frame. We used the EyeVision software to establish the 2D position of the PoG in the scene plane. Next, we used ASL Results Plus GM software (Applied Science Laboratories, Bedford, MA) to filter invalid values for the PoG. With custom-made software in MATLAB (Mathworks R2012b), we determined the 2D position of the ball in the scene plane, by hand. Finally, we computed the absolute distance in pixels between the PoG and the ball in the scene image. We assigned points of gaze to one of two categories: either on the ball (‘tracking’) or not on the ball (‘other’). A PoG was considered to be on the ball whenever the absolute distance between the ball and the PoG in the scene plane was smaller than 75 pixels (corresponding to 6.25u visual angle). Although theoretically this criterion allows for changes in distance of 150 pixels between successive frames to be assigned to the ‘tracking’ category, it turned out that these changes were smaller than 25 pixels in 95.8% of the frames identified as ‘tracking’, and that in only 0.4% of the ‘tracking’ frames the change was greater than 75 pixels. For each trial, the relative contributions of frames associated with ‘tracking’ and ‘other’ behavior will be expressed as a percentage of the total number of frames that had a valid PoG. Furthermore, we will present median distances between the PoG and the ball, as well as interquartile ranges. Results The relation between the PoG and the ball could be established in 74.7% of all video frames. In the remaining frames, the relation between the ball and the PoG could not be assessed, either because the ball could not be identified in the video frame or because the PoG was lost. Unsuccessful calibration of the Mobile Eye led us to exclude the data from two participants from further analysis. Finally, 20 trials were excluded from further analysis because participants did not catch the ball but also did not indicate that it would be uncatchable (16 trials) or because we were unable to determine whether participants had touched the ball (4 trials). Preliminary video analysis suggested that participants almost exclusively directed their point of gaze to the ball and rarely directed their gaze to locations elsewhere in the scene (for representative examples of a trial in which the ball was caught and of a trial in which a ball was judged to be uncatchable, see Figure S1 and Movies S1 and S2). The average median distance between the PoG and the ball was 24.26 pixels (with an average interquartile range of 22.67 pixels); medians (interquartile ranges) were 23.10 (20.28) pixels in the trials in which balls were caught (n = 230), and 25.85 (25.58) pixels in the trials in which balls were judged to be uncatchable (n = 168). Participants tracked the ball, on average, in 95.5% of the trial when they caught the ball, and in 92.9% of the trial when they called a ‘no’. As detailed in the Methods section, video frames in which the distance between the ball and the PoG was more than 75 pixels formed the ‘other’ category. Further investigation of this category (representing 5.7% of all frames with a valid PoG) showed that 3.1% of all ‘other’ gaze behavior constituted meaningful gaze behavior and could be classified as ‘fixations’ (operationally defined as stable gaze for three or more consecutive frames). That is, fixations on items other than the ball accounted for 0.2% of all displayed gaze behavior. Figure 1 gives the percentage of frames that were categorized as ‘tracking’ as a function of time. Data points are represented in bins of 100 ms, combining sets of three consecutive video frames. In Figure 1A, which shows the trials in which the balls were caught, the abscissa represents the time until contact with the ball. It can be seen in Figure 1A that the contributions of tracking behavior to total gaze behavior remained relatively constant throughout the Keeping Your Eyes Continuously on the Ball PLOS ONE | www.plosone.org 3 March 2014 | Volume 9 | Issue 3 | e92392 trial. Only the last 100 ms before the catch deviated substantially from this trend. Figure 1B represents the trials in which participants judged balls to be uncatchable. In Figure 1B, the abscissa represents the time until the moment that a ‘no’ was called. Also for this type of trials, the contribution of tracking behavior to total gaze behavior remained relatively constant throughout the trial. A slight deviation from this trend can be seen on the left side of Figure 1B (i.e., the two bins spanning from t = 2.2 to t = 2.0). Because these early bins included only few trials, the percentages of these bins were sensitive to the presence of single frames with a PoG that was coded as ‘other’. More particularly, one trial that had a few consecutive frames that were classified as ‘other’ early on in the trial (more than 2 s before the participant called ‘no’) was mostly responsible for the apparent decrease in tracking behavior. Discussion In his famous catch in 1954, Willie Mays looked at the ball, turned his back to the ball while running, and finally looked back at the ball again. Is this the usual way for fielders running to catch a fly ball? Do they simply know where to run from a single glance on the ball’s trajectory (cf. [1,2,10,11]), or do fielders need continuous monitoring of the ball’s position? The results of the present study show that participants running to catch an approaching fly ball continuously keep their eyes on the ball. Although the use of a predictive strategy would not preclude continuous tracking of the ball, and the use of a prospective strategy would not necessitate 100% tracking, the gaze behaviour of our participants suggests that their running is under continuous visual control, characteristic for a prospective strategy. Earlier work by Oudejans et al. [32] showed that their participants reliably tracked the ball in both a fly-ball watching and a catching task. In the majority of trials administered by Oudejans and colleagues, participants were asked to simply observe fly balls approaching head on. Watching these balls resulted in pursuit tracking, with both head and eye movement contributing to keeping the gaze on the ball. In a catching condition, participants were allowed to move and actually catch the approaching balls. Because the gaze tracker used by Oudejans et al. was wired, it restricted the participants’ mobility, such that they were only able to make a few steps to intercept a fly ball. As a consequence, balls in their catching condition had to be projected within a few meters from the participants’ initial position. Also in the catching condition, participants tracked the ball with their gaze, although the contributions of head and eye movement to directing the gaze were different than in the watching conditions (see also [29]). The present study allowed participants to move as they would naturally do when catching a fly ball. With the mobile gaze tracker that we used, we were able to study gaze in situations comparable to real catching in the outfield. Participants in the present study tracked the ball with their gaze nearly exclusively, regardless of the projected distance. Further- more, they also showed tracking of balls that they realized were uncatchable. This latter condition is not commonly part of a study into the control of interception, although it is part and parcel of the reality of outfielders. The results suggest that the information for knowing that a ball cannot be caught should not be sought in a failure to keep tracking a ball. That is to say, our participants tracked the ball up until the moment that they indicated that the ball was out of their reach. They had no problems doing so, even when running at their maximum speed. Clearly, a failure to track the ball was no indication for the participants that an approaching fly ball would be uncatchable. As discussed before, our results demonstrate that participants tracked the ball throughout its trajectory. We would like to stress that especially the fact that tracking continued to just before the actual catch speaks in favour of the use of continuous guidance rather than early prediction. Both the use of a predictive and of a prospective strategy would predict gaze pursuit during the early part of ball flight. However, when using a predictive strategy, in which the interception location and time are inferred from the first part of the ball trajectory, there seems to be no advantage of keeping an eye on the ball for the rest of its flight; continuous tracking fits more naturally with continuous visual control. Only during the very final part of the ball’s approach, approximately during the final 100 ms, did tracking become inconsistent. A reason for this might be that the participants had actually stopped tracking the ball with their gaze because it was not needed for running to the interception location anymore. It has been suggested that fly ball interception consists of two phases; locomotion to the interception point and making the actual catch (e.g., see [23,26]). The last 100 ms might reflect the latter phase. Alternatively, participants might have started to prepare for a follow-up action, such as throwing the ball to a teammate. In conclusion, the present results paint a picture that is consistent with the use of a prospective strategy in dealing with the outfielder problem. Gaze data are not able to show indisputably that fielders do not use a predictive strategy, in Figure 1. Percentage of tracking as a function of time. The number of frames in which participants tracked the ball expressed as a percentage of the total number of frames with valid data in a trial. A) Average percentages for trials in which the ball was caught; t = 0 represents the time of contact with the ball; B) Average percentages for trials in which the ball was judged to be uncatchable; t = 0 represents the time that a ‘no’ was called. doi:10.1371/journal.pone.0092392.g001 Keeping Your Eyes Continuously on the Ball PLOS ONE | www.plosone.org 4 March 2014 | Volume 9 | Issue 3 | e92392 which they know where to run from looking at the early part of ball flight. However, the finding that participants continuously kept their eye on the ball, while running several meters to catch a ball that might or might not be catchable, fits naturally with a continuous visual control on the basis of prospective information. Supporting Information Figure S1 Gaze in two representative trials. Distance between the ball image and the point of gaze as a function of time. A) Gaze for a participant who successfully caught the projected fly ball (after 2.64 s); B) Gaze for a participant who indicated that the projected fly ball was uncatchable for her (after 1.37 s). See also Movies S1 and S2, which show scene camera recordings of these trials. (PDF) Movie S1 Gaze in a trial in which the ball was caught. Eye-tracker recording of a trial in which the participant caught the ball. The point of gaze is indicated by the red crosshair that overlays the scene-camera images. Figure S1A gives the distance between the ball and the point of gaze as a function of time for this trial. (AVI) Movie S2 Gaze in a trial in which the ball was judged to be uncatchable. Eye-tracker recording of a trial in which the participant judged the ball to be uncatchable. The point of gaze is indicated by the red crosshair that overlays the scene-camera images. Figure S1B gives the distance between the ball and the point of gaze as a function of time for this trial. (AVI) Acknowledgments We thank Anna Kersten and Pawel van der Steen for their assistance in collecting the data, and Emyl Smid for his help with the ASL Mobile Eye tracker. Author Contributions Conceived and designed the experiments: DP FZ. Performed the experiments: DP FZ. Analyzed the data: DP RdO FZ. Contributed reagents/materials/analysis tools: DP RdO FZ. Wrote the paper: DP RdO FZ. References 1. Adair RK (2002) The physics of baseball. 3rd ed. New York: Perennial. 2. Chodosh LA, Lifson LE, Tabin C (1995) Play ball. Science 268: 1682–1683. 3. Arzamarski R, Harrison SJ, Hajnal A, Michaels CF (2007) Lateral ball interception: Hand movements during linear ball trajectories. Exp Brain Res 177: 312–323. doi:10.1007/s00221-006-0671-8. 4. Bastin J, Craig CM, Montagne G (2006) Prospective strategies underlie the control of interceptive actions. Hum Mov Sci 25: 718–732. doi:10.1016/ j.humov.2006.04.001. 5. Ledouit S, Casanova R, Zaal FTJM, Bootsma RJ (2013) Prospective control in catching: the persistent angle-of-approach effect in lateral interception. PLoS ONE 8: e80827. doi:10.1371/journal.pone.0080827.g011. 6. McLeod P, Dienes Z (1996) Do fielders know where to go to catch the ball or only how to get there? J Exp Psychol Human Perc Perf 22: 531–543. 7. Montagne G, Laurent M, Durey A, Bootsma RJ (1999) Movement reversals in ball catching. Exp Brain Res 129: 87–92. 8. Zaal FTJM, Bongers RM, Pepping G-J, Bootsma RJ (2012) Base on balls for the Chapman strategy: Reassessing Brouwer, Brenner, and Smeets (2002). Atten Percept Psychophys 74: 1488–1498. doi:10.3758/s13414-012-0328-6. 9. Bosco G, Delle Monache S, Lacquaniti F (2012) Catching what we can’t see: manual interception of occluded fly-ball trajectories. PLoS ONE 7: e49381. doi:10.1371/journal.pone.0049381.t003. 10. Saxberg BV (1987) Projected free fall trajectories. II. Human experiments. Biol Cybern 56: 177–184. 11. Brancazio PJ (1985) Looking into Chapman’s homer: the physics of judging a fly ball. Am J Phys 53: 849–855. 12. Bootsma RJ (2009) The (current) future is here! Perception 38: 851–851. 13. Bootsma RJ, Fayt V, Zaal FTJM, Laurent M (1997) On the information-based regulation of movement: What Wann (1996) may want to consider. J Exp Psychol Human Perc Perf 23: 1282–1289. 14. Fajen BR, Riley MA, Turvey M (2009) Information, affordances, and the control of action in sport. Int J Sport Psychol 40: 79–107. 15. McBeath MK, Shaffer DM, Kaiser M (1995) How baseball outfielders determine where to run to catch fly balls. Science 268: 569–573. 16. Shaffer DM, McBeath MK (2002) Baseball outfielders maintain a linear optical trajectory when tracking uncatchable fly balls. J Exp Psychol Human Perc Perf 28: 335–348. doi:10.1037//0096-1523.28.2.335. 17. Shaffer DM, McBeath MK, Roy WL, Krauchunas SM (2003) A linear optical trajectory informs the fielder where to run to the side to catch fly balls. J Exp Psychol Human Perc Perf 29: 1244–1250. doi:10.1037/0096-1523.29.6.1244. 18. Sugar TG, McBeath MK, Wang Z (2006) A unified fielder theory for interception of moving objects either above or below the horizon. Psychon B Rev 13: 908–917. 19. Shaffer DM, McBeath MK, Krauchunas SM, Sugar TG (2008) Evidence for a generic interceptive strategy. Percept Psychophys 70: 145–157. 20. McLeod P, Reed N, Dienes Z (2002) The optic trajectory is not a lot of use if you want to catch the ball. J Exp Psychol Human Perc Perf 28: 1499–1501. doi:10.1037//0096-1523.28.6.1499. 21. McLeod P, Reed N, Dienes Z (2006) The generalized optic acceleration cancellation theory of catching. J Exp Psychol Human Perc Perf 32: 139–148. doi:10.1037/0096-1523.32.1.139. 22. McLeod P, Reed N, Gilson SJ, Glennerster A (2008) How soccer players head the ball: A test of optic acceleration cancellation theory with virtual reality. Vision Res 48: 1479–1487. doi:10.1016/j.visres.2008.03.016. 23. Fink PW, Foo PS, Warren WH (2009) Catching fly balls in virtual reality: A critical test of the outfielder problem. J Vision 9: 14.1–.8. doi:10.1167/9.13.14. 24. Chapman S (1968) Catching a baseball. Am J Phys 36: 868–870. 25. Todd J (1981) Visual information about moving objects. J Exp Psychol Human Perc Perf 7: 975–810. 26. Michaels CF, Oudejans RR (1992) The optics and actions of catching fly balls: Zeroing out optical acceleration. Ecol Psychol 4: 199–222. 27. McLeod P, Reed N, Dienes Z (2003) How fielders arrive in time to catch the ball. Nature 426: 244–245. doi:10.1038/426244a. 28. McLeod P, Reed N, Dienes Z (2001) Toward a unified fielder theory: What we do not yet know about how people run to catch a ball. J Exp Psychol Human Perc Perf 27: 1347–1355. 29. Zaal FTJM, Michaels CF (2003) The information for catching fly balls: Judging and intercepting virtual balls in a CAVE. J Exp Psychol Human Perc Perf 29: 537–555. 30. Coren S, Ward LM, Enns JT (2004) Sensation and perception. 6 ed. John Wiley. 31. Bruce V, Green PR, Georgeson MA (2003) Visual perception: Physiology, psychology and ecology. 4 ed. Hove, UK: Psychology Press. 32. Oudejans RR, Michaels CF, Bakker F, Davids K (1999) Shedding some light on catching in the dark: Perceptual mechanisms for catching fly balls. J Exp Psychol Human Perc Perf 25: 531–542. 33. Bongers RM, Michaels CF (2008) The role of eye and head movements in detecting information about fly balls. J Exp Psychol Human Perc Perf 34: 1515– 1523. doi:10.1037/a0011974. 34. Brouwer A-M, Lo´pez-Moliner J, Brenner E, Smeets JBJ (2006) Determining whether a ball will land behind or in front of you: Not just a combination of expansion and angular velocity. Vision Res 46: 382–391. doi:10.1016/ j.visres.2005.09.002. Keeping Your Eyes Continuously on the Ball PLOS ONE | www.plosone.org 5 March 2014 | Volume 9 | Issue 3 | e92392

Keeping your eyes continuously on the ball while running for catchable and uncatchable fly balls.

03-26-2014

Postma, Dees B W,den Otter, A Rob,Zaal, Frank T J M

eng

PMC8363530

Vol.:(0123456789) Sports Medicine (2021) 51:1835–1854 https://doi.org/10.1007/s40279-021-01481-2 REVIEW ARTICLE Crossing the Golden Training Divide: The Science and Practice of Training World‑Class 800‑ and 1500‑m Runners Thomas Haugen1  · Øyvind Sandbakk2 · Eystein Enoksen3 · Stephen Seiler4 · Espen Tønnessen1 Accepted: 23 April 2021 / Published online: 21 May 2021 © The Author(s) 2021 Abstract Despite an increasing amount of research devoted to middle-distance training (herein the 800 and 1500 m events), informa- tion regarding the training methodologies of world-class runners is limited. Therefore, the objective of this review was to integrate scientific and best practice literature and outline a novel framework for understanding the training and development of elite middle-distance performance. Herein, we describe how well-known training principles and fundamental training characteristics are applied by world-leading middle-distance coaches and athletes to meet the physiological and neuromus- cular demands of 800 and 1500 m. Large diversities in physiological profiles and training emerge among middle-distance runners, justifying a categorization into types across a continuum (400–800 m types, 800 m specialists, 800–1500 m types, 1500 m specialists and 1500–5000 m types). Larger running volumes (120–170 vs. 50–120 km·week−1 during the prepara- tion period) and higher aerobic/anaerobic training distribution (90/10 vs. 60/40% of the annual running sessions below vs. at or above anaerobic threshold) distinguish 1500- and 800-m runners. Lactate tolerance and lactate production training are regularly included interval sessions by middle-distance runners, particularly among 800-m athletes. In addition, 800-m run- ners perform more strength, power and plyometric training than 1500-m runners. Although the literature is biased towards men and “long-distance thinking,” this review provides a point of departure for scientists and practitioners to further explore and quantify the training and development of elite 800- and 1500-m running performance and serves as a position statement for outlining current state-of-the-art middle-distance training recommendations. Key Points This review serves as a position statement for outlining state-of-the-art middle-distance training recommenda- tions. There are considerable gaps between science and best practice regarding how training principles and training methods should be applied for elite middle-distance run- ning performance. We identify physiological and training distinctions between world-class 800- and 1500-m runners. * Thomas Haugen [email protected] 1 School of Health Sciences, Kristiania University College, Sentrum, PB 1190, 0107 Oslo, Norway 2 Department of Neuromedicine and Movement Science, Centre for Elite Sports Research, Norwegian University of Science and Technology, 7491 Trondheim, Norway 3 Department of Physical Performance, Norwegian School of Sport Sciences, 0806 Oslo, Norway 4 Faculty of Health and Sport Sciences, University of Agder, PB 422, 4604 Kristiansand, Norway 1836 T. Haugen et al. 1 Background Middle-distance running was a central part of the Olym- pic program for men already at the first modern Games in 1896. Over the last century, quantum leaps in men’s performance have been achieved by barrier breaking ath- letes such as Paavo Nurmi, Gunder Hägg, Rudolf Harbig and Roger Bannister. The progression of female middle- distance running performances was initially slower than that observed for men [1], but this was due to social, not biological constraints. By the 1928 Olympic Games, women competed in 2 of the 13 running events contested by men, the 100 and 800 m. Unfortunately, even this small progress was halted when the International Olympic Com- mittee (IOC) received erroneous reports of female athletes collapsing after running the 800 m and decided to ban women from competing over distances longer than 200 m. The middle-distance events were not added to the Olympic program for women until 1960, after which the sex-gap in middle-distance performance declined gradually until the 1980s. Since then, male and female sex-specific perfor- mance differences have stabilized around ~ 10% [2]. Despite an increasing amount of research devoted to middle-distance training [e.g., 3–17], it is reasonable to argue that the developments in these disciplines have not been driven by sport scientists [18]. Publicly available “recipe books” and training diaries based upon the practi- cal experience and intuition of world-leading athletes and coaches have become important and popular sources of best practice training information and framework devel- opment for the international middle-distance community [19–59] (Table 1). While best practice training in athletic sprinting [60] and long-distance running [61–65] has been scientifically reported, information regarding the varying training components across the annual cycle of world-class middle-distance runners is limited. Furthermore, the train- ing characteristics of 800- and 1500-m runners have not yet been systematically compared. Such a comparison is warranted because of the marked shift towards a more dis- tinct emphasis on aerobic energy provision from 800 to 1500 m as well as the interactions between mechanical effectiveness and metabolic efficiency in this transition. Therefore, the objective of this review is to integrate sci- entific and best practice coaching literature to outline a novel framework for the training and development of elite middle-distance performance. Although the present review is anchored in the standard Olympic 800- and 1500-m dis- tances, the outlined terminology, training zone model and training principles are also relevant for other distances and sports. The present review strategy is challenging. Firstly, an initial review of the literature reveals that several biases are present, including a substantial sex bias (male domi- nance) as well as “group culture” biases across a hand- ful of successful training groups. A relative bias towards emphasis on training aerobic capacity is particularly pre- sent for the 800 m, as this discipline seems heavily influ- enced by “long-distance thinking” in the available research literature. Hence, the generalizable training recommenda- tions outlined in this review might not be optimal for all middle-distance athletes. Secondly, a potential source of misinterpretation is the lack of a common framework and terminology. Moreover, the included coaching literature cannot be controlled for possible training prescription- execution differences as exemplified by Ingham et al. [9]. Although these stories rarely gain attention, most “famous” coaches have also coached underperforming talents. We acknowledge this bias but note that the vast majority of the coaches listed in Table 1 have achieved success with multiple athletes. Finally, the widespread use of doping in international athletics must be acknowledged. All these challenges and limitations reflect today’s athlet- ics, for better and worse, and the outcomes of this review must therefore be interpreted with these caveats in mind. Sensitive to these limitations, we still contend that inte- gration of available research evidence and results-proven practice provides a valid point of departure for outlining state-of-the-art training recommendations and for genera- tion of new hypotheses to be tested in future research [60, 66]. 2 Physiological and Mechanical Determinants of Middle‑Distance Running Performance The 800- and 1500-m running disciplines are where aero- bic and anaerobic energetics converge [5]. Importantly, these classically defined disciplines are also where effec- tive maximal sprint speed (MSS) mechanics and efficient long-distance running energetics collide. While mechanics and energetics are not independent in middle-distance run- ning, we choose to examine these events with what might be called scientific bifocals and try to converge them in a logical manner. 2.1 The Energetic Side of the Middle‑Distance Coin During an 800-m run, the relative energy system contribu- tions from aerobic and anaerobic metabolism are reported to be 60–75 and 25–40%, respectively, while correspond- ing values for 1500 m are 75–85 and 15–25% [6, 7, 13]. The range in energy system contribution is greater in the 800 m compared to the 1500-m event due to the variability of the athletes presenting at 800 m. Overall, these relative aerobic 1837 Training and development of world-class middle-distance athletes energy contribution estimates overlap reasonably well with the reported type I muscle fiber distribution ranges in middle-dis- tance runners [13]. Just as has been well established for long- distance running, maximal oxygen uptake (VO2max), fractional utilization of VO2max, running economy (RE), velocity at the anaerobic threshold (vAT), and velocity at VO2max (vVO2max) are all positively correlated with middle-distance perfor- mance [5, 8, 67]. However, to optimize energy mobilization and utilization, O2 kinetics as well as anaerobic power and capacity play decisive roles in middle distance performance. As Olympic gold medalist 800-m runner Vebjørn Rodal suc- cinctly summarized the importance of O2 kinetics to one of the authors (ØS): “It does not matter if I can reach a higher VO2max in five minutes when I have to cross the finish line in 102 s.” In addition, both energy expenditure capacity and economy/efficiency likely deteriorate during middle-distance Table 1 Sources of best practice training information In addition, we have had personal communications with Vebjørn Rodal (Olympic 800-m champion in 1996) and Arturo Casado (European 1500-m champion in 2010). Novel training data from these athletes are presented in Table 6 WC world championships, EC European championships, WR former or current world-record holder a Honore Hoedt coached Sifan Hassan during her early career, not when she broke several world records Athletes [reference] Personal bests (min) International merits Type of source Alberto Juantorena [38] 800 m 1:43.44 (WR) Olympic gold 1976 Keynote speech/training log Clayton Murphy [26] 800 m 1:42.93 Olympic bronze 2016 Interview/presentation David Rudisha [50, 51] 800 m 1:40.91 (WR) Olympic gold 2012 and 2016 Web post and training log Hicham El Guerrouj [45] 1500 m 3:26.00 (WR) Olympic gold 2004 Lectures Jim Ryun [29] 800 m 1:44.3—1500 m 3:33.1 Olympic silver 1968 Chronicle and training log Joaquim Cruz [36] 800 m 1:41.77—1500 m 3:34.63 Olympic gold 1984 Chronicle and training log John Walker [28] 1500 m 3:32.4—mile 3:49.08 (WR) Olympic gold 1500 m 1976 Magazine article/interview Marty Liquori [39] Mile 3:52.2 Pan American champion 1971 Chronicle and training log Michael Rimmer [40] 800 m 1:43.89 EC silver 2010 Chronicle and training log Natalia Rodriguez [43] 1500 m 3:59.51 WC and EC gold 2010–2011 Chronicle Nick Symmonds [30] 800 m 1:42.95—1500 m 3:34.55 WC silver 2013 Training log Nick Willis [44] 1500 m 3:29.66—mile 3:49.83 Olympic medals 2008 and 2016 Training log Peter Elliott [22] 800 m 1:42.97—1500 m 3:32.69 Olympic silver 1988 Chronicle and training log Said Aouita [24] 1500 m 3:29.46 (WR)—mile 3:46.76 Olympic gold 1984, WC gold 1987 Training log Silas Kiplagat [49] 1500 m 3:27.64 WC silver 2011 Training log Taoufik Makloufi [46] 800 m 1:42.61—1500 m 3:28.75 Olympic gold 2012 Interview Coaches [reference] Successful middle-distance athletes Athlete merits Type of source Arthur Lydiard [19–21] Peter Snell (WR), Murray Halberg, Barry Magee Olympic gold 1960 and 1964 Books Bill Bowerman [53] Steve Prefontaine, Jack Hutchins, Sig Ohlemann He trained 31 Olympic athletes Book David Sunderland [52] Jane Finch, Lynsey Sharp Indoor WR 1977, EC gold 2012 Book Gianni Ghidini [37] Wilfred Bungei, Amel Tuka Olympic & WC medals since 2001 Presentation Harry Wilson [34, 57] Steve Ovett (WR) Olympic gold 1980, EC gold 1978 Chronicle/training log Honore Hoedt [41] Sifan Hassan (WR)a, Brad Som, Amoud Okken WC & EC medals since 2006 Presentation Jack Daniels [58] Coached seven athletes to the U.S. Olympic team Olympic finalists Book Jama Aden [31] Genzebe Dibaba (WR), Abdi Bile, Taoufik Makloufi Olympic & WC medals since 1987 Magazine article/interview Joe Vigil [33, 56] Coach for the US Olympic team in 1998 Olympic finalists Presentations Kim McDonald [23] Daniel Komen (WR), Noah Ngeny, Laban Rotich Multiple WC medals in the 1990s Chronicles/training logs Lee LaBadie [26] Clayton Murphy Olympic bronze 2016 Presentations Margo Jennings [32] Maria Mutola, Kelly Holmes Olympic & WC medals 1993–2004 Chronicle/interview Nic Bideau [48] Craig Mottram WC bronze 2005 Commentary Peter Coe [54, 59] Sebastian Coe (WR) Olympic gold 1980 and 1984 Books Steve Magness [42] Assistant coach and scientific advisor for elite runners Olympic & WC medals 2011–2012 E-book and presentation Tomasz Lewandowski [25] Marcin Lewandowski EC gold 2010, WC bronze 2019 Presentation Vin Lananna [35] U.S. Olympic team coach Olympic finalists Presentations 1838 T. Haugen et al. events, indicating that fatigue-resistance/resilience might have a decisive performance-impact. To this point, de Koning and colleagues have directly challenged the assumption of a stable gross efficiency during short maximal cycling efforts within the middle-distance time window [68, 69]. Using a sequence of sub-maximal-maximal-sub-maximal trials and back-extrapolation, they estimate that metabolic efficiency declines enough during 100–240 s duration cycling time tri- als to result in a ~ 30% underestimation of the anaerobic energy contribution to total energy expenditure. Unfortunately, com- prehensive quantification of running economy (total external work performed/total energy expenditure) at speeds above the lactate threshold remains elusive [12]. While traditional endurance disciplines can be described as maximization challenges (i.e., training that enhances VO2max or fractional utilization is “always positive” for per- formance), we propose that the 800-m event in particular requires an energy release optimization strategy that respects the interactions and trade-offs between anaerobic and aero- bic metabolism emerging in both training and performance. This complexity allows internationally successful middle- distance runners to present a variety of physiological pro- files [12–15]. For example, VO2max ranges from ~ 65 to 85 ml·kg·min−1 in elite men [16, 29, 70, 71]. Similar variation is seen among elite women, albeit at ~ 10% lower values [71] due to lower hemoglobin concentrations and higher rela- tive body fat percentage [72]. Consequently, correlations between isolated aerobic performance-determining factors and performance in homogeneous subsets of middle-dis- tance runners are modest at best. We find no evidence to suggest that female and male mid- dle distance athletes should not be examined as one elite population from an energetics point of view. However, the 800-m event rides an energetic “tipping point;” it sits on a portion of the velocity-duration curve where the aerobic and anaerobic contributions are particularly duration sensitive. Consequently, the additional ~ 15 s required to complete the 800 m by the best females may nudge this event towards the aerobic end of the training spectrum enough that it alters the optimal composition of their training compared to male counterparts. Lending some support to this possibility, we note that inspections of the top 200 all-time lists for the 800 and 1500 m reveal that 55 women appear on both lists, com- pared to only 38 men (http:// world athle tics. org). For com- parison, the 1500–3000 m double is more common among the 200 all-time best males and females with 51 men and 78 women appearing on both lists. 2.2 Mechanical Effectiveness: The Other Side of the Middle‑Distance Coin The role of anaerobic capacity in middle-distance running has received considerably less attention in the research literature, likely due to limitations in accurately and reli- ably quantifying anaerobic energetics [73]. Bachero-Mena et al. [3] have reported a strong relationship between 800-m performance and sprints over 20 m (r = 0.72) and 200 m (r = 0.84) in male national and international 800-m run- ners (1:43–1:58). Peter Coe [54] and Arthur Lydiard [19] have argued that world-class 800-m male athletes should be able to run 200 m in < 22.5 s prior to major competitions. Such sprint performance is determined by a combination of anaerobic energy release and the ability to transfer energy to speed over this particular distance, and this sprinting capacity requirement eliminates at least 99% of males on the planet as future world-class 800-m runners before other physiological demands are even considered. Power output and technique are considered key underlying determinants for MSS [74]. Fast male world-class middle-distance runners may approach 10 m·s−1 [12, 15], and if we assume a ~ 10% sex difference [75], corresponding females are capable of sprinting ≥ 9 m·s−1. To achieve such running velocities, maximal horizontal power outputs of ~ 21 and ~ 19 W·kg−1 are required for men and women, respectively [76]. Although the basic principles of MSS are relatively sim- ple and governed by the laws of motion, the way an athlete solves the mechanical constraints and utilizes the degrees of freedom within these constraints is far more complex [74]. Spatiotemporal variables, segment configuration at touch- down and lift-off, lower-limb segment velocities imme- diately prior to touchdown or during ground contact, leg stiffness, storage and release of elastic energy, as well as front- and back-side mechanics have received much atten- tion in research literature. However, these mechanical vari- ables are entangled, and no single variable is associated with better MSS [74]. For more information regarding running mechanics, we refer to previously published biomechanical analyses [e.g., 74, 77, 78]. Overall, middle-distance athletes must be able to reach high MSS if they are to reach an international level. How- ever, high and unfatigued MSS is not useful if a high per- centage of that velocity cannot be maintained for 100–240 s (see Sect. 3). This implies a complex integration of muscular power, metabolic efficiency, biomechanical efficiency and fatigue resistance at the muscle fiber level, as well as an optimal pacing strategy [79, 80]. 3 Athlete Profiling Due to the variety of physiological profiles among 800- and 1500-m runners, coaches typically categorize middle-dis- tance runners into distinct “types” [19–21, 41, 47, 54, 58, 59], and these types bear different labels (e.g., “speed-based” vs. “endurance-based”, “fast-typed” vs. “stamina-typed”). A simple method for athlete profiling and identification of 1839 Training and development of world-class middle-distance athletes individual strengths and weaknesses can be based on per- formance across a spread of distances below and above the main discipline (e.g., using IAAF points or percent time behind current world record). For example, 400, 800 and 1500-m performance can form the basis for analyzing an 800-m runner, presupposing that the performance level across all these distances is representative and reflects actual performance [13]. A brief review of the World Athletics all-time top lists (https:// www. world athle tics. org/ recor ds/ all- time- topli sts) clearly shows that 1500-m runners pos- sess a broader distance performance range, while a larger proportion of world-class 800-m runners appears to be “spe- cialists”. These observations are in accordance with Dan- iels [58], who argued that a strong performance relationship exists among distances ranging from 1500 m to marathon in heterogeneous subsets, while 800 and 1500 m performances are considerably less related. The concept of anaerobic speed reserve (ASR) was origi- nally introduced by Blondel et al. [81] and further developed by Sandford and associates [12–15] to provide a “first layer insight” of athlete profiling. ASR is defined as the speed zone ranging from vVO2max to MSS. MSS can be accurately measured using radar technology or timing gates [82, 83], while vVO2max (also known as maximal aerobic speed; MAS) traditionally has required laboratory-based proce- dures. However, a field method has recently been developed where a regression equation can be applied for accurate prediction of vVO2max from 1500 m time-trial performance (“gun-to-tape” or “predicted 1500-m shape”) [14]. Based on the speed reserve ratio concept (SRR = MSS/MAS), Sandford and associates classified 800-m runners into three sub-groups along a continuum as follows: 400–800 m types (SRR ≥ 1.58), 800 m specialists (SRR ≤ 1.57 to ≥ 1.47, and 800–1500 m types (SRR ≤ 1.47 to ≥ 1.36) [15]. Using the same approach, we propose that 1500-m runners can be categorized as 800–1500 m types, 1500-m specialists and 1500–5000 m types. However, the validity of this concept must be further elaborated in future research. In the fol- lowing sections of this review, the implications of athlete profile for training prescriptions will be explored in more detail, with most focus on the distinctions between 800- and 1500-m runners. 4 Expected Performance Development Among Elite Middle‑Distance Runners Middle-distance performance capacity evolves and devolves throughout life via growth, maturation, training and age- ing [84–87]. The age of peak performance in world-class middle-distance runners (mean ± SD) is 25–27 ± 2–3 years [87–90]. However, training age must also be considered, as early/late specialization may accelerate/delay age of peak performance [91]. For example, young African runners have a lifestyle that includes running to and from school from a very early age [23, 27, 92, 93], supporting the early engage- ment hypothesis [94]. However, history has also shown that late specialization and diversified experience in other sports can provide a platform for later elite performance [17, 36, 38, 39]. For the very best runners, the annual within-athlete per- formance differences are lower than the typical variation and the smallest worthwhile change is ~ 0.5% in middle- distance running [95]. Mean annual improvement scores for the world’s top 100 middle-distance runners in their early twenties are in the range of only 0.1–0.2% [87]. On aver- age, athletes must be at a very high level already in their late teens to become world-class as seniors. Haugen and co-workers calculated that middle-distance runners within the annual world top 100 lists averaged 98–99% of their peak performance result at the age of 20 [87]. However, athletes reaching the upper portion of this exclusive annual list improve their performances more than athletes of lower performance standards in the years immediately preced- ing peak performance age [87]. These differences may be explained by differences in training status, responsiveness to training, coaching quality, doping, etc. Although there is considerable variation among athletes and numerous routes to expertise under optimal conditions, a review of the best practice literature listed in Table 1 indicates that the majority of world-class 800- and 1500-m runners have specialized in the middle-distances already as juniors. 5 Training Principles 5.1 Progressive Overload The process of training adaptation is an interplay between loading and recovery, and the principle of progressive overload refers to the gradual increase of stress placed upon the body during exercise training [96–98]. Indeed, the capacity to perform and absorb large training loads is seen as both an adaptation over time and a talent. In mid- dle-distance running, commonly reported external load factors include volume, duration and intensity, while psy- chophysiological internal load factors typically include heart rate, blood lactate and session rating of perceived exertion. These variables will be examined in more detail in Sect. 6. While running distance is the most commonly reported loading factor in scientific and best practice literature, some authors argue that rating of perceived exertion (RPE) or training impulse (TRIMP; min × RPE) are more useful for the training decision-making process [99, 100]. With emerging and novel wearable technology, future training monitoring may put more emphasis on 1840 T. Haugen et al. biomechanical external load metrics such as tibial shock, foot-strike angle, ground contact time and leg stiffness to enable a more precise quantification of training stress [99]. The principle of progressive overload is envisioned to enhance performance over time and reduce the risk of injury and overtraining [96–98]. Indeed, a large pro- portion of injuries are attributed to rapid and excessive increases in training load [101, 102]. During the initial 8–12 weeks of the training year, it is therefore widely accepted in the middle-distance community that running volume must be increased gradually. In elite athletes, the initial training week is performed with ~ 40–60% of peak weekly running volume, increasing by ~ 5–15 km each week until maximal volume is reached [19–26, 28–32, 34, 36–46, 52, 54–59]. This increase is mainly achieved by increasing training frequency in the initial phase, then subsequently extended by lengthening individual training sessions. When peak running volume is achieved, the fur- ther progression in training load among middle-distance runners is normally achieved by increasing the amount or intensity of intensive training. Long-term progression rates depend on training experience and individual pre- dispositions, but total training volume and peak weekly mileage may increase up to ~ 10% per year during the late teens in well-trained athletes [17, 42, 55, 56]. A common “periodization” approach observed within best practice is that more intensive training sessions are introduced and total training volume decreases as the competition season approaches [17, 19–21, 23–25, 34, 36, 40–42, 50–52, 54–56, 58, 59] (see also Sect. 5.4). Within this context, running surface and footwear are crucial modifiers of training load for middle-distance running. It is generally assumed that the harder the surface, the higher mechanical load and reactive forces on lower limb tissues [19–21, 23, 36, 52, 54–59, 99]. Most elite athletes perform low-intensive running sessions with cushioned running shoes/trainers on forgiving surfaces (forest trails, parkland, dirt road, etc.), while high-intensive running and sprinting sessions are performed with spike shoes on a rubberized track surface. Because the latter is associated with high muscular load, such sessions rarely occur on consecutive days among leading coaches and practition- ers [17, 19–21, 23–25, 31, 34, 36, 40, 41, 50–52, 54–56, 58–60]. Although altitude training is an integrated part of mod- ern middle-distance training to increase the stress placed upon the body, this topic has received limited attention in the best practice coaching literature. We therefore refer to previously published reviews for more information regarding altitude training [e.g., 103–105]. 5.2 Specificity Training adaptations are specific to the stimulus applied, encompassing muscle groups and actions involved, speed of movement, range of motion and energy systems involved [98, 106]. Due to the performance demands underpinning middle-distance running performance, various types of training aimed to overload the aerobic and/or the anaerobic energy system while employing movement patterns specific to middle-distance running need to be performed. Based on a synthesis of best practice literature [19–59], the specific training methods for middle-distance running are described in Table 2. We refer to previously published review papers regarding physiological adaptations and responses associ- ated with such training forms [6, 7, 107–109]. Many successful athletes in typical endurance sports sup- plement their sport-specific training with alternative activity forms, so called cross-training [110–113]. Arguments sup- porting the inclusion of such non-specific training include injury prevention, aerobic capacity benefits, strengthening “weak links”, and avoidance of training monotony [113, 114]. Best practice coaching literature within middle-dis- tance running indicates that cross-training (e.g., cycling, swimming, running with floating vest or cross-country ski- ing) in most cases is employed during injury rehabilitation processes. However, it cannot be precluded that this is a part of the regular plan in certain training groups. Other “less specific” training forms such as strength, power and plyometric training are more commonly performed to tar- get the underlying anaerobic performance components (see Sect. 6.4). Although these training forms do not duplicate the holistic running movement, they may target specific components that limit performance. 5.3 Individualization The majority of training intervention studies demonstrate that considerable variability in adaptation to a given exer- cise stimulus is the norm [e.g., 115–117]. The principle of individualization refers to the notion that training prescrip- tion must be adapted and optimized according to individual predispositions (performance level, training status/age, sex, recovery/injury status and physiological and struc- tural/mechanical profiles) to maximize the effect and avoid non-responder outcomes [13, 52, 58, 98, 118]. Total train- ing load is typically higher in well-trained adult runners of higher performance standard compared to their younger, less trained and lower-performing counterparts [19–21, 56, 58]. A review of the best practice literature reveals that world-class middle-distance athletes have recorded very similar personal best times with substantial differences in training programs, and these differences are likely related to 1841 Training and development of world-class middle-distance athletes Table 2 Specific training methods for middle-distance running Training method Description Continuous running Warm up/recovery run/cool down Low-intensive running (typically 3–5 km·h−1 slower than marathon pace, i.e., 4:00–4:45 and 4:30–5:15 min·km−1 for men and women, however, the last part of the warm-up may approach marathon pace or slightly above), predominantly performed on soft surface (grass, woodland, forest paths, etc.). Typical duration is 10–30 min Long run Low-intensive steady-state running (marathon pace or 1–2 km·h−1 slower, i.e., 3:30– 4:00 and 4:00–4:30 min·km−1 for men and women) performed on forgiving surfaces such as forest trails where possible. Typical duration is 60–90 min, but 2-h runs are also performed during the preparation period Anaerobic threshold run A sustained run at moderate intensity/half-marathon pace (i.e., 2:55–3:15 and 3:10– 3:30 min·km−1 for world-class male and female middle-distance runners). Typical duration 15–40 min. The session should not be extremely fatiguing Fartlek An unstructured long-distance run in various terrains over 30–60 min. where periods of fast running are intermixed with periods of slower running. The pacing variations are determined by the athlete’s feelings and rhythms and terrain Progressive long runs A commonly used training form used by African runners. The first part of the session is identical to an easy long run. After about half the distance, the pace gradually quickens. In the final portion, the pace increases to the anaerobic threshold (half- marathon pace) or slightly past it. Athletes are advised to slow down when the pace becomes too strenuous Interval training Anaerobic threshold intervals Intervals of 3–10 min. duration at an intensity around anaerobic threshold (half- marathon pace) or slightly faster. Typical sessions: 8–12 × 800–1000-m with 1 min. recovery between intervals, 4–8 × 1500–2000 m with 1–2 min. recovery between intervals, or 2–4 × 10-min. with 2–3 min. recovery between intervals. As a rule of thumb, the recovery periods are ~ 1 min. of easy jogging per 5 min of running. Rec- ommended total time for elite runners is 25–40 min. Such intervals are advantageous because they allow the athlete to accumulate more total time than during a continu- ous anaerobic threshold run VO2max intervals Intervals of 2–4 min. duration at 3–10 K pace, with 2–3 min. recovery periods between intervals. Typical sessions: 4–7 × 800–1000 m or 2 × (6 × 400 m) with 30–60 s and 2–3 min. recovery between intervals and sets, respectively. Recommended total time for elite runners is ~ 15–20 min Lactate tolerance training Intervals typically ranging from 200 to 600 m with 800–1500 m race pace and 1–3 min. recoveries. Typical sessions: 10–16 × 200 m with 1 min. recovery between intervals, or 3 x (4 × 400 m) with 60–90 s and 3–5 min. recoveries between intervals and sets, respectively. Total accumulated distance ranges from 1500 to 5000 m in elite athletes Lactate production training Intervals typically ranging from 150 to 600 m at 200–600 m race pace and full recov- eries. Typical sessions: 5–7 × 300 m with 3–5 min. recoveries, 3–5 × 400 m with 7–15 min. recoveries, or 600–500–400–300–200 m with 6–15 min. recoveries. Total accumulated distance ranges from 800 to 2500 m in elite athletes Hill repeats The main intention is overloading horizontal propulsive muscle groups while reducing ballistic loading. Typical incline is 5–10%, and duration vary from ~ 15 s to ~ 4 min. depending on intensity, goal (aerobic intervals, lactate production or tolerance train- ing) and time of season. Typical sessions: 10–15 × 100 m with 60–90 s recoveries, or 6–8 × 800–1000 m with easy jog back recoveries. Hill repeats are mainly performed during the preparation period Sprints or time trials Time trials “All-out” efforts or trials aiming at achieving a target time. Distances are normally 50–80% of the athlete’s normal racing distance. Typically performed prior to (e.g., 10 days) an important race at the early part of the season Sprints 5–15 s runs with near-maximal to maximal effort and full recoveries. These can also be performed as strides, progressive runs or flying sprints, where the rate of accelera- tion is reduced to allow more total distance at higher velocities. The main aim of the session is to develop or maintain maximal sprinting speed without producing high levels of lactate 1842 T. Haugen et al. the varying physiological and profiles that exist within and between 800- and 1500-m runners (see Sect. 6). 5.4 Variation and Periodization The principle of variation refers to the concept that sys- tematic variation in training is most effective for eliciting long-term adaptations [98, 119]. The most commonly inves- tigated training theory involving planned training variation is periodization, an often-misused term that today refers to any form of training plan, regardless of structure [119]. Ever since Arthur Lydiard introduced his periodization system in the late 1950s [19–21], leading practitioners within middle- distance running typically divide the training year (macro- cycle) into distinct, ordered phases to peak for important competitions [23–26, 28, 31, 32, 34, 36–38, 40, 42, 43, 45, 52, 54–57, 59]. At least three phases are typically organized within a macrocycle: a preparation period, a competition period and a transition period. The transition period begins immediately after the outdoor competition season, typically consisting of 2–4 weeks with rest or recreational training. The following preparation period is typically broken up into general and specific preparation. Some athletes apply dou- ble periodization (i.e., two peaking phases), consisting of a preparation phase, an indoor season, a new preparation phase and finally an outdoor competition season [24, 32, 43]. However, most world-class middle-distance runners apply single periodization. Although they may participate in cross-country or indoor competitions during their prepara- tion phase, such competitions mainly serve as a refreshing change from daily training. The historical development underlying today’s prac- tices for variation and periodization among world-class middle-distance runners is described in Table 3. The train- ing organization models outlined in the 1950s, 1960s and 1970s are still valid, as we and others have systematically quantified the training of successful endurance athletes in a range of sports and reported a “polarized” (i.e., signifi- cant proportions of both high- and low-intensity training and a smaller proportion of threshold training) [122, 123] or pyramidal (i.e., most training is at low intensity, with gradu- ally decreasing proportions of threshold and high-intensity training) intensity distribution [124]. Modern endurance training practice among elite performers in numerous sports [110–112, 125–132] is dominated by frequent sessions and high total volumes of low intensity training combined with smaller volumes of high intensity training organized as 2–4 “key workouts” in most training weeks. This training organization also holds true for well-trained and world- leading middle-distance runners [10, 16, 17, 22–59, 133], although 800-m runners apply a greater proportion of train- ing at higher intensities than 1500-m runners (see Sect. 6.3). We argue that the ubiquitous nature of this basic intensity distribution across sports with very distinct “cultures and training histories” suggests some physiologically rooted self- organizing forces at play related to sustainably balancing cellular signaling and systemic stress over time. However, the long-term and cross-disciplinary influence of ground- breaking coaches cannot be discounted. 6 Training Characteristics 6.1 Training Quantification Considerations While training volume in typical endurance sports can be quantified in a straightforward manner using number of sessions, hours and kilometers, quantification of training intensity is more complicated. In scientific studies of elite endurance athletes, 3- or 5-zone intensity scales have been developed based on either external work rates (running pace or types of training), internal physiological responses (VO2, blood lactate and/or heart rate ranges) or how the training was perceived [62, 110–112, 125–129]. These previously developed scales are not applicable for middle-distance runners because (1) parts of their training are performed at considerably higher intensities, and (2) middle-distance ath- letes exhibit physiological training responses different from aerobic endurance athletes (e.g., higher blood lactate levels). Acknowledged and leading middle-distance practitioners have developed alternative training zone models [17, 54, 56, 58, 59], but no consensus has been established. However, describing and comparing training characteristics requires a common intensity scale. To identify the training differences between 800- and 1500-m runners in more detail, we have developed a 5- and 9-zone intensity model (Table 4) based on an integration of scientific [17, 62, 110–112, 122–129, 134] and best practice coaching literature [54, 56, 58, 59]. Standardized intensity scales can be criticized for several reasons. Firstly, they fail to account for individual varia- tion in the relationship among physiological variables (e.g., between heart rate and blood lactate concentration) [123]. Secondly, the method of training intensity quantification can affect the computation of the training intensity distribu- tion [135]. Thirdly, prescribing exercise intensity based on a fixed percentage of maximal physiological anchors (e.g., VO2max or maximal heart rate) has little merit for eliciting distinct or domain-specific homeostatic perturbations [136]. Finally, running pace can be affected by varying wind and temperature conditions, the rigors of training, “the myster- ies” of the body and day-to-day variation in recovery and readiness to train. Athletes must therefore cultivate an ability to “feel” the proper intensity, as intensity integrates three forms of feedback: running pace, physiological responses and perception of effort [55]. Intensity scales are imperfect tools, but the above-mentioned potential sources of error 1843 Training and development of world-class middle-distance athletes Table 3 An historical overview of middle-distance training organization New paradigms Key coaches and athletes driving the development 1920s Use of systematic methodologies targeting middle-distance running Paavo Nurmi was the pioneer of interval training and introduced the “even pace” strategy to running, using a stopwatch to control his speed [120]. He also developed systematic all-year-round training programs that included both long-distance work and high-intensive running [1], bringing middle- and long-distance training to a new and modern level with intelligent application of effort 1930s Introduction of interval concepts and use of heart rate for intensity control German Waldemar Gerschler (coach of e.g. Harbig and Moens) together with the physiologist Herbert Reindell refined the interval training concept [1]. The intensity in each interval was carefully controlled by heart rate and typically higher than competition pace interspersed by short breaks 1940s Introduction of “fartlek” as a training method Swedish Gösta Holmer (coach of e.g. Hägg and Anderson) developed “fartlek” as a training method [1], an unstructured long-distance run in various terrains where periods of fast running are intermixed with periods of slower running 1950s Use of high-volume low intensity running as a basis of middle-dis- tance running Gradually reduced volume and more competition-specific speed/inten- sity towards the competition period New Zealander Arthur Lydiard (coach of e.g. Snell and Halberg) broke with contemporary practice by prescribing a large volume of low intensity running to his middle-distance athletes, peppered with specific high-intensity training, hill bounding and plyometric training [19–21] The emphasis on high-volume aerobic training shifted towards less vol- ume and more specific anaerobic and race-specific workouts towards the competitive season, which remains the foundation for most mod- ern training programs. This training model bears great resemblance to Matveyev’s traditional training periodization [121] 1960s Systematic micro-periodization of hard and easy workouts Oregon and USA track and field coach Bill Bowerman popularized the hard/easy principle of running; days of hard workouts (e.g., interval training) were systematically alternated with easy days of low-inten- sive running [53] 1970–1980s Introduction of the multi-pace training concept Use of 2–3-day clustering of anaerobic sessions In the 1970s, Frank Horwill, the founder of the British Milers’ Club, formulated and innovated the multi-pace training concept [47]. This system involves training at four or five different combinations of paces and distances in a 10–14-day cycle. The distances are rotated so that over-distance, event-specific and under-distance paces are all covered. Horwill’s training philosophy deviates from Lydiard’s, both in terms of ~ 50% less weekly running volume, as well as larger amounts of anaerobic training throughout most of the macrocycle. This system has been utilized by several world-leading middle-distance athletes, including Sebastian Coe [54, 59], Said Aouita [24], Hicham El Guer- rouj [45], Maria Mutola and Kelly Holmes [32] Another characteristic feature that emerged in British middle-distance running in the 1970s and 1980s was the 2–3-day clustering of anaero- bic sessions (high-intensive intervals, strength, power and plyometric training), followed by 1–2 low-intensive (aerobic) training days [47, 54, 57, 59]. This micro-periodization model involves an alternate tax- ing of the cardiovascular and neuromuscular systems, also described as a reduced form of “crash training”. This philosophy has later been used by several world-leading middle-distance athletes [14, 37] (Table 6) 2000–2010s Introduction of the polarized and pyramidical intensity distribution concepts Several acknowledged scientists systematically quantified the training of successful endurance athletes in a range of sports and reported a “polarized” (i.e., significant proportions of both high- and low-inten- sity training and a smaller proportion of threshold training) [110, 111] or pyramidal (i.e., most training is at low intensity, with gradually decreasing proportions of threshold and high-intensity training) inten- sity distribution [112]. Accordingly, this training organization holds true for most of today’s world-leading middle-distance runners 1844 T. Haugen et al. seem to be outweighed by the improved communication between coach and athlete that a common scale facilitates [123]. The intensity scale outlined here (Table 4) can be used as a framework for both scientists and practitioners involved in middle-distance running. Still, future training studies should aim to verify whether different methods to prescribe training will affect resulting training execution and adaptation. Studies of endurance athletes have employed several methods of intensity distribution quantification. These are either anchored around different running paces, standard- ized blood lactate ranges, “time-in-zone” heart rate analysis based on quantification of the training time spent within dif- ferent heart rate ranges identified from preliminary threshold testing, or the “session goal” approach where each training session is nominally allocated to an intensity zone based on the intensity of the primary part of the workout [62, 122–124]. Based on the nature and characteristics of avail- able best practice training information [19–59], the session goal approach was used in this review to quantify the inten- sity distribution for the analyzed running sessions. 6.2 Training Volume Most world-leading middle-distance runners train about 500–600 h per year, although some 800-m runners may train for less than 400 h [25, 28, 30, 47, 54, 59]. This train- ing volume is 40–70% of what has been reported for suc- cessful endurance athletes in cross-country skiing, biathlon, cycling, triathlon, swimming and rowing [110, 112, 127, 128, 137–143]. This difference is likely explained by the fact that running is a weight-bearing locomotion modal- ity where large muscle groups in the lower limbs perform plyometric actions to overcome the vertical and horizontal ground reaction forces involved [99, 144]. The lower amount of training hours in middle-distance runners than the above- mentioned sports is mainly due to shorter training sessions with higher degree of neuromuscular loading, and not lower training frequency. Both 800- and 1500-m runners perform approximately 500 training sessions per year [25, 28, 30, 54, 59], similar to other elite endurance athletes [62, 111, 112, 127, 128]. After the competitive season, the training volume is substantially decreased in the transition period when mostly alternative activities and easy runs are per- formed. Thereafter, the training volume increases gradually, reaching a maximum in the mid-to-late preparation phase, decreasing again as the competition period approaches. The 30–40% reduction in training hours from late-preparation to competition period is in accordance with world-leading athletes in endurance sports such as orienteering, cross- country skiing and biathlon [111, 112, 127, 128]. However, while most of this reduction is related to a decrease/removal of cross-training in these sports, middle-distance runners Table 4 Intensity scale for elite middle-distance runners BLa typical blood lactate (normative blood lactate concentration values based on red-cell lysed blood), HR typical heart rate, VO2max maximal oxygen consumption, RPE rating of perceived exertion, TTF time to fatigue (single effort), AWD typical accumulated work duration, Int. interval, Rec. typical recovery time (active or passive) between repetitions, prog. progressive, lact. prod. lactate production, lact. tol lactate tolerance, hill rep. hill repeats, AT anaerobic threshold, TT time trials, LS long-sprint, MD middle-distance, LD long-distance, LIT low-intensity training, MIT moderate intensity training, HIT high-intensity training, VHIT very high-intensity training, SST short-sprint training a Warm-up is typically performed in zone 1–3, although with shorter duration, while cool downs are typically performed in zone 1–2 b Progressive runs are typically performed in zone 1–3 Scale BLa HR VO2max RPE TTF Race pace AWD Int. time Rec Training methods 9-zone 5-zone mmol·L−1 % max % 6–20 min min·session−1 min min 9 SST n/a n/a n/a n/a < 0:08 ≤ 60 m < 1 < 0:08 1–3 Accelerations, flying sprints (alactic) 8 SST n/a n/a n/a n/a 0:15 60–120 m 1–3 < 0:15 1–3 Progressive runs or maximal sprints 7 VHIT > 12 n/a 115–140 19–20 1 120–600 m 3–6 0:15–1:30 3–15 Lact. prod. training, TT, LS competitions, hill rep 6 VHIT > 12 n/a 100–114 19–20 4 800–1500 m 6–15 0:25–1:30 0:30–3 Lact. tol. training, TT, MD competitions 5 HIT 8.0–12.0 > 93 90–99 18–20 15 3000–5000 m 15–25 1–4 1–3 VO2max int., LD competitions, hill rep 4 HIT 4.0–8.0 88–92 85–89 16–18 30 10 000 m 20–35 2–7 1–3 VO2max int., hill rep 3 MIT 2.5–4.0 83–87 80–84 14–16 60 Half-marathon 20–50 3–10 1–2 AT runs, fartlek, AT int., prog. runsb 2 LIT 1.5–2.5 73–82 70–79 12–14 120 Marathon 20–90 n/a n/a Long run 1 LIT < 1.5 60–72 55–69 9–12 n/a n/a 20–150a n/a n/a Recovery run, easy long run 1845 Training and development of world-class middle-distance athletes reduce the amount of low-intensity running and strength/ power/plyometric training. Table 5 shows weekly training volume across season peri- ods for world-class middle-distance runners. While 800-m runners typically cover 50–120 km·week−1, 1500-m runners cover 120–170 km·week−1 during the mid-to-late prepara- tion period [10, 16, 17, 22–26, 28–32, 34, 36–41, 43–46, 49–51, 54, 59, 133]. The difference is explained by fewer running kilometers for each session for 800-m athletes, as the rate of training sessions are equal for both disciplines. More specifically, typical “long-run” sessions for 800- and 1500-m runners are in the range of 5–10 and 13–17 km, respectively. Although the best practice coaching literature is limited for female athletes, it is reasonable to assume that the ~ 11% slower running velocity in women is compensated for by less covered distance to ensure the same running duration as for the men. In long-distance running, men and women seem to apply the same training duration [62–65]. Table 5 should therefore be interpreted accordingly. Warm-ups and cool downs in conjunction with interval training and strength/power/plyometric sessions make up a large proportion of the total running volume for 800-m run- ners, while more training sessions for 1500-m athletes are centered around long runs at low to moderate intensity. Inter- estingly, the difference in running volume between 800- and 1500-m runners is larger than the difference between 1500- and long-distance/marathon runners. World-leading 5–10 km athletes run 120–200 km·week−1 [10, 62–64], while top- class marathon runners cover 150–250 km·week−1 [62–64]. Based on these running volume distinctions, one could argue that 1500-m runners in general are more long-distance than middle-distance athletes, although high finishing speed is required in slow races [80]. Running accounts for more than 90% of training hours in 1500-m runners, while the remaining training is typically spent on strength/power (core stability, circuits or light weights), drills, plyometrics and stretching [23, 24, 28, 31, 39, 43–45, 49, 64]. Fewer training sessions (70–80%) are dominated by running in 800-m runners, as they perform a greater amount of strength, power and plyometric training [26, 30–32, 36–38, 40, 50, 51]. 6.3 Intensity Distribution Previous studies have shown that elite endurance athletes seem to converge on a typical intensity distribution in which ~ 80% of annual training sessions are dominated by low-intensive work (< 2 mmol·L−1 blood lactate) and ~ 20% are dominated by training at or above the anaerobic thresh- old (e.g., interval training) [9, 17, 123, 124]. While this intensity distribution for running sessions also seems to apply for world-leading 1500-m athletes [23, 24, 28, 31, 39, 43–45, 49, 64], corresponding 800-m runners seem to follow a 70/30- or 60/40-distribution [26, 30–32, 36–38, 40, 50, 51]. However, although 800-m runners perform intensive training sessions more frequently, total effective interval time/distance remain relatively short due to the high inten- sities with long recovery times between intervals. Hence, approximately 90% of all running sessions for 800-m ath- letes is performed at low intensity based on the time-in-zone approach, in line with endurance sports [111, 112, 123]. Overall, 1500-m runners perform longer and more fre- quent training sessions in zone 1 and 2 (based on our 9-zone Table 5 Weekly training volume for world-class middle- distance runners across the annual cycle Short-sprint training (SST) is not included in this analysis, as this is rarely the main goal for an entire session in middle-distance runners. The numbers are based on scientific [74, 93] and best practice [2–42] literature LIT low-intensity training, MIT moderate intensity training, HIT high-intensity training, VHIT very high- intensity training a Supplementary training (strength, power, plyometric training and stretching) included b 2–4 weekly sessions in total for MIT, HIT and VHIT Variable Early prepara- tion Mid-to-late prepa- ration Pre-competition Mid-compe- tition 800 m 1500 m 800 m 1500 m 800 m 1500 m 800 m 1500 m Weekly training duration (h)a 8–13 9–13 9–15 10–15 9–14 9–14 8–13 8–13 Weekly training sessions (n)a 6–11 8–12 9–12 10–13 8–11 9–12 7–10 8–11 Weekly running volume (km) 40–80 70–120 70–120 120–170 60–100 100–150 50–80 80–140 Weekly running sessions (n) 4–7 8–12 6–10 10–13 6–10 10–12 6–9 10–12 Weekly LIT sessions (n) 3–6 6–9 3–5 8–11 3–5 7–10 2–5 4–8 Weekly MIT sessions (n)b 1–2 1–2 1–2 1–2 0–1 1–2 0–1 1–2 Weekly HIT sessions (n)b 1–3 0–2 1–3 1–3 0–2 1–3 0–2 1–3 Weekly VHIT sessions (n)b 0–1 n/a 1–2 0–2 1–3 0–2 1–3 1–3 1846 T. Haugen et al. scale) than 800-m runners throughout the training year [10, 16, 17, 22–26, 28–32, 34, 36–41, 43–46, 48–51, 54, 59]. Substantial differences are also present for the more inten- sive training sessions. More specifically, 1500-m runners typically follow a pyramidal intensity distribution, while the training pattern in 800-m runners is more clearly polarized. Both groups perform 2–4 weekly intensive training sessions during the preparation phase. These are typically executed in zone 3–5 for 1500-m runners, with a trend towards more zone-3 training (in the form of progressive long runs, anaer- obic threshold runs or interval sessions approximately twice a week) over the last 3–4 decades. The intensive training sessions for 800-m runners during the preparation phase are more evenly distributed across zone 3–6. The differences in the intensive training sessions between 800- and 1500-m runners become even more pronounced when approaching the competition period. During the late- preparation and early-competition period, 800-m runners typically perform 3–4 weekly intensive sessions in zone 3–7 [26, 30–32, 36–38, 40, 50, 51]. Zone-6 intervals are prioritized at the beginning of this period (1–2 weekly ses- sions), and then replaced with training in zone 7. Indeed, lactate tolerance and lactate production training are charac- teristic features for middle-distance athletes (800-m runners in particular), as such training rarely occurs among world- leading sprinters [60] or long-distance runners [61–65]. In contrast, 1500-m runners maintain their zone-3 training with 1–2 weekly sessions during the late-preparation and early- competition period [23, 24, 28, 31, 39, 43–45, 48, 49, 64]. Moreover, preparation-phase training for 1500-m runners in zone 4 and 5 is replaced with 1–2 weekly lactate tolerance training sessions (zone 6) in the late-preparation and early- competition period [23, 24, 28, 31, 39, 43–45, 48, 49]. Middle-distance runners perform short-sprint training (SST; zone 8–9) regularly during the annual cycle, but 800-m runners perform SST to a larger degree than 1500-m runners [22–27, 29–32, 34–54, 57–59]. SST is considered a supplement rather than the main goal of separate train- ing sessions and is typically performed during the last part of the warm-up or after easy long runs. It is generally assumed that sprint training should be performed without accumulation of lactic acid [19–21, 52, 54, 57, 59]. Hence, the distances are most commonly in the range of 60–120 m (zone 8), sometimes even shorter (30–60 m; zone 9), and the time/rest between each repetition is sufficient to ensure full recovery. The sprints are typically performed as strides, progressive runs or flying sprints, where the peak rate of acceleration is reduced to minimize lactate accumulation. The technical aspect of running is also highlighted during SST sessions [37, 41]. A widespread notion among coaches is that MSS is inborn and resistant to training adaptation [19–21, 52, 54, 57, 59], and SST is therefore performed to minimize the downsides of aerobic conditioning on MSS. However, studies have shown that well-trained middle-dis- tance runners can improve MSS [145, 146]. According to best practice literature within sprint training, an intensity of ≥ 90–95% of MSS is required to effectively stimulate adaptation [60]. In summary, world-class 800- and 1500-m runners organize their training quite differently, but with no apparent sex differences in intensity distribution within the disciplines. Table 6 shows case study examples of typ- ical training weeks across the annual cycle for an Olym- pic 800-m champion and a European 1500-m champion. We argue that the training of these two athletes reveals the main distinctions between typical 800- and 1500-m specialists. 6.4 Strength, Power and Plyometric Training A review of the best practice literature reveals that most world-class middle-distance runners perform regular strength, power and plyometric training as a supplement to their specific running conditioning [22–59]. This train- ing is typically executed as a combination of (1) core strength/stability (static or dynamic sit-ups and back exer- cises), (2) strength training with machines or free weights (e.g., half squats, cleans, lunges, step ups, leg press, leg curl, leg extension) without causing significant hyper- trophy, (3) circuit training with body mass resistance, (4) medicine ball exercises, and (5) vertical and horizontal multi-jumps on grass, inclines, stairs (e.g., bounding, skipping, squat jumps, hobbling, springing) or jumping over hurdles. Combinations of running and circuit train- ing exercises have also been applied (e.g., 8–10 exercises with 1 K running in between) [36, 53]. In general, the supplementary training is poorly described in terms of resistance loading, sets and repetitions, and caution must therefore be exercised when drawing conclusions. How- ever, two main features become apparent after reading the best practice literature: more supplementary training is performed during the preparation (typically 2–4 times per week) than competition (0–2 times per week) period, with 800-m runners of both sexes performing such train- ing more frequently than corresponding 1500-m runners. Future studies should aim to concretize more detailed recommendations for middle-distance runners regarding types of exercises, resistance loading, sets and repetitions. Based on experimental evidence, adding supplemen- tary training on 2–3 occasions per week in the form of strength, power and plyometric training appears to improve running economy, time trial performance and MSS in middle- and long-distance runners across a broad performance range [4, 147–149]. In contrast, a causal 1847 Training and development of world-class middle-distance athletes relationship between core stability, athletic performance and injury risk has not been established [150]. 7 Tapering While the training components across the annual cycle of world-class middle-distance runners are described con- siderably more in detail in best practice versus research literature, tapering represents an area where more infor- mation can be obtained from scientific studies. Although potential differences in tapering strategies between 800- and 1500-m runners cannot be identified based on current available information, it is reasonable to assume that the training volume is lower and the key workouts are shorter and more intensive for 800-m runners during this period. The general scientific guidelines for a likely effective taper in endurance-related sports are a 2- to 3-week period incorporating 40–60% reduction in training volume fol- lowing a progressive non-linear format, while training intensity and frequency are maintained or only slightly reduced [151–155]. However, although individual differ- ences are clearly present, tapering length increases with competition distance, and approximately 1 week seems sufficient for middle-distance athletes [33, 56, 156, 157]. Spilsbury and associates reported that elite middle-dis- tance runners perform three interval training sessions on average during the last tapering week [157]. Each of these interval sessions are typically executed at race pace with a total distance of ~ 2 K. This corresponds to ~ 50% of the total interval distance in the preceding weeks of the tapering period. It should be noted that a sub-group bias may have affected the outcomes in this study, as the Brit- ish middle-distance sample included twice the number of 1500-m runners (n = 12) than 800-m runners (n = 6). According to studies of well-trained endurance athletes, a realistic performance goal for the final taper should be a competition performance improvement of about 2–3%, corresponding to 2–4 and 4–6 s in world-leading 800- and 1500-m runners, and this is due to positive changes in the cardiorespiratory, metabolic, hematological, hormo- nal, neuromuscular and psychological status of the ath- letes [151–155]. However, based on annual performance changes in world-leading middle-distance contestants [87], we argue that the performance gains suggested in research literature are likely smaller for athletes of higher standards. 8 Conclusions This review integrated scientific and best practice coach- ing literature regarding the training and development of elite middle-distance performance. To this end, we have outlined a framework for specific characteristics (e.g., training methods, volume and intensity) and identified the training differences between 800- and 1500-m runners in detail. Overall, the training of 800-m athletes consists of considerably lower running volume, a higher propor- tion of interval training at or above the anerobic threshold and more supplementary work in form of strength, power and plyometric training compared to 1500-m runners. These features seem to reflect the divide in physiological demands separating these two middle-distance disciplines. Although there are many studies focusing on middle-dis- tance running, there is a considerable gap between science and best practice in how training principles and methods are applied, highlighting the need for future investigations employing a more holistic approach. For example, training differences and assessment of mechanical and physiologi- cal capacities of elite middle-distance runners through- out the training year and over several seasons should be observed. Such approaches would establish mechanistic connections between training content, changes in perfor- mance and underlying mechanical and physiological deter- minants. The conclusions drawn in this review may serve as a position statement and provide a point of departure for forthcoming studies regarding training and development of elite middle-distance runners. 1848 T. Haugen et al. Table 6 Case study examples of training weeks for an Olympic 800-m champion and European 1500-m champion across the annual cycle Day 800-m champion (Vebjørn Rodal) 1500-m champion (Arturo Casado) Mid-to-late preparation period Mon M: 8 km long run (z1–2) + 20 min drills E: 3–4 km warm-up (z1–2) + stretching + 10 min drills + 5 × 100 m strides (z8) + 8 × 1000 m VO2max intervals (z4), rec. 1 min + 1–2 km cool down (z1) + 4 × 100 m strides (z8) M: 14 km long run (z1) E: 10 km long run (z1) Tue M: Rest E: Warm-up with basketball + stretching + 10 min drills + 5 × 100 m strides (z8) + 3 × (3 × 4) vertical & (3 × 5) horizontal jumps + 20 × 20–75 steps of running and jumping in stairs with walk down rec. and 6 min set- break + core exercises 20 min M: 14 km long run (z1) + Drills + Hurdles technique + 10 × 400 m lactate toler- ance training (z6), rec. 1 min + 1 km cool down (z1) E: 10 km long run (z1) Wed M: 8 km long run (z1–2) + 5 × 100 m strides (z8) + stretching E: 4 km warm-up (z1–2) + 10 min drills + 5 × 80 m strides (z8) + 2 × 10 × 200m lactate tolerance training (z6), rec. 1 min and set-break 6 min + 1–2 km cool down (z1) + stretching M: 3 km warm-up (z1) + Strength training + 14 km long run (z1) + 18 × 100 m hill repeats (z7/8), rec. easy jog back + 3 km cool down (z1) + plyometrics E: Rest Thu M: 3 km warm-up (z1–2) + plyometrics and strength training without weights 2 × 10 exercises (20/20 s work/recovery) E: 4 km warm-up (z1–2) + stretching + 10 min drills + 5 × 100 m strides (z8) + 10 × 1 min VO2max intervals (z5), rec. 1 min + 1 km cool down (z1) + 4 × 100 m strides (z8) M: 4 km warm-up (z1) + 10 × 1000 m VO2max intervals (z4), rec. 1:30 min + 3 km cool down (z1) E: 10 km long run (z1) Fri M: Rest E: 6 km warm-up (z1–2) + stretching + 10 min drills + 5 × 100 m strides (z8) + 3 × 200 m lactate production training (z7) and 3 × 100 m sprint (z8), rec. 4 min and set-break 8 min + 1 km (z1) M: 3 km warm-up (z1) + Strength training + 16 km long run (z1/2) + drills + 6 × 100 m strides (z8) E: Rest Sat M: 3 km warm-up (z1–2) + stretching + 10 min drills + 5 × 100 m strides (z8) + 90 min. explosive weight training E: 12 km progressive run (z1–3) M: 4 km warm-up (z1) + 2 × 6000 m anaerobic threshold intervals (z3), rec. 2 min + 3 km cool down (z1) E: Rest Sun M: 2 h long run (z1) E: Rest M: 12 km long run (z1) E: Rest Weekly total of ~ 110 km (75% LIT, 17% HIT, 4% VHIT and 4% SST) Weekly total of ~ 152 km (81% LIT, 8% MIT, 7% HIT, 3% VHIT and 1% SST) Mon M: 6 km warm-up (z1–2) + 20 min. drills E: 4 km warm-up (z1–2) + 10 min. drills + 5 × 100 m strides (z8) + 2 × 10 × 200m lactate tolerance training (z6), rec. 1 min and set-break 6 min + 1–2 km cool down (z1) M: 14 km long run (z1) + Drills + Hurdles technique + 15 × 200 m lactate toler- ance training (z6) with 100 m easy jog in between (each 200 m in 29 s on average + 1 km cool down (z1) E: 10 km long run (z1) Tue M: 12 km long run (z1) E: Warm-up with basketball + stretching + 10 min drills + 5 × 100 m strides (z8) + 3 × (3 × 4) vertical and (3 × 5) horizontal jumps + 15 × 20–75 steps of running and jumping in stairs with walk down rec. and 6 min. set-break + 20 min. core exercises M: 3 km warm-up (z1) + Strength training + 5 km (z1) + Fartlek (5, 4, 3, 2, 3 and 2 min. running in z3 with easy jog z1–2 in between corresponding to half the repetition time) + 1 km cool down (z1) E: 10 km long run (z1) Pre-competition period Wed M: 6 km warm-up (z1–2) + 4 × 100 m strides (z8) + stretching E: 4 km warm-up (z1–2) + 10 min drills + 6 × 100 m strides (z8) + 4 × 300 m and 4 × 100 m lactate production training (z7), rec. 4 min and set-break 8 min + 1 km cool down (z1) + stretching M: 14 km long run (z1) + 12 × 100 m hill repeats (z7/8), rec. easy jog back + 3 km cool down (z1) + plyometrics E: Rest 1849 Training and development of world-class middle-distance athletes Table 6 (continued) Day 800-m champion (Vebjørn Rodal) 1500-m champion (Arturo Casado) Thu M: 5 km warm-up (z1) + 20 min drills E: 4 km warm-up (z1–2) + 10 min drills + 5 × 100 m strides (z8) + 10 × 1 min VO2max intervals on treadmill (z5), rec. 1 min + 1–2 km cool down (z1) M: 6 km warm-up (z1) + drills + 5 × 1200 m VO2max intervals (z5), rec. 3 min + 1 km cool down (z1) E: 10 km long run (z1) Fri M: Rest E: 3 km warm-up (z1–2) + stretching + 10 min drills + 5 × 100 m strides (z8) + 3 × (150–120–100 m) lactate production training (z7), rec. 3 min and set-break 6 min + 1–2 km cool down (z1) M: 3 km warm-up (z1) + Strength training + 16 km long run (z1) + drills + 3 × 3 × 60 m sprints (z8), rec. walk back and set-break 3 min E: Rest Sat M: 3 km warm-up (z1–2) + plyometrics and strength training without weights 2 × 10 exercises (20/20 s work/recovery) E: 4 km warm-up (z1–2) + 3 × 600 m, 3 × 400 m and 3 × 200 m lactate tolerance training (z6), rec. 4 min and set-break 10 min + 1 km cool down (z1) + stretching M: 4 km warm-up (z1) + 8 km AT run (z3) + 4 km cool down (z1) E: Rest Sun M: 15–20 km long run (z1) on road or treadmill E: Rest M: 14 km long run (z1) E: Rest Weekly total of ~ 100 km (77% LIT, 10% HIT, 10% VHIT and 3% SST) Weekly total of ~ 147 km (82% LIT, 11% MIT, 4% HIT, 2% VHIT and 1% SST) Competition period Mon M: 5 km warm-up (z1–2) + 4 × 100 m strides (z8) + stretching E: 3 km warm-up (z1–2) + 10 min drills + 5 × 100 m strides (z8) + 200–400– 600–600–400–200 m lactate tolerance training (z6) rec. 1–3 min + 1 km cool down (z1) + stretching M: 6 km warm-up (z1) + drills + 3 × 4 × 200 m lactate tolerance training (at 25 s on average; z6), rec 1 min and set-break 3 min + 1 km cool down (z1) E: 6 km recovery run (z1) Tue M: Rest E: 6 km warm-up (z1–2) + 10 min drills + 4 × 4 × 4 vertical and 6 × 30 m hori- zonal jumps, rec. 1 min + 10 × 20–30 jumps in stairs with walk down rec M: 6 km AT run at 3:10 min. per km (z3) + 6 km fartlek [4 × (1 km in 3 min and 500 m in 1:40 min), 18:40 in total (z3)] + 2 km (z1) E: Rest Wed M: 8 km long run (z1–2) + 4 × 100 m strides (z8) E: 3 km warm-up (z1–2) + 10 min drills + 5 × 100 m strides (z8) + 2 × 5 × 200 m lactate tolerance/production training (z6–7), rec. 2 min and set-break 10 min + 1 km cool down (z1) M: 6 km warm-up (z1) + drills + 4 × 1000 m VO2max intervals (2:30 min. on average, z5), rec. 3 min. + 2 km cool down (z1) E: 6 km recovery run (z1) Thu M: Rest E: 3 km warm-up (z1) + 10 min drills + 5 × 100 m strides (z8) + 5 × 100 m near-maximal sprints (z8), 1–2 min rec. + plyometrics and strength training without weights 20 min M: 12 km long run (z1) + Strength training + drills + 6 × 100 m strides (z8) E: Rest Fri M: Rest E: 3 km warm-up (z1–2) + 10 min drills + 4 × 100 m strides + 2 × 400 m lactate production training (z7), rec. 10 min + 1–2 km cool down (z1) M: 6 km warm-up (z1) + drills + 3 × 4 × 300 m lactate tolerance training, each run at 40 s on average (z6), rec. 1 and set-break 3 min + 1 km cool down (z1) E: 6 km recovery run (z1) Sat M: 5 km warm-up + 10 min drills + 4 × 100 m strides (z8) E: 4 km warm-up (z1–3) + drills, strides and speed work + 800 m competi- tion + 1–2 km cool down (z1) M: 15 km long run (z1) + drills + 6 × 100 m strides (z8) E: Rest Sun M: 8 km recovery run (z1) E: Rest M: 6 km warm up (z1) + drills + 8 × 150 m lactate production training with 4 kg ballast (at 16–17 s on average; z7), rec. 3 min. + 1 km cool down (z1) E: 6 km recovery run (z1) Weekly total of ~ 63 km (84% LIT, 10% VHIT and 6% SST) Weekly total of ~ 106 km (77% LIT, 11% MIT, 4% HIT, 7% VHIT and 1% SST) 1850 T. Haugen et al. Acknowledgements The authors want to thank Vebjørn Rodal, Arturo Casado, Arturo Martín-Tagarro, Leif Inge Tjelta and Ørjan Madsen for valuable inputs and contributions during the process. Declarations Funding Open access funding provided by Kristiania University Col- lege. No sources of funding were used to assist in the preparation of this article. Conflict of interest Thomas Haugen, Øyvind Sandbakk, Eystein Enok- sen, Stephen Seiler, and Espen Tønnessen declare that they have no conflicts of interest relevant to the content of this review. Availability of data and materials All data and materials support the published claims and comply with field standards. Code availability Not applicable. Author contributions TH, SS, ØS and ET planned the review. TH and ET retrieved the relevant literature. All authors (TH, SS, ØS, EE and ET) were engaged in drafting and revising the manuscript. All authors read and approved the final version of the manuscript. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. References 1. Quercetani RL. World history of long distance running: 1880– 2002. Track events: men & women. Prairie Striders Library Col- lection 2002. 2. Thibault V, Guillaume M, Berthelot G, Helou NE, Schaal K, Quinquis L, Nassif H, Tafflet M, Escolano S, Hermine O, Tous- saint JF. Women and men in sport performance: the gender gap has not evolved since 1983. J Sports Sci Med. 2010;9:214–23. 3. Bachero-Mena B, Pareja-Blanco F, Rodríguez-Rosell D, Yáñez-García JM, Mora-Custodio R, González-Badillo JJ. Relationships between sprint, jumping and strength abilities, and 800 m performance in male athletes of national and inter- national levels. J Hum Kinet. 2017;58:187–95. 4. Balsalobre-Fernández C, Santos-Concejero J, Grivas GV. Effects of strength training on running economy in highly trained runners: a systematic review with meta-analysis of controlled trials. J Strength Cond Res. 2016;30:2361–8. 5. Brandon LJ. Physiological factors associated with middle dis- tance running performance. Sports Med. 1995;19:268–77. 6. Billat LV. Interval training for performance: a scientific and empirical practice. Special recommendations for middle- and Zone 1 running was performed at ~ 3:40 min·km−1 for Casado and ~ 3:55 min·km−1 for Rodal. Drills for Rodal consisted of 2–3 × 30 m ankle triplings, high knees, butt kicks, straight-leg bounce and other running-specific exercises. Drills for Casado consisted of 2 × 40 m high knees, butt kicks, straight-leg bounce, forward skip, double leg bounce, forward lung and brief sprinting. Hur- dle technique for Casado consisted of four exercises over 6–8 hurdles with 1–2 steps in between. Strength training for Rodal consisted of explosive full and half squats, step-ups for each leg, bench flies, arm-movement simulations and toe lift at 30–70% of one repetition maximum, in addition to running-specific core exercises without weights. Strength training for Casado consisted of 3 × (6 × full squat, dead lift, step-ups for each leg, bench press at 50–70% of one repetition maximum, plus other exercises such as hamstring curls, calf raises and core strength. Stairwork for Rodal consisted of high-frequency running, one-leg and two-leg jumps. Plyometrics for Casado consisted of 2 × 50 m one-leg bounding for each leg and 2 × 50 m two-leg bounding on grass. Vertical plymetrics for Rodal consisted of hurdle jumps, while horizontal plyometrics consisted of one-leg jumps and various steps/”sprunglauf” exercises. The training program for Casado was designed by Arturo Martín-Tagarro M morning session, E evening session, z training zone (see Table 4), AT anaerobic threshold Table 6 (continued) 1851 Training and development of world-class middle-distance athletes long-distance running. Part I: aerobic interval training. Sports Med. 2001;31:13–31. 7. Billat LV. Interval training for performance: a scientific and empirical practice. Special recommendations for middle- and long-distance running. Part II: anaerobic interval training. Sports Med. 2001;31:75–90. 8. Ingham SA, Whyte GP, Pedlar C, Bailey DM, Dunman N, Nevill AM. Determinants of 800-m and 1500-m running per- formance using allometric models. Med Sci Sports Exerc. 2008;40:345–50. 9. Ingham SA, Fudge BW, Pringle JS. Training distribution, physiological profile, and performance for a male international 1500-m runner. Int J Sports Physiol Perform. 2012;7:193–5. 10. Kenneally M, Casado A, Gomez-Ezeiza J, Santos-Concejero J. Training intensity distribution analysis by race pace vs. physi- ological approach in world-class middle- and long-distance runners. Eur J Sport Sci. 2020:1–23 (Online ahead of print). 11. Padilla S, Bourdin M, Barthélémy JC, Lacour JR. Physiological correlates of middle-distance running performance. A com- parative study between men and women. Eur J Appl Physiol Occup Physiol. 1992;65:561–6. 12. Sandford GN, Kilding AE, Ross A, Laursen PB. Maximal sprint speed and the anaerobic speed reserve domain: the untapped tools that differentiate the world’s best male 800 m runners. Sports Med. 2019;49:843–52. 13. Sandford G, Stellingwerff T. Question your categories: the misunderstood complexity of middle-distance running profiles with implications for research methods and application. Front Sports Act Living. 2019;1:28. 14. Sandford GN, Rogers SA, Sharma AP, Kilding AE, Ross A, Laursen PB. Implementing anaerobic speed reserve testing in the field: Validation of vVO2max prediction from 1500-m race performance in elite middle-distance runners. Int J Sports Physiol Perform. 2019;14:1147–50. 15. Sandford GN, Allen SV, Kilding AE, Ross A, Laursen PB. Anaerobic speed reserve: a key component of elite male 800-m running. Int J Sports Physiol Perform. 2019;14:501–8. 16. Tjelta LI. A longitudinal case study of the training of the 2012 European 1500 m track champion. Int J Appl Sports Sci. 2013;25:11–8. 17. Tjelta LI. Three Norwegian brothers all European 1500 m champions: what is the secret? Int J Sport Sci Coach. 2019;14:694–700. 18. Krüger A. Training theory and why Roger Bannister was the first four-minute miler. Sport Hist. 2006;26:305–24. 19. Lydiard A, Gilmour G. Running to the top. Meyer & Meyer; 1997. 20. Lydiard A. Running with Lydiard: greatest running coach of all time. 3rd ed. Meyer & Meyer Sport; 2017. 21. Livingstone K. Healthy intelligent training: the proven principles of Arthur Lydiard. Meyer & Meyer Fachverlag und Buchhandel GmbH; 2010. 22. Watman M. Peter Elliott training program. https:// www. runne rprog ram. com/ produ ct/ peter- ellio tt- train ing- progr am/. Accessed 1 Aug 2020. 23. Kim McDonald training system. https:// www. runne rprog ram. com/ produ ct/ kim- mcdon ald- train ing- system/. Accessed 1 Aug 2020. 24. Said Aouita training program. https:// www. runne rprog ram. com/ produ ct/ said- aouita- train ing- progr am/. Accessed 1 Aug 2020. 25. Lewandowski T. Marcin Lewandowski 800m training program. https:// www. runne rprog ram. com/ produ ct/ marcin- lewan dowski- 800m- train ing- progr am/. Accessed 1 Aug 2020. 26. LaBadie L. Clayton Murphy from 1:54 to 1:42 800 m. https:// www. runne rprog ram. com/ produ ct/ clayt on- murphy- from- 154- to- 142- 800m- by- lee- labad ie/. Accessed 1 Aug 2020. 27. Sandrock M. Running with the legends: training and racing insights from 21 great runners. Hum Kinet. 1996. 28. John Walker training program. https:// www. runne rprog ram. com/ produ ct/ john- walker- train ing- progr am/. Accessed 1 Aug 2020. 29. Jim Ryun training by Bob Timmons. https:// www. runne rprog ram. com/ produ ct/ jim- ryun- train ing- by- bob- timmo ns/. Accessed 1 Aug 2020. 30. Nick Symmonds 2012 Olympic training log. https:// www. runne rprog ram. com/ produ ct/ nick- symmo nds- 2012- olymp ic- train ing- log/. Accessed 1 Aug 2020. 31. Genzebe Dibaba training by Jama Aden. https:// www. runne rprog ram. com/ produ ct/ genze be- dibaba- train ing- jama- aden/. Accessed 1 Aug 2020. 32. Margo Jennings training philosophy. https:// www. runne rprog ram. com/ produ ct/ margo- jenni ngs- train ing- philo sophy/. Accessed 1 Aug 2020. 33. Joe Vigil training philosophy. https:// www. runne rprog ram. com/ produ ct/ joe- vigil- train ing- philo sophy/. Accessed 1 Aug 2020. 34. Steve Ovett training by Harry Wilson. https:// www. runne rprog ram. com/ produ ct/ harry- wilson- train ing- philo sophy/. Accessed 1 Aug 2020. 35. Vin Lananna training middle and long distance. https:// www. runne rprog ram. com/ produ ct/ vin- lanan na- train ing- middle- long- dista nce/. Accessed 1 Aug 2020. 36. Joaquim Cruz training by Luiz De Oliveira. https:// www. runne rprog ram. com/ produ ct/ luiz- de- olive ira- train ing- system- 2/. Accessed 1 Aug 2020. 37. Gianni Ghidini training philosophy. https:// www. runne rprog ram. com/ produ ct/ gianni- ghidi ni- train ing- philo sophy/. Accessed 1 Aug 2020. 38. Alberto Juantorena Training Log (1971–1976). https:// www. runne rprog ram. com/ produ ct/ alber to- juant orena- train ing- log- 1971- 1972/. Accessed 1 Aug 2020. 39. Martin Liquori conditioning of elite runners. https:// www. runne rprog ram. com/ produ ct/ martin- liquo ri- condi tioni ng- of- elite- runne rs/. Accessed 1 Aug 2020. 40. Michael Rimmer training program. https:// www. runne rprog ram. com/ produ ct/ micha el- rimmer- train ing- progr am/. Accessed 1 Aug 2020. 41. Sifan Hassan training program. https:// www. runne rprog ram. com/ produ ct/ sifan- hassan- train ing- by- honore- hoedt/. Accessed 1 Aug 2020. 42. Steve Magness training philosophy. https:// www. runne rprog ram. com/ produ ct/ steve- magne ss- train ing- philo sophy/. Accessed 1 Aug 2020. 43. Natalia Rodriguez training program. https:// www. runne rprog ram. com/ produ ct/ natal ia- rodri guez- train ing- progr am/. Accessed 1 Aug 2020. 44. Nick Willis training—the road to Athens. https:// www. runne rprog ram. com/ produ ct/ nick- willis- train ing- road- athens/. Accessed 1 Aug 2020. 45. The Moroccan athletics training system – El Guerrouj case. https:// www. runne rprog ram. com/ produ ct/ moroc can- athle tics- train ing- system- el- guerr ouj- case/. Accessed 1 Aug 2020. 46. Taoufik Makhloufi training program. https:// www. runne rprog ram. com/ produ ct/ taoufi k- makhl oufi- train ing- progr am/. Accessed 1 Aug 2020. 47. Frank Horwill training philosophy. https:// www. runne rprog ram. com/ produ ct/ frank- horwi ll- train ing- philo sophy/. Accessed 1 Aug 2020. 48. Nic Bideau training philosophy. https:// www. runne rprog ram. com/ produ ct/ nic- bideau- train ing- philo sophy/. Accessed 1 Aug 2020. 1852 T. Haugen et al. 49. Silas Kiplagat training program by Renato Canova. https:// www. runne rprog ram. com/ produ ct/ silas- kipla gat- train ing- progr am- renato- canova/. Accessed 1 Aug 2020. 50. David Rudisha training program. https:// www. sweat elite. co/ david- rudis ha- train ing- progr am/. Accessed 1 Aug 2020. 51. Coach Colm and the training of David Rudisha. https:// www. runne rstri be. com/ expert- advice/ coach- colm- and- the- train ing- of- david- rudis ha/. Accessed 1 Aug 2020. 52. Sunderland D. High performance middle distance running. The Crowood Press Ltd; 2005. 53. Bowerman B. High performance training for track and field. Champaign: Leisure Press; 1991. 54. Coe P. Winning running: successful 800 m & 1500 m racing and training. Crowood Press, 1996. 55. Davis J. Modern training and physiology for middle and long- distance runners. Runn Writ. 2013. 56. Vigil JI. Road to the top: a systematic approach to training distance runners. 1st ed. Morning Star Communications; 1995. 57. Watts D, Wilson H, Horwill F. The complete middle distance runner. 4th ed. David & Charles, 1987. 58. Daniels J. Daniel’s running formula. 3rd ed. Human Kinetics; 2013. 59. Martin DE, Coe PN. Better training for distance runners. 2nd ed. Human Kinetics Publishers; 1997. 60. Haugen T, Seiler S, Sandbakk Ø, Tønnessen E. The training and development of elite sprint performance: an integration of scientific and best practice literature. Sports Med Open. 2019;5:44. 61. Casado A, Hanley B, Santos-Concejero J, Ruiz-Pérez LM. World-class long-distance running performances are best pre- dicted by volume of easy runs and deliberate practice of short- interval and tempo runs. J Strength Cond Res. 2019 (Online ahead of print). 62. Billat VL, Demarle A, Slawinski J, Paiva M, Koralsztein JP. Physical and training characteristics of top-class marathon run- ners. Med Sci Sports Exerc. 2001;33:2089–97. 63. Billat V, Lepretre PM, Heugas AM, Laurence MH, Salim D, Koralsztein JP. Training and bioenergetic characteristics in elite male and female Kenyan runners. Med Sci Sports Exerc. 2003;35:297–304. 64. Tjelta LI. The training of international level distance runners. Int J Sports Sci Coach. 2016;11:122–34. 65. Stellingwerff T. Case study: Nutrition and training periodization in three elite marathon runners. Int J Sport Nutr Exerc Metab. 2012;22:392–400. 66. Haugen T. Key success factors for merging sport science and best practice. Int J Sports Physiol Perform. 2019 (Online ahead of print). 67. Rabadán M, Díaz V, Calderón FJ, Benito PJ, Peinado AB, Maf- fulli N. Physiological determinants of specialty of elite middle- and long-distance runners. J Sports Sci. 2011;29:975–82. 68. de Koning JJ, Noordhof DA, Uitslag TP, Galiart RE, Dodge C, Foster C. An approach to estimating gross efficiency dur- ing high-intensity exercise. Int J Sports Physiol Perform. 2013;8:682–4. 69. Noordhof DA, Mulder RC, Malterer KR, Foster C, de Koning JJ. The decline in gross efficiency in relation to cycling time- trial length. Int J Sports Physiol Perform. 2015;10:64–70. 70. Boileau RA, Mayhew JL, Riner WF, Lussier L. Physiological characteristics of elite middle and long distance runners. Can J Appl Sports Sci. 1982;7:167–72. 71. Arrese AL, Izquierdo DM, Urdiales DM. A review of the maxi- mal oxygen uptake values necessary for different running per- formance levels. New Stud Athl. 2005;20:7–20. 72. Sandbakk Ø, Solli GS, Holmberg HC. Sex differences in world-record performance: the influence of sport discipline and competition duration. Int J Sports Physiol Perform. 2018;13:2–8. 73. Haugen T, Paulsen G, Seiler S, Sandbakk Ø. New records in human power. Invited review. Int J Sports Physiol Perform. 2018;13:678–86. 74. Haugen T, McGhie D, Sandbakk Ø, Ettema G. Sprint running: from fundamental mechanics to practice—a review. Eur J Appl Physiol. 2019;119:1273–87. 75. Slawinski J, Termoz N, Rabita G, Guilhem G, Dorel S, Morin JB, Samozino P. How 100-m event analyses improve our understanding of world-class men’s and women’s sprint per- formance. Scand J Med Sci Sports. 2017;27:45–54. 76. Haugen T, Breitschädel F, Seiler S. Sprint mechanical variables in elite athletes: are force-velocity profiles sport specific or individual? PLoS ONE. 2019;14:e0215551. 77. Mero A, Komi PV, Gregor RJ. Biomechanics of sprint running. A review. Sports Med. 1992;13:376–92. 78. Haugen T, Danielsen J, Alnes LO, McGhie D, Sandbakk O, Ettema G. On the importance of “front-side mechanics” in ath- letics sprinting. Int J Sports Physiol Perform. 2018;13:420–7. 79. Casado A, Hanley B, Jiménez-Reyes P, Renfree A. Pacing pro- files and tactical behaviors of elite runners. J Sport Health Sci. 2020;S2095–2546(20):30077–86. 80. Bellinger P, Derave W, Lievens E, Kennedy B, Arnold B, Rice H, Minahan C. Determinants of last lap speed in paced and maximal 1500-m time trials. Eur J Appl Physiol. 2020 (Online ahead of print). 81. Blondel N, Berthoin S, Billat V, Lensel G. Relationship between run times to exhaustion at 90, 100, 120, and 140% of vVO2max and velocity expressed relatively to critical velocity and maximal velocity. Int J Sports Med. 2001;22:27–33. 82. Haugen T, Buchheit M. Sprint running performance monitor- ing: methodological and practical considerations. Sports Med. 2016;46:641–56. 83. Haugen T, Breitschädel F, Samozino P. Power-force-velocity profiling of sprinting athletes: methodological and practical considerations when using timing gates. J Strength Cond Res. 2018;119:465–73. 84. Malina RM, Bouchard C, Bar-Or O. Growth, maturation and physical activity. 2nd ed. Champaign: Human Kinetics; 2004. 85. Ganse B, Ganse U, Dahl J, Gegens H. Linear decrease in ath- letic performance during the human life span. Front Physiol. 2018;9:1100. 86. Tønnessen E, Svendsen I, Olsen IC, Guttormsen A, Haugen T. Performance development in adolescent track and field ath- letes according to age, sex and sport discipline. PLoS ONE. 2015;10:e0129014. 87. Haugen T, Solberg PA, Morán-Navarro R, Breitschädel F, Hop- kins W, Foster C. Peak age and performance progression in world-class track-and-field athletes. Int J Sports Physiol Per- form. 2018;13:1122–9. 88. Hollings SC, Hopkins WG, Hume PA. Age at peak performance of successful track and field athletes. Int J Sports Sci Coach. 2014;9:651–62. 89. Allen SV, Hopkins WG. Age of peak competitive performance of elite athletes: a systematic review. Sports Med. 2015;45:1431–41. 90. Berthelot G, Len S, Hellard P, Tafflet M, Guillaume M, Vollmer JC, Gager B, Quinquis L, Marc A, Toussaint JF. Exponential growth combined with exponential decline explains lifetime performance evolution in individual and human species. Age. 2012;34:1001–9. 91. Larsen HB, Sheel AW. The Kenyan runners. Scand J Med Sci Sports. 2015;25:110–8. 92. Noble TJ, Chapman RF. Elite African marathoners specialize earlier than elite non-African marathoners. Int J Sports Physiol Perform. 2017;10:1–19. 1853 Training and development of world-class middle-distance athletes 93. Scott RA, Georgiades E, Wilson RH, Goodwin WH, Wolde B, Pitsiladis YP. Demographic characteristics of elite Ethiopian endurance runners. Med Sci Sports Exerc. 2003;35:1727–32. 94. Ford PR, Ward P, Hodges NJ, Williams AM. The role of delib- erate practice and play in career progression in sport: the early engagement hypothesis. High Abil Stud. 2009;20:65–75. 95. Hopkins WG. Competitive performance of elite track and field athletes. Variability and smallest worthwhile enhancements. Sportscience. 2005;9:17–20. 96. Harre D. Trainingslehre: einführung in die allgemeine Trainings- methodik. Berlin: Sportverlag; 1973. 97. Yakovlev NN. Sportbiochemie. Sportmedizinische Schriftenreihe Nr. 14. Barth, Leipzig; 1977. 98. Kraemer WJ, Adams K, Cafarelli E, Dudley GA, Dooly C, Fei- genbaum MS, et al. American College of Sports Medicine posi- tion stand. Progression models in resistance training for healthy adults. Med Sci Sports Exerc. 2002;34:364–80. 99. Paquette MR, Napier C, Willy RW, Stellingwerff T. Moving beyond weekly “distance”: optimizing quantification of training load in runners. J Orthop Sports Phys Ther. 2020;50:564–9. 100. Renfree A. Performance data: less is more? Sport Exerc Sci. 2000;66:20–1. 101. Bertelsen ML, Hulme A, Petersen J, Brund RK, Sørensen H, Finch CF, Parner ET, Nielsen RO. A framework for the eti- ology of running-related injuries. Scand J Med Sci Sports. 2017;27:1170–80. 102. Videbaek S, Bueno AM, Nielsen RO, Rasmussen S. Incidence of running-related injuries per 1000 h of running in different types of runners: a systematic review and meta-analysis. Sports Med. 2015;45:1017–26. 103. Mujika I, Sharma AP, Stellingwerff T. Contemporary periodiza- tion of altitude training for elite endurance athletes: a narrative review. Sports Med. 2019;49:1651–69. 104. Flaherty G, O’Connor R, Johnston N. Altitude training for elite endurance athletes: a review for the travel medicine practitioner. Travel Med Infect Dis. 2016;14:200–11. 105. Fudge BW, Pringle JS, Maxwell NS, Turner G, Ingham SA, Jones AM. Altitude training for elite endurance performance: a 2012 update. Curr Sports Med Rep. 2012;11:148–54. 106. Sale D, MacDougall D. Specificity in strength training: a review for the coach and athlete. Can J Appl Sport Sci. 1981;6:87–92. 107. Buchheit M, Laursen PB. High-intensity interval training, solutions to the programming puzzle. Part I: cardiopulmonary emphasis. Sports Med. 2013;43:313–38. 108. Buchheit M, Laursen PB. High-intensity interval training, solu- tions to the programming puzzle. Part II: Anaerobic energy, neuromuscular load and practical applications. Sports Med. 2013;43:927–54. 109. Barnes KR, Kilding AE. Strategies to improve running economy. Sports Med. 2015;45:37–56. 110. Fiskerstrand A, Seiler KS. Training and performance character- istics among Norwegian international rowers 1970–2001. Scand J Med Sci Sports. 2004;14:303–10. 111. Tønnessen E, Svendsen I, Rønnestad B, Hisdal J, Haugen T, Seiler S. The annual training periodization of 8 World Champi- ons in orienteering. Int J Sports Physiol Perform. 2015;10:29–38. 112. Tønnessen E, Sylta Ø, Haugen T, Hem E, Svendsen I, Seiler S. The road to gold: training and peaking characteristics in the year prior to a gold medal endurance performance. PLoS ONE. 2014;9:e101796. 113. Loy SF, Hoffmann JJ, Holland GJ. Benefits and practical use of cross-training in sports. Sports Med. 1995;19:1–8. 114. Foster C. Monitoring training in athletes with reference to over- training syndrome. Med Sci Sports Exerc. 1998;30:1164–8. 115. Gaskill SE, Serfass RC, Bacharach DW, Kelly JM. Responses to training in cross-country skiers. Med Sci Sports Exerc. 1999;31:1211–7. 116. Vollaard NB, Constantin-Teodosiu D, Fredriksson K, Rooyackers O, Jansson E, Greenhaff PL, et al. Systematic analysis of adap- tations in aerobic capacity and submaximal energy metabolism provides a unique insight into determinants of human aerobic performance. J Appl Physiol. 2009;106:1479–86. 117. Sylta Ø, Tønnessen E, Hammarstrøm D, Danielsen J, Skovereng K, Ravn T, et al. The effect of different high-intensity periodiza- tion models on endurance adaptations. Med Sci Sports Exerc. 2016;48:2165–74. 118. Bellinger P, Desbrow B, Derave W, Lievens E, Irwin C, Saba- pathy S, Kennedy B, Craven J, Pennell E, Rice H, Minahan C. Muscle fiber typology is associated with the incidence of overreaching in response to overload training. J Appl Physiol. 2020;129:823–36. 119. Kiely J. Periodization paradigms in the 21st century: evidence-led or tradition-driven? Int J Sports Physiol Perform. 2012;7:242–50. 120. Herzog W. Running slow or running fast; that is the question: the merits of high-intensity interval training. J Sport Health Sci. 2017;6(1):48. 121. Matveyev LP. Periodisierung des sportlichen Trainings. 2nd ed. Berlin: Bartels & Wernitz; 1975. 122. Seiler S, Tønnessen E. Intervals, thresholds, and long slow dis- tance: the role of intensity and duration in endurance training. Sportscience. 2009;13:32–53. 123. Seiler S. What is best practice for training intensity and duration distribution in endurance athletes? Int J Sports Physiol Perform. 2010;5:276–91. 124. Stöggl TL, Sperlich B. The training intensity distribution among well-trained and elite endurance athletes. Front Physiol. 2015;6:295. 125. Hellard P, Avalos-Fernandes M, Lefort G, Pla R, Mujika I, Toussaint JF, Pyne DB. Elite swimmers’ training patterns in the 25 weeks prior to their season’s best performances: insights into periodization from a 20-years cohort. Front Physiol. 2019;10:363. 126. Mujika I, Chatard JC, Busso T, Geyssant A, Barale F, Lacoste L. Effects of training on performance in competitive swimming. Can J Appl Physiol. 1995;20:395–406. 127. Solli GS, Tønnessen E, Sandbakk Ø. The training characteristics of the world’s most successful female cross-country skier. Front Physiol. 2017;8:1069. 128. Sandbakk Ø, Holmberg HC. Physiological capacity and train- ing routines of elite cross-country skiers: approaching the upper limits of human endurance. Int J Sports Physiol Perform. 2017;12:1003–11. 129. Orie J, Hofman N, de Koning JJ, Foster C. Thirty-eight years of training distribution in Olympic speed skaters. Int J Sports Physiol Perform. 2014;9:93–9. 130. Seiler KS, Kjerland GØ. Quantifying training intensity distribu- tion in elite endurance athletes: is there evidence for an “optimal” distribution? Scand J Med Sci Sport. 2006;16:49–56. 131. Gullich A, Seiler S, Emrich E. Training methods and intensity distribution of young world class rowers. Int J Sports Physiol Perform. 2009;4:448–60. 132. Gullich A, Seiler S. Lactate profile changes in relation to train- ing characteristics in elite cyclists. Int J Sports Physiol Perform. 2010;5:316–27. 133. Kenneally M, Casado A, Santos-Concejero J. The effect of perio- dization and training intensity distribution on middle- and long- distance running performance: a systematic review. Int J Sports Physiol Perform. 2018;13:1114–21. 1854 T. Haugen et al. 134. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc. 1982;14:377–81. 135. Bellinger P, Arnold B, Minahan C. Quantifying the training- intensity distribution in middle-distance runners: the influence of different methods of training-intensity quantification. Int J Sports Physiol Perform. 2019 (Online ahead of print). 136. Jamnick NA, Pettitt RW, Granata C, Pyne DB, Bishop DJ. An examination and critique of current methods to determine exer- cise intensity. Sports Med. 2020;50:1729–56. 137. Schumacher OY, Mueller P. The 4000-m team pursuit cycling world record: theoretical and practical aspects. Med Sci Sports Exerc. 2002;34:1029–36. 138. Zapico AG, Calderon FJ, Benito PJ, Gonzalez CB, Parisi A, Pigozzi F, Di Salvo V. Evolution of physiological and haemato- logical parameters with training load in elite male road cyclists: a longitudinal study. J Sports Med Phys Fitness. 2007;47:191–6. 139. Gao J. A study on pre-game training characteristics of Chinese elite swimmers. J Beijing Sport Univ. 2008;31:832–4. 140. Siewierski M. Volume and structure of training loads of top swimmers in direct starting preparation phase for main competi- tion. Pol J Sport Tour. 2010;17:227–32. 141. Guellich A, Seiler S, Emrich E. Training methods and intensity distribution of young world-class rowers. Int J Sports Physiol Perform. 2009;4:448–60. 142. Neal CM, Hunter AM, Galloway SD. A 6-month analysis of training intensity distribution and physiological adaptation in Ironman triathletes. J Sports Sci. 2009;29:1515–23. 143. Mujika I. Olympic preparation of a world-class female triathlete. Int J Sports Physiol Perform. 2014;9:727–31. 144. Sandbakk Ø, Haugen T, Ettema G. The influence of exercise modality on training load management. Int J Sports Physiol Per- form. 2021 (in press). 145. Maćkała K, Jozwiak Ł, Stodółka J. Effects of explosive type strength training on selected physical and technical performance characteristics in middle distance running—a case report. Polish J Sport Tour. 2015;21:228–33. 146. Bachero-Mena B, Pareja-Blanco F, González-Badillo JJ. Enhanced strength and sprint levels, and changes in blood parameters during a complete athletics season in 800 m high- level athletes. Front Physiol. 2017;8:637. 147. Rønnestad BR, Mujika I. Optimizing strength training for run- ning and cycling endurance performance: a review. Scand J Med Sci Sports. 2014;24:603–12. 148. Blagrove RC, Howatson G, Hayes PR. Effects of strength train- ing on the physiological determinants of middle- and long-dis- tance running performance: a systematic review. Sports Med. 2018;48:1117–49. 149. Berryman N, Mujika I, Arvisais D, Roubeix M, Binet C, Bosquet L. Strength training for middle- and long-distance performance: a meta-analysis. Int J Sports Physiol Perform. 2018;13:57–63. 150. Haugen T, Haugvad L, Røstad V. Effects of core-stability training on performance and injuries in competitive athletes. Sportsci- ence. 2016;20:1–7. 151. Mujika I, Padilla S. Scientific bases for precompetition tapering strategies. Med Sci Sports Exerc. 2003;35:1182–7. 152. Pyne DB, Mujika I, Reilly T. Peaking for optimal perfor- mance: research limitations and future directions. J Sports Sci. 2009;27:195–202. 153. Mujika I. The influence of training characteristics and tapering on the adaptation in highly trained individuals: a review. Int J Sports Med. 1998;19:439–46. 154. Mujika I. Intense training: the key to optimal performance before and during the taper. Scand J Med Sci Sports. 2010;20:24–31. 155. Bosquet L, Montpetit J, Arvisais D, Mujika I. Effects of taper- ing on performance: a meta-analysis. Med Sci Sports Exerc. 2007;39:1358–65. 156. Mujika I, Goya A, Ruiz E, Grijalba A, Santisteban J, Padilla S. Physiological and performance responses to a 6-day taper in middle-distance runners: influence of training frequency. Int J Sports Med. 2002;23:367–73. 157. Spilsbury KL, Fudge BW, Ingham SA, Faulkner SH, Nimmo MA. Tapering strategies in elite British endurance runners. Eur J Sport Sci. 2015;15:367–73.

Crossing the Golden Training Divide: The Science and Practice of Training World-Class 800- and 1500-m Runners.

05-21-2021

Haugen, Thomas,Sandbakk, Øyvind,Enoksen, Eystein,Seiler, Stephen,Tønnessen, Espen

eng

PMC3407019

COMMENTARY Open Access Ultramarathon is an outstanding model for the study of adaptive responses to extreme load and stress Grégoire P Millet1* and Guillaume Y Millet2 Abstract Ultramarathons comprise any sporting event involving running longer than the traditional marathon length of 42.195 km (26.2 miles). Studies on ultramarathon participants can investigate the acute consequences of ultra- endurance exercise on inflammation and cardiovascular or renal consequences, as well as endocrine/energetic aspects, and examine the tissue recovery process over several days of extreme physical load. In a study published in BMC Medicine, Schütz et al. followed 44 ultramarathon runners over 4,487 km from South Italy to North Cape, Norway (the Trans Europe Foot Race 2009) and recorded daily sets of data from magnetic resonance imaging, psychometric, body composition and biological measurements. The findings will allow us to better understand the timecourse of degeneration/regeneration of some lower leg tissues such as knee joint cartilage, to differentiate running-induced from age-induced pathologies (for example, retropatelar arthritis) and finally to assess the interindividual susceptibility to injuries. Moreover, it will also provide new information about the complex interplay between cerebral adaptations/alterations and hormonal influences resulting from endurance exercise and provide data on the dose-response relationship between exercise and brain structure/function. Overall, this study represents a unique attempt to investigate the limits of the adaptive response of human bodies. Please see related article: http://www.biomedcentral.com/1741-7015/10/78 Keywords: Cerebral adaptations, extreme environment, overload pathologies, ultra-endurance Background While the industrialized world adopts a sedentary life- style, ultramarathon running races have become increas- ingly popular in the last few years, notably in the US, Europe, Japan, Korea, and South Africa. The ability to run long distances is also considered to have played a role in human evolution [1]. This makes the issue of ultra-long distance physiology relevant. Ultramarathons are basically either performed on mostly flat roads or tracks, or run on varied terrain trails. They comprise races that are completed over the space of multiple days (for example, 6 days), with the winner being the one that covers the most distance within this set period of time or races that cover a specified distance during a single stage, which normally range from 50 km to 160.9 km or over several stages. The paper by Schütz et al. [2] explores the physiological changes that occur in runners during this latter type of events, probably one of the most demand- ing physical exercise in humans, maybe only overpassed by polar expeditions (for example, Scott’s party man- hauling their sleds across the Antarctic for 159 days in 1911/1912, [3]) or other individual challenges such as the run around Europe in 2009/2010, that is, 27,012 km in 1 year (74 km/day), by the Frenchman Serge Girard. We recently reviewed the origin of muscle fatigue after pro- longed exercises lasting from 30 minutes to several hours [4] and found that the knee extensors isometric strength loss increased in a non-linear way with exercise duration, that is, there was a plateau after approximately 20 h of running. Recent findings from our group confirm this result for an even longer mountain ultramarathon (Tor des Geants, 330 km; unpublished results). These fatigue studies and other experiments conducted on inflamma- tion [5], cardiovascular or renal consequences [6,7], * Correspondence: [email protected] 1ISSUL Institute of Sport Sciences, Department of Physiology, Faculty of Biology and Medicine, University of Lausanne, Switzerland Full list of author information is available at the end of the article Millet and Millet BMC Medicine 2012, 10:77 http://www.biomedcentral.com/1741-7015/10/77 © 2012 Millet and Millet; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. endocrine/energetic aspects (see for example [8]) help to understand acute consequences of an ultra-endurance exercise (generally shorter than 24 h, more rarely 2 or 6 days) but do allow examining recovery process, that is, tissue degeneration/regeneration over several days of extreme physical load. Trans Europe Foot Race: studying the limits of human endurance In an observational cohort study on 44 ultramarathon runners over 4,487 km in 64 stages from South Italy to North Cape, Norway (the Trans Europe Foot Race 2009), Schütz et al. [2] recorded daily sets of data from magnetic resonance imaging (MRI), psychometric, body composition and biological measurements. Beyond the logistical achievement of following the runners and moving a 30-m, 45-tonne 1.5 Tesla whole-body MRI across Europe (!), they succeeded with a high rate of test completion and data collection. This ‘field’ experiment is unique since it is impossible to expect subjects pushing to (and sometimes beyond) their limits for 64 days without any day of rest in a laboratory setting. Such commitment can be achieved only in an official competition and is absolutely necessary for exploring the adaptive responses in healthy subjects at the limit of stress. In our view, in the field of sports medicine, the longitu- dinal design of this study will allow us to better under- stand the time course of degeneration/regeneration of some lower leg tissues as knee joint cartilage or ventral tibial periosteum, to describe the adaptive responses (for example, red bone marrow hyperplasia), to differentiate running-induced from age-induced pathologies (for example, retropatelar arthritis), to understand why some painful reactions (for example, ‘shin splints’) can be ‘over-run’ whereas others lead to severe injuries (for example, stress fracture) and finally to assess the interin- dividual susceptibility to injuries. This study will also bring new information about the complex interplay between cerebral adaptations and hor- monal influences resulting from endurance exercise. To date, it is known that moderate exercise is beneficial to brain heath (for example, increased perfusion or increased brain-derived neurotrophic factor (BDNF)) [9,10]. But the potential deleterious effects (for example, atrophy, ischemia, brain lesions) of extreme loads on brain volume, plasticity and functionality are unknown. In our opinion, these data are paramount for better understanding the dose-response relationship between exercise and brain structure/function. We have shown that central fatigue was a major issue in long-distance running exercise (see, for example, [11]) yet, to the best of our knowledge, no studies have really assess cerebral alteration related to this type of exercise. This is because the observed decrease in voluntary activation does not mean that cortical alterations really occur, since periph- eral changes, that is, the combination of influences including excitatory and inhibitory reflex inputs from muscles, joints, tendons and cutaneous afferents, may inhibit central drive at the spinal and supraspinal levels. Also of interest is the investigation into pain perception and the possibility to describe interindividual differences in mechanisms of coping. Hormonal mechanisms (for example, cortisol) and neurotransmitters (for example, tryptophan, serotonin) are known to modulate the pain perception [12]. But most previous studies were limited to a single pain stimulus, whereas in the study of Schütz et al. [2] the stimuli are different among subjects and also fluctuating. The possibility of crossvalidation between the MRI, the psychometric and the biological results is pro- mising for better describing the time course of factors influencing the fluctuation of pain throughout the race. Future directions and conclusions This study provides the opportunity to explore the adaptive responses of humans submitted to the extreme load and stress induced by a 4,487-km road race. The methods used will allow investigation into various sub- systems and their interaction in terms of tissue degen- eration/regeneration, pain coping or cerebral adaptations. Future research directions can combine additional techniques such as transcranial magnetic sti- mulation (to assess cortical excitability and inhibition and supraspinal voluntary activation), cerebral multi- channel near-infrared spectroscopy (to measure tissue hemodynamics), and electroencephalography or cerebral MRI, the latter in particular to assess long term cerebral alterations as in Schütz et al. [2]. It is uncertain if/how the findings in Schütz et al. paper can be translated to the fields of pathophysiology or critical illness, since the stress induced by the run- ning load is highly specific. However, one may assume that some of the scientific knowledge accumulated will help in better understanding of adaptive responses. Since it is a road stage race, it is likely that the adap- tive responses to fatigue would be largely different in other environments/conditions such as high altitude, heat, mountainous competition or sleep deprivation. The exploration of exercising in such ‘extreme environ- ments’ that often cannot be performed in a laboratory is an extending field of sports physiology or sports medi- cine. Together with large epidemiological surveys on ultramarathon runners that still have to be conducted, the amazing experiment of Schütz et al. [2] represents a unique attempt to investigate the limits of adaptive response of human bodies. Millet and Millet BMC Medicine 2012, 10:77 http://www.biomedcentral.com/1741-7015/10/77 Page 2 of 3 Author details 1ISSUL Institute of Sport Sciences, Department of Physiology, Faculty of Biology and Medicine, University of Lausanne, Switzerland. 2Université de Lyon, F-42023, Saint Etienne, France. Authors’ contributions GPM and GYM drafted the manuscript and gave final approval for publication. Competing interests The authors declare that they have no competing interests. Received: 10 July 2012 Accepted: 19 July 2012 Published: 19 July 2012 References 1. Bramble DM, Lieberman DE: Endurance running and the evolution of Homo. Nature 2004, 432:345-352. 2. Schütz UHW, Schmidt-Trucksäss A, Knechtle B, Machann J, Wiedelbach H, Ehrhardt M, Freund W, Stefan Gröninger S, Horst Brunner H, Schulze I, et al: The Transeurope Footrace Project: Longitudinal data acquisition in a cluster randomized mobile MRI observational cohort study on 44 endurance runners at a 64-stage 4,486 km transcontinental ultramarathon. BMC Med 2012. 3. Noakes TD: The limits of endurance exercise. Basic Res Cardiol 2006, 101:408-417. 4. Millet GY: Can neuromuscular fatigue explain running strategies and performance in ultra-marathons?: the flush model. Sports Med 2011, 41:489-506. 5. Fallon KE: The acute phase response and exercise: the ultramarathon as prototype exercise. Clin J Sport Med 2001, 11:38-43. 6. Irving RA, Noakes TD, Burger SC, Myburgh KH, Querido D, van Zyl Smit R: Plasma volume and renal function during and after ultramarathon running. Med Sci Sports Exerc 1990, 22:581-587. 7. Scott JM, Esch BT, Shave R, Warburton DE, Gaze D, George K: Cardiovascular consequences of completing a 160-km ultramarathon. Med Sci Sports Exerc 2009, 41:26-34. 8. Sahlin K, Shabalina IG, Mattsson CM, Bakkman L, Fernstrom M, Rozhdestvenskaya Z, Enqvist JK, Nedergaard J, Ekblom B, Tonkonogi M: Ultraendurance exercise increases the production of reactive oxygen species in isolated mitochondria from human skeletal muscle. J Appl Physiol 2010, 108:780-787. 9. Seifert T, Brassard P, Wissenberg M, Rasmussen P, Nordby P, Stallknecht B, Adser H, Jakobsen AH, Pilegaard H, Nielsen HB, Secher NH: Endurance training enhances BDNF release from the human brain. Am J Physiol Regul Integr Comp Physiol 2010, 298:R372-377. 10. Seifert T, Rasmussen P, Brassard P, Homann PH, Wissenberg M, Nordby P, Stallknecht B, Secher NH, Nielsen HB: Cerebral oxygenation and metabolism during exercise following three months of endurance training in healthy overweight males. Am J Physiol Regul Integr Comp Physiol 2009, 297:R867-876. 11. Martin V, Kerherve H, Messonnier LA, Banfi JC, Geyssant A, Bonnefoy R, Feasson L, Millet GY: Central and peripheral contributions to neuromuscular fatigue induced by a 24-h treadmill run. J Appl Physiol 2010, 108:1224-1233. 12. al’Absi M, Petersen KL, Wittmers LE: Adrenocortical and hemodynamic predictors of pain perception in men and women. Pain 2002, 96:197-204. Pre-publication history The pre-publication history for this paper can be accessed here: http://www.biomedcentral.com/1741-7015/10/77/prepub doi:10.1186/1741-7015-10-77 Cite this article as: Millet and Millet: Ultramarathon is an outstanding model for the study of adaptive responses to extreme load and stress. BMC Medicine 2012 10:77. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Millet and Millet BMC Medicine 2012, 10:77 http://www.biomedcentral.com/1741-7015/10/77 Page 3 of 3

Ultramarathon is an outstanding model for the study of adaptive responses to extreme load and stress.

07-19-2012

Millet, Grégoire P,Millet, Guillaume Y

eng

PMC8087768

1 Vol.:(0123456789) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports Shorter heels are linked with greater elastic energy storage in the Achilles tendon A. D. Foster 1*, B. Block2, F. Capobianco III2, J. T. Peabody2, N. A. Puleo2, A. Vegas2 & J. W. Young 3 Previous research suggests that the moment arm of the m. triceps surae tendon (i.e., Achilles tendon), is positively correlated with the energetic cost of running. This relationship is derived from a model which predicts that shorter ankle moment arms place larger loads on the Achilles tendon, which should result in a greater amount of elastic energy storage and return. However, previous research has not empirically tested this assumed relationship. We test this hypothesis using an inverse dynamics approach in human subjects (n = 24) at speeds ranging from walking to sprinting. The spring function of the Achilles tendon was evaluated using specific net work, a metric of mechanical energy production versus absorption at a limb joint. We also combined kinematic and morphological data to directly estimate tendon stress and elastic energy storage. We find that moment arm length significantly determines the spring-like behavior of the Achilles tendon, as well as estimates of mass-specific tendon stress and elastic energy storage at running and sprinting speeds. Our results provide support for the relationship between short Achilles tendon moment arms and increased elastic energy storage, providing an empirical mechanical rationale for previous studies demonstrating a relationship between calcaneal length and running economy. We also demonstrate that speed and kinematics moderate tendon performance, suggesting a complex relationship between lower limb geometry and foot strike pattern. The role of the Achilles tendon (AT) in elastic energy storage with subsequent return during stance phase is well established1–7. Recovery of elastic energy imparted to the AT is potentially influenced by AT morphology in three ways: (1) material properties of the tendon, (2) cross-sectional area of the tendon, and (3) the moment arm of the calcaneal tuberosity loading the tendon against the muscle force of the m. triceps surae (i.e., foot geometry). Previous work suggests that foot geometry may explain variation in how much potential energy is stored in the tendon, finding that a shorter AT moment arm is correlated with lower mass-specific energy costs of locomotion (COL; L O2 kg−1  s−1)8, 9. This finding suggests that shorter AT moment arms are associated with greater elastic loads imparted to the tendon, which are then recovered as kinetic energy during the support phase of each gait cycle9, thereby reducing COL. Scholz et al.9 also suggest that the length of the AT moment arm is a more signifi- cant factor in explaining COL than variation in material properties of the tendon itself or size-related variation in the cost of swinging the leg forward during the aerial phase of the gait cycle9. However, assumptions about the interacting roles of AT moment arm length, tendon cross-sectional dimensions, and tendon material properties on variation in elastic energy storage have yet to tested in an integrated manner. Moreover, Scholz et al.9 doesn’t directly measure the variables in the model which predict variation in elastic energy storage, including muscle force and the external moment arm. Finally, because previous studies of how AT moment arm length influences COL have used trained runners running on a treadmill at a speed of 16 km/h, it is still unknown how variation in speed and athletic training impacts elastic loading to the tendon in relation to moment arm length. While previ- ous work has explored elastic loading of the AT at different speeds and under different loading conditions10–21, this study is the first to investigate the potential correlation between foot geometry like the AT moment arm length and spring-like behavior of this tendon in humans. In this study, we model elastic loading of the AT by characterizing the spring-like behavior over the support phase of each gait cycle using two metrics. First, we calculate specific net work (SNW) at the ankle joint. SNW is OPEN 1Department of Anatomy, School of Osteopathic Medicine, Campbell University, PO Box 4280, Buies Creek, NC 27506, USA. 2School of Osteopathic Medicine, Campbell University, PO Box 4280, Buies Creek, NC 27506, USA. 3Department of Anatomy and Neurobiology, Northeast Ohio Medical University (NEOMED), Rootstown, OH 44272, USA. *email: [email protected] 2 Vol:.(1234567890) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ a ratio of net joint work to total joint work that results in values from 0, comparable to a perfect spring, to 1, com- parable to a perfect motor or brake (see “Methods” section)22. SNW values of 1.0 are the result of a step with solely positive or negative work at the joint during stance phase, whereas a value of 0 indicates commensurate levels of positive and negative work, consistent with spring-like behavior. We use SNW to test whether AT moment arm length predicts spring-like behavior of the tendon. Second, we also measure individual AT cross-sectional area and tendon length using ultrasonography along with the muscle force impulse at the ankle over stance phase to estimate the magnitude of energy storage in the AT23 (see “Methods” section). We test the following hypotheses: H1: Subjects with shorter AT moment arms should exhibit more spring-like behavior (lower SNW values) in running gaits. H2: AT moment arm length should be negatively correlated with tendon stress (i.e., force per unit cross- sectional area). H3: AT moment arm length should be negatively correlated with tendon energy storage. Results Specific net work. AT moment arm lengths ranged from 3.12 to 5.01 cm (see boxplot in Supplementary Fig. S1) and are significantly correlated with body mass (r = 0.624, t[22] = 3.748, p < 0.001). SNW values at the hip are moderately high in all gaits, whereas values at the knee become more spring-like (i.e., lower values of SNW) in gaits lacking double-limb support periods, such as jogging, running, and sprinting. At the ankle, subjects have the lowest values of SNW when jogging and the highest values when sprinting (see Table 1 for summary statistics for each speed; see Fig. 1 for density plot and histogram of SNW values by joint and speed). We find that AT moment arm length is positively correlated with SNW for running and sprinting gaits, which supports hypothesis H1. However, this relationship is not significant for jogging gaits (see Table 2; Fig. 2). Additionally, while sprinting gaits exhibit higher mean SNW values at the ankle compared to walking gaits, the length of the AT moment arm appears to still have a significant effect on spring-like behavior at sprint speeds, but not during walking gaits. Using a mixed-effect multiple regression model with the average moment arm of the GRF vector at the ankle (RAnkle), tendon cross-sectional area (CSA), the GRF impulse (J), Froude number, and body mass as fixed effect predictors, we find that AT moment arm length does not significantly explain variation in SNW at running speeds. Nevertheless, AT moment arm length and RAnkle are significant predictors at sprinting speeds (see Table 3). Standardized partial correlation coefficients (i.e., β-weights) demonstrate that RAnkle plays a more significant role than AT moment arm length, which suggests that tendon elastic energy storage is particularly sensitive to vari- ation in postural variation across steps. However, average RAnkle is also correlated with AT moment arm length at sprint speeds (r = 0.370, t[70] = 3.336, p = 0.001). This result is consistent with the finding that AT moment arm length is correlated with foot length at sprint speeds (i.e., individuals with longer feet have longer AT moment arms; r = 0.684, t[22] = 4.394, p < 0.001), which should translate to a longer external moment arm (RAnkle) of the GRF. There were also no sex-based differences when sex was added as a fixed-effect (p = 0.680). Tendon stress and elastic energy storage. Both mass-specific tendon stress (MPa/kg) and mass-spe- cific elastic energy storage (Joules/kg) are negatively correlated with AT moment arm length at running and sprint speeds (see Fig. 3; Table 3). These variables are not correlated at walking, fast walking, and jogging speeds. Subject means are located in Supplementary Tables S1 and S2. Discussion We predicted that subjects with smaller AT moment arm lengths would exhibit lower SNW values in running gaits (i.e., more spring-like joint behavior). The overall pattern is that there is significant variation across gait cycles, but that AT moment arm length does appear to result in greater mass-specific tendon stress, and there- fore more spring-like behavior (i.e., lower SNW values), resulting in greater mass-specific energy storage at running and sprinting speeds. The design for this study was derived from a model proposed by Scholz et al.9, which predicts that tendon material properties and foot geometry explain inter-individual differences in COL, with AT moment arm length explaining most of this variation. The mixed-effect model analyzing sprint speed kinematics, subject morphology, body size, and tendon CSA, a determinant of tendon strength and stiffness, suggests that CSA does not play a significant role in determining spring-like behavior. Indeed, our results suggest Table 1. SNW values for each joint at each speed. Mean Froude numbers and specific net work (SNW) values with standard deviations (SD) for each joint at each speed. Gait Froude Hip SNW Knee SNW Ankle SNW Mean SD Mean SD Mean SD Mean SD Walk 0.130 0.041 0.669 0.251 0.653 0.170 0.450 0.199 Fast Walk 0.245 0.054 0.640 0.292 0.571 0.153 0.367 0.232 Jog 0.362 0.086 0.834 0.178 0.297 0.220 0.231 0.200 Run 0.662 0.179 0.930 0.124 0.428 0.245 0.447 0.252 Sprint 1.086 0.283 0.935 0.148 0.450 0.299 0.755 0.201 3 Vol.:(0123456789) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ that AT moment arm length and kinematics (i.e., ankle joint posture) play the largest role in determining elastic loading of the tendon. Additionally, the results of the mixed-effect model demonstrate that when holding body size constant, AT moment arm length is still a significant predictor. If subjects had similar tendon properties for their mass, mass-specific measures of elastic loading to the tendon would not vary with AT moment arm length. Therefore, these results provide robust support for the hypothesis that heel allometry is a significant factor in moderating elastic energy storage. It is also worth noting that there is variation about the regression line in ways that differ from scatter about the regression line when looking at COL in previous work8, 9. This is likely attributed to step-to-step variation among and between subjects. COL calculations are an average over two minutes, containing multiple steps, whereas values for SNW, tendon stress, and elastic strain energy in this study are averages over stance phase for multiple sequential steps. Figure 1. Density plot of individual specific net work (SNW) values for the hip, knee, and ankle for each step, for all subjects, at each speed. Peaks represent the most concentrated distribution of SNW values for each joint, at each speed. Plot rugs (vertical lines) are histograms of SNW values from all subjects and all steps. Table 2. Pearson’s correlation coefficient comparisons between ankle SNW, mass-specific stress, and mass- specific elastic energy storage and the Achilles tendon moment arm at different speeds. Bolded values indicate significance at p ≤ 0.05. Variable Speed Statistic R p-value SNW Walk t144 = − 0.796 − 0.066 0.786 Fast walk t96 = − 1.051 − 0.107 0.852 Jog t103 = 0.675 0.066 0.251 Run t85 = 3.367 0.343 0.001 Sprint t70 = 4.094 0.440 < 0.001 AT stress/BM Run t165 = − 9.185 − 0.582 < 0.001 Sprint t102 = − 5.190 − 0.457 < 0.001 AT tendon energy/BM Run t165 = − 4.938 − 0.359 < 0.001 Sprint t102 = − 2.319 − 0.224 0.011 4 Vol:.(1234567890) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ This study also builds upon previous work by exploring how speed and AT moment arm length moderates elastic energy storage. AT moment arm length does not explain variation in tendon performance during walk- ing gaits, which is consistent with predictions and results from Scholz et al.9 and Raichlen et al.8. AT moment arm length does explain variation in tendon performance in running and sprinting gaits, though not in jogging gaits, despite jogging gaits having the lowest mean value for SNW at the ankle for all subjects (0.231). The mean SNW values for running and sprinting were higher, at 0.447 and 0.755, respectively (see Table 1). However, while sprint speed SNW values are more motor/brake-like at sprinting speeds, the relationship between AT moment arm length and SNW is stronger at sprint speeds (r = 0.440, t[70] = 4.094, p < 0.001) than running speeds (r = 0.343, t[85] = 3.367, p = 0.001) (see Table 2). We interpret these findings to indicate that subjects need to generate more positive work at all lower limb joints as speed increases, but that individuals with short calcanei are nonethe- less able to harness relatively more energy from the spring-like return of the AT in running gaits. However, our results demonstrate that AT moment arm length is not correlated with spring-like behavior during walking gaits, which supports previous work showing no correlation between AT moment arm length and metabolic costs in walking gaits8. It is notable that the subject-determined sprint speeds used in this study are on average slower than the tread- mill speed used in Scholz et al.9 and Raichlen et al.8. Subject sprinting speeds in this study are 3.02 ± 0.478 m/s (Froude 1.04 ± 0.309). Subjects from previous studies ran at a sustained pace of 16 km/h (4.44 m/s) on a treadmill, while shod8, 9. However, the results from this study suggest a relationship between AT moment arm length and spring-like behavior at speeds above a jog, which may include at 4.44 m/s (16 km/h). The subject population is also different. Raichlen et al.8 used trained endurance runners whose 10 K personal best runs were under 36 min and Scholz et al.9 used subjects who self-identified as runners. In this study, subjects were recruited based on being recreationally fit but are not necessarily regular runners. Additionally, these previous studies only sampled Figure 2. Scatter plots of SNW and the Achilles tendon moment arm length at running (A) and sprinting speeds (B) against AT moment arm length. The black lines are least squares best fit lines and the gray bands represent 95% confidence intervals. Table 3. Mixed-effect model of variation in ankle SNW with morphological and kinematic variables as fixed effects at sprint speed. Bolded p-values indicate significant fixed-effect predictors. β are the partial regression coefficients (i.e., β-weights). CSA is the cross-sectional area of the AT. RAnkle is the mean external moment arm of the ground reaction force. J is the GRF impulse. Predictor variable β Statistic p-value Intercept − 0.002 F1,45 = 0.003 0.954 Ankle moment arm 0.229 F1,20 = 15.491 0.001 RAnkle 0.336 F1,45 = 7.074 0.011 Body mass 0.151 F1,20 = 0.536 0.473 Froude 0.191 F1,45 = 3.271 0.067 J − 0.236 F1,45 = 3.520 0.077 CSA 0.020 F1,20 = 0.001 0.976 5 Vol.:(0123456789) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ males, whereas this study has both males and females (and has more females than males) and a larger sample size. Broadly, our comprehensive sampling methods and sample size provide strong empirical support for previous models which suggest that heel morphology moderates AT tendon performance in running gaits. The mixed-effect model results expand on this relationship by exploring how kinematic variables and tendon CSA are correlated with spring-like behavior of the AT. When these variables are included as fixed effects, AT moment arm length is not a significant predictor of SNW at running speeds. However, at sprint speeds both the AT moment arm and the external moment arm of the GRF (RAnkle) are significant predictors, with RAnkle explaining more variation. Posture and footfall pattern (e.g., mid- vs. forefoot striking) may play a role in elastic loading, perhaps as reflected in the inter- and intra-subject variation in SNW values from step to step. RAnkle may also play a role in elastic energy storage by altering tendon stiffness depending on foot strike pattern (e.g., heel vs. fore-foot strike). Hof et al.25 found that subjects with the highest ankle moments exhibited greater stiffness in the elastic series component of the m. triceps surae. Results from tendon stress and estimates of elastic energy storage are consistent with measures of spring-like behavior (i.e., SNW). These results demonstrate that smaller AT moment arm lengths are correlated with higher mass-specific tendon stress values, which in turn result in greater amounts of mass-specific elastic energy storage. These data support previous models and empirical findings demonstrating a correlation between AT moment arm length and COL. The differences between the smallest and largest values of AT moment arm length in our sample are substantial. Values of AT moment arm lengths in our study varied from 3.12 to 5.01 cm, a 37.7% difference that can lead to as much as a 60.7% increase in mass-specific elastic energy storage between subjects with the shortest and longest moment arms. In this study, tendon stress is calculated using the force impulse (time integrated force). This has the advantage of reflecting force imparted to the AT over the entirety of stance phase. Peak values, by contrast, only reflect instantaneous loads and are more relevant to estimating safety factor and injury risk. Previous work calculated peak stresses of 111 MPa measured from turnbuckles on the AT while running26. Stress data measured in this study are similar. The mean value for peak stress from this study (using an inverse dynamics approach) from all subjects is 110.58 MPa at sprint speeds and 101.66 MPa for running speeds (see subject means in Table S1 and S2). There have been a range of estimates for failure stress in the AT26–30. Wren et al.30 found a mean failure stress of 79 MPa when straining Achilles tendons at 1–10% per second. While the mean stress values from this study exceed the mean failure stress from Wren et al.30, loading rates measured in this study were not of the same magnitude and duration as tests for plastic deformation and failure. Overall, the AT has similar material properties to other tendons, but receives a much higher load during running, with an estimated safety factor of 1.5, compared to other tendons that have safety factors of ~ 428, 31. However, because of the relationship with AT moment arm length and tendon stress, it is possible that foot geometry may be predictive of risk for ten- dinopathy. Achilles tendinopathy (i.e., pain and swelling of the Achilles tendon) is one of the most common sports-related injuries32. Ex vivo data from human ATs suggest that excessive tendon strain (the result of tendon Figure 3. Scatter plots of AT moment arm length and mass-specific tendon stress (MPa/kg) for running (a) and sprinting (b) trials. Scatter plot of AT moment arm and mass-specific tendon energy storage (Joules/kg) for running (c) and sprinting (d) trials. The black line is a least squares best fit line and the gray band represents the confidence interval. 6 Vol:.(1234567890) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ stress) is a primary factor responsible for tendon damage, and that repetitive loading contributes to AT injury and tendinopathy33. Moreover, a majority of AT ruptures are sub-clinical34, which suggests that intrinsic factors may play a significant role in explaining tendon strain and predicting risk for injury. Future work should explore this association. Additionally, athletic training is linked with changes to AT material properties which is relevant to injury prevention and may play a role in the capacity for elastic energy storage in the AT29, 35. Previous work has characterized elastic energy storage of the m. gastrocnemius and Achilles tendon during walking and running gaits using inverse dynamics and ultrasonography13, 14, 16–19, 36, 37. This is the first study to measure how AT moment arm length moderates tendon stress and elastic energy storage. However, it is impor- tant to note there are some limitations imposed by the study design and that a detailed understanding of the relationship between foot geometry and the spring-like function of the AT requires further work. In particular, future work should incorporate non-invasive imaging methods (e.g., ultrasonography) to measure instantane- ous changes in AT length and moment arm dimensions. Increased temporal precision may offer greater insight into how foot geometry may impact elastic energy storage at different points in stance phase. Indeed, previous work has demonstrated that accounting for tendon curvature at different points of stance phase may provide more precision in both measurements of tendon dimensions and moment arm length38. Rasske et al.39 found that during walking gaits, AT moment arm length changed 10% at toe-off relative to heel strike. Recent work by Harkness-Armstrong et al. found that assuming a straight (as opposed to curved) AT led to larger estimates of moment arm length than actual size38. In this study, joint work was calculated using moment arm lengths which vary with joint angle (and scaled to each subject; see “Methods” section). However, future work which measures how the AT moment arm changes with joint angle, load, and speed for each individual using ultrasonography may increase precision for estimating joint work across stance phase40–42. For statistical comparisons of spring- like behavior (SNW), joint stress, strain, and elastic energy storage, a static measure of AT moment arm length was used following Scholz et al.9. Previous work from other studies suggests that this method provides reliable measures of the AT moment arm43. However, exploring how tendon performance changes at different points of stance phase in relation to ankle moment arm length may also provide further insight into this relationship. Exploring how AT moment arm length moderates tradeoffs between muscle work and elastic energy storage would also provide further clarification. For example, holding all else equal, a shorter moment arm should result in less muscle fiber work (shortening) for a given joint rotation. Any decrease in muscle fiber work should result in a reduction in metabolic cost, which is consistent with results from previous work demonstrating a negative correlation between AT moment arm length and COL8, 9. The model proposed by Scholz et al.9 (and explored in this study), also predicts that shorter AT moment arms will result in increased tendon load. However, any increase in tendon load should be a result of an increase in muscle force. This creates a potential conflict in interpreting the relationship between AT moment arm length, metabolic cost, and muscle force. The results from this study suggest that shorter AT moment arms increase tendon load and elastic energy storage such that the balance of this tradeoff still favors a shorter AT moment arm, holding all else equal. However, future work which explores how AT moment arm length moderates muscle fiber work, muscle force, tendon load, elastic energy storage, and metabolic cost would provide further insight into the tradeoffs imposed between muscle force and elastic energy storage in individuals with shorter AT moment arm lengths. Individual measures of tendon material properties and elastic modulus may also offer greater clarification on how anatomy predicts elastic energy storage. Previous research indicates that material properties of the AT and force generating capacity of muscles varies between individuals and is correlated with elastic energy storage15, 18, 44–46. Additionally, tendon CSA may change with deformation of the tendon as force is applied throughout stance phase and should be accounted for in future studies47. The stiffness of the m. gastrocnemius aponeurosis is also a significant factor in contributing to muscle work and elastic energy storage17, 48–50. Methods which measure gearing and muscle–tendon stiffness, which vary with speed and torque development, also have been shown to influence elastic energy storage at different points of stance phase, and therefore may also be moderated by foot geometry like the AT moment arm13, 14, 19, 36. In conclusion, the results from this study suggest that there is a significant correlation between moment arm length and spring-like behavior of the AT. This spring-like behavior also corresponds with greater tendon stress and elastic energy storage in subjects with smaller AT moment arm sizes. Overall, our findings provide empirical mechanical support for the energetic model proposed by Scholz et al.9, suggesting that calcaneal length may be an important skeletal determinant of variation in COL during human bipedal running. Methods To test how AT moment arm length predicts spring-like behavior of the AT, we collected morphometric, kin- ematic, kinetic, and ultrasound data from 24 recreationally fit adults (see Table 4 for summary information). Subjects were asked to walk, fast walk, jog, run, and sprint at self-determined speeds across 6 force platforms Table 4. Subject morphometrics and summary statistics. Subject morphometrics and statistics presented as the number of males and females and all numeric variables represent the mean and the standard deviation (AT: Achilles tendon). Sex Age Body mass AT moment arm AT cross-sectional area Male Female Years kg cm cm2 7 17 23.50 ± 2.80 64.15 ± 9.84 4.27 ± 0.48 0.61 ± 0.15 7 Vol.:(0123456789) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ embedded in the floor (2 platforms in width, 3 platforms in length), which allowed for multiple support phases to be measured per trial. Subjects were unshod and repeated each speed three times. Subject speeds were calculated as the mean velocity of the marker placed on the greater trochanter of each leg during the support phase of each gait cycle, which were then standardized using Froude numbers (see Eq. (1))51: Ultrasound. The cross-sectional area of the AT was measured with B-mode ultrasound (SonoSite M-Turbo, Fujifilm SonoSite, Bothell, Washington, USA) using a 6 cm linear array transducer operating at 15 MHz (HFL50; Fujifilm SonoSite, Bothell, Washington, USA), with a depth set to 2.7 cm. The resolution of the ultrasound image is 100 pixels/cm. To measure the cross-sectional area of the AT, the probe was placed in a transverse plane at the level of the malleoli while subjects were prone on an examination table and the ankle was fixed at 160° (plantarflexed), which was measured with a goniometer. The cross-sectional area of the tendon was measured in ImageJ, by tracing the outline of the tendon using the polygon area selection tool and measuring the area52 (see Fig. 4). To measure tendon length, subjects lay prone with their foot in a neutral position (i.e., 90°). Using the same ultrasonographic technique, the myotendinous junction of the medial head of m. gastrocnemius on the left and right leg was located by placing it in the center of the image. A washable marker was used to place a dot on the skin at this location using the midpoint guide on the probe. This same procedure was used to determine the most inferior extent of the insertion of the tendon on the calcaneal tuberosity. Tendon length was measured using a flexible measuring tape and is defined as the distance of the myotendinous junction to the insertion on the calcaneal tuberosity. Tendon length for each subject is an average of the lengths of the left and right sides. Morphometrics. Subject body mass and height were recorded using a digital scale and segment lengths (hip height and foot length) were measured using a flexible measuring tape. Hip height was defined as the distance from the greater trochanter to the floor while standing. Foot length was defined as the most anterior point of the first digit to the most posterior point of the calcaneal tuberosity. Subject values for segment lengths are an average of the left and right sides. To compare tendon performance data collected in this study to previous work, we measured the moment arm of the AT using photographic methods9. Subjects stood with their ankle at an angle of 90° (where the leg is perpendicular to the foot) on a board that was affixed with a measuring tape (see Supplementary Fig. S2). The location of the footboard and camera were standardized and the malleoli were centered in the image to reduce the effects of parallax distortion. The moment arm was defined as the most medially or laterally prominent point of the medial or lateral malleolus, respectively, to the most posterior edge of the skin covering the tendon. Dis- tances were measured in ImageJ52 using the measuring tape as a scale. The medial and lateral values for both the left and right foot were averaged from three repeated measurements, then those averages from the medial and lateral side were averaged for each subject to obtain a single mean value for the AT moment arm. These moment arm values are used as the dependent variable when exploring the relationship between kinematic variables and elastic loading. See Table 4 for subject summary statistics and Supplementary Fig. S1 for a boxplot of subject AT moment arm lengths. Kinematics and kinetics. Subjects were fitted with retro-reflective markers placed on joint centers of the hip, knee, and ankle, in addition to other standardized locations (see Supplementary Fig. S3). Subjects were (1) velocity2 gravitational acceleration · hip height Figure 4. The cross-sectional area of the Achilles tendon measured using ultrasound. Example of the cross- sectional area of the Achilles tendon measured at the level of the malleoli using B-mode ultrasound. The area of the tendon outlined by the dashed white line and shaded in blue was calculated using ImageJ. A is anterior and P is posterior. The hyperechoic region anterior to the shaded ellipse around the tendon is Kager’s fat pad. 8 Vol:.(1234567890) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ asked to walk, fast walk, jog, run, and sprint at self-determined speeds. Runway distances were sufficient for subjects to reach steady state locomotion when crossing force plates. Force data were collected using six force platforms (BTS P6000D, BTS Bioengineering Corp., Quincy, MA, USA) that were embedded in the floor, with a sampling rate of 1 kHz. Kinematic data were recorded using a 12-camera motion capture system (BTS Bioen- gineering Corp., Quincy, MA, USA), with a sampling rate of 500 Hz. Force data were deprecated to 500 Hz to synchronize force and kinematic data which were processed using custom routines in MATLAB (version 2018b; Mathworks, Natick, MA, USA). A zero-lag Butterworth, low-pass filter was used to smooth kinematic data (fourth-order, with a 6 Hz cutoff) and force data (fourth-order, with a 100 Hz cutoff). An inverse dynamics approach was used to calculate joint work for the hip, knee, and ankle over the sup- port phase of each gait cycle using 2D kinematics and ground reaction forces (GRF). We calculated the exter- nal moment arm of the GRF (R) as the fore-aft and vertical distance from the center of pressure (COP). Net joint moments were calculated using the GRF vector, limb segment accelerations, and joint moments following Biewener et al.53 and Winter54. Because mediolateral forces have a negligible impact on joint moments, these moments were ignored in the calculation of biomechanical variables. Net joint moments, M, were determined for each kinematic frame, at the hip, knee, and ankle joint using the free-body method described in Winter54. Limb segment accelerations were calculated using the second-order finite differences method54. Extensor muscle forces (Fankle, Fknee, and Fhip) to generate these moments were determined by solving a system of equations from Biewener et al.53: Subtracted terms represent flexor actions of bi-articular muscles gastrocnemius (G), hamstrings (H), and rectus femoris (RF). Flexor and extensor moments were calculated assuming the force produced by each muscle is proportional to its physiological cross-sectional area (PCSA)53. Muscle force impulse at the hip, knee, and ankle was calculated as the finite integral of instantaneous muscle force throughout each support phase. Human extensor moment arms, r, for the hip, knee, and ankle, were calculated as instantaneous values that vary with joint angle using equations from the literature for the hip and knee from Visser55 and ankle from Rugg et al.56. To calculate the average moment arm of the GRF vector (R) at the ankle (RAnkle) for each support phase, we used the GRF impulse (J), AT moment arm length (rAnkle; calculated following Rugg et al.56) and the ankle force impulse (FImpulse) following equations from Biewener et al.53 [see Eq. (5)]: Using equations from Winter54, we calculated work, which is the finite integral of joint power (Watts/kg) over time, at the hip, knee, and ankle joint. Positive and negative values for work were used to calculate SNW, which is a ratio of the sum of positive and negative work to the sum of the total work, to characterize the spring- like behavior of lower limb joints [see Eq. (6)]. Negative values (instantaneous values less than 0) for work are summed as WorkNeg and positive values for work are summed as WorkPos. Tendon stress, strain, and energy storage. We calculate tendon stress, which is defined as the force impulse imparted to the AT, scaled to the cross-sectional area of the AT, for each subject: Here, FAnkle is the instantaneous value of muscle force at the ankle joint for each support phase of a gait cycle (ankle force impulse, FImpulse), and t1 and t2 represent the beginning and end of the support phase interval. Tendon CSA is calculated as the cross-sectional area and converted to square meters (m2) from ultrasound images of the AT for each subject (see ultrasound methods). Tendon strain is calculated by dividing tendon stress by the elastic modulus of the AT. Here, we use 819 MPa for the elastic modulus, which is the mean value of a sample of human ATs from Wren et al.30. Tendon strain is then multiplied by the resting tendon length for each subject (see ultrasound methods) to calculate the estimated change in the length of the tendon. Strain energy, or the amount of elastic energy storage in the tendon, is mod- eled following Hooke’s law [see Eq. (8)]. Here, FAnkle is the ankle force impulse and L is the change in length of the AT. Following Moore et al.23, we multiplied the amount of energy recovered by 0.93, which is an estimate of tendon resilience (i.e., 93%). (2) Mankle = Fankle · rAnkle (3) MKnee = FKnee · rKnee− FG,Knee · rG,Knee−FH,Knee · rH,Knee (4) MHip = FHip · rHip − FRF,Hip · rRF,Hip (5) RAnkle =  t2 t1GRF · rAnkle  t2 t1FAnkle (6) |WorkPos + WorkNeg| |WorkPos| + |WorkNeg| (7) Tendonstress =  t2 t1FAnkle TendonCSA (8) W = 1 2FAnkL 9 Vol.:(0123456789) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ Statistics. All statistical analyses for this study were conducted in the R statistical platform (version 3.6.1, ’Action of the Toes’)57. We used one-tailed Pearson product-moment correlations to test for positive association between SNW and AT moment arm lengths and negative association between AT moment arm lengths and stress and elastic energy storage. Mixed-effect models were used to explore the correlation between SNW and AT moment arm lengths with kinematic and morphometric variables as fixed effects (using the lme function)58. A mixed-effect model is used for testing hypotheses in this study as it allows for adjustment of degrees of free- dom to account for variation between individuals, and error terms to account for repeated measures of multiple steps from the same subject. Raw variates were scaled and centered (converted to z-scores) to permit direct comparisons of resulting partial regression coefficients (β weights) to allow for comparison of which predictors best explain variance in the dependent variable. Results for all tests were significant at p < 0.05. Plots were made using ggplot259. Institutional oversight and compliance. Institutional Review Board approval was obtained by Camp- bell University (Protocols 376 and 472) and all study methods and procedures in this study followed the approved protocols and all IRB guidelines. Informed consent was obtained prior to subject participation. Data availability Data and the R code used for statistical analysis and generating figures and tables in this study are available at: https:// github. com/ adfos ter/ achil leste ndon Received: 8 September 2020; Accepted: 16 April 2021 References 1. Alexander, R. M. Elastic energy stores in running vertebrates. Integr. Comp. Biol. 24, 85–94 (1984). 2. Alexander, R. M. Energy-saving mechanisms in walking and running. J. Exp. Biol. 160, 55–69 (1991). 3. Alexander, R. M. Tendon elasticity and muscle function. Comp. Biochem. Physiol. A Mol. Integr. Physiol. 133, 1001–1011 (2002). 4. Fukunaga, T. et al. In vivo behaviour of human muscle tendon during walking. Proc. R. Soc. Lond. Ser. B Biol. Sci. 268, 229–233 (2001). 5. Lichtwark, G. A. & Wilson, A. M. Is Achilles tendon compliance optimised for maximum muscle efficiency during locomotion?. J. Biomech. 40, 1768–1775 (2007). 6. Sawicki, G. S., Lewis, C. L. & Ferris, D. P. It pays to have a spring in your step. Exerc. Sport Sci. Rev. 37, 130–138 (2009). 7. Zelik, K. E. & Franz, J. R. It’s positive to be negative: achilles tendon work loops during human locomotion. PLoS ONE 12, e0179976 (2017). 8. Raichlen, D. A., Armstrong, H. & Lieberman, D. E. Calcaneus length determines running economy: implications for endurance running performance in modern humans and Neandertals. J. Hum. Evol. 60, 299–308 (2011). 9. Scholz, M. N., Bobbert, M. F., van Soest, A. J., Clark, J. R. & van Heerden, J. Running biomechanics: shorter heels, better economy. J. Exp. Biol. 211, 3266–3271 (2008). 10. Kharazi, M., Bohm, S., Theodorakis, C., Mersmann, F. & Arampatzis, A. Quantifying mechanical loading and elastic strain energy of the human Achilles tendon during walking and running. Sci. Rep. 11, 5830 (2021). 11. Peltonen, J., Cronin, N. J., Stenroth, L., Finni, T. & Avela, J. Viscoelastic properties of the Achilles tendon in vivo. Springerplus 2, 212 (2013). 12. Pożarowszczyk, B., Gołaś, A., Chen, A., Zając, A. & Kawczyński, A. The impact of post activation potentiation on achilles tendon stiffness, elasticity and thickness among basketball players. Sports 6, 117 (2018). 13. Monte, A. & Zignoli, A. Muscle and tendon stiffness and belly gearing positively correlate with rate of torque development during explosive fixed end contractions. J. Biomech. 114, 110110 (2021). 14. Monte, A., Nardello, F. & Zamparo, P. Mechanical advantage and joint function of the lower limb during hopping at different frequencies. J. Biomech. 118, 110294 (2021). 15. Maganaris, C. N. & Paul, J. P. Tensile properties of the in vivo human gastrocnemius tendon. J. Biomech. 35, 1639–1646 (2002). 16. Lichtwark, G. A., Cresswell, A. G. & Newsham-West, R. J. Effects of running on human Achilles tendon length–tension properties in the free and gastrocnemius components. J. Exp. Biol. 216, 4388–4394 (2013). 17. Lichtwark, G. A. & Wilson, A. M. Interactions between the human gastrocnemius muscle and the Achilles tendon during incline, level and decline locomotion. J. Exp. Biol. 209, 4379–4388 (2006). 18. Lichtwark, G. A. & Wilson, A. M. In vivo mechanical properties of the human Achilles tendon during one-legged hopping. J. Exp. Biol. 208, 4715–4725 (2005). 19. Lai, A. K. M., Biewener, A. A. & Wakeling, J. M. Muscle-specific indices to characterise the functional behaviour of human lower- limb muscles during locomotion. J. Biomech. https:// doi. org/ 10. 1016/j. jbiom ech. 2019. 04. 027 (2019). 20. Monte, A., Baltzopoulos, V., Maganaris, C. N. & Zamparo, P. Gastrocnemius Medialis and Vastus Lateralis in vivo muscle-tendon behavior during running at increasing speeds. Scand. J. Med. Sci. Sports 30, 1163–1176 (2020). 21. Monte, A., Maganaris, C., Baltzopoulos, V. & Zamparo, P. The influence of Achilles tendon mechanical behaviour on “apparent” efficiency during running at different speeds. Eur. J. Appl. Physiol. 120, 2495–2505 (2020). 22. Lee, D. V., McGuigan, M. P., Yoo, E. H. & Biewener, A. A. Compliance, actuation, and work characteristics of the goat foreleg and hindleg during level, uphill, and downhill running. J. Appl. Physiol. (Bethesda, Md.: 1985) 104, 130–141 (2008). 23. Moore, T. Y., Rivera, A. M. & Biewener, A. A. Vertical leaping mechanics of the Lesser Egyptian Jerboa reveal specialization for maneuverability rather than elastic energy storage. Front. Zool. 14, 32 (2017). 24. Kouno, M., Ishigaki, T., Ikebukuro, T., Yata, H. & Kubo, K. Effects of the strain rate on mechanical properties of tendon structures in knee extensors and plantar flexors in vivo. Sports Biomech. 1–14 (2019). 25. Hof, A. L., Zandwijk, J. P. V. & Bobbert, M. F. Mechanics of human triceps surae muscle in walking, running and jumping. Acta Physiol. Scand. 174, 17–30 (2002). 26. Komi, P. V., Fukashiro, S. & Järvinen, M. Biomechanical loading of Achilles tendon during normal locomotion. Clin. Sports Med. 11, 521–531 (1992). 27. Butler, D. L., Grood, E. S., Noyes, F. R., Zernicke, R. F. & Brackett, K. Effects of structure and strain measurement technique on the material properties of young human tendons and fascia. J. Biomech. 17, 579–596 (1984). 28. Ker, R. F., Alexander, R. M. C. N. & Bennett, M. B. Why are mammalian tendons so thick?. J. Zool. 216, 309–324 (1988). 29. Kongsgaard, M., Aagaard, P., Kjaer, M. & Magnusson, S. P. Structural Achilles tendon properties in athletes subjected to different exercise modes and in Achilles tendon rupture patients. J. Appl. Physiol. 99, 1965–1971 (2005). 10 Vol:.(1234567890) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ 30. Wren, T. A. L., Yerby, S. A., Beaupré, G. S. & Carter, D. R. Mechanical properties of the human achilles tendon. Clin. Biomech. 16, 245–251 (2001). 31. Biewener, A. A. & Roberts, T. J. Muscle and tendon contributions to force, work, and elastic energy savings: a comparative perspec- tive. Exerc. Sport Sci. Rev. 28, 99 (2000). 32. Raikin, S. M., Garras, D. N. & Krapchev, P. V. Achilles tendon injuries in a United States population. Foot Ankle Int. 34, 475–480 (2013). 33. Wren, T. A. L., Lindsey, D. P., Beaupré, G. S. & Carter, D. R. Effects of creep and cyclic loading on the mechanical properties and failure of human achilles tendons. Ann. Biomed. Eng. 31, 710–717 (2003). 34. Davis, J. J., Mason, K. T. & Clark, D. A. Achilles tendon ruptures stratified by age, race, and cause of injury among active duty U.S. military members. Mil Med. 164, 872–873 (1999). 35. Mersmann, F., Bohm, S., Schroll, A., Marzilger, R. & Arampatzis, A. Athletic training affects the uniformity of muscle and tendon adaptation during adolescence. J. Appl. Physiol. 121, 893–899 (2016). 36. Qiao, M. & Jindrich, D. L. Leg joint function during walking acceleration and deceleration. J. Biomech. 49, 66–72 (2016). 37. Cronin, N. J. & Lichtwark, G. The use of ultrasound to study muscle–tendon function in human posture and locomotion. Gait Posture 37, 305–312 (2013). 38. Harkness-Armstrong, C. et al. Effective mechanical advantage about the ankle joint and the effect of achilles tendon curvature during toe-walking. Front. Physiol. 11, 407 (2020). 39. Rasske, K., Thelen, D. G. & Franz, J. R. Variation in the human Achilles tendon moment arm during walking. Comput. Methods Biomech. Biomed. Eng. 20, 201–205 (2017). 40. Fath, F., Blazevich, A. J., Waugh, C. M., Miller, S. C. & Korff, T. Direct comparison of in vivo Achilles tendon moment arms obtained from ultrasound and MR scans. J. Appl. Physiol. 109, 1644–1652 (2010). 41. Maganaris, C. N. Imaging-based estimates of moment arm length in intact human muscle-tendons. Eur. J. Appl. Physiol. 91, 130–139 (2004). 42. Olszewski, K., Dick, T. J. M. & Wakeling, J. M. Achilles tendon moment arms: the importance of measuring at constant tendon load when using the tendon excursion method. J. Biomech. 48, 1206–1209 (2015). 43. Gallinger, T. L., Fletcher, J. R. & MacIntosh, B. R. Mechanisms of reduced plantarflexor function in Cerebral palsy: smaller triceps surae moment arm and reduced muscle force. J. Biomech. 110, 109959 (2020). 44. Fukunaga, T. et al. Muscle volume is a major determinant of joint torque in humans. Acta Physiol. Scand. 172, 249–255 (2001). 45. Hof, A. L. In vivo measurement of the series elasticity release curve of human triceps surae muscle. J. Biomech. 31, 793–800 (1998). 46. Muraoka, T., Muramatsu, T., Fukunaga, T. & Kanehisa, H. Elastic properties of human Achilles tendon are correlated to muscle strength. J. Appl. Physiol. 99, 665–669 (2005). 47. Reeves, N. D. & Cooper, G. Is human Achilles tendon deformation greater in regions where cross-sectional area is smaller?. J. Exp. Biol. 220, 1634–1642 (2017). 48. Maganaris, C. N., Kawakami, Y. & Fukunaga, T. Changes in aponeurotic dimensions upon muscle shortening: in vivo observations in man. J. Anat. 199, 449–456 (2001). 49. Magnusson, S. P., Aagaard, P., Rosager, S., Dyhre-Poulsen, P. & Kjaer, M. Load-displacement properties of the human triceps surae aponeurosis in vivo. J. Physiol. 531, 277–288 (2001). 50. Magnusson, S. P. et al. Differential strain patterns of the human gastrocnemius aponeurosis and free tendon, in vivo. Acta Physiol. Scand. 177, 185–195 (2003). 51. Alexander, R. M. & Jayes, A. S. A dynamic similarity hypothesis for the gaits of quadrupedal animals. J. Zool. Lond. 201, 135–152 (1983). 52. Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH Image to ImageJ: 25 years of image analysis. Nat. Methods 9, 671–675 (2012). 53. Biewener, A. A., Farley, C. T., Roberts, T. J. & Temaner, M. Muscle mechanical advantage of human walking and running: implica- tions for energy cost. J. Appl. Physiol. 97, 2266–2274 (2004). 54. Winter, D. A. Biomechanics and Motor Control of Human Movement (Wiley, 2009). 55. Visser, J. J., Hoogkamer, J. E., Bobbert, M. F. & Huijing, P. A. Length and moment arm of human leg muscles as a function of knee and hip-joint angles. Eur. J. Appl. Physiol. 61, 453–460 (1990). 56. Rugg, S. G., Gregor, R. J., Mandelbaum, B. R. & Chiu, L. In vivo moment arm calculations at the ankle using magnetic resonance imaging (MRI). J. Biomech. 23, 495–501 (1990). 57. R Core Team. R: A language and environment for statistical computing. (R Foundation for Statistical Computing, 2019). 58. Pinheiro, J., Bates, D., DebRoy, S., Sarkar, D. & R Core Team. nlme: Linear and Nonlinear Mixed Effects Models. R package version 3.1–137 (2018). 59. Wickham, H. ggplot2: Elegant Graphics for Data Analysis. (Springer, 2009). Acknowledgements The authors wish to thank the Campbell University School of Osteopathic Medicine and the Campbell Univer- sity Medical Student Summer Research Scholars program. We also thank Dr. Jennifer Bunn and the Campbell University College of Pharmacy & Health Sciences for use of the Advanced Interdisciplinary Movement Science (AIMS) lab. Finally, we wish to thank the subjects for their participation in this research. Author contributions Conceptualization: A.D.F., J.W.Y.; Methodology: A.D.F., J.W.Y.; Software: A.D.F.; Analysis: A.D.F., J.W.Y.; Data Collection: A.D.F., F.C., J.T.P., N.A.P., B.B., V.A.; Writing – original draft: A.D.F., J.W.Y.; Writing – editing: A.D.F., J.W.Y., F.C., J.T.P., N.A.P., B.B., V.A. Funding Funding for this study was provided by the Campbell University School of Osteopathic Medicine and the Medical Student Summer Research Scholars program. Competing interests The authors declare no competing interests. Additional information Supplementary Information The online version contains supplementary material available at https:// doi. org/ 10. 1038/ s41598- 021- 88774-8. Correspondence and requests for materials should be addressed to A.D.F. 11 Vol.:(0123456789) Scientific Reports | (2021) 11:9360 | https://doi.org/10.1038/s41598-021-88774-8 www.nature.com/scientificreports/ Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. © The Author(s) 2021

Shorter heels are linked with greater elastic energy storage in the Achilles tendon.

04-30-2021

Foster, A D,Block, B,Capobianco, F,Peabody, J T,Puleo, N A,Vegas, A,Young, J W

eng

PMC10415251

Supplementary Table S1 Chinese Athletics Association Marathon Runners Level Evaluation Standards (Time). Levels Race Gender Age groups (years) 18- 29 30- 34 35- 39 40- 44 45- 49 50- 54 55- 59 60- 64 65+ Elite Half Female 1h 49m 1h 50m 1h 51m 1h 52m 1h 53m 1h 54m 1h 55m 1h 56m 2h 03m male 1h 34m 1h 35m 1h 36m 1h 37m 1h 38m 1h 39m 1h 40m 1h 41m 1h 48m Full Female 3h 48m 3h 49m 3h 50m 3h 51m 3h 52m 3h 53m 3h 54m 4h 04m 4h 30m male 3h 24m 3h 25m 3h 26m 3h 27m 3h 28m 3h 29m 3h 33m 3h 39m 3h 50m Level 1 Half Female 2h 11m 2h 12m 2h 13m 2h 14m 2h 15m 2h 16m 2h 17m 2h 18m 2h 19m male 1h 51m 1h 52m 1h 53m 1h 54m 1h 55m 1h 56m 1h 57m 1h 58m 2h 05m Full Female 4h 25m 4h 26m 4h 27m 4h 28m 4h 29m 4h 32m 4h 34m 4h 41m 4h 30m male 4h 03m 4h 04m 4h 05m 4h 06m 4h 07m 4h 08m 4h 09m 4h 14m 3h 50m Level 2 Half Female 2h 31m 2h 32m 2h 33m 2h 34m 2h 35m 2h 36m 2h 37m 2h 38m 2h 41m male 2h 14m 2h 15m 2h 16m 2h 17m 2h 18m 2h 19m 2h 17m 2h 21m 2h 29m Full Female 5h 17m 5h 18m 5h 19m 5h 20m 5h 21m 5h 22m 5h 23m 5h 26m 5h 27m male 4h 53m 4h 54m 4h 55m 4h 56m 4h 57m 4h 58m 4h 59m 5h 04m 5h 16m Level 3 Half Female 3h 3h 3h 3h 3h 3h 3h 3h 3h male 3h 3h 3h 3h 3h 3h 3h 3h 3h Full Female 6h 6h 6h 6h 6h 6h 6h 6h 6h male 6h 6h 6h 6h 6h 6h 6h 6h 6h Note: The runner classification criteria are based on participants completing the race within or equal to the time indicated in the table. More information can be found at: https://www.athletics.org.cn/bulletin/hygd/mls/2023/0323/464400.html

Gender differences in footwear characteristics between half and full marathons in China: a cross-sectional survey.

08-10-2023

Xia, Yuyu,Shen, Siqin,Jia, Sheng-Wei,Teng, Jin,Gu, Yaodong,Fekete, Gusztáv,Korim, Tamás,Zhao, Haotian,Wei, Qiang,Yang, Fan

eng

PMC4473093

Incidence of Running-Related Injuries Per 1000 h of running in Different Types of Runners: A Systematic Review and Meta-Analysis.

[]

Videbæk, Solvej,Bueno, Andreas Moeballe,Nielsen, Rasmus Oestergaard,Rasmussen, Sten

eng

PMC7379642

Supplement Table 5. Change in VO2max (ml·min-1·kg-1) from 1995-1997 to 2016-2017 in relation to region. L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 Year n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change n Mean (SD) Change Mean (SD) Change 95-97 882 2.72 (0.05) Ref 37.4 (0.60) Ref 2241 2.82 (0.05) Ref 38.2 (0.54) Ref 1451 2.88 (0.05) Ref 39.7 (0.57) Ref 98-99 2017 2.72 (0.06) 0,2% 37.3 (0.68) -0,3% 2565 2.86 (0.05) 1,4% 38.3 (0.53) 0,4% 1961 2.78 (0.04) -3,5% 37.7 (0.46) -5,1% 00-01 4629 2.76 (0.05) 1,6% 37.2 (0.62) -0,5% 3171 2.83 (0.04) 0,2% 37.5 (0.48) -1,9% 4744 2.78 (0.04) -3,5% 37.6 (0.50) -5,3% 02-03 9973 2.62 (0.05) -3,6% 35.5 (0.54) -5,1% 5659 2.70 (0.04) -4,2% 35.6 (0.56) -6,8% 6996 2.73 (0.04) -5,4% 36.8 (0.53) -7,4% 04-05 15177 2.64 (0.04) -2,8% 35.3 (0.54) -5,8% 9660 2.66 (0.04) -5,9% 35.5 (0.47) -6,9% 12580 2.71 (0.04) -6,0% 36.9 (0.46) -7,0% 06-07 17130 2.65 (0.04) -2,4% 35.6 (0.47) -4,9% 8053 2.68 (0.05) -5,0% 35.3 (0.48) -7,6% 13332 2.68 (0.04) -6,8% 35.9 (0.50) -9,6% 08-09 22928 2.69 (0.04) -1,0% 35.9 (0.47) -4,1% 8028 2.66 (0.04) -5,6% 34.5 (0.51) -9,6% 12519 2.68 (0.04) -7,0% 35.5 (0.50) -10,5% 10-11 20374 2.69 (0.04) -0,9% 35.6 (0.48) -4,9% 6129 2.70 (0.04) -4,3% 35.5 (0.42) -7,1% 12669 2.68 (0.04) -7,0% 35.2 (0.48) -11,4% 12-13 31735 2.64 (0.04) -2,9% 35.0 (0.49) -6,5% 7736 2.68 (0.04) -4,9% 34.7 (0.43) -9,1% 17768 2.66 (0.04) -7,6% 35.0 (0.49) -11,8% 14-15 28977 2.60 (0.04) -4,2% 34.4 (0.45) -8,1% 9383 2.64 (0.04) -6,3% 34.2 (0.43) -10,5% 17220 2.65 (0.04) -8,1% 34.6 (0.45) -12,8% 16-17 13652 2.62 (0.03) -3,5% 34.5 (0.40) -7,8% 4589 2.64 (0.03) -6,5% 34.2 (0.40) -10,5% 7317 2.61 (0.04) -9,4% 34.2 (0.39) -14,0% Urban counties Rural counties All other counties

Decline in cardiorespiratory fitness in the Swedish working force between 1995 and 2017.

11-15-2018

Ekblom-Bak, Elin,Ekblom, Örjan,Andersson, Gunnar,Wallin, Peter,Söderling, Jonas,Hemmingsson, Erik,Ekblom, Björn

eng

PMC3737354

The Power of Auditory-Motor Synchronization in Sports: Enhancing Running Performance by Coupling Cadence with the Right Beats Robert Jan Bood1, Marijn Nijssen1, John van der Kamp1,2, Melvyn Roerdink1* 1 MOVE Research Institute Amsterdam, Faculty of Human Movement Sciences, VU University Amsterdam, Amsterdam, the Netherlands, 2 Institute of Human Performance, University of Hong Kong, Hong Kong SAR Abstract Acoustic stimuli, like music and metronomes, are often used in sports. Adjusting movement tempo to acoustic stimuli (i.e., auditory-motor synchronization) may be beneficial for sports performance. However, music also possesses motivational qualities that may further enhance performance. Our objective was to examine the relative effects of auditory-motor synchronization and the motivational impact of acoustic stimuli on running performance. To this end, 19 participants ran to exhaustion on a treadmill in 1) a control condition without acoustic stimuli, 2) a metronome condition with a sequence of beeps matching participants’ cadence (synchronization), and 3) a music condition with synchronous motivational music matched to participants’ cadence (synchronization+motivation). Conditions were counterbalanced and measurements were taken on separate days. As expected, time to exhaustion was significantly longer with acoustic stimuli than without. Unexpectedly, however, time to exhaustion did not differ between metronome and motivational music conditions, despite differences in motivational quality. Motivational music slightly reduced perceived exertion of sub-maximal running intensity and heart rates of (near-)maximal running intensity. The beat of the stimuli –which was most salient during the metronome condition– helped runners to maintain a consistent pace by coupling cadence to the prescribed tempo. Thus, acoustic stimuli may have enhanced running performance because runners worked harder as a result of motivational aspects (most pronounced with motivational music) and more efficiently as a result of auditory-motor synchronization (most notable with metronome beeps). These findings imply that running to motivational music with a very prominent and consistent beat matched to the runner’s cadence will likely yield optimal effects because it helps to elevate physiological effort at a high perceived exertion, whereas the consistent and correct cadence induced by auditory-motor synchronization helps to optimize running economy. Citation: Bood RJ, Nijssen M, van der Kamp J, Roerdink M (2013) The Power of Auditory-Motor Synchronization in Sports: Enhancing Running Performance by Coupling Cadence with the Right Beats. PLoS ONE 8(8): e70758. doi:10.1371/journal.pone.0070758 Editor: Ramesh Balasubramaniam, University of California, Merced, United States of America Received January 29, 2013; Accepted June 21, 2013; Published August 7, 2013 Copyright:
2013 Bood et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: The contribution of Melvyn Roerdink was supported by the Netherlands Organization for Scientific Research (NWO, Veni Grant 451-09-024). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected] Introduction On February 18th 1998, the Ethiopian athlete Haile Gebrse- lassie astonished sport spectators when he achieved a world best time of 4:52.86 min in the 2000 m. Shortly after the race, Gebrselassie indicated that he had coupled his running cadence with the beat of the pop song Scatman by the late Scatman John, which was played throughout his race at Birmingham’s National Indoor Arena, UK. This anecdote of Gebrselassie and Scatman testifies to the fact that rhythmic bodily movements are often coupled with external acoustic stimuli, such as acoustic metronomes and music, a phenomenon known as auditory-motor synchronization [1,2,3]. Dancing to music, for example, involves the synchronization of whole-body movements to the beat [1]. Another striking example is our natural tendency to tap our fingers, hands, or feet along to a beat when listening to music [2,3,4]. In fact, research demonstrates that even young infants spontaneously sway and wiggle around with rhythmic acoustic stimuli [5], which supports the notion that humans have a predisposition for auditory-motor synchronization [5,6,7]. This apparent predisposition for auditory-motor synchroniza- tion has led to the exploitation of acoustic rhythms as a potential means to enhance performance in practical settings, including rehabilitation [8,9], exercise [10,11], and sports [12,13]. With respect to the latter, pundits and experts alike consider acoustic rhythms particularly useful in sports that are cyclic in nature, like running, rowing, and cycling. In coxed rowing, for example, the coxswain in the stern coordinates the stroke rate and rhythm for the rowers to follow. Musical tempo also dictates movement rate in popular exercise-to-music classes, like spinning, aerobics, and Zumba. In spinning classes for example, variations in the tempo of the music determine the pedaling rate and thereby the work rate (i.e., music functions as an external pacemaker). The potential scope of auditory-motor synchronization as a performance enhancement tool has been demonstrated by the recent development of products that help attune musical content to the actual or desired work rate during running, cycling, rowing, and circuit-class training (e.g., [14,15,16,17]) as well as by recent PLOS ONE | www.plosone.org 1 August 2013 | Volume 8 | Issue 8 | e70758 scientific findings in the context of sports or exercises (i.e., [18,19,20,21]). Specifically, Simpson and Karageorghis demon- strated superior performance (i.e., faster times) of 400 m sprints completed in a synchronous music condition in comparison to a no-music control condition [20]. Likewise, Terry and colleagues reported longer times to volitional exhaustion with than without synchronous music in a group of elite triathletes during high- intensity treadmill running [21]; see also Karageorghis et al. for similar effects in treadmill walking [19]. Karageorghis and Terry aptly summarized these insights as follows: ‘‘When athletes work in time to music, they often work harder for longer’’ [13]. Finally, Bacon and colleagues found that cycle ergometry at a fixed pedaling rate was more efficient (i.e., lower oxygen consumption) when performed synchronously with music than when the musical tempo was set slightly slower than a visually controlled, fixed pedaling rate [18]. These findings underscore the potential of synchronous music and auditory-motor synchronization to enhance work-rate, endurance, and efficiency in cyclic sports. Music not only provides a stimulus for synchronization, but often also possesses motivational qualities that may enhance performance. The body of evidence on the beneficial effects of motivational music in sports and exercise mainly stems from research on the use of asynchronous, background music (i.e., without an explicit synchronization of movements to the beat). Research suggests that motivational music can enhance sports performance, for example, by lifting mood and arousal levels and by dissociation from feelings of pain and fatigue [13]. With respect to the latter, attending to motivational music during low-to- moderate levels of physical exertion typically reduces perceived exertion by roughly 10% [22]. At higher intensities, attending to music does not reduce perceived exertion. That is, at higher intensities music does not appear to influence what one feels (i.e., a similarly high perceived exertion) but only how one feels it (i.e., it positively shapes the interpretation of exertion symptoms like fatigue and pain; cf. [13]). Interestingly, not all music seems to be equally effective in terms of motivational quality [23]. For example, loud, fast, percussive music with accentuated bass frequencies has stimulative effects, which increase arousal and associated physiological responses (e.g., heart rate). In contrast, soft, slow music has sedative effects (i.e., reducing arousal). Attending to such music during sports may therefore adversely affect performance. Validated, objective methods, including the Brunel Music Rating Inventory 2 (BMRI-2; [24]), have now also been developed to select music that is likely to yield optimal effects for the task at hand. It is evident that sport and exercise performance can benefit from synchronous motivational music in terms of ergogenic (i.e., work-enhancing), psychophysical (e.g., perceived exertion), and physiological (e.g., heart rate) effects. It remains unclear, however, whether the beneficial effects are primarily mediated by auditory- motor synchronization, by motivational quality of music, or by a combination of both factors. The opening anecdote concerning Gebrselassie and Scatman highlights the potential of auditory- motor synchronization in enhancing running performance, but motivational quality of the pop song Scatman may also play a role. Interestingly, in a comparative study, Terry et al. [21] observed that elite triathletes ran 18.1% and 19.7% longer, respectively, when running in time to motivational and motivationally-neutral music compared to exhaustive treadmill running in a no-music control condition (see also [20]). This observation suggests that the motivational quality of music may be less important than the prominence of the beat of music and the degree to which participants are able to synchronize their movements to this beat [21]. However, the latter was not quantified. This study aims to examine the relative effects of auditory- motor synchronization and motivational quality of music on running performance. Therefore, participants will perform three running-to-exhaustion conditions on a treadmill: 1) a control condition without music, 2) a metronome condition with a sequence of beeps matching participants’ cadence (synchroniza- tion) and 3) a music condition with synchronous motivational music matching participants’ cadence (synchronization+motiva- tion). The effects of condition will be quantified using a set of complementary ergogenic (time to volitional exhaustion), psycho- physical (ratings of perceived exertion), physiological (heart rate), and behavioral (cadence consistency) outcome measures. The latter measure is included to assess the degree to which cadence corresponded with the rhythmic acoustic stimuli (i.e., metronome, music), a methodological advancement to quantify auditory-motor synchronization that has not been embraced by past studies on performance-enhancing effects of synchronous music [18,19,20,21]. An increased time to exhaustion is expected for running with acoustic stimuli (i.e., both metronome and motiva- tional music conditions) in comparison with the control condition without acoustic stimuli. This effect is expected because acoustic stimuli with a tempo matching participants’ cadence promote auditory-motor synchronization, which will likely enhance running efficiency through a more consistent cadence. We further expect a longer time to exhaustion for the motivational music condition (synchronization+motivation) than for the metronome condition (synchronization) because music not only provides a stimulus for auditory-motor synchronization but also possesses motivational qualities that may further enhance running performance. Methods Participants To establish sample size, a power analysis for a repeated- measures design was conducted using G*Power 3.1.6 (cf. [25]). Based on the effect sizes reported in comparable studies (e.g., g2 p = 0.38 [19], g2 p = 0.24 [20]), the analysis indicated that minimally 16 participants for an a of 0.05 and a power of 0.80 would be required. We recruited 19 students (10 males and nine females) from the Faculty of Human Movement Sciences, VU University Amsterdam, to participate in the study (age: 22.5 years of age, range 19–27 years; height: 180 cm, range 163–198 cm; weight: 69 kg, range 50–82 kg). Participants were recreational runners in good physical condition. Ethics Statement The research met all applicable standards for the ethics of experimentation and was approved by the Ethics Committee of the Faculty of Human Movement Sciences of VU University Amsterdam (ECB 2010-02). Participants provided written in- formed consent prior to the experiment. Experimental Procedure and Setup Participants reported to the laboratory on three occasions at the same time of day, at least 48 h and up to one week apart. In the pre-experimental phase of the first session, participants were acquainted to the laboratory setting and to treadmill running. Subsequently, belt speed was progressively increased every 30 seconds from 9 km/h upwards in a 5-min protocol. At each speed plateau, participants were asked to estimate how long they could maintain running at that speed. The belt speed increments depended on participants’ estimates: +3 km/h for .120 min, +2 km/h for 60–120 min, +1 km/h for 30–60 min, and +0.5 km/ h for 15–30 min. The test speed was defined as the speed that Auditory-Motor Synchronization to Improve Running PLOS ONE | www.plosone.org 2 August 2013 | Volume 8 | Issue 8 | e70758 participants perceived they would be able continue running at for 7–15 min. On average, test speed was 13.25 km/h (range: 9.5– 17.5 km/h). As soon as the test speed was defined, participants completed this 5-min protocol at 9 km/h. Subsequently, partic- ipants rested for 15 min until the start of the experiment proper, which was similar for all three sessions. The experiment (see Figure 1) started with a 3-min warm-up phase in which participants ran at 9 km/h. In the subsequent observational phase, belt speed was increased to the test speed and participants’ cadence was determined by counting strides for a 1- min interval in the first two minutes at the test speed. Then, participants were instructed to run to volitional exhaustion in one of three experimental conditions: 1) a control condition without acoustic stimuli, 2) a metronome condition, 3) a motivational music condition. In the latter two conditions, participants were additionally advised to synchronize their steps to the beat of the acoustic stimuli, yet to give priority to running as long as possible. Experimenters did not encourage participants. The test ended when the participant gave a stop signal. Prior to the motivational music condition, participants were invited to select a song from a motivational music Top 5 (see Table 1). This Top 5 was created as follows. First, a panel of 71 students from VU University Amsterdam were asked to list three artists or bands that produce motivational music for high-intensity sports, like running at high intensity. Then, from each of the top- 10 listed performers, two fast songs were selected with a beat per minute (bpm) of 130 bpm or higher following evidence-based recommendations for motivational music (cf. [23]). The motiva- tional quality of the 20 selected songs was rated by an independent panel consisting of four students (mean age: 21.75 years, range 21– 22 years) using the 6-item BMRI-2 [24]). The Top 5 comprised the songs that received the highest average BMRI-2 rating score (cf. Table 1). The tempo of the motivational music and the metronome was matched to the participant’s cadence, with a beat for each footfall to enhance auditory-motor synchronization (cf. [26,27]). To this end, we used disk-jockey software that enabled us to alter the tempo of the motivational music without changing other aspects of the music (Virtual DJ Pro, Atomix Productions) and a digital metronome (Metronome Plus 2.0.0.1, M & M - Systeme), respectively. Acoustic stimuli were played using a stereo system (Akai QX5690UFX micro music system) at a standardized 80– 84 dB volume (verified with Extech HD600), which is loud but still within acceptable noise levels for working environments [28]. Participants performed the three conditions in counterbalanced order on separate occasions. For the exhaustion phase of the experiment we recorded for all three conditions: 1) time to exhaustion (TTE in seconds) using a stopwatch (Oregon Scientific C510 Digital Stopwatch); 2) heart rate every five seconds using a heart rate monitor (Polar S610); 3) rating of perceived exertion (RPE) every minute using Borg’s 15-grade scale positioned at eye- level in front of the treadmill [29]; and 4) cadence on a stride-by- stride basis using a footswitch sensor placed under the left shoe (sampling rate 500 Hz; MA-153 Event Switches, Motion Lab Systems, Baton Rouge, USA). Data Preparation, Outcome Measures and Statistical Analysis We selected the RPE value and the mean heart rate (averaged over the 12 samples) corresponding to the first, central, and final 1- minute segment of the exhaustion phase for each trial for further statistical analyses. The data collected with the footswitch sensor were processed using custom-written Matlab software. After determining event onsets from the footswitch-sensor data, the inverse of event onset intervals was taken to reconstruct cadence time series. To increase the reliability of cadence estimates, they were extracted from n moving windows containing 19 intervals (i.e., 20 strides) from which the average cadence was taken (in steps/min). From the so-obtained set of n average-cadence observations, we quantified cadence consistency by taking the Figure 1. Schematic overview of the experimental design, the experimental phases, and the corresponding timeline. Note that control, metronome, and music conditions are performed in counterbalanced order on separate days, at least 48 h up to one week apart. TTE represents time to volitional exhaustion, which may vary across conditions. doi:10.1371/journal.pone.0070758.g001 Table 1. Performer and song title of the motivational music Top 5. Performer Song title bpm BMRI-2 #selected @bpm Black Eyed Peas Pump It 153.62 32.50 2 181 The Prodigy Omen 140.00 31.00 7 173 DJ Tie¨sto He’s A Pirate 140.01 30.00 3 178 Red Hot Chili Peppers Higher Ground 140.78 29.75 5 171 David Guetta feat. Juliet Do Something Love 134.00 29.75 2 158 bpm indicates the song’s tempo in beats per minute as verified using Virtual DJ Pro (Atomix Productions), BMRI-2 scores the song’s motivational quality, #selected represents the number of times that the song was selected by participants and @bpm indicates the average played tempo of the song to match participants’ cadence. doi:10.1371/journal.pone.0070758.t001 Auditory-Motor Synchronization to Improve Running PLOS ONE | www.plosone.org 3 August 2013 | Volume 8 | Issue 8 | e70758 mean absolute difference between each of the n average-cadence observations and the average cadence of the observation phase preceding the exhaustion phase of the experiment (cf. Figure 1). Note that the tempo of the acoustic stimuli was also based on the observed cadence recorded during the observation phase. The so-obtained TTE, cadence-consistency, RPE, and heart- rate data were first checked for univariate outliers (using an absolute z-score criterion of 3.29) as well as normality to ensure that parametric analyses were appropriate. Repeated-measures ANOVAs with the within-subjects factor Condition (three levels: control, metronome, music) were performed for TTE and cadence consistency, with post-hoc paired-samples t-tests for significant main effects. Repeated-measures ANOVAs with the within- subjects factors Condition (three levels) and Segment (three levels: first, central, final 60-sec segments of the exhaustion phase) were performed for RPE and heart rate, again with post-hoc paired- samples t-tests for significant main or interaction effects. In view of the fact that participants are expected to run longer in the two acoustic stimuli conditions than in the control condition, we furthermore conducted Condition6Time repeated-measures ANO- VAs for RPE and heart-rate data. To this end, RPE and heart-rate values of the metronome and music conditions were anchored to the specific time points of the first, central, and final 60-sec segments of the control condition. Degrees of freedom were adjusted when the sphericity assumption was violated, using either Huynh-Feldt (if Greenhouse-Geisser e.0.75) or Greenhouse- Geisser (if Greenhouse-Geisser e,0.75) adjustments [30]. Partial eta squared (g2 p) was used to determine effect size. Because effects of conditions on RPE and heart rate are generally subtle in nature, we also determined Cohen’s d effect sizes for both acoustic stimuli conditions against the control condition (see Terry et al. [21]). Results The statistical analyses for cadence consistency were based on 16 out of 19 participants as the data from three participants were not available for the ANOVA due to a malfunctioning footswitch sensor in one or more conditions. Time to Exhaustion was Longer with than without Acoustic Stimuli The time to exhaustion differed significantly across conditions (F(2, 36) = 5.05, p = 0.012, g2 p = 0.219; Figure 2a). Compared to the control condition (TTE = 624 seconds), participants ran signifi- cantly longer with acoustic stimuli (metronome: TTE = 746 seconds, t(18) = 2.97, p = 0.008, music: TTE = 733 seconds, t(18) = 2.43, p = 0.026). The time to exhaustion did not differ significantly between metronome and music conditions (t(18) = 0.318, p = 0.75). Cadence was most Consistent in the Metronome Condition Cadence consistency differed significantly across conditions (F(2, 30) = 3.84, p = 0.033, g2 p = 0.204; Figure 2b). Cadence consistency differed significantly between control (4.33 steps/min) and metronome (2.90 steps/min, t(15) = 2.30, p = 0.036) conditions. The difference in cadence consistency between control (4.33 steps/ min) and music (3.17 steps/min) conditions was non-significant (t(15) = 1.78, p = 0.095). Cadence consistency did not differ significantly between metronome and music conditions (t(15) = 0.876, p = 0.39). Perceived Exertion and Heart Rate Revealed that Participants Ran to Exhaustion in all Conditions A significant main effect of Segment was observed for both RPE and heart rate (F(1.15, 20.68) = 70.94, p,0.001, g2 p = 0.798 and F(1.09, 19.57) = 51.09, p,0.001, g2 p = 0.739, respectively). Post- hoc analysis for RPE and heart rate revealed significant differences between each segment (all t(18)9s.6.58, all p9s ,0.001; Figure 3). No main or interaction effects involving the Condition factor were observed (all F9s ,1.99, all p9s.0.151, all g2 p9s ,0.100). Nevertheless, close inspection of Figure 3 suggests that RPE and heart rate tend to vary somewhat across conditions. Indeed, small- to-moderate reductions in RPE were observed for the first (metronome: d = 20.30; music: d = 20.45) and the central segment (metronome: d = 20.21; music: d = 20.43), but not for the final segment (metronome: d = 0.13; music: d = 20.10; see also Figure 3a). Furthermore, small-to-moderate increments in heart rate were observed for the final (metronome: d = 0.28; music: d = 0.49) and the central segment (metronome: d = 0.17; music: d = 0.35), but not for the first segment (metronome: d = 0.02; music: d = 0.20; see also Figure 3b). When anchoring RPE values of the metronome and music conditions to the time points of the three segments of the control condition, we observed a significant main effect for Condition (F(1.88, 33.91) = 7.708, p = 0.002, g2 p = 0.300), with post-hoc analyses revealing that RPE was significantly lower in the music condition (15.9) than in the control condition (17.1; t(18) = 1.281, p = 0.02). RPE for the metronome condition (16.4) did not differ significantly from control and music conditions (t(18) = 0.754, p = 0.08 and t(18) = 0.526, p = 0.165, respectively). Furthermore, a significant main effect for Time was observed (F(2, 36) = 70.079, p,0.001, g2 p = 0.796); similar to the main analyses, RPE values increased significantly as a function of time (all t(18)9s.2.018, p9s ,0.001). The Condition 6 Time interaction was non-significant (F(2.26, 40.66) = 1.014, p = 0.380, g2 p = 0.053). Anchoring heart- rate values of metronome and music conditions to the time points of the three segments of the control condition again resulted in a significant main effect Time (F(1.04, 18.66) = 44.61, p,0.001, g2 p = 0.713; heart rate increased significantly as a function of time [all t(18)9s.3.25, p9s ,0.001]). Main and interaction effects involving the factor Condition were non-significant (F(2, 36) = 0.814, p = 0.451, g2 p = 0.043; F(2.00, 35.99) = 1.946, p = 0.158, g2 p = 0.098, respectively). Discussion The current experiment sought to examine the relative effects of the motivational quality of music and auditory-motor synchroni- zation with the beat on running performance. Participants ran to exhaustion on a treadmill without acoustic stimuli (control condition), with a metronome beat matched to participants’ cadence (synchronization), and with motivational music with a beat that matched participants’ cadence (synchronization+motiva- tion). Time to exhaustion differed significantly across conditions. In line with our hypothesis we found that the time to exhaustion was longer in the metronome and motivational music conditions than in the control condition without acoustic stimulation. Specifically, participants ran approximately two minutes longer with acoustic stimuli in comparison with a control condition (Figure 2a). Comparable effects have recently been reported for elite triathletes who were instructed to run to self-selected synchronous motivational music, synchronous oudeterous music (i.e., motivationally neutral music) and in a no-music control Auditory-Motor Synchronization to Improve Running PLOS ONE | www.plosone.org 4 August 2013 | Volume 8 | Issue 8 | e70758 condition ([21] see also [19] for a similar study on treadmill walking to exhaustion). Time to exhaustion was longer with than without music, regardless of its motivational quality. In combina- tion, these results suggest that the motivational quality of music Figure 2. TTE in seconds (A) and cadence consistency in steps/min (B) data of the exhaustion phase for control (black), metronome (dark gray), and motivational music (light gray) conditions. Error bars represent the standard error while asterisks indicate significant differences across conditions. doi:10.1371/journal.pone.0070758.g002 Figure 3. Perceived exertion (A) and heart rate (B) data for the first, central, and final 1-minute segments of the exhaustion phase for control (black), metronome (dark gray), and motivational music (light gray) conditions. Error bars represent the standard error. RPE and heart rate of each segment differed significantly from each other. doi:10.1371/journal.pone.0070758.g003 Auditory-Motor Synchronization to Improve Running PLOS ONE | www.plosone.org 5 August 2013 | Volume 8 | Issue 8 | e70758 may be less important than the prominence of the beat of music, which allows participants to synchronize their pace to the prescribed tempo of the acoustic stimulus. We additionally expected differential effects of the two acoustic stimuli conditions on time to exhaustion in view of the clear difference in motivational quality between metronome and motivational-music conditions. However, this was not the case (Figure 2a). In the following, we will discuss motivation, synchronization, and dissociation effects associated with the two types of acoustic stimuli on psychophysical (e.g., perceived exertion), physiological (e.g., heart rate), and behavioral (e.g., cadence consistency) outcome measures to explain why we did not find the expected superior time to exhaustion for the synchronous motivational music condition (i.e., synchronization+motivation). Psychophysical and Physiological outcome Measures are Affected more by Motivational Music than by Metronomes With regard to psychophysical outcome measures, Terry and colleagues [21] recently reported evidence for lower perceived exertion during sub-maximal running to neutral and motivational music compared to a control condition without music (Cohen’s d- values ranging from 0.19 to 0.39). In the present study, ratings of perceived exertion varied significantly with Segment (Figure 3a). Similar to Terry et al. [21], we found indications for small-to- moderate reductions in perceived exertion with acoustic stimuli, particularly for the first 1-minute segment of sub-maximal running, and most prominently, for motivational music (i.e., d = 20.45). This segment-dependent effect is in agreement with previous studies, which indicate that motivational music reduced perceived exertion for sub-maximal intensities [19,31] but not for maximal intensities [22,32,33]. We further observed that the physiological outcome measure heart rate was affected by acoustic stimuli. In the final, near-maximal segment (Figure 3b), small-to- moderate increments in heart rate were observed with acoustic stimuli, particularly during the final 1-minute segment and again most prominently for the motivational music condition (i.e., d = 0.49). The combined effects of acoustic stimuli on psychophysical and physiological outcome measures suggest that for a given sub- maximal heart rate (as observed for the first segment) the presence of acoustic stimuli lowered participants’ perceived exertion, a finding in line with Terry and colleagues [21]. Interestingly, for a given (near-)maximal perceived exertion (as observed for the final segment) the presence of acoustic stimuli may have helped to elevate the attainable physiological load (i.e., higher heart rate). With acoustic stimuli, and in particular with motivational music, participants appeared to be able to work at a higher intensity (higher heart rate at the final segment) for longer (increased TTE) at a comparably high-level rating of perceived exertion. When RPE values were anchored to the time points of the three segments of the control condition, we indeed found that attending to acoustic stimuli reduced perceived exertion considerably, again most prominently for the motivational music condition. This effect was present at all three time points of the running-to-exhaustion phase. Consistent with previous studies on the control of physiological strain during strenuous endurance exercises (e.g., [34]), these findings suggest that athletes actively regulate their relative physiological strain, that is, relative to their perceived exertion. Assuming similar auditory-motor synchronization effects for both acoustic pacing conditions, one would therefore expect a superior effect of the synchronous motivational music condition on the time to exhaustion (allowing participants to work harder for longer) because motivational music had a stronger effect on physiological and psychophysical outcome measures than the metronome given the evident difference in motivational quality between metronome beeps and motivational music. This was, however, not the case, implying that other performance enhancing factors were involved that were - as discussed in the following two paragraphs - seemingly: 1) more effective for the metronome condition than for the motivational music condition (i.e., synchronization); and 2) also effective for the metronome condition (i.e., dissociation). Cadence is more Consistent for Running with a Metronome than for Running with Motivational Music For the metronome condition, running cadence was most consistent, as evidenced by a significant difference in cadence consistency between control and metronome conditions. En- hanced cadence consistency may help to improve running economy because energy loss associated with accelerations and decelerations in cadence is reduced. Moreover, the runner is forced to maintain the right cadence, that is, the cadence that was adopted by the runner in the pre-experimental phase prior to the exhaustion phase of the trial (see Figure 1), which is likely to be near to the optimal running cadence for the imposed running speed. Recently, Bacon et al. [18] reported that a cyclic exercise was performed more efficiently when executed synchronously with music than when the musical tempo was set slightly slower than the cyclical movement rate. Unlike the current study, in which we quantified cadence consistency, Bacon et al. [18] did not register the revolutions per minute. Hence it is unclear whether the enhanced cycling efficiency with synchronous music was due to: 1) differences in the consistency of the movement relative to the beat of the music, 2) differences in movement tempo between slower, synchronous, and faster music conditions, or 3) an interaction between these two factors. As far as we know, the current study is the first to demonstrate that movement consistency was signifi- cantly affected by acoustic stimuli. Specifically, we showed that the type of acoustic stimuli affected auditory-motor synchronization. That is, running cadence was most consistent in the presence of a metronome beat that was matched to the runners’ preferred cadence at the imposed speed. In contrast, the difference in cadence consistency between control and motivational music conditions only tended toward significance (p = 0.095). This is presumably due to the retrospective observation that the beat was not as constant, apparent, and prominent throughout the song as the beat in a sequence of metronome beeps (see Figure 4). Hence, from a research point of view (but see also the Practical Recommendations section), a limitation in the study may be that, despite tempo matching, the motivational music only provided a sub-optimal template for auditory-motor synchronization. As a consequence, running cadence was more variable. It is important to note, however, that this is an inherent feature of music. Dissociation through Auditory-Motor Synchronization Another well-known way in which motivational music may influence running performance is by narrowing attention, specif- ically by diverting it from running-induced feelings of fatigue and discomfort [13,35]. Focusing attention on motivational music for its distraction effect is a known and effective dissociation technique, especially in athletes who prefer to be distracted from physiological signals in shaping their performance (i.e., so-called dissociators [36]). The idea behind dissociation is that people can only process a limited amount of information at any given time. Thus, dissociation induced by focusing on motivational music, including its lyrical content, may alter the perception of effort, Auditory-Motor Synchronization to Improve Running PLOS ONE | www.plosone.org 6 August 2013 | Volume 8 | Issue 8 | e70758 allowing runners to work harder for longer [13]. Likewise, the very act of auditory-motor synchronization may also contribute to dissociation because auditory-motor synchronization is known to be an attention demanding process [3]. Peper and colleagues [37], for example, recently employed a stimulus-response reaction-time task to quantify the attentional costs of walking with and without acoustic pacing. They showed that reaction times were signifi- cantly longer with acoustic pacing, emphasizing the elevated attentional demands for auditory-motor synchronization (see also [38]). The same may be true for running with acoustic stimulation with a beat that matches the participant’s preferred cadence, regardless of the motivational quality of the stimuli (viz. motivational music vs. metronome). This is an area of research that deserves more attention, especially for its profound practical implications in enhancing performance in sport, exercise, and rehabilitation settings. Practical Recommendations Given that both the motivational quality of music and the beat of acoustic stimuli appear to have an effect on running performance, it is of practical importance to optimize both of these aspects. Indeed, this is exactly what we did in the motivational music and metronome conditions of the present study. With regard to the former, following existing research findings, we selected loud, fast, percussive music with accentuated bass frequencies [23]. We employed the validated and objective BMRI-2 [24] to optimize the selection of motivational music, resulting in a Top 5 (see Table 1) based on the motivational music preferences for high-intensity sports of a large student panel. A limitation of this procedure is that we could not guarantee that participants actually liked the selected music, which might have diminished the motivational effect of the selected music. However, we chose this procedure because it allowed us to select songs with at least 130 bpm, thus adhering to evidence-based recommenda- tions for motivational music (cf. [23]). Note that in previous research, the music selection in the study of Terry et al. (online appendix in [21]) may have been suboptimal because in both neutral and motivational music conditions, the beats per minute of the music were adjusted to the stride frequency of the participant (i.e., a beat for one step per stride), which is much slower than recommended for motivational music [23]. In contrast, in the present study, we adjusted the beats per minute of the music to the step frequency of the participant (i.e., a beat for both steps per stride), resulting in a much faster beat, as recommended for motivational music [23]. A second benefit of adjusting the tempo of the acoustic stimuli to the cadence of the runner was that it created optimal conditions for auditory-motor synchronization in general [3,25,39] and for bipedal locomotion in particular [9,26,40], where the beat should pace both footfalls per stride [26] and match the preferred cadence [9,40]. For the metronome condition, these recommendations for optimal auditory-motor synchronization was easily fulfilled by creating a regular beat sequence. In contrast, in the motivational music condition, we successfully modified the average tempo of the song. However, this modification did not necessarily mean that the resultant beat was constant, readily apparent, and prominent throughout the song due to, for example, the fact that musical intermezzos and other tempo irregularities are inherent to music (cf. Figure 4). Thus, even though the beats per minute of the motivational music were adjusted to match the preferred cadence of the runner, the music still may have provided a less effective reference for synchronization than the metronome. That is, in the latter condition, the beat was always prominent, constant, and readily apparent throughout the running-to-exhaustion phase of the experiment. Figure 4. Spectrograms of acoustic stimuli, both at 140 bpm, as indicated by the dashed white line. The ‘hotter’ the color, the more prominent the beat. (A) Spectrogram of the metronome, showing a constant beat. (B) Spectrogram of the motivational music track ‘He’s A Pirate’ by DJ Tie¨sto, where the tempo is not as constant, readily apparent, and prominent throughout the song as the beat in the metronome. doi:10.1371/journal.pone.0070758.g004 Auditory-Motor Synchronization to Improve Running PLOS ONE | www.plosone.org 7 August 2013 | Volume 8 | Issue 8 | e70758 Conclusions Motivational quality of music positively influences perceived exertion of sub-maximal running intensity as well as heart rates for a (near-)maximal running intensity, which may enhance running performance because it allows runners to work harder. Next to motivational quality, the music’s beat may help runners to maintain a consistent pace if they couple their cadence to the prescribed tempo of the acoustic stimulus, which may enhance running performance by helping athletes work more efficiently. Therefore, running to motivational music with a very prominent and consistent beat that is matched to the runner’s preferred cadence will likely yield optimal effects because it helps elevate physiological strain at a very high perceived exertion, while the consistent and correct cadence induced by auditory-motor synchronization helps to facilitate running economy. Motivational music with the right beat may therefore help runners to work harder and more efficiently, which is likely to enhance their running performance. Author Contributions Conceived and designed the experiments: RB MN JK MR. Performed the experiments: RB MN. Analyzed the data: RB MN MR. Contributed reagents/materials/analysis tools: MR. Wrote the paper: RB MN JK MR. References 1. Miura A, Kudo K, Ohtsuki T, Kanehisa H (2011) Coordination modes in sensorimotor synchronization of whole-body movement: a study of street dancers and non-dancers. Hum Mov Sci 30: 1260–1271. 2. Large EW (2000) On synchronizing movements to music. Hum Mov Sci 19: 527–566. 3. Repp B (2005) Sensorimotor synchronization: A review of the tapping literature. Psychon Bull Rev 12: 969–992. 4. Wilson EM, Davey NJ (2002) Musical beat influences corticospinal drive to ankle flexor and extensor muscles in man. Int J Psychophysiol 44: 177–184. 5. Zentner M, Eerola T (2010) Rhythmic engagement with music in infancy. Proc Natl Acad Sci U S A 107: 5768–5773. 6. Patel A (2008) Music, Language, and the Brain. New York, NY: Oxford University Press. 7. Zatorre RJ, Chen JL, Penhune VB (2007) When the brain plays music: Auditory-motor interactions in music perception and production. Nat Rev Neurosci 8: 547–558. 8. Thaut MH (2008) Rhythm, Music, and the Brain: Scientific Foundations and Clinical Applications. New York, NY: Routledge. 9. Roerdink M, Bank PJM, Peper CE, Beek PJ (2011) Walking to the beat of different drums: Practical implications for the use of acoustic rhythms in gait rehabilitation. Gait Posture 33: 690–694. 10. Karageorghis CI, Priest DL (2012) Music in the exercise domain: a review and synthesis (Part I). Int Rev Sport Exerc Psychol 5: 44–66. 11. Karageorghis CI, Priest DL (2012) Music in the exercise domain: a review and synthesis (Part II). Int Rev Sport Exerc Psychol 5: 67–84. 12. Bishop DT (2010) ‘Boom Boom How’: Optimising performance with music. Sport Exerc Psychol Rev 6: 35–47. 13. Karageorghis CI, Terry PC (2011) Inside sport psychology. Champaign, IL: Human Kinetics. 14. Moens B, van Noorden L, Leman M (2010) D-Jogger: syncing music with walking. Available: http://smcnetwork.org/files/proceedings/2010/66.pdf. Ac- cessed 2013 Jan 08. 15. van der Vlist B, Bartneck C, Ma¨ueler S (2011) moBeat: Using interactive music to guide and motivate users during aerobic exercising. Appl Psychophysiol Biofeedback 36: 135–145. 16. Wijnalda G, Pauws S, Vignoli F, Stuckenschmidt H (2005) A personalized music system for motivation in sport performance. IEEE Pervasive Comput 4: 26–32. 17. Westerink JH, Claassen A, Schuurmans T, IJsselsteijn W, de Kort Y, et al. (2011) Runners’ experience of implicit coaching through music. In: Westerink J, Krans M, Ouwerkerk M, editors. Sensing emotions: The impact of context on experience measurements. Philips Research Book Series 12, 121–134. Springer Book Archives, ISBN 978-90-481-3258-4. 18. Bacon CJ, Myers TR, Karageorghis CI (2012) Effect of music-movement synchrony on exercise oxygen consumption. J Sports Med Phys Fitness 52: 359– 365. 19. Karageorghis CI, Mouzourides D, Priest DL, Sasso T, Morrish D, et al. (2009) Psychophysical and ergogenic effects of synchronous music during treadmill walking. J Sport Exerc Psychol 31: 18–36. 20. Simpson SD, Karageorghis CI (2006) The effects of synchronous music on 400- m sprint performance. J Sports Sci 24: 1095–1102. 21. Terry PC, Karageorghis CI, Saha AM, D’Auria S (2012) Effects of synchronous music on treadmill running among elite triathletes. J Sci Med Sport 15: 52–57. 22. Boutcher SH, Trenske M (1990) The effects of sensory deprivation and music on perceived exertion and affect during exercise. J Sport Exerc Psychol 12: 167– 176. 23. Karageorghis CI, Terry PC, Lane AM, Bishop DT, Priest DL (2012) The BASES Expert Statement on use of music in exercise. J Sports Sci 30: 953–956. 24. Karageorghis CI, Priest DL, Terry PC, Chatzisarantis NLD, Lane AM (2006) Development and validation of an instrument to assess the motivational qualities of music in exercise: The Brunel Music Rating Inventory-2. J Sports Sci 24: 899– 909. 25. Faul F, Erdfelder E, Lang AG, Buchner A (2007) G*Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behav Res Methods 39: 175–191. 26. Kudo K, Park H, Kay BA, Turvey MT (2006) Environmental coupling modulates the attractors of rhythmic coordination. J Exp Psychol Hum Percept Perform 32: 599–609. 27. Roerdink M, Lamoth CJC, van Kordelaar J, Elich P, Konijnenbelt M, et al. (2009) Rhythm perturbations in acoustically paced treadmill walking after stroke. Neurorehabil Neural Repair 23: 668–678. 28. Williams W (2005) Noise exposure levels from personal stereo use. Int J Audiol 44: 231–236. 29. Borg G (1998) Borg’s perceived exertion and pain scales. Champaign, IL: Human Kinetics. 30. Girden ER (1992) ANOVA: repeated measures. Newbury Park, CA: Sage Publications. 31. Nethery VM (2002) Competition between internal and external sources of information during exercise: Influence on RPE and the impact of the exercise load. J Sports Med Phys Fitness 42: 172–178. 32. Schwartz SE, Fernhall B, Plowman SA (1990) Effects of music on exercise performance. J Cardiopulm Rehabil 10: 312–316. 33. Tenenbaum G, Lidor R, Lavyan N, Morrow K, Tonnel S, et al. (2004) The effect of music type on running perseverance and coping with effort sensations. Psychol Sport Exerc 5: 89–109. 34. Esteve-Lanao J, Lucia A, de Koning JJ, Foster C (2008) How do humans control physiological strain during strenuous endurance exercise? PLoS ONE doi:10.1371/journal.pone.0002943. 35. Karageorghis CI, Terry PC (1997) The psychophysical effects of music in sport and exercise: A review. J Sport Behav 20: 54–68. 36. Morgan WP, Pollock ML (1977) Psychological characterization of the elite distance runner. Ann NY Acad Sci 301: 382–403. 37. Peper CE, Oorthuizen JK, Roerdink M (2012) Attentional demands of cued walking in healthy young and elderly adults. Gait Posture 36: 378–382. 38. Lamoth CJC, Roerdink M, Beek PJ (2007) Acoustically-paced treadmill walking requires more attention than unpaced treadmill walking in healthy young adults. Gait Posture 26: S96–S97. 39. Roerdink M, Ridderikhoff A, Peper CE, Beek PJ (2012) Informational and neuromuscular contributions to anchoring in rhythmic wrist cycling. Ann Biomed Eng, In press. 40. Roerdink M, Lamoth CJC, Kwakkel G, van Wieringen PCW, Beek PJ (2007) Gait coordination after stroke: Benefits of acoustically paced treadmill walking. Phys Ther 87: 1009–1022. Auditory-Motor Synchronization to Improve Running PLOS ONE | www.plosone.org 8 August 2013 | Volume 8 | Issue 8 | e70758

The power of auditory-motor synchronization in sports: enhancing running performance by coupling cadence with the right beats.

08-07-2013

Bood, Robert Jan,Nijssen, Marijn,van der Kamp, John,Roerdink, Melvyn

eng

PMC7763525

International Journal of Environmental Research and Public Health Article Maximal Time Spent at VO2max from Sprint to the Marathon Claire A. Molinari 1,2, Johnathan Edwards 3 and Véronique Billat 1,* 1 Unité de Biologie Intégrative des Adaptations à l’Exercice, Université Paris-Saclay, Univ Evry, 91000 Evry-Courcouronnes, France; [email protected] 2 BillaTraining SAS, 32 rue Paul Vaillant-Couturier, 94140 Alforville, France 3 Faculté des Sciences de la Motricité, Unité d’enseignement en Physiologie et Biomécanique du Mouvement, 1070 Bruxelles, Belgium; [email protected] * Correspondence: [email protected]; Tel.: +33-(0)-786117308 Received: 6 November 2020; Accepted: 8 December 2020; Published: 10 December 2020



 Abstract: Until recently, it was thought that maximal oxygen uptake (VO2max) was elicited only in middle-distance events and not the sprint or marathon distances. We tested the hypothesis that VO2max can be elicited in both the sprint and marathon distances and that the fraction of time spent at VO2max is not significantly different between distances. Methods: Seventy-eight well-trained males (mean [SD] age: 32 [13]; weight: 73 [9] kg; height: 1.80 [0.8] m) performed the University of Montreal Track Test using a portable respiratory gas sampling system to measure a baseline VO2max. Each participant ran one or two different distances (100 m, 200 m, 800 m, 1500 m, 3000 m, 10 km or marathon) in which they are specialists. Results: VO2max was elicited and sustained in all distances tested. The time limit (Tlim) at VO2max on a relative scale of the total time (Tlim at VO2max%Ttot) during the sprint, middle-distance, and 1500 m was not significantly different (p > 0.05). The relevant time spent at VO2max was only a factor for performance in the 3000 m group, where the Tlim at VO2max%Ttot was the highest (51.4 [18.3], r = 0.86, p = 0.003). Conclusions: By focusing on the solicitation of VO2max, we demonstrated that the maintenance of VO2max is possible in the sprint, middle, and marathon distances. Keywords: VO2max; performance; running 1. Introduction Classically, the solicitation of the maximal uptake of oxygen (VO2max) was thought only to be possible in the middle-distance (1500 m) events, and not the sprint or the marathon distances [1]. (1) Power output may be high (greater than critical speed), but insufficient to elicit VO2max (i.e., the average marathon speed). (2) Power may be very high or maximal, and sufficient to drive VO2 to its maximum before exhaustion (i.e., middle-distance events). (3) Power may be extremely high, such that the subject becomes exhausted before sufficient time has elapsed for VO2 to reach its maximum (i.e., sprint events) [2]. This classification is the basis of the century-old constant-speed paradigm applied in laboratories since the discovery of VO2max by AV Hill in 1923 [3]. Today, innovative technologies such as the portable breath-by-breath gas exchange systems allows researchers to investigate the solicitation of VO2max during 100 and 200 m sprints in elite runners. By assessing the fundamental physiology, it has been shown that the change in tissue oxygen uptake is directly proportional to changes in creatine (Cr) content [4]. This close reciprocal relationship between pulmonary VO2 and phosphocreatine (Pcr) has been demonstrated at the systemic level during high-intensity constant power output exercises [5]. Hence, there is a close relationship between oxygen uptake kinetics and changes in Cr/Pcr ratios. Int. J. Environ. Res. Public Health 2020, 17, 9250; doi:10.3390/ijerph17249250 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2020, 17, 9250 2 of 11 The rapid depletion of creatine phosphate during a sprint may be a signal for a rapid increase in VO2 and possibly until VO2max. Therefore, our first hypothesis is that VO2max can be reached during a sprint, but also that the relative time spent at VO2max may be of the same order during middle distances, and possibly a discriminant factor of performance. The marathon is the longest Olympic endurance distance. Previous research has estimated that the marathon only elicits a fractional utilization of VO2max [6]. However, technological advances now allow breath-by-breath VO2 measurements during an entire marathon. In the past, it was only possible to measure VO2 over 1 or 2 km using Douglas bags from the back of a moving vehicle, as performed by Michael Maron. These pioneering experiments highlighted marathon training and performance, as he showed that VO2max was reached during the marathon and our research confirms his results. Indeed, the paradigm of constant (constant vs average) velocity still endures today as determined by the ratio of energy output and the cost of running [6]; this all comes from the treadmill experiments of constant speed physiology. It is generally thought that VO2max is not elicited in the marathon and that it must be run below maximal aerobic speed (vVO2max) in order to maintain a sub lactate threshold VO2 steady state [7,8]. One obvious consequence of the slow component response is that it creates a range of velocities, all which elicit VO2max, provided the exercise is continued to exhaustion. VO2max can be elicited during constant power exercise, over a range of intensities that may be higher or lower than the minimum value for which it occurs during incremental exercise [9]. Maron’s pioneering research reported that VO2max could be elicited during a marathon; however, we did not have portable gas exchange measurements to confirm this remarkable result [10]. Today, portable breath-by-breath gas exchange analyzers have minimal measurement delays and can be easily worn in competition. The plateau in VO2 at the end of an incremental exercise test is used as an important criterion to validate that VO2max has been achieved [6]; however, the duration that subjects can sustain that plateau has largely been ignored. The time limit at PVO2max (Tlim@PVO2max), while reproducible, has been reported to be highly variable between subjects (3–8 min) [11]; it is negatively correlated with PVO2max and VO2max but positively correlated with the maximal oxygen deficit, which is an index of the ability to generate energy from anaerobic metabolism (i.e., anaerobic capacity) [12,13]. Hence, while debates continue around the central versus peripheral limiting factors of VO2max [14,15], the limiting factors of VO2max and of the ability to sustain VO2max remain to be investigated independently of PVO2max [13]. It was shown that VO2max can be sustained for a longer duration when exercise is controlled by the maintenance of VO2max, and that the limiting cardiovascular factors of endurance at VO2max are unrelated to its value. The examination of the time limit at VO2max in different running events is a more ecological approach to the time to plateau at VO2max as it relates to the total time run from sprint to the marathon. Real-world races are not run at constant speeds [16,17], and we wish to reverse the paradigm of power around PVO2max or constant VO2 in order to examine the plateau at VO2max as a common performance factor when expressed as a percentage of total race time. Indeed, the underlying idea is that the greater the energy at VO2max (maximum oxidation rate), the more Adenosine Triphosphate resynthesized from creatine and lactic acid contributes to sprint and marathon performances. Hence, the more relative time run at VO2max, the better the performance, independent of the distance. The concept of relative time to exhaustion at VO2max could be a central energy concept independent of whether the dominant metabolism is aerobic or anaerobic. We hypothesize that this concept could lead to a new method of high intensity interval training that uses very short sprints around the average marathon speed in accordance with the target distance (from 100 to 42,195 m). Therefore, our primary hypothesis is that VO2max can be sustained from the sprint to the marathon and independent of the distance run, the time spent relative to exhaustion at VO2max, as expressed as a percentage of the total performance time, is a discriminant factor for performance. Int. J. Environ. Res. Public Health 2020, 17, 9250 3 of 11 2. Materials and Methods Seventy-eight well-trained male athletes (training 4 days per week) participated in the study (mean ± standard deviation [SD] age: 32 [13]; weight: 73 [9] kg; height: 1.80 [0.8 m]. The participants’ preferred racing distances were as follows: 100 m (n = 13), 200 m (n = 13), 800 m (n = 8), 1500 m (n = 16), 3000 m (n = 9), 10 km (n = 7), and the marathon (n = 12). All of the participants were experienced in their respective full effort race distances and VO2max tests (University of Montreal Track Test, UMTT). All subjects gave their informed consent for inclusion before they participated in the study. The study was conducted in accordance with the Declaration of Helsinki, and the protocol was approved by an independent ethics committee (CPP Sud-Est V, Grenoble, France; reference: 2018-A01496-49). All participants were provided with study information and gave their written consent before participation. All participants performed the University of Montreal Track Test (UMTT), to determine individual VO2max values. After 7 to 14 days, they ran one or two different race simulation efforts in which they are specialists (100 m, 200 m, 800 m, 1500 m, 3000 m, 10 km or the marathon). A portable breath-by-breath sampling system (K5 [18], COSMED Srl, Rome, Italy) that continuously measured respiratory gases (oxygen uptake [VO2], ventilation [VE], and the respiratory exchange ratio) was worn in both the UMTT and race efforts. During the 7 to 14 day period between the UMTT and the running effort, the participants were instructed to continue their training activities as normal. A global positioning system watch (Garmin, Olathe, KS, USA) was used to measure the heart rate and the speed responses (5 s averaged data) of each effort. In the UMTT, the rating of perceived exertion (RPE), on a scale from 6 (least exertion) to 20 (greatest exertion) [19], was recorded 15 s before the end of each stage [20]. 2.1. Determination of Maximal Oxygen Uptake and Velocity Associated with VO2max—The UMTT The UMTT was conducted on a 400 m track with cones placed every 20 m. Pre-recorded sound beeps indicated when the subject needed to be near a cone to maintain the imposed speed. A longer sound marked speed increments. The first step was set to 8.5 km·h−1, with a subsequent increase of 0.5 km·h−1 every minute. When the runner was unable to maintain the imposed pace and thus failed to reach the cone in time for the beep on two consecutive occasions, the test was terminated. The speed corresponding to the last completed step was recorded as the vVO2max (km·h−1). During the UMTT, VO2max was confirmed by a visible plateau in VO2 (≤2 mL·kg−1·min−1) with a standard increase in exercise intensity, and any indicative secondary criteria (visible signs of exhaustion; HRmax ±10 beats·min−1) around the point of volitional exhaustion and an RPE of 19–20. 2.2. Determination of The Time Limit at VO2max (Tlim at VO2max) Oxygen uptake is not a simple function of power output or velocity, for it is a function of time as well. Even steady-state oxygen uptake is not a linear function of power output beyond a certain level [2]. The slow component of oxygen uptake and increasing oxygen cost of exercise at higher powers outputs complicates the issue [21]. The slow component has, however, been successfully modeled, both theoretically [22] and empirically [23], and the energy cost of running can safely be assumed to be constant (or very nearly so) provided the power or velocity range is narrow [2]. Perhaps, then, these difficulties can be largely overcome by considering endurance at a fixed value of oxygen uptake, say at its maximum (VO2max) [2]. This time limit at VO2max depends on the duration of the subject’s exhaustion time (time limit = Tlim) and the time to reach VO2max (TA VO2max), both of which decrease with increasing exercise intensity (Tlim VO2max = Tlim − TA VO2max) [12]. Steady-state VO2 was defined when the subject reached 95% of incremental VO2max [12] during an incremental test. During each race effort, the VO2max Tlim was therefore computed by calculating the difference between the total running time (Tlim) and the time taken to reach 95% incremental VO2max (TA VO2max) [12]. Tlim at VO2max is also defined as the time (seconds) spent at maximal oxygen consumption during the completed distance. Knowing that VO2max was the maximal oxygen consumption during the Int. J. Environ. Res. Public Health 2020, 17, 9250 4 of 11 UMTT (mL·kg−1·min−1), we then processed the data to test the effect of the Tlim VO2max on the relative exercise duration for each distance. We normalized the duration of the run on a relative scale of total time (%Ttot) by comparing the time to the distance. For each effort, the Tlim at VO2max, (assuming that VO2max was reached and maintained) is the Tlim at VO2max%Ttot and is determined to be the ratio between Tlim at VO2max and total time of the effort. 2.3. Calculation of the Intensity of Race in the Percentage of Vvo2max (Intensity of Exercise %Vvo2max) We also calculated exercise intensity (average speed) as a percentage of vVO2max (km·h−1), since it would appear that the factors limiting time spent at VO2max are different depending on whether the intensity is greater or less than vVO2max [13]. 2.4. Statistical Analysis All statistical analyses were performed using XLSTAT software (version 1 January 2019, Addinsoft, Paris, France). For each variable, the normality and homogeneity of the data distribution were examined using a Shapiro–Wilk test. A one-way analysis of variance (ANOVA) was applied to assess the various race distances in terms of performance variables: International Association of Athletics Federations (IAAF) score, running time (s), vVO2max (km·h−1), VO2max (mL·kg−1·min−1), and post-run blood lactate level (mM). A one-way analysis of variance (ANOVA) was also used to assess the time at VO2max and the intensity of exercise. Pearson’s coefficient (r) was used to measure the correlations between performances, Tlim at VO2max%Ttot, and intensity of exercise %vVO2max. 3. Results The descriptive physiological responses in UMTT are summarized in Table 1. Sprinters and 800 m runners have significantly lower VO2max than the middle- and long-distance runners (3000 m and 10 km) (Table 1). There were significant differences in VO2max between participants who ran the 800 m and those who ran the sprints, 3000 m, and 10 km (p < 0.0001, p < 0.0001, and p = 0.0002, respectively). VO2max was significantly higher in the participants who ran the 10 km than in the sprinters and the 3000 m runners (p < 0.0001 and p < 0.0001, respectively). Table 1. Descriptive physiological responses in UMTT. Runners n vVO2max (km·h−1) VO2max (mL·kg−1·min−1) HRmax (Beat·min−1) RPE Last Stage of UMTT 100 m 13 15.4 ± 1.6 53.1 ± 5.5 196.3 ± 4.5 19.5 ± 0.5 200 m 13 15.4 ± 1.6 53.1 ± 5.5 196.3 ± 4.5 19.5 ± 0.5 800 m 8 19.3 ± 0.7 ab 64.6 ± 3.4 ab 196.9 ± 6.4 19.7 ± 0.5 1500 m 16 17.8 ± 2.2 ab 59.0 ± 10.5 188.6 ± 12.6 19.8 ± 0.4 3000 m 9 16.2 ± 1.0 abc 51.1 ± 5.3 cd 181.9 ± 11.7 abc 19.9 ± 0.3 10,000 m 7 19.1 ± 1.8 abe 67.0 ± 6.5 abef 183.4 ± 11.2 abc 19.3 ± 0.5 de 42,195 m 12 17.0± 0.9 abc 55.4 ± 4.7 c 189.1 ± 8.2 abc 19.5 ± 0.5 Abbreviations: VO2max, maximal oxygen consumption; vVO2max, running speed associated with their maximal level of oxygen consumption maximal aerobic velocity; HRmax, maximal heart rate and RPE, rating of perceived exertion and UMTT, University of Montreal Track Test. Note: a indicates a significant difference (p < 0.05) vs. 100 m, b 200 m, c 800 m, d 1500 m, e 3000 m and f marathon. The data are quoted as the mean ± SD. The 100, 200, and 800 m were run at much higher values than their vVO2max (209 ± 25, 206 ± 25, and 116 ± 8% of vVO2max, respectively. p < 0.001). All other distances were run at or below vVO2max, 102, and 80% of vVO2max in the 1500 m and the marathon, respectively (Figure 1). Due to the large difference in relative speed to vVO2max, Tlims at VO2max%Ttot during the sprint, middle-distance, 800 m, and the 1500 m were not significantly different (Table 2). The highest Tlim at VO2max%Ttot was measured in the 3000 m race, while the lowest was measured in the marathon Int. J. Environ. Res. Public Health 2020, 17, 9250 5 of 11 (Figure 2). The 3000 m runners spent their half of the time at VO2max (51 ± 18% of Ttot), while all of the marathon runners all reached VO2max, but only for 5% of the time (Table 2). Int. J. Environ. Res. Public Health 2020, 17, x 5 of 11 Figure 1. Exercise intensity (average speed) as a percentage of vVO2max at each race. Due to the large difference in relative speed to vVO2max, Tlims at VO2max%Ttot during the sprint, middle-distance, 800 m, and the 1500 m were not significantly different (Table 2). The highest Tlim at VO2max%Ttot was measured in the 3000 m race, while the lowest was measured in the marathon (Figure 2). The 3000 m runners spent their half of the time at VO2max (51 ± 18% of Ttot), while all of the marathon runners all reached VO2max, but only for 5% of the time (Table 2). Table 2. Performance (IAAF score and racing time), number of subjects having reached VO2max and Tlim at V̇O2max during the specific running distance Distance n IAAF Score Race Time (hh:min:sec) V̇O2max Reached (n, %) Tlim at V̇O2max (s) Tlim at V̇O2max%Ttot Post-Run Lactate (mmol∙L−1) 100 m 13 799.0 ± 143.5 11″ ± 0.5″ 10 (76%) 3 ± 2.1 25.6 ± 18.5 14.0 ± 2.8 200 m 13 795.5 ± 135.5 23″ ± 1″1 11 (85%) 6 ± 4.0 28.5 ± 17.7 14.9 ± 1.5 800 m 8 563.0 ± 131.0 ab 2′09″ ± 6″4 f 8 (100%) 28 ± 19.7 aef 22.0 ± 15.8 15.9 ± 1.7 1500 m 16 474.6 ± 191.8 ab 4′40″ ± 24″7 acd 15 (94%) 129 ± 92.2 abe 41.7 ± 28.6 12.4 ± 1.8 bc 3000 m 9 472.2 ± 218.8 ab 10′07″ ± 1′9″ ab 8 (89%) 341 ± 103.3 abcd 51.4 ± 18.3 abc 11.7 ± 2.3 bc 10,000 m 7 522.4 ± 242.5 ab 36′22″ ± 4′19″ ab 7 (100%) 680 ± 590.6 abcd 30.6 ± 27.2 f / 42,195 m 12 385.6 ± 190.7 ab 3h7′17″ ± 18′41″ abcd 10 (83%) 479 ± 497.9 abc 4.1 ± 4.0 abcde 6.6 ± 2.1 abcde Abbreviations: IAAF, International Association of Athletics Federations; V̇O2max, maximal oxygen consumption; Tlim, Time limit; Ttot, Total race time. Note: a indicates a significant difference (p < 0.05) vs. 100 m, b 200 m, c 800 m, d 1500 m, e 3000 m and f marathon. The data are quoted as the mean ± SD. The relative time spent at VO2max was only a factor predicting performance in the groups for which the Tlim at VO2max%Ttot was the highest and the lowest, the 3000 m and the marathon, respectively. Indeed, the 3000 m race was the distance eliciting the highest Tlim at VO2max%Ttot (more than half of the effort) and the distance for which the Tlim at VO2max%Ttot was significantly correlated with the performance (r = 0.86, p = 0.003, Figure 2). 0 50 100 150 200 250 42195 10000 3000 1500 800 200 100 percentage % Distance (m) intensity of exercise %vVO2max VO2max Figure 1. Exercise intensity (average speed) as a percentage of vVO2max at each race. Table 2. Performance (IAAF score and racing time), number of subjects having reached VO2max and Tlim at VO2max during the specific running distance. Distance n IAAF Score Race Time (hh:min:sec) VO2max Reached (n, %) Tlim at VO2max (s) Tlim at VO2max%Ttot Post-Run Lactate (mmol·L−1) 100 m 13 799.0 ± 143.5 11” ± 0.5” 10 (76%) 3 ± 2.1 25.6 ± 18.5 14.0 ± 2.8 200 m 13 795.5 ± 135.5 23” ± 1”1 11 (85%) 6 ± 4.0 28.5 ± 17.7 14.9 ± 1.5 800 m 8 563.0 ± 131.0 ab 2′09” ± 6”4 f 8 (100%) 28 ± 19.7 aef 22.0 ± 15.8 15.9 ± 1.7 1500 m 16 474.6 ± 191.8 ab 4′40” ± 24”7 acd 15 (94%) 129 ± 92.2 abe 41.7 ± 28.6 12.4 ± 1.8 bc 3000 m 9 472.2 ± 218.8 ab 10′07” ± 1′9” ab 8 (89%) 341 ± 103.3 abcd 51.4 ± 18.3 abc 11.7 ± 2.3 bc 10,000 m 7 522.4 ± 242.5 ab 36′22” ± 4′19” ab 7 (100%) 680 ± 590.6 abcd 30.6 ± 27.2 f / 42,195 m 12 385.6 ± 190.7 ab 3h7′17” ± 18′41” abcd 10 (83%) 479 ± 497.9 abc 4.1 ± 4.0 abcde 6.6 ± 2.1 abcde Abbreviations: IAAF, International Association of Athletics Federations; VO2max, maximal oxygen consumption; Tlim, Time limit; Ttot, Total race time. Note: a indicates a significant difference (p < 0.05) vs. 100 m, b 200 m, c 800 m, d 1500 m, e 3000 m and f marathon. The data are quoted as the mean ± SD. Int. J. Environ. Res. Public Health 2020, 17, x 6 of 11 Figure 2. Correlation between the Tlim VO2max on the relative exercise duration (Tlim at VO2max%Ttot) and the performance in the 3000 m race effort. Seventy-four percent of the 3000 m performance variance could be predicted by the relative time limit at VO2max (Tlim at VO2max%Ttot), higher than with vVO2max (69%). Furthermore, as highlighted above, even if the relative time spent at VO2max was low (5%) during the marathon, the fraction of vVO2max was a significant predictor of marathon performance (r² = 0.81). 4. Discussion Classically, it was thought that neither the sprint nor the marathon elicited VO2max. Our results 0 10 20 30 40 50 60 70 80 90 100 500 550 600 650 700 750 Tlim VO2max (s)% Tltot Performance (s) Tlim VO2max (s)% Tot VO2max %Ttot Figure 2. Correlation between the Tlim VO2max on the relative exercise duration (Tlim at VO2max%Ttot) and the performance in the 3000 m race effort. Int. J. Environ. Res. Public Health 2020, 17, 9250 6 of 11 The relative time spent at VO2max was only a factor predicting performance in the groups for which the Tlim at VO2max%Ttot was the highest and the lowest, the 3000 m and the marathon, respectively. Indeed, the 3000 m race was the distance eliciting the highest Tlim at VO2max%Ttot (more than half of the effort) and the distance for which the Tlim at VO2max%Ttot was significantly correlated with the performance (r = 0.86, p = 0.003, Figure 2). Seventy-four percent of the 3000 m performance variance could be predicted by the relative time limit at VO2max (Tlim at VO2max%Ttot), higher than with vVO2max (69%). Furthermore, as highlighted above, even if the relative time spent at VO2max was low (5%) during the marathon, the fraction of vVO2max was a significant predictor of marathon performance (r2 = 0.81). 4. Discussion Classically, it was thought that neither the sprint nor the marathon elicited VO2max. Our results show that VO2max can be elicited and sustained in the sprint, marathon, and middle-distance events. Furthermore, we found that the time spent at VO2max represents a high fraction of the distance run in the sprint and middle-distances (800–3000 m). However, this time spent at VO2max was only correlated with the 3000 m event. We believe that this is the first study focusing on the solicitation of VO2max during the sprint (100, 200 m). The solicitation of VO2max is brief, given that both oxygen kinetics and the delay of achieving VO2max depends heavily on the acceleration phase [24]. Indeed, the time constant values of the fundamental amplitude for VO2, the muscle phosphocreatine response to exercise, and VO2 dynamics cohere during both the moderate and high-intensity exercise [25]. We showed that VO2max is elicited in the marathon, even though the time spent at VO2max is only 5 percent. The results reported by Michael Maron (1976) agree with our results. Even if the Tlim at VO2max%Ttot was the lower in the marathon (4 ± 4%), most marathon runners reached VO2max during the effort in Maron’s study. The relative time runners spent at VO2max were not significantly different between the sprint and short middle-distance events (800 and 1500 m). Our group of elite national level sprinters possess an exceptionally high maximal aerobic capacity that must be considered when examining our results [26]. Indeed, this ability to rapidly reach VO2max during a sprint allows an athlete to perform sprint repeats during training and racing [27]. In a recent study, the authors investigated the aerobic contribution to isolated sprints within a repeated-sprint bout involving 5 × 6 s sprints [28]. The findings have shown that the aerobic contribution to the first sprint is ∼10%, while during the fifth sprint, it is ∼40%. The aerobic contribution to the final sprint of each bout was also significantly related to VO2max [28]. This is supported by the VO2 attained during the final sprint of each bout, which was not different from VO2max (p = 0.448). Due to the incomplete recovery between sprints, it is possible that the progressive increases in PCr breakdown and Pi accumulation over the course of the 5 × 6 s sprints would also have driven the increase in VO2 from the first to the final sprint [28]. Thus, the significantly greater VO2 in the fifth sprint of each bout can probably be attributed to starting from an elevated baseline [29], priming as a consequence of the previous sprints, and an ADP-mediated stimulation of VO2 [28]. Their findings suggest that the aerobic contribution to repeated-sprint exercise may be limited by VO2max and that by increasing this capacity a greater aerobic contribution may be achieved during latter sprints, potentially improving performance [28,29]. it is likely that all sprints after the first were initiated from an elevated baseline [30], which would have elevated the VO2 during subsequent sprints [28]. Aerobic metabolism provides nearly 50% of the energy during the second sprint of 10 or 30 s, whereas the phosphocreatine (PCr) availability is essential for high power output during the initial 10 s [27]. Peak oxygen deficit is also an important factor of performance in the sprint and middle-distance events. Furthermore, multiple regression analyses indicate that the peak oxygen deficit is the strongest metabolic predictor of performance in the 800, 1500, and 5000 m events [31]. Int. J. Environ. Res. Public Health 2020, 17, 9250 7 of 11 Likewise, Billat et al. reported that a high peak oxygen consumption and the ability to run fast over a 1000 m section of the marathon determined the difference between an elite marathon performance (2 h 6 min–2 h 11 min) and a non-elite marathon time (2 h 12 min–2 h 16 min) [32]. Force-velocity characteristics and maximal anaerobic power are of great interest, especially in elite runners [33]. Successful elite runners possess the ability to run at high speeds over periods of a few seconds to several minutes [34]. This is likely mediated by the ability to rapidly deplete phosphocreatine (PCr) [28], accelerate the oxygen kinetics, and increase the relative time spent at VO2max. Indeed, evidence suggests that PCr depletion is related to sprint duration and subjects’ training status [35]. Hirvonen et al. (1987) suggested that sprint performance is related to depleting a more significant amount of high-energy phosphates and at faster rates during the initial stages of exercise; he demonstrated that PCr depletion was greater in a group of elite national level 100 m track sprinters [36]. The elite sprinters depleted significantly higher amounts of PCr than the slower sprinters during 80 and 100 m sprints (76 and 71%) [36]. The rapid depletion of PCr could also induce faster oxygen kinetics and, therefore, a more extended time spent at VO2max. Korzeniewski and Zoladz (2004) (this last one being a prior high 800 m level) clearly demonstrated that the half–transition time of VO2 kinetics is determined by the amount of PCr that has been transformed into creatine during the rest-to-work transition [37]. A fast-start during a running effort has been reported to increase VO2 kinetics and to improve exercise tolerance [38–40]. Sahlin (2004) highlighted that the ATP turnover rate during a 100 m sprint is estimated to be three-fold higher than during a marathon and 50 times higher than at rest [41]. Acceleration corresponds to about 10 and 40% of the total energy demand during 400 and 100 m running, respectively [41]. During a 5000 m effort, Sahlin (2004) considered that the total energy demand is significant, and that the contribution from kinetic energy becomes negligible. If we consider that the time to reach VO2max contributes to the relative time spent at VO2max, our results show that until the 10 km, the time spent at VO2max is not negligible (50% on 3000 m and 31% on 10 km). Furthermore, once VO2max is reached in a sprint to the 10 km, it is maintained until the end of the effort, and this contributes to the relative time to exhaustion at VO2max. This contrasts with prior studies that found a systematic decrease in VO2 in the last 100 m of a 400 and 800 m effort after VO2max was reached, but they did not observe this systematic decrease at the end of the 1500 m effort [42]. We can explain this difference in VO2 observed in the last 100 m between the 800 and the 1500 m efforts are due to the difference in speeds and the fact that the 1500 m effort is run at a steady-state pace just above vVO2max, whereas the 800 m is an all-out effort [1]. The highest Tlim at VO2max%Ttot measured was in the 3000 m effort, while the lowest was measured in the marathon. Indeed, the 3000 m runners spent half of their time at VO2max (51 ± 18% of Ttot), while the marathon runners reached VO2max, but only for 5% of the time. Maron et al. confirmed that VO2max was reached during 4% of the marathon in his research using Douglas bags [10]. We recently analyzed the pacing strategy of the world record marathon performance of Eluid Kipchoge at the 2019 Berlin marathon, 2h01 [43]. Kipchoge implemented a fast start near vVO2max, then allowed himself to “recover” during the following two-thirds of the marathon by running below his threshold and running above vVO2max km before the finish [43]. Many marathons are now won in a final sprint; Kenya’s Lawrence Cherono won the 2019 Boston Marathon in such a manner. The 3000 m effort is a true balance between aerobic and anaerobic contributions, with high energy production at VO2max. This corresponds to the average power at which the longest time to exhaustion at VO2max is obtained, based on a model of the maximal endurance time at VO2max [2] and experimental data from 90% to 140% of vVO2max [12,44]. This relative endurance time spent at VO2max was only a factor of performance in the group for which the Tlim at VO2max%Ttot was the highest and the lowest, i.e., the 3000 m and marathon, respectively). Indeed, the 3000 m effort was the distance eliciting the highest Tlim at VO2max%Ttot Int. J. Environ. Res. Public Health 2020, 17, 9250 8 of 11 (more than half of the time), and the race for which the Tlim at VO2max%Ttot was significantly correlated with the performance. Previously, our laboratory studied the concept of time spent at VO2max by observing the speeds that elicit the longest time to exhaustion at VO2max [44,45]. However, we now appreciate that this approach is flawed because it was based upon the model of constant power or speed, and not according to variable pace running. It would be better to study this concept using variable pace running, which is how humans run naturally. Indeed, the time spent at vVO2max was accurately predicted when the vVO2max was expressed as a percentage of the maximal speed reserve (i.e., the difference between maximal sprint velocity and the “critical speed” [44]. In our study, the average speed during the 3000 m was the closest to the critical speed at VO2max. This “critical speed” is that speed between at which vVO2max and maximal lactate are reached. This is significant because critical speed corresponds to the highest metabolic rate at which energy is supplied through substrate-level phosphorylation and reaches a steady-state at VO2max. The critical speed represents the highest metabolic rate at which the energy supply produced via substrate-level phosphorylation reaches a steady-state below VO2max, and represents the greatest rate of energy production via “pure oxidative” just above the maximal lactate steady state [46,47]. However, this critical speed model was developed to find the speed that elicits the maximal time spent at VO2max. Billat et al. (1999) developed the concept of the critical speed at VO2max (CP’) and defined it as the speed that can be maintained while running at VO2max [45]. The authors used a test with progressively increasing speeds to determine the subjects’ vVO2max, which is defined as the speed at which VO2max is attained. Therefore CP’, i.e., the speed eliciting the maximal time spent at VO2max, was higher than the traditional critical speed and was then defined as the speed between the velocity at maximal lactate steady state and vVO2max (equal to 87% of vVO2max in Morton and Billat, 2000). Therefore, CP’ was sufficient to drive VO2 to its maximum and elicit the maximal time before exhaustion [2]. Expressing running intensity as a percentage of the difference between maximal velocity (measured from an individual 60 m effort) and the critical velocity allowed better prediction of the time limit at VO2max compared to the critical speed VO2max model [48]. This work confirmed prior studies performed on different exercises (swimming, cycling, kayaking, and running) by Faina et al. (1997), who have demonstrated that the anaerobic capacity was a significant factor of the time spent at VO2max [49]. However, this approach was based on the constant speed paradigm. In addition, we know that interval training protocols, alternating speed above and below the critical speed, allow a doubling of the time limit at VO2max in comparison with the time limit at vVO2max (14 ± 5 vs. 4 ± 1 min) [50,51]. Surprisingly, extending this endurance time was shown to be possible using descending speed cardiorespiratory test protocols after having reached VO2max until the maximal lactate steady state speed while maintaining VO2max for almost 30 min [52]. 5. Conclusions In conclusion, our study showed that VO2max is clearly elicited in all distances from the sprint to the marathon. A fast start and the time to reach VO2max is important in increasing VO2 kinetics and to improve exercise tolerance. Human locomotion naturally uses a variable pace running strategy, and it is time to break down the barriers between the so-called aerobic and anaerobic metabolisms. We can only achieve this by moving the laboratory outdoors and performing studies in real-world environments and racing conditions. In this way, a new paradigm of applied physiology will be developed to provide new training and racing insights. Author Contributions: Conceptualization, V.B.; methodology, V.B. and C.A.M.; software, C.A.M.; validation, V.B. and C.A.M.; formal analysis, C.A.M.; investigation, V.B. and C.A.M.; resources, V.B.; data curation, V.B.; writing—original draft preparation, V.B. and C.A.M.; writing—review and editing, V.B., J.E. and C.A.M.; visualization, V.B., J.E. and C.A.M.; supervision, V.B.; project administration, V.B., J.E. and C.A.M. All authors have read and agreed to the published version of the manuscript. Int. J. Environ. Res. Public Health 2020, 17, 9250 9 of 11 Funding: Claire Molinari received a CIFRE(Conventions Industrielles de Formation par la Recherche). fellowship, funded by BillaTraining. The work did not receive any significant funding that could have influenced its outcome Acknowledgments: The authors wish to thank the study participants for their collaboration, and Jean-Pierre Koralsztein for helpful advice. Conflicts of Interest: The authors declare no conflict of interest. References 1. Billat, V.; Hamard, L.; Koralsztein, J.P.; Morton, R.H. Differential modeling of anaerobic and aerobic metabolism in the 800-m and 1500-m run. J. Appl. Physiol. 2009, 107, 478–487. [CrossRef] [PubMed] 2. Morton, R.H.; Billat, V. Maximal endurance time at VO2max. Med. Sci. Sports Exerc. 2000, 32, 1496–1504. [CrossRef] [PubMed] 3. Hill, A.V.; Lupton, H. Muscular Exercise, Lactic Acid, and the Supply and Utilization of Oxygen. QJM Int. J. Med. 1923, 62, 135–171. [CrossRef] 4. Saks, V.A.; Kongas, O.; Vendelin, M.; Kay, L. Role of the creatine/phosphocreatine system in the regulation of mitochondrial respiration. Acta Physiol. Scand. 2000, 168, 635–641. [CrossRef] [PubMed] 5. Korzeniewski, B.; Zoladz, J.A. Possible mechanisms underlying slow component of Vo2 on-kinetics in skeletal muscle. J. Appl. Physiol. 2015, 118, 1240–1249. [CrossRef] 6. Di Prampero, P.E. The energy cost of human locomotion on land and in water. Int. J. Sports Med. 1986, 7, 55–72. [CrossRef] 7. Zinner, C. Training Aspects of Marathon Running. In Marathon Running: Physiology, Psychology, Nutrition and Training Aspects; Zinner, C., Sperlich, B., Eds.; Springer International Publishing: Cham, Switzerland, 2016; pp. 153–171. [CrossRef] 8. Billat, V.; Bernard, O.; Pinoteau, J.; Petit, B.; Koralsztein, J.P. Time to exhaustion at VO2max and lactate steady state velocity in sub elite long-distance runners. Arch. Int. Physiol. Biochim. Biophys. 1994, 102, 215–219. [CrossRef] 9. Whipp, B.J. The slow component of O2 uptake kinetics during heavy exercise. Med. Sci. Sports Exerc. 1994, 26, 1319–1326. [CrossRef] 10. Maron, M.B.; Horvath, S.M.; Wilkerson, J.E.; Gliner, J.A. Oxygen uptake measurements during competitive marathon running. J. Appl. Physiol. 1976, 40, 836–838. [CrossRef] 11. Billat, V.; Dalmay, F.; Antonini, M.T.; Chassain, A.P. A method for determining the maximal steady state of blood lactate concentration from two levels of submaximal exercise. Eur. J. Appl. Physiol. 1994, 69, 196–202. [CrossRef] 12. Billat, V.L.; Morton, R.H.; Blondel, N.; Berthoin, S.; Bocquet, V.; Koralsztein, J.P.; Barstow, T.J. Oxygen kinetics and modelling of time to exhaustion whilst running at various velocities at maximal oxygen uptake. Eur. J. Appl. Physiol. 2000, 82, 178–187. [CrossRef] [PubMed] 13. Billat, V.; Petot, H.; Karp, J.R.; Sarre, G.; Morton, R.H.; Mille-Hamard, L. The sustainability of VO2max: Effect of decreasing the workload. Eur. J. Appl. Physiol. 2013, 113, 385–394. [CrossRef] [PubMed] 14. Bergh, U.; Ekblom, B.; Astrand, P.O. Maximal oxygen uptake ‘classical’ versus ‘contemporary’ viewpoints. Med. Sci. Sports Exerc. 2000, 32, 85–88. [CrossRef] [PubMed] 15. Ekblom, B. Counterpoint: Maximal oxygen uptake is not limited by a central nervous system governor. J. Appl. Physiol. 2009, 106, 339–341. [CrossRef] [PubMed] 16. Billat, V.; Vitiello, D.; Palacin, F.; Correa, M.; Pycke, J.R. Race Analysis of the World’s Best Female and Male Marathon Runners. Int. J. Environ. Res. Public Health 2020, 17, 1177. [CrossRef] [PubMed] 17. Di Prampero, P.E.; Botter, A.; Osgnach, C. The energy cost of sprint running and the role of metabolic power in setting top performances. Eur. J. Appl. Physiol. 2015, 115, 451–469. [CrossRef] [PubMed] 18. Perez-Suarez, I.; Martin-Rincon, M.; Gonzalez-Henriquez, J.J.; Fezzardi, C.; Perez-Regalado, S.; Galvan-Alvarez, V.; Juan-Habib, J.W.; Morales-Alamo, D.; Calbet, J.A. Accuracy and Precision of the COSMED K5 Portable Analyser. Front. Physiol 2018, 9, 1764. [CrossRef] 19. Borg, G. Borg’s Perceived Exertion and Pain Scales; Human Kinetics: Champaign, IL, USA, 1998. 20. Hogg, J.S.; Hopker, J.G.; Mauger, A.R. The Self-Paced VO2max Test to Assess Maximal Oxygen Uptake in Highly Trained Runners. Int. J. Sports Physiol. Perform. 2015, 10, 172–177. [CrossRef] Int. J. Environ. Res. Public Health 2020, 17, 9250 10 of 11 21. Gaesser, G.A.; Poole, D.C. The Slow Component of Oxygen Uptake Kinetics in Humans. Exerc. Sport Sci. Rev. 1996, 24, 35. [CrossRef] 22. Morton, R.H. A three component model of human bioenergetics. J. Math. Biol. 1986, 24, 451–466. [CrossRef] 23. Barstow, T.J.; Mole, P.A. Linear and nonlinear characteristics of oxygen uptake kinetics during heavy exercise. J. Appl. Physiol. 1991, 71, 2099–2106. [CrossRef] [PubMed] 24. Glaister, M. Multiple Sprint Work. Sports Med. 2005, 35, 757–777. [CrossRef] [PubMed] 25. Rossiter, H.B.; Ward, S.A.; Kowalchuk, J.M.; Howe, F.A.; Griffiths, J.R.; Whipp, B.J. Dynamic asymmetry of phosphocreatine concentration and O(2) uptake between the on- and off-transients of moderate- and high-intensity exercise in humans. J. Physiol. 2002, 541, 991–1002. [CrossRef] [PubMed] 26. Volkov, N.I.; Shirkovets, E.A.; Borilkevich, V.E. Assessment of aerobic and anaerobic capacity of athletes in treadmill running tests. Eur. J. Appl. Physiol. 1975, 34, 121–130. [CrossRef] [PubMed] 27. Bogdanis, G.C.; Nevill, M.E.; Boobis, L.H.; Lakomy, H.K. Contribution of phosphocreatine and aerobic metabolism to energy supply during repeated sprint exercise. J. Appl. Physiol. 1996, 80, 876–884. [CrossRef] [PubMed] 28. McGawley, K.; Bishop, D.J. Oxygen uptake during repeated-sprint exercise. J. Sci. Med. Sport 2015, 18, 214–218. [CrossRef] [PubMed] 29. Buchheit, M.; Ufland, P. Effect of endurance training on performance and muscle reoxygenation rate during repeated-sprint running. Eur. J. Appl. Physiol. 2011, 111, 293–301. [CrossRef] 30. Belfry, G.R.; Paterson, D.H.; Murias, J.M.; Thomas, S.G. The effects of short recovery duration on VO2 and muscle deoxygenation during intermittent exercise. Eur. J. Appl. Physiol. 2012, 112, 1907–1915. [CrossRef] 31. Weyand, P.; Curcton, K.; Conley, D.; Sloniger, M. Percentage Anaerobic Energy Utilized During Track Running Events. Med. Sci. Sports Exerc. 1993, 25, S105. [CrossRef] 32. Billat, V.L.; Demarle, A.; Slawinski, J.; Paiva, M.; Koralsztein, J.P. Physical and training characteristics of top-class marathon runners. Med. Sci. Sports Exerc. 2001, 33, 2089–2097. [CrossRef] 33. Nikolaidis, P.T.; Knechtle, B. Do Fast Older Runners Pace Differently From Fast Younger Runners in the “New York City Marathon”? J. Strength Cond. Res. 2019, 33, 3423–3430. [CrossRef] [PubMed] 34. Bundle, M.W.; Hoyt, R.W.; Weyand, P.G. High-speed running performance: A new approach to assessment and prediction. J. Appl. Physiol. 2003, 95, 1955–1962. [CrossRef] [PubMed] 35. Spencer, M.; Bishop, D.; Dawson, B.; Goodman, C. Physiological and Metabolic Responses of Repeated-Sprint Activities. Sports Med. 2005, 35, 1025–1044. [CrossRef] [PubMed] 36. Hirvonen, J.; Rehunen, S.; Rusko, H.; Härkönen, M. Breakdown of high-energy phosphate compounds and lactate accumulation during short supramaximal exercise. Eur. J. Appl. Physiol. 1987, 56, 253–259. [CrossRef] [PubMed] 37. Korzeniewski, B.; Zoladz, J.A. Factors determining the oxygen consumption rate (V.o2) on-kinetics in skeletal muscles. Biochem. J. 2004, 379, 703–710. [CrossRef] [PubMed] 38. Heubert, R.A.P.; Billat, V.L.; Chassaing, P.; Bocquet, V.; Morton, R.H.; Koralsztein, J.P.; Di Prampero, P.E. Effect of a previous sprint on the parameters of the work-time to exhaustion relationship in high intensity cycling. Int. J. Sports Med. 2005, 26, 583–592. [CrossRef] [PubMed] 39. Jones, A.M.; Wilkerson, D.P.; Vanhatalo, A.; Burnley, M. Influence of pacing strategy on O2 uptake and exercise tolerance. Scand. J. Med. Sci. Sports 2008, 18, 615–626. [CrossRef] [PubMed] 40. Sandals, L.E.; Wood, D.M.; Draper, S.B.; James, D.V. Influence of Pacing Strategy on Oxygen Uptake During Treadmill Middle-Distance Running. Int. J. Sports Med. 2006, 27, 37–42. [CrossRef] 41. Sahlin, K. High-Energy Phosphates and Muscle Energetics. Princ. Exerc. Biochem. 2004, 46, 87–107. [CrossRef] 42. Hanon, C.; Thomas, C. Effects of optimal pacing strategies for 400-, 800-, and 1500-m races on the [Vdot]O2 response. J. Sports Sci. 2011, 29, 905–912. [CrossRef] 43. Billat, V.; Carbillet, T.; Correa, M.; Pycke, J.R. Detecting the marathon asymmetry with a statistical signature. Phys. Stat. Mech. Appl. 2019, 515, 240–247. [CrossRef] 44. Blondel, N.; Berthoin, S.; Billat, V.; Lensel, G. Relationship between Run Times to Exhaustion at 90, 100, 120, and 140 % of vV·O2max and Velocity Expressed Relatively to Critical Velocity and Maximal Velocity. Int. J. Sports Med. 2001, 22, 27–33. [CrossRef] [PubMed] 45. Billat, V.L.; Blondel, N.; Berthoin, S. Determination of the velocity associated with the longest time to exhaustion at maximal oxygen uptake. Eur. J. Appl. Physiol. 1999, 80, 159–161. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2020, 17, 9250 11 of 11 46. Poole, D.C.; Burnley, M.; Vanhatalo, A.; Rossiter, H.B.; Jones, A.M. Critical Power: An Important Fatigue Threshold in Exercise Physiology. Med. Sci. Sports Exerc. 2016, 48, 2320–2334. [CrossRef] [PubMed] 47. Jones, A.M.; Vanhatalo, A.; Burnley, M.; Morton, R.H.; Poole, D.C. Critical power: Implications for determination of V·O2max and exercise tolerance. Med. Sci. Sports Exerc. 2010, 42, 1876–1890. [CrossRef] 48. Blondel, N.; Billat, V.; Berthoin, S. Relation entre le temps limite de course et l’intensité relative de l’exercice, exprimée en fonction de la vitesse critique et de la vitesse maximale. Sci. Sports 2000, 15, 242–244. [CrossRef] 49. Faina, M.; Billat, V.; Squadrone, R.; De Angelis, M.; Koralsztein, J.P.; Dal Monte, A. Anaerobic contribution to the time to exhaustion at the minimal exercise intensity at which maximal oxygen uptake occurs in elite cyclists, kayakists and swimmers. Eur. J. Appl. Physiol. 1997, 76, 13–20. [CrossRef] 50. Billat, V.L.; Slawinski, J.; Bocquet, V.; Demarle, A.; Lafitte, L.; Chassaing, P.; Koralsztein, J.P. Intermittent runs at the velocity associated with maximal oxygen uptake enables subjects to remain at maximal oxygen uptake for a longer time than intense but submaximal runs. Eur. J. Appl. Physiol. 2000, 81, 188–196. [CrossRef] 51. Billat, V.L.; Slawinksi, J.; Bocquet, V.; Chassaing, P.; Demarle, A.; Koralsztein, J.P. Very short (15 s–15 s) interval-training around the critical velocity allows middle-aged runners to maintain VO2 max for 14 minutes. Int. J. Sports Med. 2001, 22, 201–208. [CrossRef] 52. Petot, H.; Meilland, R.; Le Moyec, L.; Mille-Hamard, L.; Billat, V.L. A new incremental test for VO2max accurate measurement by increasing VO2max plateau duration, allowing the investigation of its limiting factors. Eur. J. Appl. Physiol. 2012, 112, 2267–2276. [CrossRef] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Maximal Time Spent at VO_2max from Sprint to the Marathon.

12-10-2020

Molinari, Claire A,Edwards, Johnathan,Billat, Véronique

eng

PMC7432325

International Journal of Environmental Research and Public Health Article Right Ventricular Diastolic Dysfunction after Marathon Run Zuzanna Lewicka-Potocka 1,2 , Alicja D ˛abrowska-Kugacka 1,* , Ewa Lewicka 1 , Rafał Gał ˛aska 2, Ludmiła Daniłowicz-Szymanowicz 1, Anna Faran 1, Izabela Nabiałek-Trojanowska 1,2 , Marcin Kubik 1, Anna Maria Kaleta-Duss 1 and Grzegorz Raczak 1 1 Department of Cardiology and Electrotherapy, Medical University of Gda´nsk, 80-210 Gda´nsk, Poland; [email protected] (Z.L.-P.); [email protected] (E.L.); [email protected] (L.D.-S.); [email protected] (A.F.); [email protected] (I.N.-T.); [email protected] (M.K.); [email protected] (A.M.K.-D.); [email protected] (G.R.) 2 First Department of Cardiology, Medical University of Gda´nsk, 80-210 Gda´nsk, Poland; [email protected] * Correspondence: [email protected]; Tel.: +48-601-910-480 Received: 19 June 2020; Accepted: 21 July 2020; Published: 24 July 2020



 Abstract: It has been raised that marathon running may significantly impair cardiac performance. However, the post-race diastolic function has not been extensively analyzed. We aimed to assess whether the marathon run causes impairment of the cardiac diastole, which ventricle is mostly affected and whether the septal (IVS) function is altered. The study included 34 male amateur runners, in whom echocardiography was performed two weeks before, at the finish line and two weeks after the marathon. Biventricular diastolic function was assessed not only with conventional Doppler indices but also using the heart rate-adjusted isovolumetric relaxation time (IVRTc). After the run, IVRTc elongated dramatically at the right ventricular (RV) free wall, to a lesser extent at the IVS and remained unchanged at the left ventricular lateral wall. The post-run IVRTc_IVS correlated with IVRTc_RV (r = 0.38, p < 0.05), and IVRTc_RV was longer in subjects with IVS hypertrophy (88 vs. 51 ms; p < 0.05). Participants with measurable IVRT_RV at baseline (38% of runners) had longer post-race IVRTc_IVS (102 vs. 83 ms; p < 0.05). Marathon running influenced predominantly the RV diastolic function, and subjects with measurable IVRT_RV at baseline or those with IVS hypertrophy can experience greater post-race diastolic fatigue. Keywords: marathon run; amateur runners; diastolic function; right ventricle; relaxation; isovolumic relaxation time; myocardial performance index 1. Introduction There is increasing evidence that prolonged intense exercise, such as marathon running, can affect cardiac performance [1]. Recently, this form of sports activity has gained popularity, but the question remains about its safety for amateur non-elite runners, who are often middle-aged [2]. It has been raised that the right ventricle (RV) may be the “Achilles heel” of the competing heart, because RV enlargement and reduction in RV contractility was observed following marathon running [3]. An increase in oxygen demand during prolonged exercise, followed by the augmentation of cardiac output and a proportionate rise in pulmonary artery pressure, ultimately affects RV, which is not adapted to high vascular resistance [4,5]. In the presence of common septum and pericardium constraint, dysfunction of one ventricle affects the other in the process of ventricular interdependence. Apart from their dominant role in this interaction, the septal muscle fibers that obliquely bind both Int. J. Environ. Res. Public Health 2020, 17, 5336; doi:10.3390/ijerph17155336 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2020, 17, 5336 2 of 14 ventricles also fulfil a major function in the overall movement of the RV [6,7]. Exercise-induced cardiac fatigue has been studied mainly in terms of depressed systolic function [8]. In contrast, a small number of studies have assessed the cardiac diastole among amateur marathon runners from pre-event to in-run testing, especially regarding RV. According to the theory of myocardial damage progression, the diastole tends to become impaired before the systole, and, in several diseases, the slowing of relaxation was shown as an early sign of myocardial damage [9–11]. The prolongation of the isovolumic relaxation time (IVRT) may reflect the delayed relaxation filling pattern and therefore be the first marker of developing diastolic dysfunction [12]. In this study, we analyzed the pathophysiology of heart exhaustion associated with a marathon competition, with special attention to diastolic function and interventricular septum involvement. 2. Materials and Methods After having received acceptance of the study protocol from the Independent Bioethics Commission for Research of the Medical University of Gdansk (NKBBN/104/2016), via advertisements sent to local sport clubs, we recruited male amateur marathon runners planning to run in the 2nd PZU Marathon in Gda´nsk, Poland. The study was conducted in accordance with the Declaration of Helsinki. Detailed study information was provided to all volunteers and written consent was obtained from each participant before entering the study. We enrolled subjects who were at defined age (20–55 years old) and showed no chronic diseases. The study protocol consisted of three stages and participants were examined two weeks before the marathon run (stage I), at the marathon finish line (stage II) and two-weeks after the competition (stage III). At each stage, physical and echocardiographic examination (ECHO) was performed. Moreover, at baseline (stage I), the training history was collected and a cardiopulmonary exercise test (CPET) was performed. Detailed characteristics of the examined amateur marathon runners have been previously presented [13]. During the competition, participants were allowed to rehydrate on a whim and no food intake restrictions were advised. The temperature was around 12 ◦C and the wind speed was around 21.5 km/h. The ECHO was carried out with a commercially available system (Vivid E9, GE Healthcare, Horten, Norway), in accordance with the recommendations of the American Society of Echocardiography and European Association of Cardiovascular Imaging [14]. All analyses and measurements (averaged from three consecutive beats) were performed off-line by two researchers, using echocardiographic quantification software (EchoPac 201, GE Healthcare, Norway). By means of two-dimensional (2D) and M-mode ECHO, cardiac dimensions were obtained, including diastolic interventricular septum diameter (IVSd) in the parasternal long-axis view (PLAX). The end-systolic right atrial (RA) area was calculated in the apical four-chamber (A4C) view and the left atrial (LA) volume indexed to body surface area (LAVI) was obtained from the apical two-chamber and A4C views. At end-diastole, the basal LV and RV transversal dimensions (LVd BAS and RVd BAS) were acquired in the A4C view and, subsequently, the RVd/LVd BAS ratio was calculated to assess diastolic ventricular interaction. The transverse RV diameter was also measured in the middle of the RV inflow (RVd MID) in the RV-focused apical view, as described in detail previously [14]. The diastolic function of LV and RV was assessed with pulsed wave Doppler (PWD) and spectral Doppler tissue imaging (DTI) indices according to the guidelines [15,16]. In the 4C view, the PWD transmitral (MV) and transtricuspid (TV) flow velocities were obtained: peak early (E), peak atrial (A) and E/A ratio. Spectral DTI mitral annular velocities, namely peak systolic (S’), peak early diastolic (E’), peak atrial diastolic (A’), E’/A’ ratio, were assessed at basal septum (S’IVS, E’IVS, A’IVS) and LV lateral wall (S’LW, E’LW, A’LW). Then, the following ratios were calculated: E_MV/E’_LW, E_MV/E’_IVS and E_MV/E’_AVG (E’ averaged from IVS and LW) [16]. Corresponding tricuspid annular velocities were obtained (S’RV, E’RV, A’RV), along with E_TV/E’_RV ratio. Spectral DTI isovolumic relaxation time (IVRT) was estimated at the lateral mitral (IVRT_LW), tricuspid RV (IVRT_RV) and septal (IVRT_IVS) level, as the interval between the end of the S’-wave and the beginning of the E’-wave, as described previously [17,18]. In order to compensate for the augmented heart rate after exercise, the heart Int. J. Environ. Res. Public Health 2020, 17, 5336 3 of 14 rate-adjusted IVRT (IVRTc) was calculated as the ratio of the IVRT and square RR (interval between two subsequent beats, given in seconds). RV and LV global performance was assessed by the DTI-derived myocardial performance index (MPI), which was analyzed in the 4C view at the lateral tricuspid (MPI_RV), lateral mitral (MPI_LW) and septal annulus (MPI_IVS). MPI was calculated with the formula (ICT + IVRT)/ET, where ICT is the isovolumic contraction time, measured from the cessation of A′ -wave to the onset of S’-wave, and ET is the ejection time, measured as the width of S’-wave. The DTI-derived MPI was chosen over the PWD as it is derived from one cardiac cycle and therefore is more reliable [19]. The RV four-chamber longitudinal strain (including ventricular septum), here also referred to as “RV global strain” (RV 4CSL) was measured in the RV-focused 4C view, in accordance with the consensus document on deformation imaging [20]. Along with the guidelines, tricuspid annular plane systolic excursion (TAPSE) and RV fractional area change (RV FAC) were additionally calculated to measure the RV systolic function, and regarding the LV ejection fraction (LV EF) and LV global longitudinal strain (LV GLS), assessment was performed [14,15]. The study subjects were divided into 2 groups on the basis of the baseline IVRT_RV: (1) IVRT_RV = 0 ms or (2) IVRT_RV > 0 ms. Another division was performed according to the presence of IVS hypertrophy, defined as IVSd > 11 mm. Data analysis was performed using Statistica 13.3 software (Statsoft Inc., Tulsa, Oklahoma, United States). The normality of variables was tested with the Shapiro–Wilk test. Data are presented as mean ± SD (if normally distributed) or median with first and third quartile (25th; 75th percentile) (if non-normally distributed). The comparison between 3 stages was performed with ANOVA analysis and the post-hoc Tukey test for normally distributed data. Non-normally distributed measurements were compared with Friedman ANOVA and post-hoc for Friedman ANOVA. Comparisons between predefined groups were performed by Student’s t-test for independent samples or the Mann–Whitney U test, where appropriate. The p-value of < 0.05 was considered significant. Spearman’s correlation was calculated to determine the dependency of ECHO measurements and parameters of cardiorespiratory fitness obtained in the CPET. 3. Results Thirty-four amateur marathon runners, with a mean age of 40 ± 8 years, who finished the competition, were enrolled in the study. All participants were men of Caucasian race with no relevant medical history. Training history and CPET results have been recently published [13]. In brief, the mean training distance was 56.5 ± 19.7 km/week, and the mean marathon finishing time was 3.7 ± 0.4 h. The mean oxygen uptake at anaerobic threshold (VO2AT) was 39.7 ± 6.9 mL/kg/min. Tables 1–3 provide the comparison of 2D ECHO measurements between three stages, including measures of both ventricles (Table 1), Doppler parameters of LV performance (Table 2) and the RV Doppler indices (Table 3). There were no significant differences between parameters from stages I and III, apart from E_MV and E’_RV (Tables 2 and 3). Running the marathon had no impact on the LV systolic function (Table 1). There were no relevant valvular regurgitations among marathon runners at any stage of the study. Reduced ratios of E/A_MV and E/A_TV were observed after the competition (Tables 2 and 3). The results obtained by PWD were consistent with DTI-derived E’/A’ in both RV and LV (Tables 2 and 3). On the contrary, the E/E’ ratios remained at the same level after the race (Tables 2 and 3). The most striking evidence of impaired RV relaxation was shown on the basis of the IVRT analysis (Figure 1). Int. J. Environ. Res. Public Health 2020, 17, 5336 4 of 14 Table 1. Echocardiographic parameters of the left and right ventricle obtained in amateur marathon runners. Parameter Stage I Stage II Stage III ANOVA p-Value Post-Hoc p-Value Mean ± SD 1 or Median (1st; 3rd Quartile) 2 Stage I vs. II Stage I vs. III LV EF (%) 61.8 ± 4.9 60.5 ± 4.4 60.7 ± 4.5 * 0.38 - - LV GLS (%) −19.9 ± 2.3 −19.4 ± 2.1 −19.7 ± 2.2 * 0.41 - - RV 4CSL (%) −22.0 ± 2.8 −20.80 ± 2.6 −21.49 ± 2.5 * <0.05 <0.05 0.46 TAPSE (mm) 25.0 ± 3.6 24.0 ± 3.7 25.0 ± 2.7 * 0.56 - - RV FAC (%) 43 (37; 45) 39 (35; 44) 41 (36; 45) ˆ 0.19 - - RVd MID (cm) 3.4 ± 0.6 3.7 ± 0.5 3.5 ± 0.5 * <0.01 <0.01 0.08 RVd BAS (cm) 3.8 ± 0.4 3.8 ± 0.5 3.9 ± 0.5 * 0.44 - - LVd BAS (cm) 4.8 ± 0.4 4.6 ± 0.3 4.9 ± 0.3 * <0.001 <0.01 0.88 RVd/LVd BAS 0.77 ± 0.1 0.82 ± 0.1 0.79 ± 0.1 * <0.05 <0.05 0.59 1—when normally distributed; 2—when non-normally distributed; Stage I—two weeks before the marathon run; Stage II—at the marathon finish line; Stage III—two weeks after the marathon run; LV —left ventricular; EF—ejection fraction; GLS—global longitudinal strain; RV—right ventricular; 4CSL—four-chamber longitudinal strain = global strain; TAPSE—tricuspid annular plane systolic excursion; FAC—fractional area change; RVd MID—RV mid-cavity end-diastolic dimension; LVd BAS—LV basal end-diastolic diameter; RVd BAS—RV basal end-diastolic diameter; RVd/LVd BAS—basal RV to LV end-diastolic diameter ratio; SD—standard deviation; * ANOVA with post-hoc Tukey test if applicable; ˆ Friedman ANOVA with post-hoc average rank test if applicable. Int. J. Environ. Res. Public Health 2020, 17, 5336 5 of 14 Table 2. Left ventricular parameters obtained by means of the pulsed wave Doppler and spectral Doppler tissue imaging in amateur marathon runners. Parameter Stage I Stage II Stage III ANOVA p-Value Post-Hoc p-Value Mean ± SD 1 or Median (1st; 3rd Quartile) 2 Stage I vs. II Stage I vs. III S’_LW (cm/sec) 11 ± 3 11 ± 3 11 ± 3 * 0.88 - - E’_LW (cm/sec) 15 (12; 17) 12 (10; 15) 14 (13; 16) ˆ <0.001 <0.05 ns A’_LW (cm/sec) 8 (7;9) 10 (9; 11) 8 (7; 9) ˆ <0.001 <0.05 ns E’/A’_LW 1.8 (1.4; 2.1) 1.2 (1.0; 1.5) 1.9 (1.6; 2.3) ˆ <0.001 <0.05 ns IVRT_LW (ms) 53 ± 17 54 ± 19 56 ± 15 * 0.99 - - IVRTc_LW 53 ± 17 59 ± 23 54 ± 15 * 0.28 - - MPI_LW 0.41 ± 0.08 0.45 ± 0.17 0.42 ± 0.07 * 0.20 - - S’_IVS (cm/sec) 8 (8; 9) 9 (8; 10) 9 (7; 10) ˆ 0.24 - - E’_IVS (cm/sec) 11 ± 2 10 ± 2 11 ± 2 * <0.001 <0.01 0.73 A’_IVS (cm/sec) 8 (8; 10) 10 (9; 11) 8 (7; 10) ˆ <0.001 <0.05 ns E’/A’_IVS 1.3 ± 0.4 1.0 ± 0.3 1.4 ± 0.4 * <0.001 <0.001 0.26 IVRT_IVS (ms) 82 (65; 95) 80 (68; 94) 78 (64; 86) ˆ 0.46 - - IVRTc_IVS 78 (66; 92) 92 (77; 108) 73 (65; 85) ˆ <0.001 <0.05 ns MPI_IVS 0.55 (0.44; 0.59) 0.53 (0.44; 0.6) 0.47 (0.44; 0.54) ˆ 0.09 - - E_MV (cm/sec) 71 (67; 87) 67 (55; 77) 78 (68; 92) ˆ <0.01~ ns ns A_MV (cm/sec) 51 ± 10 65 ± 14 56 ± 11 * <0.001 <0.001 0.29 E/A_MV 1.5 ± 0.4 1.1 ± 0.3 1.5 ± 0.4 * <0.001 <0.001 0.7 E_MV/E’_LW 5.4 ± 1.2 5.5 ± 1.6 5.6 ± 1.8 * 0.66 - - E_MV/E’_IVS 7.1 ± 1.5 7.0 ± 1.8 7.6 ± 2.0 * 0.34 - - E_MV/E’_AVG 6.3 (5.2; 7.0) 5.8 (5.0; 7.2) 6.2 (5.4; 7.3) ˆ 0.74 - - 1—when normally distributed; 2—when non-normally distributed; LW—parameter measured at the lateral mitral annulus; IVS—parameter measured at the septal mitral annulus; S’—peak systolic tissue velocity; E’—peak early diastolic tissue velocity; A’—peak atrial diastolic tissue velocity; IVRT—isovolumic relaxation time; IVRTc—IVRT adjusted for heart rate; MPI—myocardial performance index; MV—mitral inflow; E—peak early flow velocity; A—peak atrial flow velocity; AVG—averaged for parameters obtained at IVS and LW; ns-p-value of >0.05 of post-hoc average rank test for Friedman ANOVA; ~ ANOVA test p <0.05, but post-hoc test revealed the difference between stages II and III, which was not the question of our study. * ANOVA with post-hoc Tukey test if applicable; ˆ Friedman ANOVA with post-hoc average rank test if applicable. For other abbreviations, see Table 1. Int. J. Environ. Res. Public Health 2020, 17, 5336 6 of 14 Table 3. Right ventricular parameters obtained by means of the pulsed wave Doppler and spectral Doppler tissue imaging in amateur marathon runners. Parameter Stage I Stage II Stage III ANOVA p-Value Post-Voc p-Value Mean ± SD 1 or Median (1st; 3rd Quartile) 2 Stage I vs. II Stage I vs. III S’_RV (cm/sec) 14 (13; 16) 14 (13.5; 16) 15 (13; 16) ˆ 0.51 - - E’_RV (cm/sec) 12 (11; 15) 12 (9; 14) 14 (13; 16) ˆ <0.05 ~ ns ns A’_RV (cm/sec) 13 (10; 14) 16 (13; 20) 13 (12; 16) ˆ <0.01 <0.05 ns E’/A’_RV 1.0 (0.9; 1.2) 0.7 (0.6; 0.9) 1.2 (0.9; 1.3) ˆ <0.001 <0.05 ns IVRT_RV (ms) 0 (0; 29) 52 (32; 70) 21 (9; 34) ˆ <0.001 <0.05 ns IVRTc_RV (ms) 0 (0; 27) 58 (39; 78) 20 (0; 35) ˆ <0.001 <0.05 ns MPI_RV 0.28 (0.22; 0.37) 0.48 (0.35; 0.64) 0.33 (0.25; 0.41) ˆ <0.001 <0.05 ns E_TV (cm/sec) 55 ± 13 49 ± 11 56 ± 11 * 0.18 - - A_TV (cm/sec) 33 ± 10 46 ± 15 31 ± 8 * <0.001 <0.001 0.99 E/A_TV 1.7 ± 0.4 1.2 ± 0.3 1.9 ± 0.6 * <0.001 <0.001 0.55 E_TV/E’_RV 4.5 ± 1.1 4.2 ± 1.7 3.9 ± 1.0 * 0.44 - - 1—when normally distributed; 2—when non-normally distributed; RV—right ventricular parameters measured at the lateral tricuspid annulus; TV—transtricuspid inflow. * ANOVA with post-hoc Tukey test if applicable; ˆ Friedman ANOVA with post-hoc average rank test if applicable. For other abbreviations, see Tables 1 and 2. Int. J. Environ. Res. Public Health 2020, 17, 5336 7 of 14 Int. J. Environ. Res. Public Health 2020, 17, x 6 of 13 Figure 1. Changes in the right ventricular isovolumic relaxation time (IVRT_RV) between the three study stages. (a) IVRT_RV was undetectable at stage I (two weeks before the marathon run) and (c) at stage III (two weeks after the competition); (b) in contrast, the appearance of IVRT_RV and its prolongation up to 104 ms at stage II (at the marathon finishing line). The marathon run resulted in significant prolongation of IVRTc_RV, which was found in 24 out of 34 runners (71%). The IVRT_RV appeared in 58% of subjects, in whom it was undetectable at the baseline. The marathon impact on the relaxation was different for the RV and LV: immediately after Figure 1. Changes in the right ventricular isovolumic relaxation time (IVRT_RV) between the three study stages. (a) IVRT_RV was undetectable at stage I (two weeks before the marathon run) and (c) at stage III (two weeks after the competition); (b) in contrast, the appearance of IVRT_RV and its prolongation up to 104 ms at stage II (at the marathon finishing line). The marathon run resulted in significant prolongation of IVRTc_RV, which was found in 24 out of 34 runners (71%). The IVRT_RV appeared in 58% of subjects, in whom it was undetectable at the Int. J. Environ. Res. Public Health 2020, 17, 5336 8 of 14 baseline. The marathon impact on the relaxation was different for the RV and LV: immediately after the run, IVRTc elongated dramatically at the RV free wall, to a lesser extent at the IVS and remained at the same level at the LV lateral wall (Figure 2). Int. J. Environ. Res. Public Health 2020, 17, x 7 of 13 the run, IVRTc elongated dramatically at the RV free wall, to a lesser extent at the IVS and remained at the same level at the LV lateral wall (Figure 2). Figure 2. Changes in the heart rate-adjusted isovolumic relaxation time between the three stages of the study, assessed in the spectral Doppler tissue imaging (a) at the lateral tricuspid annulus (IVRTc_RV), (b) at septum (IVRTc_IVS) and (c) at the lateral mitral annulus (IVRTc_LW). In post-marathon analysis, there was a correlation between IVRTc_RV and IVRTc_IVS (r = 0.38, p < 0.05), and IVRTc_RV was significantly longer in runners with IVS hypertrophy (88 vs. 51 ms; p < Figure 2. Changes in the heart rate-adjusted isovolumic relaxation time between the three stages of the study, assessed in the spectral Doppler tissue imaging (a) at the lateral tricuspid annulus (IVRTc_RV), (b) at septum (IVRTc_IVS) and (c) at the lateral mitral annulus (IVRTc_LW). In post-marathon analysis, there was a correlation between IVRTc_RV and IVRTc_IVS (r = 0.38, p < 0.05), and IVRTc_RV was significantly longer in runners with IVS hypertrophy (88 vs. 51 ms; Int. J. Environ. Res. Public Health 2020, 17, 5336 9 of 14 p < 0.05). At the baseline, 13 (38%) participants presented measurable IVRT_RV, and in this group, a longer post-race IVRTc_IVS was revealed (102 vs. 83 ms; p < 0.05). MPI measurements paralleled the IVRT results. After the race, a significant prolongation at the RV free wall was observed, accompanied by a trend towards its increase at IVS (p = 0.09) but unchanged MPI at the LV lateral wall (Tables 2 and 3). There was a post-marathon reduction in RV deformation, but other 2D parameters assessing RV systolic function did not change after the race (Table 1). The cardiorespiratory fitness assessed in CPET was a predictor of RV performance during marathon running. Participants with higher VO2AT had lower MPI_RV post-race (r = −0.41, p < 0.05). After the race, the RV enlargement and diminishment of the LV diameter was observed: RVd MID and RVd/LVd BAS became significantly larger and LVd BAS significantly decreased (Table 1). As presented previously, the median IVSd was 11 mm, with a range of 7–17 mm [13], and in nine participants who were significantly older (44 ± 4 vs. 37 ± 9 years, p < 0.05), IVSd was > 11 mm. There were no differences in the pre- and post-marathon left and right atrial sizes, expressed as LAVI and RA area. Nevertheless, there was a correlation between the post-marathon RA enlargement and IVRT_RV (r = 0.48, p < 0.05) and MPI_RV (r = 0.58, p < 0.05). Moreover, in the post-race analysis, larger RA areas were found in subjects with more evident diastolic ventricular interaction, as indicated by an increased RVd/LVd BAS ratio (r = 0.49, p < 0.05). Notably, participants with IVRT_RV > 0 ms at baseline and those with IVS hypertrophy (IVSd > 11 mm) presented significantly larger LAVI and RA areas after the run (p < 0.05). 4. Discussion There is an ongoing debate regarding whether intense exercise, such as running a marathon, is harmful to an overloaded heart [1]. Even if marathon-induced changes in cardiac function develop, they are difficult to register and show, as documented by the discrepancies between reports on this topic. In this study, we demonstrated the post-run RV enlargement, the increase in the RV/LV ratio and ambiguous findings on the RV systolic function (manifested in the decline in RV global strain but not in TAPSE, FAC or S’RV). The analysis of MPI, which reflects both systolic and diastolic function, showed changes only at the RV free wall. Our study revealed RV diastolic dysfunction manifesting as the post-run prolonged relaxation, especially at the RV free wall and to a lesser degree at the IVS. Although we found differences in the post-race E/A ratios for both ventricles, we do not interpret these results as diastolic impairment but rather as the marathon-induced impact of increased heart rate. All abnormalities were transient and not observed in the control examination performed two weeks after the marathon run. DTI-derived IVRTc proved to be the most sensitive marker of the RV diastolic failure among amateur marathon runners. We suggest that the threshold level of exercise that causes myocardial damage is different for RV and LV, as there was no change in IVRTc at the LV lateral wall. With the variety of possible diastolic parameters, a question arises which should be considered in amateur marathon runners. The diastole consists of isovolumic relaxation, rapid filling, diastasis and atrial contraction, and each can be assessed depending on clinical indications [12]. With reference to RV, the guidelines suggest the assessment of transtricuspid E/A ratio, E/E’ ratio and RA size [15]. The recognition of abnormal LV diastolic function in patients with preserved LV EF relies on E’ velocities, average E/E’ ratio, LAVI and peak tricuspid regurgitation velocity [16]. Nevertheless, due to the preload and afterload-dependence of the diastolic indices, their measurements should be interpreted with caution [16]. With reference to diastolic function, the majority of previous studies among marathon runners have focused mainly on the LV and documented generally the post-race decrease in E/A ratio [21–24]. In our study we revealed similar findings for both ventricles, with mitral and transtricuspid inflow velocities having acted similarly. However, Doppler measurements have some limitations as they are strongly load-dependent [15,25]. Thus, the hydration status of runners is relevant, as fluid loss causes the decrease in E-, A-velocities and E/A ratio [26]. Moreover, as the duration of diastole is inversely correlated with heart rate, the atrial contraction gains significance over the rapid filling in case of Int. J. Environ. Res. Public Health 2020, 17, 5336 10 of 14 tachycardia, resulting in decreased E/A ratio [15,27]. Undoubtedly, such results do not necessarily reflect diastolic dysfunction, as the E/E’ ratio after the run remained at the same level. Therefore, we argue for the utility of PWD indices in recognition of diminished ventricular relaxation among marathon runners. IVRT is a more reliable parameter and among Doppler indices is the earliest to alter in the case of impaired relaxation [12]. Furthermore, it is reproducible and easy to obtain. It has been shown that inadequate IVRT_LV shortening during dobutamine stress echocardiography was the only Doppler diastolic parameter able to discriminate patients with residual ischemia after myocardial infarction [18]. With reference to RV, IVRT_RV correlates well with the RV systolic pressure and this non-invasive parameter is able to distinguish patients with elevated pulmonary pressure from those without [17]. Additionally, invasively measured early RV relaxation is abnormal in patients with pulmonary hypertension with preserved RV contractility [11]. In previous reports on marathon runners, no significant post-race alterations were noticed in IVRT_LV, which is consistent with our results [28,29]. However, in this study, IVRT was measured additionally for the RV free wall and septum, and to preclude the impact of heart rate increase during exercise, the heart rate-adjusted IVRT (IVRTc) was calculated. To our knowledge, our study is the first to report on the pre- and post-marathon IVRTc within the LV and RV. We showed the acute post-marathon impairment of the RV relaxation, manifesting by IVRTc_RV prolongation. In contrast to the LV, in a well-functioning RV, we rather expect IVRT to be undetectable. IVRT becomes measurable when RV impairment occurs or when RV has to overcome a significant increase in pulmonary artery pressure [19]. The fact that more than half of runners who had IVRT_RV undetectable at baseline developed it after the race is alarming and indicates that marathon running can seriously overload the RV and alter its performance. The prolongation of RV relaxation probably reflects the exercise-induced augmentation of pulmonary artery pressure. No increment in tricuspid or pulmonary regurgitation was registered in ECHO performed several minutes after the race; therefore, the pressure increase seems transient. Exercise ECHO studies in patients with pulmonary hypertension do confirm that the increase in peak tricuspid regurgitation velocity returns to baseline values briefly after exercise cessation [30]. The analysis of IVRT_RV allows us to monitor changes in pulmonary vasculature, independently of tricuspid regurgitation velocity. The recognition of detectable IVRT_RV at the baseline in 38% of the studied runners suggests the existence of pre-marathon RV overload or even subclinical damage in these subjects. Most strikingly, they showed more evident post-race diastolic impairment, with a significant increase in atrial sizes and longer IVS relaxation. The presence of measurable resting IVRT_RV raises questions about the individual’s upper limit of endurance exercise and their predisposition to training-induced cardiac fatigue. We hypothesize about the possibility of a “cumulative” alteration of the RV diastole due to repetitive RV exhaustion. Previously, it was demonstrated that greater marathon-induced decrement of RV contractility happened in less-trained runners [24]. Though we did not find a direct correlation between changes in IVRT_RV and the amount of training, cardiorespiratory fitness assessed in CPET was a predictor of global RV performance during the marathon run. The influence of the marathon run on the LV relaxation was complex. There was no significant change in IVRTc_LW, but interestingly, the IVRTc_IVS was extended. Therefore, the impairment of septal relaxation mirrored abnormalities in the RV free wall. Changes in septal function may be explained as alterations in continuity with RV damage. Although IVS is mainly a constituent part of the LV, it also shares muscle fibers with RV [31]. Therefore, IVS is involved in the functioning of both RV and LV and also transmits the altered load conditions from one ventricle to another [6]. Our study demonstrates the post-marathon increase in diastolic ventricular interaction with enlarged RV and reduced LV. In accordance with former research that showed the post-run right to left IVS displacement [32], we speculate that another factor that could account for impaired septal relaxation is enhanced ventricular dependence. Int. J. Environ. Res. Public Health 2020, 17, 5336 11 of 14 Due to its complicated structure, an accurate 2D analysis of the RV’s performance remains challenging and often requires multiparametric assessment [33]. We revealed the post-marathon decline in global RV function, demonstrated by MPI, which combines systolic and diastolic function and does not depend on heart rate or ventricular geometry [34]. In previous studies, MPI was shown to correlate well with RV EF derived from cardiac magnetic resonance (CMR) imaging and was used to assess ventricular function in many diseases [34–36]. In pulmonary arterial hypertension, MPI_RV is successful in determining the severity of the disease and in patient monitoring [36]. To our knowledge, our study is the first reporting on the pre- and post-endurance MPI and for both LV and RV [37,38]. In contrast to RV, we found no significant change in MPI_LV, which confirms the notion that marathon running does not alter the LV diastole. Notably, only an insignificant trend of MPI_IVS prolongation was observed (p = 0.09). However, it is unclear whether this is due to the IVS involvement in RV diastolic impairment or whether it reflects a decrease in septal contractility. This requires further examination to determine whether septal diastolic dysfunction precedes its systolic impairment and whether a drop in IVS contractility appears at some point during repetitive marathon attendance. With reference to septum, the negative effect of “cumulative exercise dose” was previously proved in a CMR study that analyzed the late gadolinium enhancement and found that the occurrence of myocardial fibrosis in IVS near the RV attachment correlated with the longevity of sport competition [32]. In our study, we demonstrated that the effort-induced remodeling of IVS and the degree of its hypertrophy is relevant to the RV’s performance during the marathon run. Though IVS hypertrophy is generally considered a physiological adaptation to endurance exercise, we found that the presence of IVS with a width of > 11 mm appeared to be a predictor of more evident post-race diastolic fatigue. Our results showed the impaired relaxation of the RV, with inconclusive observations regarding RV systolic function, as only RV deformation decreased. Additionally, we demonstrated that, after the marathon run, RV enlargement and LV diminishment appear. We presume that the sequence of cardiac changes started with elevated pulmonary pressure due to the exercise-induced increase in cardiac output, which influenced RV relaxation, as the first marker of developing dysfunction. Consequently, the RV enlarged and, due to pericardium constraint, the LV diminished its volume. The observation regarding post-race RA size is also very interesting. Although we did not observe changes in atrial size in the whole group, post-race RA enlargement was greater in runners with longer post-race IVRT_RV, MPI_RV and larger RV volume. Additionally, in runners with impaired RV function at the baseline (IVRT_RV > 0 ms) and in those with hypertrophied septum, the atria after the run were bigger than in runners without these abnormalities. Our study has an important limitation. As we did not include women in the study group, we cannot apply our findings to the population of female runners. Therefore, we are not able to discuss possible gender differences in RV diastolic fatigue caused by marathon running. 5. Conclusions In amateur participants, the marathon run influences predominantly the RV diastole, and post-race IVRT assessment reveals the dramatic impairment of relaxation at the RV free wall, with concomitant alteration of the IVS. The division of LV_IVRTc into IVRTc_LW and IVRTc_IVS and their separate analyses is crucial, as marathon running does not influence the relaxation of the LV lateral wall. The RV MPI and RV strain analysis, showing the decline in RV global function, follows the IVRTc observations, demonstrating that marathon attendance overburdens mainly the RV. Nevertheless, RV overload through enhanced diastolic ventricular interaction also affects LV, causing a decrease in LV cavity and alteration of IVS function. At this point, we also indicate that there is a group of runners predisposed to occurrence of diastolic dysfunction after the marathon. Participants with evidence of detectable IVRT_RV at the baseline or those with IVS hypertrophy are endangered by greater post-race cardiac exhaustion. These subjects may require constant cardiac monitoring if they continue to exercise intensively. Int. J. Environ. Res. Public Health 2020, 17, 5336 12 of 14 Author Contributions: Conceptualization, Z.L.-P., A.D.-K. and E.L.; Data curation, Z.L.-P., A.D.-K., E.L., R.G., L.D.-S., A.F., I.N.-T., M.K. and A.M.K.-D.; Formal analysis, Z.L.-P. and A.D.-K.; Funding acquisition, A.D.-K., E.L. and G.R.; Investigation, Z.L.-P., A.D.-K., E.L., R.G., L.D.-S., A.F., I.N.T., M.K., A.M.K.-D. and G.R.; Methodology, Z.L.-P., A.D.-K. and E.L.; Project administration, A.D.-K. and G.R.; Resources, A.D.-K., E.L. and G.R.; Software, Z.L.-P. and A.D.-K. Supervision, A.D.-K., E.L. and G.R.; Validation, Z.L.-P., A.D.-K., R.G., L.D.-S., A.F., I.N.T., M.K. and A.M.K.-D.; Visualization, Z.L.-P.; Writing—original draft, Z.L.-P.; Writing—review and editing, A.D.-K. and E.L. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflict of interest. References 1. Shave, R.; Oxborough, D. Exercise-induced cardiac injury: Evidence from novel imaging techniques and highly sensitive cardiac troponin assays. Prog. Cardiovasc. Dis. 2012, 54, 407–415. [CrossRef] [PubMed] 2. Leyk, D.; Erley, O.; Gorges, W.; Ridder, D.; Rüther, T.; Wunderlich, M.; Sievert, A.; Essfeld, D.; Piekarski, C.; Erren, T. Performance, training and lifestyle parameters of marathon runners aged 20–80 years: Results of the PACE-study. Int. J. Sports Med. 2009, 30, 360–365. [CrossRef] 3. La Gerche, A.; Rakhit, D.J.; Claessen, G. Exercise and the right ventricle: A potential Achilles’ heel. Cardiovasc. Res. 2017, 113, 1499–1508. [CrossRef] [PubMed] 4. Herve, P.; Lau, E.M.; Sitbon, O.; Savale, L.; Montani, D.; Godinas, L.; Lador, F.; Jaïs, X.; Parent, F.; Günther, S.; et al. Criteria for diagnosis of exercise pulmonary hypertension. Eur. Respir. J. 2015, 46, 728–737. [CrossRef] [PubMed] 5. La Gerche, A.; Heidbüchel, H.; Burns, A.T.; Mooney, D.J.; Taylor, A.J.; Pfluger, H.B.; Inder, W.J.; Macisaac, A.I.; Prior, D.L. Disproportionate exercise load and remodeling of the athlete’s right ventricle. Med. Sci. Sports Exerc. 2011, 43, 974–981. [CrossRef] [PubMed] 6. Axell, R.G.; Hoole, S.P.; Hampton-Till, J.; White, P.A. RV diastolic dysfunction: Time to re-evaluate its importance in heart failure. Heart Fail. Rev. 2015, 20, 363–373. [CrossRef] 7. Naeije, R.; Badagliacca, R. The overloaded right heart and ventricular interdependence. Cardiovasc. Res. 2017, 113, 1474–1485. [CrossRef] 8. Elliott, A.D.; La Gerche, A. The right ventricle following prolonged endurance exercise: Are we overlooking the more important side of the heart? A meta-analysis. Br. J. Sports Med. 2015, 49, 724–729. [CrossRef] 9. Brutsaert, D.L.; Rademakers, F.E.; Sys, S.U. Triple control of relaxation: Implications in cardiac disease. Circulation 1984, 69, 190–196. [CrossRef] 10. Hirota, Y. A clinical study of left ventricular relaxation. Circulation 1980, 62, 756–763. [CrossRef] 11. Murch, S.D.; La Gerche, A.; Roberts, T.J.; Prior, D.L.; MacIsaac, A.I.; Burns, A.T. Abnormal right ventricular relaxation in pulmonary hypertension. Pulm. Circ. 2015, 5, 370–375. [CrossRef] [PubMed] 12. Silbiger, J.J. Pathophysiology and echocardiographic diagnosis of left ventricular diastolic dysfunction. J. Am. Soc. Echocardiogr. 2019, 32, 216–232. [CrossRef] [PubMed] 13. Lewicka-Potocka, Z.; D ˛abrowska-Kugacka, A.; Lewicka, E.; Kaleta, A.M.; Dorniak, K.; Daniłowicz- Szymanowicz, L.; Fijałkowski, M.; Nabiałek-Trojanowska, I.; Ratkowski, W.; Potocki, W.; et al. The “athlete’s heart” features in amateur male marathon runners. Cardiol. J. 2020. [CrossRef] [PubMed] 14. Lang, R.M.; Badano, L.P.; Mor-Avi, V.; Afilalo, J.; Armstrong, A.; Ernande, L.; Flachskampf, F.A.; Foster, E.; Goldstein, S.A.; Kuznetsova, T.; et al. Recommendations for cardiac chamber quantification by echocardiography in adults: An update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J. Am. Soc. Echocardiogr. 2015, 28, 1–39. [CrossRef] [PubMed] 15. Rudski, L.G.; Lai, W.W.; Afilalo, J.; Hua, L.; Handschumacher, M.D.; Chandrasekaran, K.; Solomon, S.D.; Louie, E.K.; Schiller, N.B. Guidelines for the echocardiographic assessment of the right heart in adults: A report from the American Society of Echocardiography endorsed by the European Association of Echocardiography, a registered branch of the European Society of Cardiology, and the Canadian Society of Echocardiography. J. Am. Soc. Echocardiogr. 2010, 23, 685–713. 16. Nagueh, S.F.; Smiseth, O.A.; Appleton, C.P.; Byrd, B.F.; Dokainish, H.; Edvardsen, T.; Flachskampf, F.A.; Gillebert, T.C.; Klein, A.L.; Lancellotti, P.; et al. Recommendations for the evaluation of left ventricular diastolic function by echocardiography: An update from the American Society of Echocardiography and the European Association of Cardiovascular Imaging. J. Am. Soc. Echocardiogr. 2016, 29, 277–314. [CrossRef] Int. J. Environ. Res. Public Health 2020, 17, 5336 13 of 14 17. Zimbarra Cabrita, I.; Ruísanchez, C.; Grapsa, J.; Dawson, D.; North, B.; Pinto, F.J.; Gibbs, J.S.; Nihoyannopoulos, P. Validation of the isovolumetric relaxation time for the estimation of pulmonary systolic arterial blood pressure in chronic pulmonary hypertension. Eur. Heart J. Cardiovasc. Imaging 2013, 14, 51–55. [CrossRef] 18. Dabrowska-Kugacka, A.; Claeys, M.J.; Rademakers, F.E. Diastolic indexes during dobutamine stress echocardiography in patients early after myocardial infarction. J. Am. Soc. Echocardiogr. 1998, 11, 26–35. [CrossRef] 19. Jurcut, R.; Giusca, S.; La Gerche, A.; Vasile, S.; Ginghina, C.; Voigt, J.U. The echocardiographic assessment of the right ventricle: What to do in 2010? Eur. J. Echocardiogr. 2010, 11, 81–96. [CrossRef] 20. Badano, L.P.; Kolias, T.J.; Muraru, D.; Abraham, T.P.; Aurigemma, G.; Edvardsen, T.; D’Hooge, J.; Donal, E.; Fraser, A.G.; Marwick, T.; et al. Standardization of left atrial, right ventricular, and right atrial deformation imaging using two-dimensional speckle tracking echocardiography: A consensus document of the EACVI/ASE/Industry Task Force to standardize deformation imaging. Eur. Heart J. Cardiovasc. Imaging 2018, 19, 591–600. [CrossRef] 21. Oxborough, D.; Shave, R.; Middleton, N.; Whyte, G.; Forster, J.; George, K. The impact of marathon running upon ventricular function as assessed by 2D, Doppler, and tissue-Doppler echocardiography. Echocardiography 2006, 23, 635–641. [CrossRef] [PubMed] 22. Roeh, A.; Schuster, T.; Jung, P.; Schneider, J.; Halle, M.; Scherr, J. Two dimensional and real-time three dimensional ultrasound measurements of left ventricular diastolic function after marathon running: Results from a substudy of the BeMaGIC trial. Int. J. Cardiovasc. Imaging 2019, 35, 1861–1869. [CrossRef] [PubMed] 23. Neilan, T.G.; Yoerger, D.M.; Douglas, P.S.; Marshall, J.E.; Halpern, E.F.; Lawlor, D.; Picard, M.H.; Wood, M.J. Persistent and reversible cardiac dysfunction among amateur marathon runners. Eur. Heart J. 2006, 27, 1079–1084. [CrossRef] [PubMed] 24. Neilan, T.G.; Januzzi, J.L.; Lee-Lewandrowski, E.; Ton-Nu, T.T.; Yoerger, D.M.; Jassal, D.S.; Lewandrowski, K.B.; Siegel, A.J.; Marshall, J.E.; Douglas, P.S.; et al. Myocardial injury and ventricular dysfunction related to training levels among nonelite participants in the Boston marathon. Circulation 2006, 114, 2325–2333. [CrossRef] [PubMed] 25. O˘guzhan, A.; Arinç, H.; Abaci, A.; Topsakal, R.; Kemal, E.N.; Ozdo˘gru, I.; Basar, E.; Ergin, A. Preload dependence of Doppler tissue imaging derived indexes of left ventricular diastolic function. Echocardiography 2005, 22, 320–325. [CrossRef] [PubMed] 26. Alarrayed, S.; Garadah, T.S.; Alawdi, A.A. The impact of left ventricular preload reduction on cardiac pulsed Doppler indices during hemodialysis and its relation to intra-dialysis hypotension: A pulsed Doppler study. Saudi J. Kidney Dis. Transplant. 2009, 20, 201–207. 27. Burns, A.T.; Connelly, K.A.; La Gerche, A.; Mooney, D.J.; Chan, J.; MacIsaac, A.I.; Prior, D.L. Effect of heart rate on tissue Doppler measures of diastolic function. Echocardiography 2007, 24, 697–701. [CrossRef] 28. Alshaher, M.; El-Mallakh, R.; Dawn, B.; Siddiqui, T.; Longaker, R.A.; Stoddard, M.F. Cardiac manifestations of exhaustive exercise in nonathletic adults: Does cardiac fatigue occur? Echocardiography 2007, 24, 237–242. [CrossRef] 29. Oxborough, D.; Whyte, G.; Wilson, M.; O’Hanlon, R.; Birch, K.; Shave, R.; Smith, G.; Godfrey, R.; Prasad, S.; George, K. A depression in left ventricular diastolic filling following prolonged strenuous exercise is associated with changes in left atrial mechanics. J. Am. Soc. Echocardiogr. 2010, 23, 968–976. [CrossRef] 30. Wierzbowska-Drabik, K.; Picano, E.; Bossone, E.; Ciampi, Q.; Lipiec, P.; Kasprzak, J.D. The feasibility and clinical implication of tricuspid regurgitant velocity and pulmonary flow acceleration time evaluation for pulmonary pressure assessment during exercise stress echocardiography. Eur. Heart J. Cardiovasc. Imaging 2019, 20, 1027–1034. [CrossRef] 31. Buckberg, G.; Hoffman, J.I. Right ventricular architecture responsible for mechanical performance: Unifying role of ventricular septum. J. Thorac. Cardiovasc. Surg. 2014, 148, 3166–3171. [CrossRef] [PubMed] 32. La Gerche, A.; Burns, A.T.; Mooney, D.J.; Inder, W.J.; Taylor, A.J.; Bogaert, J.; Macisaac, A.I.; Heidbüchel, H.; Prior, D.L. Exercise-induced right ventricular dysfunction and structural remodelling in endurance athletes. Eur. Heart J. 2012, 33, 998–1006. [CrossRef] [PubMed] 33. Smolarek, D.; Gruchała, M.; Sobiczewski, W. Echocardiographic evaluation of right ventricular systolic function: The traditional and innovative approach. Cardiol. J. 2017, 24, 563–572. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2020, 17, 5336 14 of 14 34. Meric, M.; Yesildag, O.; Yuksel, S.; Soylu, K.; Arslandag, M.; Dursun, I.; Zengin, H.; Koprulu, D.; Yilmaz, O. Tissue doppler myocardial performance index in patients with heart failure and its relationship with haemodynamic parameters. Int. J. Cardiovasc. Imaging 2014, 30, 1057–1064. [CrossRef] [PubMed] 35. Hamilton-Craig, C.R.; Stedman, K.; Maxwell, R.; Anderson, B.; Stanton, T.; Chan, J.; Yamada, A.; Scalia, G.M.; Burstow, D.J. Accuracy of quantitative echocardiographic measures of right ventricular function as compared to cardiovascular magnetic resonance. Int. J. Cardiol. Heart Vasc. 2016, 23, 38–44. [CrossRef] 36. Ogihara, Y.; Yamada, N.; Dohi, K.; Matsuda, A.; Tsuji, A.; Ota, S.; Ishikura, K.; Nakamura, M.; Ito, M. Utility of right ventricular Tei-index for assessing disease severity and determining response to treatment in patients with pulmonary arterial hypertension. J. Cardiol. 2014, 63, 149–153. [CrossRef] 37. Kasikcioglu, E.; Oflaz, H.; Akhan, H.; Kayserilioglu, A. Right ventricular myocardial performance index and exercise capacity in athletes. Heart Vessels 2005, 20, 147–152. [CrossRef] 38. Tüzün, N.; Ergün, M.; Alio˘glu, E.; Edem, E.; Tengiz, I.; Aytemiz, F.; Ercan, E.; ˙I¸sle˘gen, Ç. TEI Index in elite sprinters and endurance athletes. J. Sports Med. Phys. Fit. 2015, 55, 988–994. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Right Ventricular Diastolic Dysfunction after Marathon Run.

07-24-2020

Lewicka-Potocka, Zuzanna,Dąbrowska-Kugacka, Alicja,Lewicka, Ewa,Gałąska, Rafał,Daniłowicz-Szymanowicz, Ludmiła,Faran, Anna,Nabiałek-Trojanowska, Izabela,Kubik, Marcin,Kaleta-Duss, Anna Maria,Raczak, Grzegorz

eng

PMC9986668

1 Vol.:(0123456789) Scientific Reports | (2023) 13:3692 | https://doi.org/10.1038/s41598-023-30932-1 www.nature.com/scientificreports Changes in pacing variation with increasing race duration in ultra‑triathlon races Mirko Stjepanovic 1, Beat Knechtle 1,2*, Katja Weiss 2, Pantelis Theodoros Nikolaidis 3, Ivan Cuk 4, Mabliny Thuany 5 & Caio Victor Sousa 6 Despite the increasing scientific interest in the relationship between pacing and performance in endurance sports, little information is available about pacing and pacing variation in ultra‑endurance events such as ultra‑triathlons. Therefore, we aimed to investigate the trends of pacing, pacing variation, the influence of age, sex, and performance level in ultra‑triathlons of different distances. We analysed 969 finishers (849 men, 120 women) in 46 ultra‑triathlons longer than the original Ironman® distance (e.g., Double‑, Triple‑, Quintuple‑ and Deca Iron ultra‑triathlons) held from 2004 to 2015. Pacing speed was calculated for every cycling and running lap. Pacing variation was calculated as the coefficient of variation (%) between the average speed of each lap. Performance level (i.e., fast, moderate, slow) was defined according to the 33.3 and 66.6 percentile of the overall race time. A multivariate analysis (two‑way ANOVA) was applied for the overall race time as the dependent variable with ‘sex’ and ‘age group’ as independent factors. Another multivariate model with ‘age’ and ‘sex’ as covariates (two‑way ANCOVA) was applied with pacing variation (cycling and running) as the dependent variable with ‘race’ and ‘performance level’ as independent factors. Different pacing patterns were observed by event and performance level. The general pacing strategy applied was a positive pacing. In Double and Triple Iron ultra‑triathlon, faster athletes paced more evenly with less variation than moderate or slower athletes. The variation in pacing speed increased with the length of the race. There was no significant difference in pacing variation between faster, moderate, and slower athletes in Quintuple and Deca Iron ultra‑triathlon. Women had a slower overall performance than men. The best overall times were achieved at the age of 30–39 years. Successful ultra‑triathlon athletes adapted a positive pacing strategy in all race distances. The variation in pacing speed increased with the length of the race. In shorter ultra‑triathlon distances (i.e., Double and Triple Iron ultra‑triathlon), faster athletes paced more evenly with less variation than moderate or slower athletes. In longer ultra‑triathlon distances (i.e., Quintuple and Deca Iron ultra‑triathlon), there was no significant difference in pacing variation between faster, moderate, and slower athletes. An ultra-endurance performance is defined as any endurance performance of six hours in duration or longer1. A triathlon is characterized by the successive completion of three disciplines (e.g., swimming, cycling, and run- ning). Therefore, an Ironman® triathlon with 3.8 km swimming, 180 km cycling, and 42.2 km running must be considered an ultra-endurance performance considering the men’s world record being at 7:21:12 h:min:s2. In addition to the single Ironman® distance, ultra-triathlon races of x-times the Ironman® distance exist, such as the Double Iron ultra-triathlon (7.6 km swimming, 360 km cycling and 84 km running), the Triple Iron ultra-triathlon (11.4 km swimming, 540 km cycling and 126.6 km running), the Quintuple Iron ultra-triathlon (19 km swimming, 900 km cycling and 211 km running), and the Deca Iron ultra-triathlon (38 km swimming, 1800 km cycling and 422 km running). The popularity of ultra-triathlons is increasing, indicated by a rising number of participants3–7. For example, participation trends during 1985–2009 showed an increase in both Double and Triple Iron (i.e., 17–98 per year in Double and 7–41 per year in Triple Iron, respectively)8. Similar results were shown by Sigg et al., in which the number of triathletes from Europe has increased for both sexes, while the numbers for North America increased only for the woman athletes9. OPEN 1Institute of Primary Care, University of Zurich, Zurich, Switzerland. 2Medbase St. Gallen Am Vadianplatz, Vadianstrasse 26, 9001 St. Gallen, Switzerland. 3School of Health and Caring Sciences, University of West Attica, Athens, Greece. 4Faculty of Sport and Physical Education, University of Belgrade, Belgrade, Serbia. 5Faculty of Sports, University of Porto, CIFI2D Porto, Portugal. 6Health and Human Sciences, Loyola Marymount University, Los Angeles, CA, USA. *email: [email protected] 2 Vol:.(1234567890) Scientific Reports | (2023) 13:3692 | https://doi.org/10.1038/s41598-023-30932-1 www.nature.com/scientificreports/ An important aspect regarding performance in endurance events is the distribution of exercise intensity. Athletes need to distribute their metabolic energy as fuel for their activity to avoid premature fatigue by going too fast early on10. This has been often termed as ‘pacing’11,12. There are six different pacing strategies such as positive pacing (i.e., slowing over time), negative pacing (i.e. increase in speed over time), all-out pacing (i.e., maximal speed possible), even pacing (i.e. same speed over time), variable pacing (i.e., pacing with multiple fluctuations) and parabolic-shaped pacing (i.e., positive and negative pacing in different segments of the race)13. Research regarding pacing has increased over the past years10–12,14–20. Positive pacing has been shown as the adequate strat- egy adapted in many races and disciplines (i.e., for elite Ironman® triathletes, ultra-cyclists, and swimmers)21–23. Race distance should be considered a factor that influences pacing strategy13. It has been shown that athletes pace positively in long and ultra-long races22. Considering triathlon, in both an Olympic distance and an Iron- man® 70.3 distance triathlon, pacing strategies were strongly influenced by both the distance and the discipline17. Regarding longer triathlon distances, finishers of a one per day Deca Iron ultra-triathlon applied positive pacing20. At the ‘Ultraman Hawaii’, women paced differently than men (i.e. applied a more even pace after increasing speed at the start of the race)16. However, no study systematically investigates the effect of race distances on pacing strategies applied in very long triathlon distances such as Double Iron-, Triple Iron-, Quintuple Iron- and Deca Iron ultra-triathlon. Another aspect is the performance level, where faster finishers pace differently than slower finishers and elite athletes pace differently than recreational athletes24–26. In ultra-endurance running such as 100-km ultra- marathon running, faster runners had fewer changes in running speed and started at a higher running speed while maintaining running speed longer27. Among marathoners, slower finishers showed a greater pace variability than faster finishers26. To the best of our knowledge, there is no study investigating the influence of performance level on pacing in ultra-triathlon. Several physiological and psychological factors can influence both performance and pacing during endur- ance and an ultra-endurance performance, such as age5, sex16, race distance11, performance level28, mechanical damage to muscle fibers29, increasing body temperature30, reduction in neuromuscular activity31 and muscle glycogen depletion8. Additionally, gender seems to influence pacing during a triathlon race32,33. In elite Ironman® triathletes, women were significantly slower when applying the same positive pacing strategy15. However, at the ‘Ultra- man Hawaii’, a multistage event consisting of 10 km swimming, 165 km cycling (day 1), 261 km cycling (day 2) and 85 km running (day 3), women applied a more even pacing strategy compared to men. The fastest women decreased performance on day 1 and could then maintain on days 2 and 3, whereas the fastest men impaired performance on days 1 and 2 but improved on day 316. A comparison of top performers from the Ironman® to the Double Deca Iron ultra-triathlon distance showed that men were faster than women and that the sex differ- ence increased for swimming, running, and overall race time but not for cycling34. There is, however, no study investigating all finishers (i.e., not only the top athletes). Age has an influence on endurance performance and pacing. Regarding triathlon, performance and age- related trends in elite triathletes competing in age group classes have been investigated for an Ironman® triath- lon such as ‘Ironman® Hawaii’16,35 and short distances such as the Olympic distance triathlon36. An age-related performance decline has been shown for all investigated distances37. In Ironman® triathletes, the age-related performance decline started at a higher age than in short distance triathletes3. Another important aspect is the age of peak performance. When Olympic, Ironman® 70.3 and Ironman® distance races were compared, the age of peak triathlon performance was higher in the longer triathlon race distances38. For ultra-triathlon, Deca Iron ultra-triathletes were older than Triple Iron ultra-triathletes39. How- ever, the age-related performance and the age of peak performance for the Double and Quintuple Iron ultra- triathlon distance are unknown. Based on the limited existing literature regarding the effect of the race distance, age, gender, and performance level on pacing in ultra-triathlon races we hypothesized, firstly, that a positive pacing strategy would be applied on the shorter distances (e.g., Double and Triple Iron ultra-triathlon) and an even pacing on the longer ones (e.g., Quintuple and Deca Iron ultra-triathlon). Secondly, we hypothesized that slower finishers apply more variable pacing. Thirdly, we hypothesized to find differences in pacing between men and women. We hypothesized that the age-related declines would occur later at the longer distances (e.g., Quintuple and Deca Iron ultra-triathlon) compared to the shorter distances (e.g., Double and Triple Iron ultra-triathlon). Materials and methods Ethical approval. The Institutional Review Board of Kanton St. Gallen, Switzerland approved all procedures used in the study with a waiver of the requirement for informed consent of the participants given the fact that the study involved the analysis of publicly available data (01/06/2010). The study was conducted in accordance with recognized ethical standards according to the Declaration of Helsinki adopted in 1964 and revised in 2013. Subjects. We focused on distances longer than the original Ironman® distance, like the Double Iron, Tri- ple Iron, Quintuple Iron and Deca Iron ultra-triathlon. Overall results are available on the International Ultra Triathlon Association homepage (IUTA, https:// www. iutas port. com/). We contacted all race directors directly for the lap times in running, cycling and for the contestants’ age. Lap times were provided electronically in in spreadsheets. Furthermore, the age of all finishers was provided. The athletes were sorted into age groups of 10-year age intervals. Lap distances were different for each race as every race has its course. Tables 1 and 2 provide an overview of the number of finishers for each race. We excluded all non-finishers. The number of calculated laps for each race type was provided additionally. 3 Vol.:(0123456789) Scientific Reports | (2023) 13:3692 | https://doi.org/10.1038/s41598-023-30932-1 www.nature.com/scientificreports/ The final analysis included all finishers, in total 969 athletes (849 men, 120 women) competing in a Double Iron, Triple Iron, Quintuple Iron, and Deca Iron ultra-triathlon. We calculated the pacing speed from lap distance and time per lap for every cycling and running lap. We excluded swimming times for the analysis because no lap times were recorded. However, evidence suggests that variations of pacing during swimming are not apparent11. One race had to be excluded because of missing split times (Triple Iron ultra-triathlon in Bad Blumau, Austria). We included only official finishers. Finishers with improbable speed values (i.e., cycling speed > 48 km/h and running speed > 20 km/h) were also excluded. The last races included were the Double and Triple Iron ultra- triathlon in Florida in 2020, as the other races had been canceled due to the COVID-19 pandemic. Statistical analyses. Data normality and homogeneity were confirmed with Kolmogorov–Smirnov and Levene’s test, respectively. Pacing variation was calculated individually as the coefficient of variation (%) between the average speed of each lap in the close circuit the athletes performed their ultra-triathlons. Each cycling lap ranges from 4 to 16.4 km and the running lap ranges from 1.4 to 3.5 km. Total distance is standard for all athletes. The performance level was defined according to the 33.3 and 66.6 percentile of overall race time for women and men. The first percentile (≤ 33.3 percentile) was named as ‘fast’, the last percentile (≥ 66.6 percentile) was named as ‘slow’, and the intermediate was named as ‘moderate’26. A multivariate model with two factors (two-way ANOVA) was applied for race time performance (overall, swimming, cycling, running) as the dependent vari- able with ‘sex’ and ‘age group’ as independent factors. Quintuple and Deca Iron ultra-triathlon were not included in the previous model because of the low number of participants across the age groups. Another multivariate model with two factors with ‘age’ and ‘sex’ as covariates (two-way ANCOVA) was applied with pacing variation (cycling and running) as dependent variable, ‘race’ and ‘performance level’ as independent factors. Partial eta squared (ηp 2) was calculated for the ANOVAs where the values of the effect size 0.01, 0.06, and above 0.14 were considered small, medium, and large, respectively40. Statistical significance was defined as p < 0.05. All statistical analyses were carried out with Statistical Software for the Social Sciences (IBM® SPSS v.25, Chicago, Ill, USA). Table 1. Number of finishers for each race type and number of excluded finishers. Races Double Triple Quintuple Deca Excluded Total Male 346 499 23 23 42 849 Finishers Female 58 52 6 6 2 120 Overall 404 551 29 29 44 969 Laps Bike 22,184 38,334 3560 5745 69,823 Run 19,231 51,404 3311 9818 83,764 Table 2. All races included in the statistical analysis sorted by race type and year. Double Triple Quintuple Deca 2014 Virginia (USA) 2004 Lehnsahn (GER) 2016 Virginia (USA) 2017 Buchs (CH) 2015 Emsdetten (GER) 2005 Lehnsahn (GER) 2017 Buchs (CH) 2018 Buchs (CH) 2015 Florida (USA) 2006 Lehnsahn (GER) 2018 Buchs (CH) 2015 Virginia (USA) 2007 Lehnsahn (GER) 2019 Virginia (USA) 2015 Oregon (USA) 2008 Lehnsahn (GER) 2016 Florida (USA) 2009 Lehnsahn (GER) 2016 Oregon (USA) 2010 Lehnsahn (GER) 2016 Virginia (USA) 2011 Lehnsahn (GER) 2017 Florida (USA) 2012 Lehnsahn (GER) 2017 Emsdetten (GER) 2013 Lehnsahn (GER) 2017 Oregon (USA) 2014 Lehnsahn (GER) 2017 Panevėžys (LT) 2015 Virginia (USA) 2017 Virginia (USA) 2015 Lehnsahn (GER) 2018 Florida (USA) 2016 Lehnsahn (GER) 2018 Oregon (USA) 2016 Virginia (USA) 2019 Florida (USA) 2017 Virginia (USA) 2019 Emsdetten (GER) 2018 Lehnsahn (GER) 2019 Virginia (USA) 2019 Lehnsahn (GER) 2020 Florida (USA) 2019 Lehnsahn (GER) 2019 Virginia (USA) 2020 Florida (USA) 4 Vol:.(1234567890) Scientific Reports | (2023) 13:3692 | https://doi.org/10.1038/s41598-023-30932-1 www.nature.com/scientificreports/ Results The final analysis included 386 athletes competing in a Double Iron, 539 in a Triple Iron, 15 in a Quintuple Iron, and 29 in Deca Iron ultra-triathlon (n = 969). For all race distances, the majority of athletes were aged 40–49 years, followed by 30–39 years (Fig. 1). The multivariate model for Double Iron ultra-triathlon with overall race time as the dependent variable showed that ‘age group’ was a significant factor for swimming (F = 2.48, p = 0.044, ηp 2 = 0.026), and running (F = 5.99, p < 0.001, ηp 2 = 0.060), but not for cycling or overall performance. ‘Sex’ showed significant effects for swimming (F = 4.17, p = 0.042, ηp 2 = 0.011), cycling (F = 6.36, p = 0.012, ηp 2 = 0.017), overall (F = 6.15, p = 0.014, ηp 2 = 0.016), and a borderline effect for running (F = 3.03, p = 0.082, ηp 2 = 0.008). There were no significant (p > 0.05) interactions for ‘age group × sex’. Pairwise comparisons showed that swimming performance for men was faster in age groups 30–39 and 40–49 years. The female age group 30–39 years was faster than older age groups, and the male age group ≥ 60 years was slower than age groups 40–49 and 50–59 years (Fig. 2A). Men’s cycling performance in age groups 30–39 and 40–49 years was faster than in the same age groups for women (Fig. 2B). Men in the age group 30–39 years ran faster than others, followed by 40–49 years (Fig. 2C). For women, athletes in the age group 30–39 years were faster than athletes in age groups ≤ 29 and 50–59 years (Fig. 2D). Women had a slower overall performance in age groups 30–39 and 40–49 years. Men in the age group 30–39 years were faster than all the others, except the age group 40–49 years (Fig. 2D). For the Triple Iron ultra-triathlon, the multivariate model showed that ‘age group’ was a significant factor only for cycling (F = 2.66, p = 0.032, ηp 2 = 0.020). ‘Sex’ showed significant effects for cycling (F = 16.35, p < 0.001, ηp 2 = 0.030), running (F = 6.55, p = 0.011, ηp 2 = 0.012), overall (F = 13.03, p < 0.001, ηp 2 = 0.024), and a border- line effect for swimming (F = 2.91, p = 0.089, ηp 2 = 0.005). There were no significant (p > 0.05) interactions ‘age- group × sex’. Pairwise comparisons showed that swimming performance for men was faster in the age group 40–49 years. Men aged 50 or older swam slower than men in other age groups (Fig. 2E). Women’s cycling perfor- mance was slower than for men, and men aged 60 years or older were slower than all others (Fig. 2F). Similarly, running and overall performance were lower in women, and men aged ≥ 60 years were slower than all others, whereas athletes in the age group 30–39 were faster than in age groups 40–49 and 50–59 years (Fig. 2G and H). The multivariate model with pacing variation as the dependent variable showed that ‘performance level’ was a significant factor for both cycling (F = 10.51, p < 0.001, ηp 2 = 0.022), and running (F = 63.63, p < 0.001, ηp 2 = 0.118). Similarly, ‘race’ was also significant for both cycling (F = 173.12, p < 0.001, ηp 2 = 0.352), and running (F = 62.97, p < 0.001, ηp 2 = 0.165). The interaction ‘performance level × race’ was significant for cycling (F = 2.31, p = 0.032, ηp 2 = 0.014) and running (F = 36.15, p < 0.001, ηp 2 = 0.185). Pairwise comparisons showed that pacing variation was lower for faster athletes cycling in Double and Triple Iron ultra-triathlon and running in Triple Iron ultra-triathlon. Cycling pacing variation in Double Iron ultra-triathlon was the lowest across all performance levels. It was lower in Triple Iron ultra-triathlon than in Quintuple and Deca Iron ultra-triathlon. For running, all performance levels of Double Iron ultra-triathlon showed a lower pacing variation than in the Triple Iron ultra-triathlon. Moderate and fast athletes were also different from the Quintuple Iron ultra-triathlon and slow and fast athletes were different from the Deca Iron ultra-triathlon (Fig. 3). Discussion This study intended to investigate the trends in pacing and the effects of performance level, sex and age in ultra- triathlons with the primary hypothesis of a positive pacing strategy in Double and Triple ultra-triathlon and an even pacing strategy in Quintuple and Deca ultra-triathlon. Figure 1. Number of participants in each age-group for Double (A), Triple (B), Quintuple (C), and Deca (D) ultra-triathlon. 5 Vol.:(0123456789) Scientific Reports | (2023) 13:3692 | https://doi.org/10.1038/s41598-023-30932-1 www.nature.com/scientificreports/ Figure 2. Race performance (minutes) by age groups of men and women competing in Double Iron (A–D) and Triple Iron (E–H) ultra-triathlon. *: different between sex, within the same age group, #: different from all the other age groups, a: different from age group ≤ 29 years, b: different from age group 30–39 years, c: different from age group 40–49 years, d: different from age group 50–59 years, e: different from age group ≥ 60 years. Figure 3. Cycling and running pacing variation in Double, Triple, Quintuple, and Deca Iron ultra-triathlon by different performance levels. *: different from all other performance levels within the same race, b: different from Triple within the same performance level, c: different from Quintuple within the same performance level, d: different from Deca within the same performance level. 6 Vol:.(1234567890) Scientific Reports | (2023) 13:3692 | https://doi.org/10.1038/s41598-023-30932-1 www.nature.com/scientificreports/ The first important finding was a positive pacing strategy applied in all distances confirming our hypothesis for Double and Triple ultra-triathlon. This is consistent with other studies, where positive pacing strategies were observed in long-distance events. In 2014, the top 100 Ironman® finishers adopted a positive pacing strategy in most races15. Two studies investigating 6 and 23 finishers competing in Deca Iron ultra-triathlon showed that split and overall race times increased linearly across the ten days20,41. However, several studies showed athletes adopting an even pacing strategy41–43. Male Triple Deca Iron ultra-triathletes competing for 30 days finishing an Ironman® distance triathlon daily could maintain their performance5. Additionally, a male triathlete competing for 33 days and finishing an Ironman® distance triathlon daily was able to maintain his performance in cycling, running, and overall race times42. The best performing female in Quintuple and Deca Iron ultra-triathlon applied an even pacing strategy during her two world record races43. A potential explanation for the consistent pacing could be her background as an elite cyclist and her previous experience43. Another important aspect is that we compared continuous races instead of one per day races. In one-a-day events, the change between disciplines might give enough rest to allow a steady pace throughout the race. A second important finding was the difference in pacing variation. We hypothesized that slower finishers would apply more variable pacing. Faster finishers showed a significantly lower variation in cycling pacing in Double and Triple Iron ultra-triathlon and running pacing in the Triple Iron ultra-triathlon, confirming our hypothesis. Although not significant, faster athletes running in a Double Iron ultra-triathlon also showed the low- est pacing variation. This is most likely explained by more ‘moderate’ athletes being split into three even groups. A recent study regarding elite triathletes showed that a lower variability in race pacing during a10-km run also reflected more successful run times44. Keeping a low pacing variation seems to be the appropriate strategy for most distances and disciplines up to a Triple Iron ultra-triathlon. Pacing variation is also different within disciplines. The variation in pace was lower in cycling than in run- ning. This result corroborates with data of the world’s best female ultra-triathlete in Quintuple and Deca Iron ultra-triathlon where she showed a higher pacing variation in running than in cycling43. Drinking and eating in such events is often reported to be harder in running than in cycling, as a quick stop in running reduces the speed to zero, whereas in cycling the speed drops slowly even when one stops pedaling43. In the longer race distances, the pacing variation increased. There were no significant changes in pacing vari- ation between the different performance levels in the Quintuple and Deca Iron ultra-triathlon. These contrasting findings might be due to the low participation number in Quintuple and Deca Iron ultra-triathlon. There is no systematic research comparing pacing variation in ultra-triathlon to this day. With increasing distance, exercise economy seems harder to maintain or limited (i.e., stops for sleeping and nutrition), resulting in a more signifi- cant pacing variation in longer distances45. A further important finding was that men paced faster than women confirming our hypothesis of different pacing between men and women. Women in age groups 30–39 and 40–49 years were overall significantly slower than men. The difference between sexes increased with increasing race distance for swimming, running, and overall race time. However, there was no significant difference in swimming time between men and women in age group 30–39 years in Triple Iron ultra-triathlon. This finding is consistent with an analysis of open-water ultra- distance swimmers from 5 to 25 km, where the difference between sexes was lowest in 10 km compared to 5 km and 25  km46. For ultra-swimming, women seemed to achieve a similar or even better performance than men34,47. The difference between sexes in endurance performance is primarily caused by physiological differences in VO2 max48 and anthropometric characteristics such as skeletal muscle mass and body fat49. It has been shown that female ultra-runners have lower skeletal muscle mass and a higher percentage of body fat than male ultra- runners, which may disadvantage women in ultra-running performance50. Regarding age, most participants were in age group 40–49 years, followed by the age group 30–39 years for all race distances. We hypothesized that ultra-triathletes would reach peak performance at a higher age compared to Ironman® triathletes. We could confirm this in the Double Iron ultra-triathlon, where men achieved the best performances overall in age groups 30–39 and 40–49 years while in Ironman® triathlon, peak performance for men was reported to be at 32.2 ± 1.5  years51. Regarding running performance in Double Iron ultra-triathlon, athletes in the age group 30–39 years were the fastest. Several studies showed that endurance and ultra-endurance performance in disciplines such as Ironman® triathlon and ultra-marathon appeared to be maintained until the age of 35–40 years, followed by modest decreases until 50 years of age and a progressive decrease in performance thereafter37,52,53. Women in Double Iron ultra-triathlon showed no difference in overall performance across all age groups. In swimming, athletes in the age group 40–49 and 50–59 years were significantly faster than athletes in the age group 30–39 years. In a study regarding the age-related performance decline in Ironman® triathlon, the decline in swimming performance already started in the age group 25–29 years for both women and men3. A possible explanation for our finding could be the low participation number of women in long endurance events such as Iron ultra-triathlon or ultra-marathons. A study investigating participation trends in Triple Iron ultra-triathlon from 1988 to 2011 showed a stable participation rate of ∼8%54. In other ultra-endurance distances such as 100- km ultra-marathons, female participation was higher at ∼13% but still very low50. Consequently, the athletes competing have a good estimation of the extent of their performance due to their accumulated experience and therefore applied a pacing strategy that was right and correct. In addition, athletes might stop competing after an injury, so older athletes result from a further selection process that could be described as ’survival of the fittest3. Another factor for the increase in performance in women in older age groups could be that the motivation in achieving best times has more importance for women competing in these races55. In contrast to our hypothesis, in Triple Iron ultra-triathlon, athletes in the age group 30–39 years were faster than athletes in the age group 40–49 years, suggesting a similar age of peak performance as reported in Ironman® triathlon for men51. A previous study investigating both Triple Iron ultra-triathletes and Deca Iron ultra-triathletes showed that athletes were able to maintain their best performances for ages comprised between 7 Vol.:(0123456789) Scientific Reports | (2023) 13:3692 | https://doi.org/10.1038/s41598-023-30932-1 www.nature.com/scientificreports/ 25 and 44 years, independent of the race distance39. To summarize, athletes in only one age group (10-years interval) between 40 and 49 years might hide an age-related decline in the age group 44–49 years (e.g., sampling bias). Peak performance might still be achievable between 40 and 44 years. There were no significant differences between the age groups in women participating in the Triple Iron ultra-triathlon. This is primarily explained by the low participation number of women. Quintuple and Deca Iron ultra-triathletes were not included because of the low number of participants across the age groups. Age-related declines in endurance and ultra-endurance performance have been well described in the literature for running56, cycling57, swimming58 and for triathlon37,59. The duration of a triathlon race exerts an important influence on the age-related changes in triathlon performance60,61. The age-related performance decline in Ironman® triathlon starts with 25–29 years in swimming for both women and men and in the age group 35–39 years for men respective age 30–34 years for women in cycling, running and overall performance3. We hypothesized that the age-related declines would occur later in the longer distances (e.g., Quintuple and Deca Iron ultra-triathlon) compared to the shorter distances (e.g., Double and Triple Iron ultra-triathlon). We could confirm our hypothesis for overall performance as in Double Iron ultra- triathlon the decline started with age group 50–59 years. However, our hypothesis could not be confirmed for the Triple Iron ultra-triathlon where the age group 40–49 years was significantly slower than 30–39 years. In swimming, there were no significant differences for men until an age-related decline in the age group 50–59 years in Double Iron ultra-triathlon and age > 60 years in Triple Iron ultra-triathlon. No significant differences were observed in cycling times for all age groups. The age-related decline for running started for both Double and Triple Iron ultra-triathletes in the age group 40–49 years. This is consistent with other studies, finding that the age-related decline in swimming and cycling was less pronounced39,59,62. The decrease in running performance for men in Triple Iron ultra-triathlon seems to be much more impactful on overall performance than in Double Iron ultra-triathlon. Quintuple and Deca Iron ultra-triathlon were not included because of the low number of participants across the age groups. Strength, weakness, limitations and implications for future research. The strength of this study is the large data set including all finishers of ultra-triathlon races held worldwide between 2004 and 2020, with a total of 849 male and 120 female finishers in 46 ultra-triathlons. However, some races might not have been documented on the official IUTA website. In earlier years, lap times were not recorded electronically. Another strength of this study is its novelty, as it is the first one containing a detailed analysis of the pacing variation within the laps of the races. A weakness is that we were not able to consider environmental conditions63–65 and individual factors such as anthropometric49 and nutritional characteristics66,67 and previous experience49. Besides environmental conditions, other race characteristics can influence pacing, such as elevation in cycling and running, swimming in open water (sea, lake) or a pool. Another weakness is the low participation number of women and the low participation numbers in Quintuple and Deca Iron ultra-triathlon. This is due to the low number of races held for these distances. We did not include non-finishers, so including those might give further insight into which pacing strategies result in failure. Future studies would need to include more female participants and a higher number of participants across all age groups in Quintuple and Deca ultra-triathlon. Conclusion Ultra-triathletes adapted a positive pacing strategy, i.e., speed decreased over the duration of the race. In addition, faster athletes show less variation in pacing then moderate and slower athletes in Double and Triple Iron ultra- triathlon. Pacing variation differed between the disciplines, with the lowest variation in cycling. Consequently, ultra-triathletes should be advised to adopt less variable pacing while maintaining a positive pacing strategy. For professional athletes, an even pacing strategy might be achievable. Men pace faster than women. The sex differ- ence increased with increasing race distance for swimming, running, and overall race time. It seems unlikely that women will outperform men with increasing distance. The best age for men in Double Iron ultra-triathlon is between 30 and 49 years and in Triple Iron ultra- triathlon between 30 and 39 years. The age-related performance decline in the age group 40–49 years is due to the lower running performance. This leads to a significant decrease in overall performance in the Triple Iron ultra-triathlon. Training should emphasize that finding by focusing more on improvement in running times. Data availability The athletes’ data was downloaded from the official IUTA website (www. iutas port. com/). Received: 16 April 2022; Accepted: 3 March 2023 References 1. Zaryski, C. & Smith, D. J. Training principles and issues for ultra-endurance athletes. Curr. Sports Med. Rep. 4, 165–170 (2005). 2. Guinness World Records, Official site, 2023. 3. Käch, I. W., Rüst, C. A., Nikolaidis, P. T., Rosemann, T. & Knechtle, B. The age-related performance decline in ironman triathlon starts earlier in swimming than in cycling and running. J. Strength Cond. Res. 32, 379–395 (2018). 4. Whyte, G. Age, sex and (the) race: Gender and geriatrics in the ultra-endurance age. Extrem. Physiol. Med. 3, 1 (2014). 5. Knechtle, B., Zingg, M. A., Rosemann, T., Stiefel, M. & Rüst, C. A. What predicts performance in ultra-triathlon races? A com- parison between Ironman distance triathlon and ultra-triathlon. Open Access J. Sports Med. 6, 149–159 (2015). 6. Meili, D., Knechtle, B., Rüst, C. A., Rosemann, T. & Lepers, R. Participation and performance trends in “Ultraman Hawaii” from 1983 to 2012. Extrem. Physiol. Med. 2, 25 (2013). 7. Sigg, K. et al. Sex difference in Double Iron ultra-triathlon performance. Extrem. Physiol. Med. 2, 12 (2013). 8 Vol:.(1234567890) Scientific Reports | (2023) 13:3692 | https://doi.org/10.1038/s41598-023-30932-1 www.nature.com/scientificreports/ 8. Rauch, H. G., St Clair Gibson, A., Lambert, E. V. & Noakes, T. D. A signalling role for muscle glycogen in the regulation of pace during prolonged exercise. Br. J. Sports Med. 39, 34–38 (2005). 9. Sigg, K. et al. Central European triathletes dominate Double Iron ultratriathlon - analysis of participation and performance 1985–2011. Open Access J. Sports Med. 3, 159–168 (2012). 10. Skorski, S. & Abbiss, C. R. The manipulation of pace within endurance sport. Front. Physiol. 8, 102 (2017). 11. Wu, S. S., Peiffer, J. J., Brisswalter, J., Nosaka, K. & Abbiss, C. R. Factors influencing pacing in triathlon. Open Access J. Sports Med. 5, 223–234 (2014). 12. Le Meur, Y. et al. Relationships between triathlon performance and pacing strategy during the run in an international competition. Int. J. Sports Physiol. Perform. 6, 183–194 (2011). 13. Abbiss, C. R. & Laursen, P. B. Describing and understanding pacing strategies during athletic competition. Sports Med. (Auckland, N.Z.) 38, 239–252 (2008). 14. Nikolaidis, P. T. & Knechtle, B. Pacing strategies by age in marathon cross-country skiing. Phys. Sportsmed. 1–7 (2018). 15. Angehrn, N., Rüst, C. A., Nikolaidis, P. T., Rosemann, T. & Knechtle, B. Positive pacing in elite IRONMAN triathletes. Chin. J. Physiol. 59, 305–314 (2016). 16. Knechtle, B. & Nikolaidis, P. T. Sex differences in pacing during “Ultraman Hawaii”. PeerJ 4, e2509 (2016). 17. Wu, S. S. et al. Pacing strategies during the swim, cycle and run disciplines of sprint, Olympic and half-Ironman triathlons. Eur. J. Appl. Physiol. 115, 1147–1154 (2015). 18. Roelands, B., de Koning, J., Foster, C., Hettinga, F. & Meeusen, R. Neurophysiological determinants of theoretical concepts and mechanisms involved in pacing. Sports Med. (Auckland, N.Z.) 43, 301–311 (2013). 19. Jones, H. S. et al. Physiological and psychological effects of deception on pacing strategy and performance: A review. Sports Med. (Auckland, N.Z.) 43, 1243–1257 (2013). 20. Herbst, L. et al. Pacing strategy and change in body composition during a Deca Iron Triathlon. Chin. J. Physiol. 54, 255–263 (2011). 21. Casado, A., Hanley, B., Jiménez-Reyes, P. & Renfree, A. Pacing profiles and tactical behaviors of elite runners. J. Sport Health Sci. 10, 537–549 (2021). 22. Heidenfelder, A., Rosemann, T., Rüst, C. A. & Knechtle, B. Pacing strategies of ultracyclists in the "Race Across AMerica". Int. J. Sports Physiol. Perform. 11 (2016). 23. Wu, S. S., et al. Improvement of sprint triathlon performance in trained athletes with positive swim pacing. Int. J. Sports Physiol. Perform. 11 (2016). 24. Thuany, M. et al. Discriminant analysis of anthropometric and training variables among runners of different competitive levels. Int. J. Environ. Res. Public Health 18, 4248 (2021). 25. Lorenz, D. S., Reiman, M. P., Lehecka, B. J. & Naylor, A. What performance characteristics determine elite versus nonelite athletes in the same sport?. Sports Health 5, 542–547 (2013). 26. Haney, T. A. & Mercer, J. A. A description of variability of pacing in marathon distance running. Int. J. Exerc. Sci. 4, 133–140 (2011). 27. Lambert, M. I., Dugas, J. P., Kirkman, M. C., Mokone, G. G. & Waldeck, M. R. J. Sports Sci. Med. 3, 167–173 (2004). 28. C. A. Rüst, R. Lepers, M. Stiefel, T. Rosemann, B. Knechtle, Performance in Olympic triathlon: Changes in performance of elite female and male triathletes in the ITU World Triathlon Series from 2009 to 2012. Springerplus 2, (2013). 29. Del Coso, J. et al. Running pace decrease during a marathon is positively related to blood markers of muscle damage. PLoS ONE 8, e57602 (2013). 30. Tucker, R., Marle, T., Lambert, E. V. & Noakes, T. D. The rate of heat storage mediates an anticipatory reduction in exercise intensity during cycling at a fixed rating of perceived exertion. J. Physiol. 574, 905–915 (2006). 31. St Clair Gibson, A., Schabort, E. J. & Noakes, T. D. Reduced neuromuscular activity and force generation during prolonged cycling. Am. J. Physiol. Regul. Integr. Comp. Physiol. 281, R187-196 (2001). 32. Le Meur, Y. et al. Influence of gender on pacing adopted by elite triathletes during a competition. Eur. J. Appl. Physiol. 106, 535–545 (2009). 33. Vleck, V. E., Bentley, D. J., Millet, G. P. & Burgi, A. Pacing during an elite Olympic distance triathlon: Comparison between male and female competitors. J. Sci. Med. Sport 11, 424–432 (2008). 34. Knechtle, B., Zingg, M. A., Rosemann, T. & Rüst, C. A. Sex difference in top performers from Ironman to double Deca iron ultra- triathlon. Open Access J. Sports Med. 5, 159–172 (2014). 35. Lepers, R. Analysis of Hawaii ironman performances in elite triathletes from 1981 to 2007. Med. Sci. Sports Exerc. 40, 1828–1834 (2008). 36. Wonerow, M., Rüst, C. A., Nikolaidis, P. T., Rosemann, T. & Knechtle, B. Performance trends in age group triathletes in the Olympic Distance Triathlon at the World Championships 2009–2014. Chin. J. Physiol. 60, 137–150 (2017). 37. Lepers, R., Knechtle, B. & Stapley, P. J. Trends in triathlon performance: Effects of sex and age. Sports Med. 43, 851–863 (2013). 38. Knechtle, R., Rüst, C. A., Rosemann, T. & Knechtle, B. The best triathletes are older in longer race distances—A comparison between Olympic, Half-Ironman and Ironman distance triathlon. Springerplus 3, 538 (2014). 39. Knechtle, B., Rust, C. A., Knechtle, P., Rosemann, T. & Lepers, R. Age-related changes in ultra-triathlon performances. Extrem. Physiol. Med. 1, 5 (2012). 40. Cohen, J. Statistical Power Analysis for the Behavioral Sciences (Lawrence Erlbaum Associates, 2nd, 1988). 41. Knechtle, B., Rosemann, T., Lepers, R. & Rüst, C. A. A comparison of performance of Deca Iron and Triple Deca Iron ultra- triathletes. Springerplus 3, 1–13 (2014). 42. Knechtle, B., Rüst, C. A., Rosemann, T. & Martin, N. 33 Ironman triathlons in 33 days–a case study. Springerplus 3, 1–9 (2014). 43. Sousa, C. V., Nikolaidis, P. T., Clemente-Suárez, V. J., Rosemann, T. & Knechtle, B. Pacing and performance analysis of the World’s Fastest Female Ultra-Triathlete in 5x and 10x Ironman. Int. J. Environ. Res. Public Health 17, 1543 (2020). 44. Etxebarria, N., Wright, J., Jeacocke, H., Mesquida, D. & Pyne, B. Running your best triathlon race. Int. J. Sports Physiol. Perform. 16, 744–747 (2022). 45. Roecker, K., Schotte, O., Niess, A. M., Horstmann, T., Dickhuth, H. H. Predicting competition performance in long-distance run- ning by means of a treadmill test. Med. Sci. Sports Exercise 30, (1998). 46. M. A. Zingg, C. A. Rüst, T. Rosemann, R. Lepers, B. Knechtle, Analysis of sex differences in open-water ultra-distance swimming performances in the FINA World Cup races in 5 km, 10 km and 25 km from 2000 to 2012. BMC sports science, medicine & reha- bilitation 6, (2014). 47. C. A. Rüst, B. Knechtle, T. Rosemann, R. Lepers, Women reduced the sex difference in open-water ultra-distance swimming [For- mula: see text] La Traversée Internationale du Lac St-Jean, 1955–2012. Applied physiology, nutrition, and metabolism = Physiologie appliquee, nutrition et metabolisme 39, (2014). 48. B. Knechtle et al., Fat oxidation in men and women endurance athletes in running and cycling. International journal of sports medicine 25, (2004). 49. Knechtle, B., Wirth, A., Baumann, B., Knechtle, P. & Rosemann, T. Personal best time, percent body fat, and training are differently associated with race time for male and female ironman triathletes. Res. Q. Exerc. Sport 81, 62–68 (2010). 50. B. Knechtle, C. A. Rüst, T. Rosemann, R. Lepers, Age-related changes in 100-km ultra-marathon running performance. Age (Dordrecht, Netherlands) 34, (2012). 51. Stiefel, M., Knechtle, B., Rüst, C. A., Rosemann, T. & Lepers, R. The age of peak performance in Ironman triathlon: A cross-sectional and longitudinal data analysis. Extrem. Physiol. Med. 2, 27 (2013). 9 Vol.:(0123456789) Scientific Reports | (2023) 13:3692 | https://doi.org/10.1038/s41598-023-30932-1 www.nature.com/scientificreports/ 52. Reaburn, P. & Dascombe, B. Endurance performance in Masters athletes. Eur. Rev. Aging Phys. Activ. 5, 31–42 (2008). 53. M. Zingg, C. A. Rüst, R. Lepers, T. Rosemann, B. Knechtle, Master runners dominate 24-h ultramarathons worldwide-a retrospec- tive data analysis from 1998 to 2011. Extreme Physiol. Med. 2, (2013). 54. Rüst, C. A., Knechtle, B., Knechtle, P., Rosemann, T. & Lepers, R. Participation and performance trends in triple iron ultra- triathlon—A cross-sectional and longitudinal data analysis. Asian J. Sports Med. 3(3), 145–152 (2012). 55. Krouse, R. Z., Ransdell, L. B., Lucas, S. M. & Pritchard, M. E. Motivation, goal orientation, coaching, and training habits. J. Strength Cond. Res. 25, 2835–2842 (2011). 56. N. PT, A.-C. JR, V. E, R. T, K. B, The Age-Related Performance Decline in Marathon Running: The Paradigm of the Berlin Marathon. International journal of environmental research and public health 16, (2019). 57. B. J, P. CR, B. SR, D. RC, Age-related changes in maximal power and maximal heart rate recorded during a ramped test in 114 cyclists age 15–73 years. Journal of aging and physical activity 13, (2005). 58. T. H, S. DR, Endurance exercise performance in Masters athletes: age-associated changes and underlying physiological mechanisms. The Journal of physiology 586, (2008). 59. Lepers, R., Sultana, F., Bernard, T., Hausswirth, C. & Brisswalter, J. Age-related changes in triathlon performances. Int J Sports Med 31, 251–256 (2010). 60. E. F et al., Age and gender interactions in short distance triathlon performance. Journal of sports sciences 31, (2013). 61. E. F, K. B, R. CA, R. T, L. R, The age-related decline in Olympic distance triathlon performance differs between males and females. The Journal of sports medicine and physical fitness 53, (2013). 62. T. Bernard, F. Sultana, R. Lepers, C. Hausswirth, J. Brisswalter, Age-related decline in olympic triathlon performance: effect of locomotion mode. Experimental aging research 36, (2010). 63. J. A. Wegelin, M. D. Hoffman, Variables associated with odds of finishing and finish time in a 161-km ultramarathon. European journal of applied physiology 111, (2011). 64. Dallam, G. M., Jonas, S., Miller, T. K. Medical considerations in triathlon competition: recommendations for triathlon organisers, competitors and coaches. Sports Med. 35 (2005). 65. Vihma, T. Effects of weather on the performance of marathon runners. Int. J. Biometeoroll 54 (2010). 66. Applegate, E. Nutritional concerns of the ultraendurance triathlete. Med. Sci. Sports Exercise 21 (1989). 67. Bentley, D. J., Cox, G. R., Green, D. & Laursen, P. B. Maximising performance in triathlon: applied physiological and nutritional aspects of elite and non-elite competitions. J. Sci. Med. Sport 11, 407–416 (2008). Author contributions Conceptualization: M.S., B.K. Data curation: M.S. Formal analysis: C.V.S. Writing- original draft: M.S., B.K. All authors have read and agreed to the published version of the manuscript. All authors consent to the publication of this manuscript. Competing interests The authors declare no competing interests. Additional information Correspondence and requests for materials should be addressed to B.K. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. © The Author(s) 2023

Changes in pacing variation with increasing race duration in ultra-triathlon races.

03-06-2023

Stjepanovic, Mirko,Knechtle, Beat,Weiss, Katja,Nikolaidis, Pantelis Theodoros,Cuk, Ivan,Thuany, Mabliny,Sousa, Caio Victor

eng

PMC8651147

RESEARCH ARTICLE Elastic energy savings and active energy cost in a simple model of running Ryan T. SchroederID1*, Arthur D. Kuo1,2 1 Faculty of Kinesiology, University of Calgary, Alberta, Canada, 2 Biomedical Engineering Program, University of Calgary, Alberta, Canada * [email protected] Abstract The energetic economy of running benefits from tendon and other tissues that store and return elastic energy, thus saving muscles from costly mechanical work. The classic “Spring-mass” computational model successfully explains the forces, displacements and mechanical power of running, as the outcome of dynamical interactions between the body center of mass and a purely elastic spring for the leg. However, the Spring-mass model does not include active muscles and cannot explain the metabolic energy cost of running, whether on level ground or on a slope. Here we add explicit actuation and dissipation to the Spring-mass model, and show how they explain substantial active (and thus costly) work during human running, and much of the associated energetic cost. Dissipation is modeled as modest energy losses (5% of total mechanical energy for running at 3 m s-1) from hyster- esis and foot-ground collisions, that must be restored by active work each step. Even with substantial elastic energy return (59% of positive work, comparable to empirical observa- tions), the active work could account for most of the metabolic cost of human running (about 68%, assuming human-like muscle efficiency). We also introduce a previously unappreci- ated energetic cost for rapid production of force, that helps explain the relatively smooth ground reaction forces of running, and why muscles might also actively perform negative work. With both work and rapid force costs, the model reproduces the energetics of human running at a range of speeds on level ground and on slopes. Although elastic return is key to energy savings, there are still losses that require restorative muscle work, which can cost substantial energy during running. Author summary Running is an energetically economical gait whereby the legs bounce like pogo sticks. Leg tendons act elastically to store and return energy to the body, thus saving the muscles from costly work with each running step. Although elasticity is known to save energy, it does not explain why running still requires considerable effort, and why the muscles still do substantial work. We use a simple computational model to demonstrate two possible reasons why. One is that small amounts of energy are lost when the leg collides with the ground and when the tendons are stretched, and muscles must restore that energy during PLOS COMPUTATIONAL BIOLOGY PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 1 / 22 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Schroeder RT, Kuo AD (2021) Elastic energy savings and active energy cost in a simple model of running. PLoS Comput Biol 17(11): e1009608. https://doi.org/10.1371/journal. pcbi.1009608 Editor: Barbara Webb, The University of Edinburgh, UNITED KINGDOM Received: May 10, 2021 Accepted: November 2, 2021 Published: November 23, 2021 Peer Review History: PLOS recognizes the benefits of transparency in the peer review process; therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pcbi.1009608 Copyright: © 2021 Schroeder, Kuo. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: The source code and data used to produce the results and analyses presented in this manuscript are available from the Bitbucket Git repository: https://bitbucket.org/hbcl/ runoptsol/src/main/. steady running. A second reason is that muscles may perform work to avoid turning on and off rapidly, which may be even more energetically costly. The resulting muscle work, while small in quantity, may still explain most of the energetic cost of running. Economy may be gained from elasticity, but running nonetheless requires muscles to do active work. Introduction Running is distinguished by the spring-like, energy-saving behavior of the stance limb [1–4], analogous to a pogo stick (Fig 1). It is modeled well by a classic analogy, the Spring-mass model, where the limb acts elastically to support and redirect the body center of mass (CoM) between flight phases, and all mechanical energy is conserved throughout each step. This sim- ple model can reproduce the motion and forces of running remarkably well and explains how series elastic tissues such as tendon can improve running economy. It applies to bipeds and even polypeds, making it one of the most universal and elegant models for running. However, it does not include muscles that actively contract against series elasticity, and it fails to explain Human Spring- mass B. Force C. Power D. V. work loop V. disp. (cm) V. GRF (Mg) Time (s) 0 0.1 0.2 0.3 0 2 4 Power (kW) 0 0.1 0.2 0.3 -2 2 0 Time (s) V. acc. (g) -6 -1 0 1 2 0 3 -3 Ground Slope Cost of Transport -1.2 E. Cost vs. slope -0.4 0 0.4 0.8 1.2 0 0.4 0.25 Human Spring- mass Touchdown Take-off A. Trajectory Stance Flight Point-mass body Spring-like stance leg Foot-ground collision Fig 1. Human running and the Spring-mass model. (A) Human stance phases resemble motion of a spring-mass system with no energy loss, alternating with parabolic Flight phases. Body mass is lumped into a single point center of mass (CoM). Traces of (B) vertical ground reaction forces (V. GRF) vs. Time, (C) leg Power vs. Time, (D) and vertical acceleration vs. displacement (V. acc. vs. V. disp.; termed vertical work loop curve) are all shown for both human data (gray lines) and the Spring-mass model (dark solid lines). (E) The energetic Cost of Transport (cost per unit weight and distance) for humans running on slopes (after [11]) is not explained by the Spring-mass model, which only operates at zero ground slope and zero energetic cost. The spring-like behaviors (B-D) should be regarded as pseudo-elastic, because humans and other animals experience dissipation such as in foot-ground collision, and thus, require active muscle actuation. https://doi.org/10.1371/journal.pcbi.1009608.g001 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 2 / 22 Funding: This work was funded by the Dr. Benno Nigg Research Chair (ADK) and the National Sciences and Engineering Research Council of Canada (https://www.nserc-crsng.gc.ca/; NSERC CRC, Tier 1 to ADK; NSERC Discovery to ADK). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. the substantial metabolic cost measured during running. An extension of the Spring-mass model to include active actuation may help explain organismal running energetics and how best to exploit series elasticity for economy. The Spring-mass model agrees well with a wide body of experimental evidence [4]. It repro- duces mechanical characteristics such as the body’s trajectory in space (Fig 1A), ground reac- tion forces (Fig 1B), leg mechanical power (Fig 1C), and even the leg’s vertical work loop curve (vertical acceleration vs. vertical displacement similar to an elastic spring’s work loop, Fig 1E; [4]). The spring can passively store and return mechanical energy to the CoM, reducing the active work otherwise required of active muscle, and thus, improve running economy. Some have therefore proposed that more compliance or longer tendons are key to running economy [5,6]. For example, the energetic cost of human running is less than half of that expected if muscles alone performed work on the CoM [7]. In turkeys, tendon contributes over 60% of the shortening work performed by the lateral gastrocnemius [8]. Although the simple Spring- mass model (Fig 2A) applies mainly to bipedal running or polypedal trotting, multiple leg springs can reproduce galloping, and indeed, practically all of the running gaits observed in nature [9,10]. Few other models reproduce so many behaviors with such simplicity. There are also important aspects of running not captured by the Spring-mass model. A crit- ical feature is metabolic energy expenditure by muscle [12], considered important for selecting gait and speed [13,14], and more generally for a variety of animal behaviors [15]. The Spring- mass model is conservative of mechanical energy and predicts no such expenditure. Lacking muscle actuation, it is incapable of accelerating from rest or running on sloped ground. Steeper slopes in particular have energetic costs approaching that expected from muscles performing positive and negative work against gravity at their respective efficiencies (Fig 1E; [11]). Even steady running on the level entails substantial muscle shortening work, as shown in turkeys (e.g., 40% of muscle-tendon work; [8]), and in human running [16,17]. Some of that work is fundamentally necessary because of dissipation, for example by tendons with hysteresis (26% loss per cycle in Achilles tendon during hopping; [18]), by the heel pad [19] and other soft tis- sues that deform (33% per step of human running at 3 m s-1, [20]). Restoration of those losses alone could account for up to 29% of the energetic cost of human running [20], and the overall active work of muscle for as much as 76% of the energetic cost of human running [21]. The spring-like mechanics of running (Fig 1B–1D) should therefore be regarded as pseudo-elastic, as opposed to purely elastic. Beyond the conceptual illustration of energy savings, the Spring- mass does not account for dissipation and is not predictive of actual energy costs observed in nature. Other simple models of running have included elements other than springs. Perhaps the simplest of these has only an active, extending actuator (Actuator-only model, Fig 2B; [14]). Minimization of its work alone is sufficient for both walking and running to emerge as optimal gaits, with running more economical at faster speeds and exhibiting a pseudo-elastic, bounc- ing-like stance phase despite no passive elasticity [14]. In addition to mechanical work as a cost, we have proposed that muscles also expend energy for a force-rate cost, associated with rapid production of force [22–25], which helps to explain human-like ground reaction forces [26]. Others have optimized the actuation of robots, including both elasticity and dissipative elements, and have shown a variety of running gaits to emerge [27,28]. For organisms, similar models have been used to explore stability [29] and economical strategies [30] of terrain navi- gation. Still, while dissipation has been characterized empirically (e.g., [31]), most running energetics models have not included such dissipation with series elasticity and actuation to explain energy cost. Here we propose a running model that combines series elasticity with active actuation and passive dissipation (Fig 2). The classic Spring-mass (Fig 2A) and pure Actuator-only (Fig 2B) PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 3 / 22 models serve as opposing reference points that can produce running mechanics (Fig 1A–1D) with and without elasticity. We propose to combine the spring and actuator in series, along with dissipation, in an Actuated Spring-mass model (Fig 2C) that may be more representative of running in organisms. We expect that such a model will leverage series elasticity to perform minimal work, as needed to restore dissipative losses. We test whether such a model is suffi- cient to explain both the mechanics and metabolic cost of running (Fig 1E), with work mini- mization as the sole objective, or work plus the proposed force-rate cost. Such a model may help determine whether more compliant tendon is indeed economical [6], and provide insight on the energy expended by muscle. We use human data for running at different speeds and ground slopes as experimental comparison, but the principles revealed by the model are intended to help explain running across a range of animal species. Methods We used dynamic optimization to determine optimal actuation strategies for the proposed Actuator-Spring-mass running model. The model extends the classic Spring-mass model by adding an active series actuator and dissipative losses (Fig 2C). The actuator can perform posi- tive and/or negative work, in part to compensate for two modes of passive energy dissipation: collision loss associated with foot-ground contact and hysteresis of the tendon spring. The model was optimized for energy economy, as defined by the energetic costs of that active actu- ator work, briefly summarized here. Two of the most basic elements of the model are the mass and elastic spring. The point mass M was supported by a spring with stiffness k, which was varied as a free parameter to pro- duce a wide spectrum of gaits. These included low stiffnesses ranging from grounded running with no flight phase, to more impulsive running with a brief stance period and a relatively long flight phase. The limiting case of impulsive running has infinitesimal stance and an infinitely stiff spring, where the body’s motion is almost entirely described by its parabolic trajectory during flight. M F g Actuator-only B. M g k m LlLl LlLl C. Actuated Spring-mass Spring-mass A. Actuator Hysteresis M c k Fm Lm Lt g Collision Fig 2. Simple running models with and without elastic spring, active actuator, and dissipative elements. (A) The Spring-mass model comprises a point-mass body and a massless spring for a leg. (B) The Actuator-only model replaces the spring with a massless, active actuator producing extension forces in the leg [14]. (C) The proposed Actuated Spring-mass model combines an actuator and a spring (analogous to a muscle-tendon unit), along with two passive, dissipative elements: a damper in parallel with the spring to model tendon hysteresis, and collision loss to model dissipation of kinetic energy at touchdown. In the models, g is gravitational acceleration and M is body mass. Ll(t), Lt(t), and Lm(t) are time-varying lengths of the leg, spring and actuator, respectively. Parameters k and c are spring stiffness and damping coefficient. Fm(t) is the active actuator force in the leg’s extension direction. https://doi.org/10.1371/journal.pcbi.1009608.g002 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 4 / 22 Two types of dissipation were included in the model, representing losses from collisions and hysteresis. Collisions model the kinetic energy dissipated when the body impacts the ground. For example, humans lose momentum associated with 2.6–7.8% body mass [31]. We modeled this as a simple discontinuity in the CoM velocity vector magnitude at touchdown, defining collision fraction (CF) as the fraction of momentum lost in the collision. A nominal collision fraction of 3% resulted in a 5.9% loss in kinetic energy (see S1 Text for details). Hysteresis was included to model the imperfect energy return of tendons and other series elastic tissues. Hysteretic energy losses of 10–35% per stretch-shortening cycle have been esti- mated in vivo for tissues such as the human Achilles tendon [32]. Estimates of soft tissue defor- mation suggest that much of the actual dissipation occurs during the first half of stance [20], modeled with a viscous damper (in parallel with the spring) only dissipating energy during spring loading (Fig 2C). Damping was parameterized by damping ratio, z ¼ c= ffiffiffiffiffiffiffiffiffi 4Mk p , where c is the damping coefficient. A nominal damping value of z = 0.1 was selected to yield 26% hys- teresis, roughly within the estimated range of humans. Each stance phase was computed with dynamic optimization for energy economy. The main control variable was the time-varying actuator length during the leg’s stance (specifically, its third derivative L . . . mðtÞ was used to allow for calculations of force rate during implementa- tion), treating the stance and swing phases as periodic and symmetric between legs. The objec- tive function to be minimized was the energy cost per step E, E ¼ EW þ ER; ð1Þ as the sum of a cost EW for work, and another cost ER for force rate. The work cost depends on positive and negative efficiencies for muscle, η+ and η- respectively (25% and -120% from [11]), defined as work divided by metabolic energy (superscript + or − for positive and nega- tive work, respectively). The energetic cost was therefore defined by actuator work (Wm per step, and power Pm for work per time, superscripts for positive and negative work), EW ¼ R T 0 Pþ m Zþ the leg’s posture at touchdown and did not allow slipping of that contact. Model states included the position (xb, yb) and velocity (_xb, _yb) of the point-mass body and the first and sec- ond derivatives of the actuator’s length ( _Lm, €Lm) to facilitate inclusion of work and force-rate in the objective function. All optimizations were conducted with the MATLAB software GPOPS-II [42] and the resulting nonlinear problem was solved using SNOPT [43]. All variables and equations were non-dimensionalized with parameter combinations (L = 0.90 m, M = 70 kg and g = 9.81 m s2) during optimization and outputs were subsequently re-dimensionalized as indicated in figures. Further details regarding the model and implementation of the optimization problem can be found in S1 Text. Running parameters were chosen to represent human-like gait. Speed v was varied over a range of 2–4 m s-1, with empirical preferred step frequency given by f = 0.26v + 2.17 [44]. Step length s was defined as the distance travelled over one periodic step, and step frequency f as the inverse of time duration, T, per step (i.e. T = 1/f), such that v = s f. Gaits were produced while varying parameters such as tendon stiffness k (4.93–122 kN m-1 or equivalently, 6.46–160 MgL-1) and force-rate coefficient ε (0–2
10−2 M-1g-1.5L1.5). Furthermore, a single set of nominal parameter values (force-rate coefficient ε, spring stiffness k and negative and positive work efficiency η- and η+) were selected for comparisons with human ground reaction forces and metabolic data (see Table 1). For comparison with our model, we included representative human data to qualitatively illustrate well-established patterns for ground reaction forces and other trajectories. The data consist of one representative subject from a separate published study [45]: a male (25 years, body mass = 75.3 kg, leg length = 0.79 m) running on an instrumented force treadmill at 3.9 m s-1. An average step was determined from 20 s of steady-state ground reaction force data and used in plots including CoM power [46] and vertical acceleration vs. displacement [2,4], as comparison against the model. These plots reproduce patterns from accepted literature, and so no statistical analysis was performed. Results Optimization results are presented in two parts, first examining the effects of individual model components (Fig 2), and then combining them into a single, unified model. Part I presents Table 1. Model parameters and values. Symbol Description Range Nominal Units η- negative work efficiency -1.05 η+ positive work efficiency 0.32 CF collision fraction 0–0.06 0.03 z damping ratio 0–0.2 0.1 k spring stiffness 6.46–160 46.7 MgL-1 ε force-rate coefficient 0–2
10−2 5
10−4 M-1g-1.5L1.5 v running speed 0.67–1.35 1.01 g0.5L0.5 M body mass 70 kg g gravitational acceleration 9.81 m s-2 L maximum leg length 0.9 m Parameters used in the Actuated Spring-mass model, along with the ranges of values examined for parameter sensitivity analysis, and nominal values for comparison with human. Range is left empty if the parameter was not varied in optimizations. Units are left empty if the parameter has no units. The model was implemented in normalized units, with M, g and L as base units (nominal human values shown). https://doi.org/10.1371/journal.pcbi.1009608.t001 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 6 / 22 individual sensitivity studies, beginning with the Spring-mass and Actuator-only models, which have been examined in prior literature: e.g. [2,4] and [26,36,40], respectively. Next, the Actuated Spring-mass model is evaluated as an alternative, since it still uses a spring but also requires an actuator to account for passive energy dissipation occurring with each step. Ini- tially, this model is evaluated with the cost of work only (i.e. zero force-rate coefficient ε) over varying speeds and spring stiffnesses. Next, non-zero force-rate coefficients ε are introduced so changes in actuation strategies may be independently evaluated. Finally, in Part II, a single, unified set of parameters is applied to the Actuated Spring-mass model, which is then used to simulate gait over a range of running speeds, spring stiffnesses and ground slopes to assess its utility in predicting locomotion energetics and optimal spring-actuator coordination patterns. Part I: Individual model components and their contributions to running behavior Spring-mass and Actuator-only models produce similar pseudo-elastic running gaits. Both models (Fig 3) can produce similar gaits, ranging from very flat to very bouncy. For the Spring-mass model at a given speed and step length, the spring stiffness determines the gait trajectory, as described by CoM trajectories, vertical ground reaction force profiles, mechanical power profiles, and vertical work loop curves illustrating the spring-like leg behavior [4]. As reported by others [3], the Spring-mass model can run with a vast range of stiffnesses k (Fig 3, top). A less stiff (or more compliant) spring can produce grounded running, in which the stance phase occupies the entire step (blue curves in Fig 3). With greater stiffness comes a flight phase, yielding gaits that resemble more typical human running, where both stance and flight phases are finite in duration (redder curves, Fig 3). These gaits generally include a single- peaked ground reaction force profile, with higher peak forces and powers with increasing Actuator-only Spring-mass V. acc. (g) -5 V. disp. (cm) 0 2 4 6 8 0 5 V. disp. (cm) -5 0 5 0 2 4 6 8 V. acc. (g) V. GRF (Mg) Time (s) 0 0.1 0.2 0.3 0 2 4 6 8 0 0.1 0.2 0.3 Time (s) 0 2 4 6 8 V. GRF (Mg) Increasing stiffness k k=∞ low k Power (kW) 0 0.1 0.2 0.3 -5 5 0 Time (s) Power (kW) 0 0.1 0.2 0.3 -5 5 0 Time (s) ε=0 high ε min work & force-rate (EW+ER) A. Trajectories B. Force C. Power D. V. work loop Increasing k Increasing k Increasing ε Increasing ε Increasing ε one step one step Fig 3. Running gaits from the Spring-mass (top) and Actuator-only (bottom) models. They are illustrated by (A) center of mass (CoM) trajectory, (B) vertical ground reaction forces vs. time, (C) leg power performed on the CoM vs. time, and (D) vertical acceleration vs. vertical displacement of the body (or vertical work loop curve). A range of running gaits are shown, varying stiffness k in the Spring-mass model, and the force-rate cost coefficient ε in the Actuator-only model, for a single running speed v (3.5 m s-1) and step frequency f (3 Hz). In the limiting case of infinite spring stiffness or zero force-rate cost, touchdown forces become perfectly impulsive (red arrows). https://doi.org/10.1371/journal.pcbi.1009608.g003 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 7 / 22 stiffness, as well as steeper vertical work loop curves during stance. In the limit toward infinite stiffness, the model produces impulsive running [14], where flight takes up nearly the entire step and stance occurs as an instantaneous impulse (red arrows, Fig 3). In all cases, the Spring- mass model is purely elastic and has no actuator and no losses. Thus, no single spring stiffness, and no single spring-like gait, can be considered beneficial over another in terms of energy cost. The Actuator-only model can produce a very similar range of running gaits, despite the complete lack of an elastic spring (Fig 3, bottom). The optimization produces pseudo-elastic behavior resembling a spring, and tuned by a single parameter: the force-rate coefficient ε. With a coefficient of zero (i.e. work cost only), impulsive running is optimal, because least work is performed with least displacement, albeit with infinite force [14]. A greater force-rate cost results in increasing stance time and shorter flight time, more similar to humans. Increas- ing that cost further eventually causes the flight phase to disappear, producing grounded run- ning similar to a very compliant spring. For any non-zero force-rate cost, the model consistently produces an approximately linear vertical work loop curve, similar to the Spring- mass model. However, this behavior is purely active and requires substantial positive and neg- ative work. As a result, the Actuator-only model does incur an energy cost for work but has no passive elasticity to reduce that work. We thus find that the two diametrically opposed models can reproduce the pseudo-elastic behaviors similar to humans, whether or not there is true passive elasticity. There is certainly strong evidence that elasticity is important for running in humans and other animals, but spring-like ground reaction forces and vertical work loop curves are not necessarily indicative that elasticity is the dominant mechanism in running. If it were, the energetic cost of running might be expected to be close to zero. Conversely, the pure Actuator-only model also obviously cannot demonstrate that humans are purely inelastic. The work performed by humans, if there were no elasticity, would result in unrealistically high muscle efficiencies of at least 45% [7]. It is more realistic to regard the human as having some combination of series elasticity and active actuation, both contributing to the actual energetic cost of human running. Dissipative energy losses require compensatory, active positive mechanical work We next consider the effect of passive energy dissipation in the Actuated Spring-mass model (Fig 4), optimizing for the cost of work alone (with zero force-rate coefficient). Again, gaits roughly similar to those of humans are produced, for either stiff or compliant springs. How- ever, the actuator must perform positive work to restore the lost energy. With a stiff spring, it is optimal to produce a relatively “bouncy” CoM trajectory where the body spends more time in the air and thus, reaches greater heights above the ground (Fig 4A). The optimum also favors a spikier vertical force and power over shorter stance durations (Fig 4B and 4C). Con- versely, a compliant spring makes it optimal to produce a “flatter” CoM trajectory with briefer flight time (Fig 4A), with lower peak vertical force, longer stance duration, and less stiff vertical work loop curves (Fig 4D). The accompanying leg angle at touchdown also varies, with a more vertical orientation for increasing spring stiffness. With zero force-rate coefficient ε, it is generally optimal to perform only positive actuator work. For steady motion, the energy lost to hysteresis and collision must be restored with an equal magnitude of positive work, to yield zero net work per step. The optimization reveals this is performed most economically in the second half of stance (Fig 4C, blue shaded areas), in concert with elastic energy return from the spring. It is also generally economical to completely avoid active negative work, which would also require an additional amount of PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 8 / 22 positive work to be performed. Thus, actuator work is minimized when it is performed only to restore dissipative losses. Dissipative losses are minimized by increasing spring stiffness with running speed We next evaluate a range of spring stiffnesses and running speeds to determine how the cost of actuator work may be minimized (Fig 5). Here we find that optimal spring stiffness increases with running speed (Fig 5A), while generally preserving the timing of positive work within the second half of stance (Fig 5B). At lower speed v (2.5 m s-1; Fig 5A left), both hysteresis and col- lision losses are reduced with less spring stiffness. However, at higher speed (3.5 m s-1; Fig 5A right), hysteresis losses are reduced with greater stiffness whereas collision losses are relatively unchanged. For a moderate speed (3.0 m s-1; Fig 5A middle) both hysteresis and collision trade off over stiffness, and an intermediate stiffness is optimal. Overall, hysteresis losses change mainly with spring stiffness, and collision losses increase mainly with running speed, so that losses are generally minimized by a stiffness increasing with speed. The collision and hysteresis losses have distinct dependencies on running speed and/or stiffness. The model’s collision loss increases with touchdown velocity and thus, running speed, but is relatively insensitive to spring stiffness. Stiffness does affect the CoM trajectory and the distribution between horizontal and vertical velocity components but has relatively lit- tle effect on the vector magnitude. Overall, collision losses increase with speed but are rela- tively unaffected by spring stiffness (Fig 5). In contrast, hysteresis loss occurs as a fraction of elastic strain energy, which is largely deter- mined by the angle of the leg during stance. For example, a vertical leg posture is used in con- junction with a stiff spring (Fig 4), and this results in greater strain (and hysteresis losses) to redirect vertical CoM velocity of the bouncier gait. Alternatively, a less vertical leg posture is used with a compliant spring and results in greater strain to redirect horizontal velocity of the body. As such, stiff springs allow for efficient gait at higher speeds, since the vertical leg is effec- tive at mitigating excessive strain to redirect high horizontal CoM velocity. At lower speeds, compliant springs are better since a less vertical leg is better at mitigating higher vertical veloc- ity associated with bouncy running at these speeds. The overall effects of dissipation are as follows. At low speeds, both collision and hysteresis losses are reduced with relatively low spring stiffness and a shallower leg touchdown angle, A. Actuated Spring-mass V. GRF (Mg) Time (s) 0 0.1 0.2 0.3 0 2 4 6 8 B. Force Power (kW) Time (s) 0 0.1 0.2 0.3 -5 5 0 Actuator Collision loss Leg C. Power V. disp. (cm) V. acc. (g) -5 0 2 4 6 0 5 D. V. work loop min work (EW), zero force rate Fig 4. Effects of stiff vs. compliant springs on Actuated Spring-mass model minimizing cost of work (with zero force-rate cost). Optimal running gaits (speed v of 3.0 m s-1) are shown for the Actuated Spring-mass model, including (A) CoM trajectory, (B) vertical ground reaction forces (V. GRF) vs. time, (C) leg power vs. time, and (D) vertical acceleration vs. vertical displacement (V. acc. vs. V. disp.). The model includes passive dissipation (hysteresis and collision), optimized for two spring stiffnesses k (13.7 kN m-1 for compliant and 109.5 kN m-1 for stiff). The stiffer spring yields a more vertical leg, shorter stance time and bouncier gait, with higher peak forces and leg power. Net actuator work is similar in both cases. https://doi.org/10.1371/journal.pcbi.1009608.g004 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 9 / 22 thereby also reducing actuator work. But at higher speeds, hysteresis losses are actually reduced by greater spring stiffness and steeper leg touchdown angle, so that actuator work is minimized with relatively high stiffness. These effects together cause the optimal spring stiff- ness for minimizing actuator work to increase with running speed. An added force-rate cost favors active actuator dissipation We have thus far found that the model with zero force-rate cost avoids active negative work, whereas some animal muscles are observed to perform non-negligible negative work during running [5]. It may seem uneconomical to perform any amount of active negative work, because it only adds to the costly positive work needed to restore the losses. Perhaps there is some indirect energetic advantage to active negative work, not explained by a cost for work alone. In fact, the addition of a force-rate cost with coefficient ε (Fig 6) makes it favorable for the actuator to perform both negative and positive work. This distributes ground reaction forces over a longer stance duration with lower peak forces and reduced force rate. However, it also comes at the expense of additional actuator work, which is made worthwhile by its capac- ity to reduce force rate (Fig 6). Overall, the force-rate cost yields less impulsive forces and smoother CoM trajectories, at the expense of active dissipation and increased work. Model Work per Step (x10-2) 101 102 0 2 4 6 8 10 3.0 m s-1 101 102 0 2 4 6 8 10 3.5 m s-1 101 102 0 2 4 6 8 10 2.5 m s-1 10 J Power P(t) 5 kW Leg Spring Actuator Collision Hysteresis Actuator 0.1 s More work Less work Running Speed Collision Collision Hysteresis Actuator Time 101 102 101 102 101 102 Spring Stiffness k (kN m-1) Spring Stiffness k (kN m-1) Spring Stiffness k (kN m-1) A. B. min work Fig 5. Actuator work and power as a function of spring stiffness and running speed in the Actuated Spring-mass model, minimizing cost of work (with zero force-rate cost). (A) Work vs. stiffness for speeds of 2.5–3.5 m s-1. Shown are active Actuator work (black), Collision work magnitude (red), and Hysteresis work magnitude (blue). Spring diagrams (inset) illustrate touchdown angles for each stiffness. (B) Power vs. time for very compliant and very stiff springs, for each running speed. Shown are net Leg power (black lines), Spring power (orange shaded area), Actuator power (blue shaded area). Results are for spring stiffness ranging 13.7–109.5 kN m-1. https://doi.org/10.1371/journal.pcbi.1009608.g005 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 10 / 22 The added force-rate cost yields a shift in timing of positive work so that it is optimally per- formed late in stance. Experiments have shown that the triceps surae muscles undergo sub- stantial shortening throughout stance, but particularly late in stance, during human running [16,17]. On the other hand, negative work is optimally performed early in stance as ground reaction forces rise [5,47]. Part II: Unified actuated spring-mass model of human-like running and energy expenditure We next apply the Actuated Spring-mass model with a single, unified set of parameter values selected to produce human-like running in a variety of conditions. The full model therefore includes an elastic spring, both hysteresis and collision losses, and an objective to minimize both work and force rate costs. A single force-rate coefficient is selected (ε of 0.5
10−3) to approximately match the model’s output to human data, along with stiffness k of 35.6 kN m-1, positive work efficiency η+ of 32%, and negative work efficiency η- of -105%. The resulting model, with parameters thus fixed, is then applied to three comparisons with human data: mechanics of a nominal gait, energetic cost as a function of running speed, and energetic cost as a function of ground slope. Unified model produces human-like gait mechanics The resulting model qualitatively matches the human CoM trajectory (Fig 7A), vertical ground reaction forces (Fig 7B), leg power vs. time (Fig 7C), and vertical acceleration vs. displacement 10-2 Force-Rate Coefficient, ε (M –1g –1.5L 1.5) Energy Cost per Step Power V. GRF V. disp. 10-3 10-4 10 cm 2 Mg 2 kW 0.1 s Leg Spring Actuator Work cost, EW Collision Total cost, EW+ER 2 J kg-1 Force-rate cost, ER Unified ε* min work & force-rate (EW+ER) Fig 6. Effect of work and force-rate costs on running using Actuated Spring-mass model. (top:) Vertical CoM displacement vs. time, vertical GRF vs. time, and leg Power vs. time, for varying force-rate cost coefficient ε. (bottom:) Actuator work cost EW (thin blue line), Total cost EW + ER (work and force rate, solid black line), and Force-rate cost ER (difference between lines) vary with the coefficient. Impulsive actions (red arrows, V. GRF and leg Power) occur at Collision, and overall leg power (solid black line) includes contributions from the Spring and Actuator. All solutions are shown for v of 3 m s-1 and f of 2.94 Hz. https://doi.org/10.1371/journal.pcbi.1009608.g006 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 11 / 22 (Fig 7D). The model also reproduces some features that the classical Spring-mass model can- not, such as the brief initial peak at touchdown (from collision [31], Fig 7B) and a more grad- ual decrease in force than the increase (temporal asymmetry [30], Fig 7B). The model collision produces a transient burst of negative power [46] followed by elastic energy storage and return (about 53%; Fig 7C), as well as slightly non-linear vertical work loop curves (as in [4], Fig 7D), qualitatively similar to human. Whereas the Spring-mass model produces a nearly linear curve that retraces itself almost perfectly (Fig 1D), the human curve has an initial transient and self- intersecting profile resembling a tilted figure-eight. The shape indicates hysteresis with a dissi- pative counter-clockwise loop, followed by a (positive work) clockwise loop. The present model crudely reproduces these broad features, even if imperfect in detail. Unified model has increasing energy cost with speed for level running The model’s energetic cost per time (Fig 8A) increases with running speed at a rate similar to human data [48]. Here, the model’s step frequency was constrained to the empirical human preferred step frequency, but other parameters were kept fixed. The increasing overall cost with speed may be explained by the constituent force-rate and work costs (Fig 8B), evaluated as a function of spring stiffness and speed. The work cost is considerably greater in magnitude than the force-rate cost (e.g., 68% vs. 32%, respectively at 3 m s-1) and increases more as a func- tion of speed, primarily for restoring collision losses. Thus, most of the model’s overall cost for running at higher speeds is due to increased actuator work, which is not included in the Spring-mass model. Nevertheless, the force-rate cost has a large influence on the model’s gait as a function of spring stiffness (Fig 8B). Greater stiffness is associated with more impulsive ground reaction forces and briefer stance durations (as in Fig 4), thus resulting in higher force-rate cost. Fur- thermore, the actuator performs additional negative (and therefore also positive) work with greater stiffness as a trade-off against even higher force-rate costs (like in Fig 6). Overall, the Leg Actuator Human V. disp. (cm) V. GRF (Mg) Time (s) 0 0.1 0.2 0.3 0 2 4 Power (kW) 0 0.1 0.2 0.3 -2 2 0 Time (s) V. acc. (g) -1 0 1 2 0 3 -3 V. disp. (cm) V. GRF (Mg) Time (s) 0 0.1 0.2 0.3 0 2 4 Power (kW) 0 0.1 0.2 0.3 -2 2 0 Time (s) V. acc. (g) -1 0 1 2 0 3 -3 Collision Collision A. Unified model B. Force C. Power D. V. work loop 53% Elastic return min work & force-rate (EW+ER) Fig 7. Comparison of Unified Actuated Spring-mass model (top) including work and force-rate costs against human data (bottom). Shown are (A) CoM trajectories, (B) vertical ground reaction forces, (C) leg mechanical power, and (D) vertical acceleration vs. displacement. Initial force transients are highlighted (red impulse arrow for model, red line for human). In (C), spring (orange shaded area) and actuator work (blue shaded area) contributions are shown. Gait parameters v and f are 3.9 m s-1 and 3 Hz, respectively. Stiffness and force-rate coefficient in the model are selected to approximately match stance time duration: k is 35.6 kN m-1 and ε is 0.5
10−3. https://doi.org/10.1371/journal.pcbi.1009608.g007 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 12 / 22 presence of a force-rate cost makes particularly stiff springs more costly to the model and may indicate benefits of some compliance when running. Unified model explains energetic cost of running on inclines The model may also be applied to uphill and downhill running. In humans, metabolic cost asymptotically converges toward the costs of muscle performing positive and negative work (about 25% and -120%, respectively; [11]). At intermediate slopes, the cost smoothly transi- tions between these two extremes, passing through the cost for level running. The unified model produces a similar cost curve (Fig 9A), with similar asymptotes. However, the force- rate cost adds to work costs in such a way that the human asymptotes are actually achieved with slightly different positive and negative actuator efficiencies (32% and -105%), though con- sistent with estimates on cross-bridge efficiency [49]. The model’s energetic cost is dominated by positive and negative work at steeper upward and downward slopes, respectively (Fig 9B). Of course, increasing work is required of steeper slopes, but force-rate becomes less costly at those extremes. Additionally, for slopes surround- ing zero, the force-rate cost contributes to the relatively smooth transition from positive to negative efficiency tangent lines identified by Margaria [11]. The minimum of the cost curve occurs approximately where passive energy dissipation approaches the net negative mechani- cal work of descending the ground slope (about -0.08 slope) and is consistent with simple colli- sion models indicating optimal running slopes [50]. The force-rate cost is relatively high for shallow slopes and level ground, because it favors more impulsive forces that take advantage of passive dissipation to reduce active negative work. While even more passive dissipation at steeper negative slopes could reduce work costs further, this would come at a higher force-rate cost. In fact, it is less costly overall to actively dissipate energy at steeper slopes to avoid a high force-rate, but at increased work cost. A. Running Speed (m s-1) 2 3 4 0 4 8 12 16 20 Energy Cost per Time (W kg-1) Model Human 101 102 0 4 8 12 16 20 Energy Cost per Time (W kg-1) 2.5 m s-1 101 102 Spring Stiffness k (kN m-1) 0 4 8 12 16 20 3.0 m s-1 101 102 0 4 8 12 16 20 3.5 m s-1 B. Running Speed Force rate Work cost Off- set k* 68% 32% Total cost Total cost min work & force-rate (EW+ER) Fig 8. Unified model energy cost vs. speed and spring stiffness, including force-rate cost. (A) Energetic cost per time versus speed is shown for the Actuated Spring-mass model (k of 35.6 kN m-1, ε of 0.5
10−3; red curve) and for empirical metabolic data of human subjects running on a treadmill (mean ± standard deviation; [48]). Model cost includes costs for work and force-rate, plus a constant offset associated with human resting metabolism (dashed horizontal line). (B) The model’s energetic cost is shown for three speeds v (2.5–3.5 m s-1) and with parameter variation of spring stiffnesses k (13.7–109.5 kN m-1), with total cost (black lines), force-rate cost (difference between offset and magenta lines), and actuator work cost (difference between total and magenta lines). The unified model’s spring nominal stiffness k is indicated (red line in A, red symbol in B). https://doi.org/10.1371/journal.pcbi.1009608.g008 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 13 / 22 Parameter variation was used to assess the model’s cost sensitivity to spring stiffness k (Fig 9C), force-rate coefficient ε (Fig 9D), damping ratio z (Fig 9E) and collision fraction CF (Fig 9F). Costs generally increase with each of these parameters, particularly for shallow and level slopes. For example, increasing spring stiffness resulted in greater energy cost, largely due to increased force-rate cost associated with more impulsive forces. Greater force-rate coefficient was also more costly, since total cost is proportional to the coefficient (Fig 6). Increasing either dissipation parameter resulted in relatively modest increases in cost, mainly because more dis- sipation must be offset by more active work. But the overall result is that, even for extreme parameter variations, the cost of running up or down steeper slopes still tends to asymptote toward positive and negative work efficiencies. On shallow and level slopes, each parameter contributed toward a non-zero cost, resulting in an overall cost comparable to human (Fig 9A). Discussion We have proposed several additions to the Spring-mass model that help to explain the energy expenditure of running. The Actuated Spring-mass model includes passive energy dissipation, active work, and an additional energetic cost related to force-rate. In combination, these ele- ments show how running can still cost substantial energy, even though series elasticity acts Ground Slope Cost of Transport -0.4 -0.2 0 0.2 0.4 Ground Slope -0.4 -0.2 0 0.2 0.4 Force rate Increasing CF Increasing damping, ζ Increasing ε Cost of Transport Cos t con irt bu it ons . B Energy cost . A Increasing k Ground Slope .F Co ill sion D . Force r ate ness ffit C . S E. Hysteresis Ground Slope Ground Slope Ground Slope Model Positive work Negative work -120% 25% Human Model min work & force-rate (EW+ER) Offset Cost of Transport 0.2 0.2 min work & force-rate (EW+ER) Fig 9. Energetic cost of running vs. ground slope for unified Actuated Spring-mass model. (A) Model Cost of Transport (solid line) compared to humans (circles; [11]). Also shown are asymptotes (thin lines) for muscle efficiency of positive and negative mechanical work (25% and -120%, respectively). (B) Contributors to model Cost of Transport: positive work cost, negative work cost, force-rate cost ER, and a constant offset. Parameter sensitivities are included for varying (C) stiffness k, (D) force-rate coefficient ε, (E) hysteresis (damping ratio z), and (F) collision fraction CF. Each trace indicates variation from lowest to highest parameter values: k ranging 13.7 kN m-1 –1, ε ranging 0–2
10−3, z ranging 0–0.2, CF ranging 0–0.06. All model results are for nominal running at speed of 3 m s-1 and step frequency of 2.94 Hz. https://doi.org/10.1371/journal.pcbi.1009608.g009 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 14 / 22 conservatively to reduce the active work required of muscle. We next re-examine each of the elements individually, to consider both the justification and the contribution of each to an overall model of running. Active work dominates energy expenditure despite elastic return The primary energetic cost observed in the models considered here was for active work. Empirical estimates based on work during human running, along with assumed 50% elastic energy return, suggest that active work could account for 76% of the energetic cost [21]. Our model also had substantial elastic energy return, for example 53% at a speed of 3.9 m s-1 (Fig 7), yet mechanical work still accounted for 66–70% of the overall energetic cost over the range of speeds considered. The unified model actively performed both positive and negative work (Fig 7). Empirical measurements reveal modest active lengthening (and thus negative power) during early stance in humans (vastus lateralis; [47]) and in turkeys (lateral gastrocnemius; [5]). The perspective provided by our model is that such active negative work should generally be avoided if work were the only energetic cost. But active dissipation may be justified by opposing costs such as for force rate (Fig 8), which justify performing active negative work and extending stance durations (e.g. Fig 7C), but at the cost of yet more positive work to offset the active dissipation. The performance of active positive and negative work on level ground also explains the smooth transition in cost between uphill and downhill slopes. A leg that actively performs negative work on the level should simply perform more such dissipation for downward slopes, and less for upward slopes, and similarly for positive work, with cost asymptotes defined by work per- formed against gravity alone [11]. Even if passive elasticity performs most of the work of run- ning, the remaining active work by muscles [16,17,47] could still explain much of the overall energetic cost. The model also reproduces empirical correlations with energy cost. Kram and Taylor [51] observed energy costs increasing with speed, and proportional to body weight divided by ground contact time. The present unified model also yields similar correlations (cost propor- tional to inverse contact time, and inverse contact time to speed, R2 = 0.95 and R2 = 0.98 respectively), but as an outcome of optimizing costs for work and force rate. In fact, a simple analysis demonstrates that mechanical work of a series actuator can explain this proportional- ity explicitly (in supplementary material of [21]). We do not consider contact time to be a direct determinant of cost, because running (with constant body weight) with an especially flat CoM trajectory results in both greater contact time and greater energy cost [52]. Rather, a mechanistic energy cost from actuator work and force can potentially explain why quantities such as contact time can appear correlated (or not) with cost. Passive dissipation is a major determinant of mechanics and energetics of running There are several features of running that are reproduced by the inclusion of dissipation. Dissi- pative losses are the primary driver of active mechanical work, which is needed to offset losses and obtain the zero net work of a periodic gait cycle. Even a relatively small amount of dissipa- tion can be costly. For example, the unified model passively dissipated less than 5% of the body’s mechanical energy and passively returned 59% of positive shortening work at 3 m s-1 (compared to 60% in turkey; [8]), yet the remaining active positive work can still explain 61% of the net metabolic cost of equivalent human running. We also found dissipation (particularly collision loss) to increase substantially with running speed, and therefore contribute to greater energy expenditure rate (Fig 8). PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 15 / 22 Passive dissipation also helps to explain some time-asymmetries in running. For example, in humans the vertical ground reaction force increases faster early in stance than it decreases later in stance (Fig 7B). Such asymmetries have previously been attributed to the muscle force- velocity relationship [53] and to dissipation [54]. The present model predicts such asymmetry to be energetically optimal (Fig 7B). Given the leg extends after mid-stance when forces are already decreasing, the leg must undergo net extension by the time of take-off (Fig 7D). The model also predicts a more vertical leg orientation during touchdown versus at take-off to reduce dissipation similar to humans [54]. Other models have also demonstrated the economy of asymmetrical trajectories in bird running [30] but do not include a brief transient in ground reaction force and work at touchdown (Fig 7B and 7C), observed prominently in human foot contact. Asymmetrical trajectories are not observed in the models without passive dissipation. The Spring-mass model produces more time-symmetric trajectories (Fig 3) lacking initial tran- sients and predicts zero active work. The Actuator-only model also produces symmetric trajec- tories since it has no passive dissipation. However, it lacks passive elasticity resulting in cost over twice that of equivalent human running (Fig 3, bottom column). The time asymmetries in force and work profiles might seem like minor details, but here they are emergent from energy optimality. Passive dissipation helps to explain these asymmetries and is a major contri- bution to the energetic cost modeled here. Elastic tendon is critical to energy economy but not the pseudo-elastic mechanics of running We next re-examine whether long, compliant tendons are helpful for locomotion economy [5,6,55]. If the total work per step from muscle and tendon were fixed, a more compliant ten- don could indeed perform more of the work passively, and thus reduce the energetic cost. But in running, the total work need not be fixed, and changing the series stiffness also yields a dif- ferent optimal trajectory (Fig 4), and a different amount of total work for a given speed and step length. In our model with zero force rate cost coefficient ε, a stiffer spring is actually more economical at higher speeds (3.5 m s-1, Fig 5). The optimal trajectory yields lower spring dis- placement and hysteresis loss, and thus, less active work. Another potential factor for compliance is the proposed force-rate cost. Avoidance of that cost can favor more active work (Fig 8), and overall cost may indeed be reduced with more compliance. More complex running models have also suggested that the optimal compliance may actually be different for each muscle [56]. Of course, there may also be other benefits to tendon compliance beyond economy [57]. But for running economy alone, there is no general prescription that favors greater tendon length or compliance. Tendon compliance has long been implicated in the spring-like mechanics of running. This is manifested in CoM trajectories, ground reaction forces, power, and vertical work loop curves (Fig 3), which are all suggestive of elastic behavior. But all the models considered here, including those with dissipation and actuation, and even those with no elasticity at all (Actua- tor-only Model, Fig 3) also exhibit similar pseudo-elastic behavior. In models, springs are also not critical to the economical advantage of running over walking at high speeds [14], nor to the general cost trends for running on slopes (Fig 9C). This is not to diminish the importance of elasticity, which allows leg muscles to operate at lower and more efficient shortening veloci- ties [5], and store and return substantial energy. But the basic resemblance of running to elastic bouncing, and the associated pseudo-elastic mechanics (e.g., Fig 3), should not be regarded as evidence of true elasticity in running. PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 16 / 22 Force-rate cost contributes to mechanics and energetics of running In addition to the cost of active work, our model also includes an energetic cost for force rate. Such a cost has been experimentally observed in tasks such as cyclic muscle contractions [24,35], and was included as an energetic penalty for rapid transients in force production. We found that work as the only energetic cost tended to favor overly impulsive running motions (Fig 6). The force-rate cost acts as a trade-off against work, resulting in reduced peak ground reaction forces and longer stance durations, more similar to human data (Fig 7). The trade-off also makes it economical to perform active negative work, and to favor more compliant series elasticity, contributing to more human-like mechanics (Fig 7) and energetics (Fig 8). Work alone also reveals a particularly economical strategy for running downhill, where only a mini- mal amount of active negative work is performed to dissipate gravitational potential energy. However, humans expend more than the predicted amount of energy, perhaps because the force-rate makes such a strategy more costly, resulting in a smoother transition toward the negative work asymptote observed experimentally [11]. The force-rate cost produces these effects despite a relatively modest energetic cost. In the unified model, force-rate accounts for only 32% of energetic cost, compared to 68% for work (Fig 8). And across slopes, work alone predicts too low an energetic cost for running (Fig 9D). A cost such as force-rate is thus helpful for explaining human-like energetics. Implications for legged robots The present study may be relevant to running robots that employ spring-like behavior. In the SLIP (spring-loaded inverted pendulum) paradigm, a controller causes the overall leg to behave like a spring, despite internal dissipation. Robotic dissipation includes actuator heat and transmission losses (analogous to actuator work efficiency and hysteresis in our model) as well as from interactions with the environment (e.g. collisions) [58]. With SLIP control, the stiffness is sometimes selected to resemble animal gait [59], just as our model resembles human. Our results suggest that SLIP may actually be reasonably economical, because our model, similar to others modeling dissipation with more detail [27,28], yield optimal force pro- files that are still remarkably spring-like. However, closer examination also reveals that better economy is achieved with small but significant differences such as force asymmetry (Fig 7). We also note that matching a gait to human or animal is not necessarily the best option for economy, which might favor a quite different stiffness (Fig 8B). There are also potential bene- fits to including passive dissipation, which can yield improved stability [29] and velocity esti- mation, which is considered important for robust control [60,61]. Robotic controllers can take advantage of passive dissipation by modeling it explicitly (e.g., [62]) and will respond differ- ently when optimized for economy. Limitations The running model proposed here has a number of limitations, with regard to passive dissipa- tion, actuator costs and running dynamics. For dissipation, we modeled hysteresis during leg compression alone to reproduce empirically estimated energy losses, without a detailed model of hysteresis mechanics. Similarly, we modeled collision losses with a simple reduction in momentum at touchdown, without considering the direction and mechanical properties of the body mass experiencing impact. To our knowledge, most models of running do not include explicit dissipation. We consider the present model to be an indicator that dissipation is important, but also in need of better-informed dissipation mechanics. The same is true for our model of energetic cost, intended to extend Spring-mass models with a highly simplified dependence on active mechanical work. Our model also stands to be improved with other PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 17 / 22 features potentially relevant to running, such as muscle moment arms and force-velocity rela- tionships (e.g., [63]). There are also limitations of the proposed energetic cost for force rate, which is not included in most other running models. This cost is intermediate in complexity between abstract “effort” costs such as force squared [64,65], and more detailed models of muscle ener- getics [66,67], while also being supported by empirical data (e.g. [23–25,34,35]). But the spe- cific dynamics of such a cost in muscle are not well understood, with considerable uncertainty in the appropriate formulation. For example, we have found different derivative order and exponent (e.g., 2 and 1, respectively for Eq 3) to yield similar forces [26] and agree well with empirical energy costs [23]. We therefore regard the force-rate cost as a provisional implemen- tation that stands for considerable improvement. Similar to the Spring-mass model, the present model is also a gross simplification of human body dynamics. It is intended only to model basic features of running, such as pseudo-elastic mechanics and overall energy expenditure. More complexity would be required to address motion of multiple-jointed models, intersegmental dynamics, and activation or co-contraction of multiple muscles. For example, the model neglects a swing leg, whose active motion may also cost energy [22], and be exploited to modulate collision losses via active leg retraction just before touchdown [68,69]. We also fixed step length and frequency with respect to running speed, whereas these could also be included in optimization for preferred gait parameters (e.g., [38]). The present model only provides basic suggestions, that dissipation and energetic costs for work and force rate may be important for running. These suggestions are intended to apply to more complex models of running, but this remains to be tested. Conclusions The energetic cost of running is not addressed by the classic Spring-mass model of running. Although elastic tissues store and return energy during stance, there is still some dissipation due to touchdown collision and hysteresis. For steady gait, these losses must be restored by active, positive work from muscle. The present Actuated Spring-mass model shows that even a relatively small amount of work can still incur a substantial energetic cost. It is also particularly costly to perform active negative work, because the associated losses must be restored by addi- tional positive work. Muscles may also expend energy for high rates of force development that make it economical to perform some active negative work, ultimately helping to explain the energetics of running at different speeds and slopes. Spring-like forces and other mechanics emerge from an actively controlled model optimized for economy, even if there were no elas- ticity. Series elasticity may be critical to saving energy, but active work and passive dissipation appear important for determining the energetic cost of running. Supporting information S1 Text. Dynamic optimization model details. (DOCX) Acknowledgments The authors would like to thank Arash Khassetarash (University of Calgary) for sharing exper- imental data. Author Contributions Conceptualization: Ryan T. Schroeder, Arthur D. Kuo. PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 18 / 22 Data curation: Ryan T. Schroeder. Formal analysis: Ryan T. Schroeder. Funding acquisition: Arthur D. Kuo. Investigation: Ryan T. Schroeder, Arthur D. Kuo. Methodology: Ryan T. Schroeder, Arthur D. Kuo. Project administration: Ryan T. Schroeder, Arthur D. Kuo. Resources: Arthur D. Kuo. Software: Ryan T. Schroeder. Supervision: Arthur D. Kuo. Validation: Ryan T. Schroeder, Arthur D. Kuo. Visualization: Ryan T. Schroeder, Arthur D. Kuo. Writing – original draft: Ryan T. Schroeder, Arthur D. Kuo. Writing – review & editing: Ryan T. Schroeder, Arthur D. Kuo. References 1. Alexander RM. A model of bipedal locomotion on compliant legs. Philosophical Transactions of the Royal Society of London Series B: Biological Sciences. 1992; 338(1284):189–98. https://doi.org/10. 1098/rstb.1992.0138 PMID: 1360684 2. Blickhan R. The spring-mass model for running and hopping. Journal of Biomechanics. 1989 Jan; 22 (11–12):1217–27. https://doi.org/10.1016/0021-9290(89)90224-8 PMID: 2625422 3. Geyer H, Seyfarth A, Blickhan R. Compliant leg behaviour explains basic dynamics of walking and run- ning. Proceedings of the Royal Society B-Biological Sciences. 2006 Nov 22; 273(1603):2861–7. https:// doi.org/10.1098/rspb.2006.3637 PMID: 17015312 4. McMahon TA, Cheng GC. The mechanics of running: How does stiffness couple with speed? Journal of Biomechanics. 1990;23, Supplement 1:65–78. https://doi.org/10.1016/0021-9290(90)90042-2 PMID: 2081746 5. Biewener AA, Roberts TJ. Muscle and tendon contributions to force, work, and elastic energy savings: a comparative perspective. Exerc Sport Sci Rev. 2000 Jul; 28(3):99–107. PMID: 10916700 6. Bramble DM, Lieberman DE. Endurance running and the evolution of Homo. Nature. 2004 Nov; 432 (7015):345–52. https://doi.org/10.1038/nature03052 PMID: 15549097 7. Cavagna GA, Kaneko M. Mechanical work and efficiency in level walking and running. J Physiol. 1977 Jun 1; 268(2):467–81. https://doi.org/10.1113/jphysiol.1977.sp011866 PMID: 874922 8. Roberts TJ, Marsh RL, Weyand PG, Taylor CR. Muscular Force in Running Turkeys: The Economy of Minimizing Work. Science. 1997 Feb 21; 275(5303):1113–5. https://doi.org/10.1126/science.275.5303. 1113 PMID: 9027309 9. Gan Z, Remy CD. A passive dynamic quadruped that moves in a large variety of gaits. In: 2014 IEEE/ RSJ International Conference on Intelligent Robots and Systems. 2014. p. 4876–81. 10. Gan Z, Yesilevskiy Y, Zaytsev P, Remy CD. All common bipedal gaits emerge from a single passive model. Journal of The Royal Society Interface. 2018 Sep 30; 15(146):20180455. https://doi.org/10. 1098/rsif.2018.0455 PMID: 30257925 11. Margaria R. Positive and negative work performances and their efficiencies in human locomotion. Int Z Angew Physiol. 1968 May 28; 25(4):339–51. https://doi.org/10.1007/BF00699624 PMID: 5658204 12. Taylor CR, Heglund NC, Maloiy GM. Energetics and mechanics of terrestrial locomotion. I. Metabolic energy consumption as a function of speed and body size in birds and mammals. J Exp Biol. 1982 Apr 1; 97(1):1–21. PMID: 7086334 13. Long LL, Srinivasan M. Walking, running, and resting under time, distance, and average speed con- straints: optimality of walk–run–rest mixtures. Journal of The Royal Society Interface. 2013 Apr 6; 10 (81):20120980. PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 19 / 22 14. Srinivasan M, Ruina A. Computer optimization of a minimal biped model discovers walking and running. Nature. 2006; 439(7072):72–5. https://doi.org/10.1038/nature04113 PMID: 16155564 15. Alexander RM. Optima for Animals [Internet]. Princeton, NJ: Princeton University Press; 1996 [cited 2020 Sep 5]. Available from: https://press.princeton.edu/books/paperback/9780691027982/optima-for- animals 16. Lai A, Lichtwark GA, Schache AG, Lin Y-C, Brown NAT, Pandy MG. In vivo behavior of the human soleus muscle with increasing walking and running speeds. Journal of Applied Physiology. 2015 Mar 26; 118(10):1266–75. https://doi.org/10.1152/japplphysiol.00128.2015 PMID: 25814636 17. Lichtwark GA, Bougoulias K, Wilson AM. Muscle fascicle and series elastic element length changes along the length of the human gastrocnemius during walking and running. Journal of Biomechanics. 2007 Jan 1; 40(1):157–64. https://doi.org/10.1016/j.jbiomech.2005.10.035 PMID: 16364330 18. Lichtwark GA, Wilson AM. In vivo mechanical properties of the human Achilles tendon during one-leg- ged hopping. Journal of Experimental Biology. 2005 Dec 15; 208(24):4715–25. https://doi.org/10.1242/ jeb.01950 PMID: 16326953 19. Ker RF. The time-dependent mechanical properties of the human heel pad in the context of locomotion. J Exp Biol. 1996; 199:1501–8. PMID: 8699155 20. Riddick RC, Kuo AD. Soft tissues store and return mechanical energy in human running. Journal of Bio- mechanics. 2016 Feb 8; 49(3):436–41. https://doi.org/10.1016/j.jbiomech.2016.01.001 PMID: 26806689 21. Riddick RC, Kuo AD. Mechanical Work Accounts for Most of the Energetic Cost in Human Running. bioRxiv. 2020; 22. Doke J, Donelan JM, Kuo AD. Mechanics and energetics of swinging the human leg. The Journal of experimental biology. 2005; 208(3):439–45. https://doi.org/10.1242/jeb.01408 PMID: 15671332 23. Doke J, Kuo AD. Energetic cost of producing cyclic muscle force, rather than work, to swing the human leg. Journal of Experimental Biology. 2007; 210(13):2390–8. https://doi.org/10.1242/jeb.02782 PMID: 17575044 24. van der Zee TJ, Kuo AD. The high energetic cost of rapid force development in cyclic muscle contrac- tion. Journal of Experimental Biology [Internet]. 2020 [cited 2020 Sep 8]; Available from: https://www. biorxiv.org/content/10.1101/2020.08.25.266965v1 25. Wong JD, Cluff T, Kuo AD. The energetic basis for smooth human arm movements [Internet]. Neurosci- ence; 2020 Dec [cited 2021 Mar 17]. Available from: http://biorxiv.org/lookup/doi/10.1101/2020.12.28. 424067 26. Rebula JR, Kuo AD. The Cost of Leg Forces in Bipedal Locomotion: A Simple Optimization Study. PLOS ONE. 2015 Feb 23; 10(2):e0117384. https://doi.org/10.1371/journal.pone.0117384 PMID: 25707000 27. Xi W, Remy CD. Optimal gaits and motions for legged robots. In: 2014 IEEE/RSJ International Confer- ence on Intelligent Robots and Systems. 2014. p. 3259–65. 28. Xi W, Yesilevskiy Y, Remy CD. Selecting gaits for economical locomotion of legged robots. The Interna- tional Journal of Robotics Research. 2016 Aug 1; 35(9):1140–54. 29. Heim S, Millard M, Le Mouel C, Badri-Spro¨witz A. A little damping goes a long way: a simulation study of how damping influences task-level stability in running. Biol Lett. 2020 Sep; 16(9):20200467. https:// doi.org/10.1098/rsbl.2020.0467 PMID: 32961093 30. Birn-Jeffery AV, Hubicki CM, Blum Y, Renjewski D, Hurst JW, Daley MA. Don’t break a leg: running birds from quail to ostrich prioritise leg safety and economy on uneven terrain. Journal of Experimental Biology. 2014 Nov 1; 217(21):3786–96. https://doi.org/10.1242/jeb.102640 PMID: 25355848 31. Chi KJ, Schmitt D. Mechanical energy and effective foot mass during impact loading of walking and run- ning. J Biomech. 2005; 38:1387–95. https://doi.org/10.1016/j.jbiomech.2004.06.020 PMID: 15922749 32. Finni T, Peltonen J, Stenroth L, Cronin NJ. Viewpoint: On the hysteresis in the human Achilles tendon. Journal of Applied Physiology. 2013 Feb 15; 114(4):515–7. https://doi.org/10.1152/japplphysiol.01005. 2012 PMID: 23085961 33. Bergstrom M, Hultman E. Energy cost and fatigue during intermittent electrical stimulation of human skeletal muscle. Journal of Applied Physiology. 1988; 65:1500–5. https://doi.org/10.1152/jappl.1988. 65.4.1500 PMID: 3182513 34. Chasiotis D, Bergstro¨m M, Hultman E. ATP utilization and force during intermittent and continuous mus- cle contractions. J Appl Physiol. 1987 Jul; 63(1):167–74. https://doi.org/10.1152/jappl.1987.63.1.167 PMID: 3624122 35. Hogan MC, Ingham E, Kurdak SS. Contraction duration affects metabolic energy cost and fatigue in skeletal muscle. Am J Physiol. 1998; 274:E397–402. https://doi.org/10.1152/ajpendo.1998.274.3.E397 PMID: 9530120 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 20 / 22 36. Croft JL, Schroeder RT, Bertram JE. Determinants of optimal leg use strategy: horizontal to vertical transition in the parkour wall climb. Journal of Experimental Biology. 2019; 222(1):jeb190983. https:// doi.org/10.1242/jeb.190983 PMID: 30446542 37. Handford ML, Srinivasan M. Robotic lower limb prosthesis design through simultaneous computer opti- mizations of human and prosthesis costs. Sci Rep. 2016 Apr; 6(1):19983. https://doi.org/10.1038/ srep19983 PMID: 26857747 38. Kuo AD. A simple model of bipedal walking predicts the preferred speed-step length relationship. J Bio- mech Eng. 2001 Jun; 123(3):264–9. https://doi.org/10.1115/1.1372322 PMID: 11476370 39. Polet DT, Bertram JEA. An inelastic quadrupedal model discovers four-beat walking, two-beat running, and pseudo-elastic actuation as energetically optimal. Donelan M, editor. PLoS Comput Biol. 2019 Nov 21; 15(11):e1007444. https://doi.org/10.1371/journal.pcbi.1007444 PMID: 31751339 40. Schroeder RT, Bertram JE. Minimally Actuated Walking: Identifying Core Challenges to Economical Legged Locomotion Reveals Novel Solutions. Front Robot AI. 2018 May 22; 5(58):58. https://doi.org/ 10.3389/frobt.2018.00058 PMID: 33644120 41. Schroeder RT, Bertram JEA, Son Nguyen V, Vinh Hac V, Croft JL. Load carrying with flexible bamboo poles: optimization of a coupled oscillator system. J Exp Biol. 2019 Dec 1; 222(23):jeb203760. https:// doi.org/10.1242/jeb.203760 PMID: 31801848 42. Patterson MA, Rao AV. GPOPS-II: A MATLAB Software for Solving Multiple-Phase Optimal Control Problems Using hp-Adaptive Gaussian Quadrature Collocation Methods and Sparse Nonlinear Pro- gramming. ACM Trans Math Softw. 2014 Oct 27; 41(1):1–37. 43. Gill PE, Murray W, Saunders MA. SNOPT: An SQP Algorithm for Large-Scale Constrained Optimiza- tion. SIAM Rev. 2005 Jan; 47(1):99–131. 44. Gutmann AK. Constrained optimization in human running. Journal of Experimental Biology. 2006 Feb 15; 209(4):622–32. https://doi.org/10.1242/jeb.02010 PMID: 16449557 45. Khassetarash A, Vernillo G, Kruger RL, Edwards WB, Millet GY. Neuromuscular, biomechanical, and energetics adjustments following repeated bouts of downhill running. J Sport Health Sci. 2021; 46. Arampatzis A, Knicker A, Metzler V, Bru¨ggemann G-P. Mechanical power in running: a comparison of different approaches. Journal of Biomechanics. 2000 Apr; 33(4):457–63. https://doi.org/10.1016/s0021- 9290(99)00187-6 PMID: 10768394 47. Bohm S, Marzilger R, Mersmann F, Santuz A, Arampatzis A. Operating length and velocity of human vastus lateralis muscle during walking and running. Scientific Reports. 2018 Mar 22; 8(1):5066. https:// doi.org/10.1038/s41598-018-23376-5 PMID: 29567999 48. Kipp S, Grabowski AM, Kram R. What determines the metabolic cost of human running across a wide range of velocities? Journal of Experimental Biology [Internet]. 2018 Sep 15 [cited 2020 Jul 9]; 221(18). Available from: https://doi.org/10.1242/jeb.184218 PMID: 30065039 49. Barclay CJ. Chapter 6—Efficiency of Skeletal Muscle. In: Zoladz JA, editor. Muscle and Exercise Physi- ology [Internet]. Academic Press; 2019 [cited 2021 Feb 3]. p. 111–27. Available from: http://www. sciencedirect.com/science/article/pii/B9780128145937000062 50. Dewolf AH, Willems PA. Running on a slope: A collision-based analysis to assess the optimal slope. Journal of Biomechanics. 2019 Jan; 83:298–304. https://doi.org/10.1016/j.jbiomech.2018.12.024 PMID: 30611540 51. Kram R, Taylor CR. Energetics of running: a new perspective. Nature. 1990 Jul 19; 346(6281):265–7. https://doi.org/10.1038/346265a0 PMID: 2374590 52. McMahon TA, Valiant G, Frederick EC. Groucho running. Journal of Applied Physiology. 1987 Jun 1; 62 (6):2326–37. https://doi.org/10.1152/jappl.1987.62.6.2326 PMID: 3610929 53. Cavagna GA. The landing–take-off asymmetry in human running. Journal of Experimental Biology. 2006 Oct 15; 209(20):4051–60. https://doi.org/10.1242/jeb.02344 PMID: 17023599 54. Dewolf AH, Willems PA. A collision-based analysis of the landing-takeoff asymmetry during running. Computer Methods in Biomechanics and Biomedical Engineering. 2017 Oct 30; 20(sup1):S65–6. https://doi.org/10.1080/10255842.2017.1382863 PMID: 29088651 55. Kielty M, Radke H. Radiolab: Man Against Horse [Internet]. [cited 2021 Mar 16]. Available from: https:// www.wnycstudios.org/podcasts/radiolab/articles/man-against-horse 56. Uchida TK, Hicks JL, Dembia CL, Delp SL. Stretching Your Energetic Budget: How Tendon Compliance Affects the Metabolic Cost of Running. Zadpoor AA, editor. PLoS ONE. 2016 Mar 1; 11(3):e0150378. https://doi.org/10.1371/journal.pone.0150378 PMID: 26930416 57. Daley MA, Usherwood JR. Two explanations for the compliant running paradox: reduced work of bounc- ing viscera and increased stability in uneven terrain. Biol Lett. 2010 Jun 23; 6(3):418–21. https://doi.org/ 10.1098/rsbl.2010.0175 PMID: 20335198 PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 21 / 22 58. Seok S, Wang A, Meng Yee Chuah, Otten D, Lang J, Kim S. Design principles for highly efficient quad- rupeds and implementation on the MIT Cheetah robot. In: 2013 IEEE International Conference on Robotics and Automation [Internet]. Karlsruhe, Germany: IEEE; 2013 [cited 2021 Aug 26]. p. 3307–12. Available from: http://ieeexplore.ieee.org/document/6631038/ 59. Zhang C, Zou W, Ma L, Wang Z. Biologically inspired jumping robots: A comprehensive review. Robot- ics and Autonomous Systems. 2020 Feb; 124:103362. 60. Johnson M, Shrewsbury B, Bertrand S, Wu T, Duran D, Floyd M, Abeles P, Stephen D, Mertins N, Les- man A, Carff J, Rifenburgh W, Kaveti P, Straatman W, Smith J, Griffioen M, Layton B, de Boer T, Koolen T, Neuhaus P, Pratt J. Team IHMC’s Lessons Learned from the DARPA Robotics Challenge Trials: Team IHMC’s Lessons Learned from the DARPA Robotics Challenge Trials. J Field Robotics. 2015 Mar; 32(2):192–208. 61. Xinjilefu X, Feng S, Atkeson CG. A distributed MEMS gyro network for joint velocity estimation. In: 2016 IEEE International Conference on Robotics and Automation (ICRA) [Internet]. Stockholm, Sweden: IEEE; 2016 [cited 2021 Aug 26]. p. 1879–84. Available from: http://ieeexplore.ieee.org/document/ 7487334/ 62. Hutter M, Remy CD, Hopflinger MA, Siegwart R. SLIP running with an articulated robotic leg. In: 2010 IEEE/RSJ International Conference on Intelligent Robots and Systems [Internet]. Taipei: IEEE; 2010 [cited 2021 Mar 16]. p. 4934–9. Available from: http://ieeexplore.ieee.org/document/5651461/ 63. McMahon TA, Greene PR. The influence of track compliance on running. Journal of Biomechanics. 1979; 12(12):893–904. https://doi.org/10.1016/0021-9290(79)90057-5 PMID: 528547 64. Ackermann M, van den Bogert AJ. Optimality principles for model-based prediction of human gait. Jour- nal of Biomechanics. 2010 Apr; 43(6):1055–60. https://doi.org/10.1016/j.jbiomech.2009.12.012 PMID: 20074736 65. Martin AE, Schmiedeler JP. Predicting human walking gaits with a simple planar model. Journal of Bio- mechanics. 2014 Apr; 47(6):1416–21. https://doi.org/10.1016/j.jbiomech.2014.01.035 PMID: 24565183 66. Ma SP, Zahalak GI. A distribution-moment model of energetics in skeletal muscle. J Biomech. 1991; 24 (1):21–35. https://doi.org/10.1016/0021-9290(91)90323-f PMID: 2026631 67. Zahalak GI, Motabarzadeh I. A re-examination of calcium activation in the Huxley cross-bridge model. J Biomech Eng. 1997 Feb; 119(1):20–9. https://doi.org/10.1115/1.2796060 PMID: 9083845 68. Haberland M, Karssen JGD, Kim S, Wisse M. The effect of swing leg retraction on running energy effi- ciency. In: 2011 IEEE/RSJ International Conference on Intelligent Robots and Systems [Internet]. San Francisco, CA: IEEE; 2011 [cited 2021 Mar 16]. p. 3957–62. Available from: http://ieeexplore.ieee.org/ document/6094627/ 69. Hasaneini SJ, Bertram JEA, Macnab CJB. Energy-optimal relative timing of stance-leg push-off and swing-leg retraction in walking. Robotica. 2017 Mar; 35(3):654–86. PLOS COMPUTATIONAL BIOLOGY A simple optimization model of running PLOS Computational Biology | https://doi.org/10.1371/journal.pcbi.1009608 November 23, 2021 22 / 22

Elastic energy savings and active energy cost in a simple model of running.

11-23-2021

Schroeder, Ryan T,Kuo, Arthur D

eng

PMC10001845

Citation: Römer, C.; Wolfarth, B. Prediction of Relevant Training Control Parameters at Individual Anaerobic Threshold without Blood Lactate Measurement. Int. J. Environ. Res. Public Health 2023, 20, 4641. https://doi.org/10.3390/ ijerph20054641 Academic Editor: Lana Ruži´c Received: 30 January 2023 Revised: 24 February 2023 Accepted: 28 February 2023 Published: 6 March 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). International Journal of Environmental Research and Public Health Article Prediction of Relevant Training Control Parameters at Individual Anaerobic Threshold without Blood Lactate Measurement Claudia Römer * and Bernd Wolfarth Department of Sports Medicine, Charité—Universitätsmedizin Berlin, Humboldt-University of Berlin, 10117 Berlin, Germany * Correspondence: [email protected] Abstract: Background: Active exercise therapy plays an essential role in tackling the global burden of obesity. Optimizing recommendations in individual training therapy requires that the essential parameters heart rate HR(IAT) and work load (W/kg(IAT) at individual anaerobic threshold (IAT) are known. Performance diagnostics with blood lactate is one of the most established methods for these kinds of diagnostics, yet it is also time consuming and expensive. Methods: To establish a regression model which allows HR(IAT) and (W/kg(IAT) to be predicted without measuring blood lactate, a total of 1234 performance protocols with blood lactate in cycle ergometry were analyzed. Multiple linear regression analyses were performed to predict the essential parameters (HR(IAT)) (W/kg(IAT)) by using routine parameters for ergometry without blood lactate. Results: HR(IAT) can be predicted with an RMSE of 8.77 bpm (p < 0.001), R2 = 0.799 (|R| = 0.798) without performing blood lactate diagnostics during cycle ergometry. In addition, it is possible to predict W/kg(IAT) with an RMSE (root mean square error) of 0.241 W/kg (p < 0.001), R2 = 0.897 (|R| = 0.897). Conclusions: It is possible to predict essential parameters for training management without measuring blood lactate. This model can easily be used in preventive medicine and results in an inexpensive yet better training management of the general population, which is essential for public health. Keywords: performance diagnostics; blood lactate; individual anaerobic threshold; training recommendation 1. Introduction Reluctance to undertake physical activity and obesity are associated with an increase in cardiovascular diseases, in particular in coronary heart disease, diabetes mellitus and a higher level of inflammation [1–3]. Research has demonstrated that engaging in regu- lar physical activity leads to a reduction in both morbidity and mortality rates [4,5]. As our society continues to face an increasing burden of disease from conditions such as diabetes mellitus, arterial hypertension, and obesity, the cost of treating these cardiovas- cular diseases will also become a growing financial strain on the healthcare system in the future [6–11]. Especially, during the COVID-19 pandemic, regular exercising decreased significantly [12]. Consequences may be not only increasing obesity and cardiovascular diseases but also mental health conditions [13,14]. The WHO Guidelines recommend regular activity (150–300 min per week of moderate intensity, or 150 min per week of intensive physical activity) [15]. In order to achieve comprehensive prevention, simple and inexpensive training recommendations and the prescription of physical activity are required [16]. Optimizing training intensity recommendations in cardiopulmonary training requires that the essential parameters heart rate (HR) and training load (W/kg) at individ- ual anaerobic threshold (IAT) are known [17]. It is necessary to define individual training parameters, as several studies have confirmed that training adherence depends on training intensity [18,19]. Adherence to physical activity is one of the most relevant factors to better health [20]. Realization of exercise recommendations by health workers is reported to be insufficient [21,22], which might be caused by the lack of personalized trainings programs. Int. J. Environ. Res. Public Health 2023, 20, 4641. https://doi.org/10.3390/ijerph20054641 https://www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2023, 20, 4641 2 of 12 Overexertion, defined as the transition from the aerobic to the anaerobic metabolism [23], may therefore reduce training adherence. To perform optimal training, knowing the heart rate at the individual anaerobic threshold is essential for better training control. Measuring these important parameters (HR(IAT) and W/kg(IAT)) is largely limited to competitive athletes in dedicated sports medicine centers. Performance diagnostics with blood lactate is one of the most established methods for these kinds of diagnostics [24,25], but it is time- and cost-intensive [26]. The prediction of IAT was early performed by Conconi et al. by determining heart rate threshold which shows a high correlation in runners [27]. There are only a few studies using linear regression models to predict anaerobic threshold on cycling ergometry in the general population. Mostly athletes were examined, and the number of examined subjects is low [27–30]. Simple methods for measuring performance are crucial for allowing the gen- eral population access to the appropriate training parameters, particularly for individuals who are new to sports. There is a lack of studies examining prediction models for essential training parameters in the general population using non-invasive methods [31–34]. The aim of this study was to assess HR(IAT) and W/kg(IAT) by linear regression models to establish an easy access training recommendation for the general population, as physical inactivity is an important predictor of mortality [35]. 2. Methods In this study, a retrospective analysis was performed. Secondary data of the Sports Medicine Institute of the University Medical Center Charité Berlin were analyzed for the prediction of HR(IAT) and W/kg(IAT) without lactate measurement. All of the ergometry protocols conducted between 2015 and 2017 were obtained from the institutional sports medicine information system. Patients who were not included in the study were excluded for specific reasons: For the present analysis, the following inclusion criteria were applied: patients (I) with missing lactate data, (II) with missing heart rate data, and (III) with insufficient protocols and implausible data. Exclusion criteria were cardio-pulmonary and musculoskeletal diseases. The study was conducted in accordance with the Declaration of Helsinki and with the approval of the local ethics committee of Humboldt University Berlin. 2.1. Peak Performance Test The performance test on the cycle ergometer started at 50 Watt (W) and was raised in 25 W steps after 3 min. Resting heart rate, blood pressure and blood lactate were measured before the lactate step test was initiated. During the test, heart rate was continuously measured by electrocardiogram. Blood pressure, blood lactate and RPE (rate of perceived exertion) were measured in the last thirty seconds of each step. Determination of lactate threshold (LT = first significant increase in blood lactate during exercise test starting from the resting lactate values) and individual lactate threshold (IAT = second significant increase in blood lactate and transition from aerobic to anaerobic metabolism) were assessed using the method of Dickhuth et al. [36]. 2.2. Statistical Analysis The Kolmogorov–Smirnov test was used to determine whether the continuous vari- ables were normally distributed, and a descriptive analysis was carried out. The power per kilogram body weight at the individual anaerobic threshold (W/kg(IAT)) was used as a measure of individual physical performance. The Pearson correlation coefficient and root mean square error (RMSE) were used to assess correlation. A two-sided significance level of α = 0.001 was set as the threshold for determining statistical significance. Before performing multiple regression analysis of HR (IAT), all parameters were checked individually for their respective correlation and linear regression with a very high level of significance p < 0.001. Descriptive analysis was performed and is shown in Table 1. Minimum, mean and maximum HR and HR after one-, three- and five-minutes post-workout were examined. All parameters with a significance level p > 0.001 were removed. All statistical analyses Int. J. Environ. Res. Public Health 2023, 20, 4641 3 of 12 were performed using the SPSS software (IBM Corp. Released 2016. IBM SPSS Statistics for Windows, Version 25.0. Armonk, NY, USA: IBM Corp.) and Matlab (MATLAB and Statistics Toolbox Release 2022b, The MathWorks, Inc., Natick, MA, USA). Table 1. Descriptive characteristics of female and male individuals. Female (n = 469) Male (n = 765) Mean ± SD Mean ± SD Age (years) 48.24 ± 19.36 46.61 ± 18.60 Height (cm) 166.25 ± 7.79 180.33 ± 7.93 Weight (kg) 67.12 ± 12.37 84.77 ± 13.53 Pmax/kg (W/kg) 2.16 ± 0.91 2.78 ± 1.02 Pmean (W) Mean power 89.85 ± 35.44 141.19 ± 48.68 HRmin Minimum HR 73.30 ± 12.53 70.38 ± 12.41 HRmean Mean HR 124.00 ± 16.83 123.32 ± 15.94 HRmax Maximum HR 166.99 ± 20.93 170.13 ± 20.78 HRpw1 HR 1 min post workout 142.43 ± 22.63 144.85 ± 20.04 HRpw3 HR 3 min post workout 115.66 ± 20.04 119.26 ± 17.44 HRpw5 HR 5 min post workout 103.80 ± 18.57 108.59 ± 16.60 HRR delta HRmax − HRpw5 63.19 ± 14.04 61.54 ± 15.02 P/kg(IAT) (W/kg) 1.58 ± 0.64 1.98 ± 0.77 HR(IAT) 143.99 ± 18.66 141.12 ± 19.46 Study Population The population consisted of 188 competitive athletes (football, handball, athletics, volleyball, etc.), 226 prevention and rehabilitation athletes (with various chronic diseases, e.g., orthopedic, rheumatological or other autoimmune diseases) and 820 recreational athletes. None of the athletes had known coronary artery disease or heart failure. A total of 579 had a BMI greater than 25. Overall, 141 individuals of the 226 prevention and rehabilitation athletes had a BMI greater than 25, 52 individuals in this subgroup had a BMI greater than 30, and 13 individuals had a BMI greater than 35. Descriptive analysis is shown in Table 1. 3. Results We performed multiple linear regression analyses for both HR(IAT) and W/kg(IAT) using personal parameters such as gender, age, height and weight as well as performance measurements such as heart rate and power as input parameters. After each multiple regression analysis, we removed one parameter with the highest p-value until the desired significance level of p < 0.001 was met by all remaining input parameters. After completing this process, the following input parameters are included in the multi- ple linear regression analysis for determining the HR(IAT); see equation in Figure 1: gender; weight; mean power (Pmean); maximum power (Pmax); mean HR (HRmean); and minimal HR (HRmin). Using these parameters in multiple linear regression, the determination of HR(IAT) is possible with an RMSE = 8.77 bpm. The adjusted R-squared is 0.798. The proposed linear regression model for determining HR at IAT was compared to the Karvonen formula (Figure 2). The proposed method shows a lower RMSE (8.77 bpm) than the Karvonen formula (RMSE of 11.2 bpm), and HR determination at IAT is more exact using linear regression. The essential parameter W/kg(IAT), which is especially important for determining changes in performance, was also examined. As explained above, the respective input parameters were iteratively removed unless they met a level of significance p < 0.001. This includes the removal of the heart rate recovery (HRR = HRmax-HR after 5 min of recovery) parameter, as its significance level was p = 0.057. As a result, only the following four parameters were included for multiple linear regression analysis to determine W/kg(IAT): gender; body weight (kg); mean power (Pmean); maximum power (Pmax); maximum HR Int. J. Environ. Res. Public Health 2023, 20, 4641 4 of 12 (HRmax). Using these parameters in multiple linear regression (Figure 3), the determination of W/kg(IAT) is possible with a root mean square error, RMSE = 0.241 W/kg. The adjusted R-squared was 0.897. Figure 3 shows the comparison between W/kg(IAT) values on the horizontal axis determined by means of blood lactate values and the W/kg(IAT) values on the vertical axis determined by means of multiple linear regression. Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 4 of 12 HRS(IAS) = 25.464 + 1.0302 × HRmean [bpm] − 0.11889×HRmin [bpm] − 0.23561 × Pmean [W] + 0.16634 × Pmax [W] + 3.8084 × gender [f = 1, m = 0] − 0.09512 × weight [kg] Figure 1. The comparison between HR(IAT) values on the horizontal axis determined by means of blood lactate values and the HR(IAT) values on the vertical axis determined by means of multiple linear regression. The proposed linear regression model for determining HR at IAT was compared to the Karvonen formula (Figure 2). The proposed method shows a lower RMSE (8.77 bpm) than the Karvonen formula (RMSE of 11.2 bpm), and HR determination at IAT is more exact using linear regression. Figure 2. Comparison between regression model and Karvonen formula. The essential parameter W/kg(IAT), which is especially important for determining changes in performance, was also examined. As explained above, the respective input pa- rameters were iteratively removed unless they met a level of significance p < 0.001. This includes the removal of the heart rate recovery (HRR = HRmax-HR after 5 min of recov- ery) parameter, as its significance level was p = 0.057. As a result, only the following four parameters were included for multiple linear regression analysis to determine W/kg(IAT): gender; body weight (kg); mean power (Pmean); maximum power (Pmax); maximum HR (HRmax). Using these parameters in multiple linear regression (Figure 3), the determination of W/kg(IAT) is possible with a root mean square error, RMSE = 0.241 W/kg. The adjusted R-squared was 0.897. Figure 3 shows the comparison between W/kg(IAT) values on the horizontal axis determined by means of blood lactate values and the W/kg(IAT) values on the vertical axis determined by means of multiple linear regression. 70 90 110 130 150 170 190 210 70 90 110 130 150 170 190 210 HR(IAT) determined using multiple linear regression HR(IAT) determined by blood lactate values 70 90 110 130 150 170 190 210 70 90 110 130 150 170 190 210 HR(IAT) determined using Karvonen formula HR(IAT) determined by blood lactate values 70 90 110 130 150 170 190 210 70 90 110 130 150 170 190 210 HR(IAT) determined using multiple linear regression HR(IAT) determined by blood lactate values Karvonen formula Proposed method Figure 1. The comparison between HR(IAT) values on the horizontal axis determined by means of blood lactate values and the HR(IAT) values on the vertical axis determined by means of multiple linear regression. Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 4 of 12 HRS(IAS) = 25.464 + 1.0302 × HRmean [bpm] − 0.11889×HRmin [bpm] − 0.23561 × Pmean [W] + 0.16634 × Pmax [W] + 3.8084 × gender [f = 1, m = 0] − 0.09512 × weight [kg] Figure 1. The comparison between HR(IAT) values on the horizontal axis determined by means of blood lactate values and the HR(IAT) values on the vertical axis determined by means of multiple linear regression. The proposed linear regression model for determining HR at IAT was compared to the Karvonen formula (Figure 2). The proposed method shows a lower RMSE (8.77 bpm) than the Karvonen formula (RMSE of 11.2 bpm), and HR determination at IAT is more exact using linear regression. Figure 2. Comparison between regression model and Karvonen formula. The essential parameter W/kg(IAT), which is especially important for determining changes in performance, was also examined. As explained above, the respective input pa- rameters were iteratively removed unless they met a level of significance p < 0.001. This includes the removal of the heart rate recovery (HRR = HRmax-HR after 5 min of recov- ery) parameter, as its significance level was p = 0.057. As a result, only the following four parameters were included for multiple linear regression analysis to determine W/kg(IAT): gender; body weight (kg); mean power (Pmean); maximum power (Pmax); maximum HR (HRmax). Using these parameters in multiple linear regression (Figure 3), the determination of W/kg(IAT) is possible with a root mean square error, RMSE = 0.241 W/kg. The adjusted R-squared was 0.897. Figure 3 shows the comparison between W/kg(IAT) values on the horizontal axis determined by means of blood lactate values and the W/kg(IAT) values on the vertical axis determined by means of multiple linear regression. 70 90 110 130 150 170 190 210 70 90 110 130 150 170 190 210 HR(IAT) determined using multiple linear regression HR(IAT) determined by blood lactate values 70 90 110 130 150 170 190 210 70 90 110 130 150 170 190 210 HR(IAT) determined using Karvonen formula HR(IAT) determined by blood lactate values 70 90 110 130 150 170 190 210 70 90 110 130 150 170 190 210 HR(IAT) determined using multiple linear regression HR(IAT) determined by blood lactate values Karvonen formula Proposed method Figure 2. Comparison between regression model and Karvonen formula. Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 5 of 12 W/kg(IAS) = 2.2306 − 3.8992 ×10−3 × HRmax [bpm] −5.0916 × 10−3 × Pmean [W] + 6.5394 × 10−3 × Pmax [W] + 71.384 × 10−3 × gender [f = 1, m = 0] − 21.407 × 10−3 × weight [kg] Figure 3. Impact of the individual input parameters on the W/kg (IAT). To better understand the impact of individual input parameters on the W/kg (IAT), we have visualized the regression parameters using an effect plot in Figure 4. For this, we multiplied the weights of the formula in Figure 3 with the actual values in our database. The latter are normalized by subtracting their respective mean values, as this offset is al- ready modeled in linear regressions in our case 2 2306 W/kg 0 1 2 3 4 5 0 1 2 3 4 5 W/kg(IAT) determined using multiple linear regression W/kg(IAT) determined by blood lactate values Figure 3. Impact of the individual input parameters on the W/kg(IAT). Int. J. Environ. Res. Public Health 2023, 20, 4641 5 of 12 To better understand the impact of individual input parameters on the W/kg(IAT), we have visualized the regression parameters using an effect plot in Figure 4. For this, we multiplied the weights of the formula in Figure 3 with the actual values in our database. The latter are normalized by subtracting their respective mean values, as this offset is already modeled in linear regressions, in our case 2.2306 W/kg. 21.407 10 weight [kg] Figure 3. Impact of the individual input parameters on the W/kg (IAT). To better understand the impact of individual input parameters on the W/kg (IAT), we have visualized the regression parameters using an effect plot in Figure 4. For this, we multiplied the weights of the formula in Figure 3 with the actual values in our database. The latter are normalized by subtracting their respective mean values, as this offset is al- ready modeled in linear regressions, in our case 2.2306 W/kg. Figure 4. Effect plot for individual input parameters on W/kg (IAT) with parameter values normal- ized to a mean of zero for better comparison. 4. Discussion This retrospective analysis of this dataset was examined to predict heart rate at IAT as well as training load (W/kg) at IAT without measuring blood lactate values for cycle ergometry. Both heart rate and the number of watts at the individual anaerobic threshold are essential parameters for training control. These parameters are currently best deter- mined via blood lactate diagnostics during ergometry performance testing. A total of 1234 performance protocols with blood lactate in cycle ergometry were analyzed. Multiple lin- ear regression analyses were performed to predict the essential parameters heart rate at individual anaerobic threshold (HR(IAT)) and workload at individual anaerobic thresh- old (W/kg(IAT)) by using routine parameters for ergometry without blood lactate. HR(IAT) can be predicted with a root mean square error, RMSE of 8.77 bpm (p < 0.001). The intention of this regression model is the acceleration of preventive medicine by using 0 1 0 1 2 3 4 5 W/kg(IAT) W/kg(IAT) determined by blood lactate values Figure 4. Effect plot for individual input parameters on W/kg(IAT) with parameter values normalized to a mean of zero for better comparison. 4. Discussion This retrospective analysis of this dataset was examined to predict heart rate at IAT as well as training load (W/kg) at IAT without measuring blood lactate values for cycle ergometry. Both heart rate and the number of watts at the individual anaerobic threshold are essential parameters for training control. These parameters are currently best deter- mined via blood lactate diagnostics during ergometry performance testing. A total of 1234 performance protocols with blood lactate in cycle ergometry were analyzed. Multi- ple linear regression analyses were performed to predict the essential parameters heart rate at individual anaerobic threshold (HR(IAT)) and workload at individual anaerobic threshold (W/kg(IAT)) by using routine parameters for ergometry without blood lactate. HR(IAT) can be predicted with a root mean square error, RMSE of 8.77 bpm (p < 0.001). The intention of this regression model is the acceleration of preventive medicine by using every ergometry to compile an individual training recommendation in primary and sec- ondary prevention. At once, the greatest challenge and the utmost benefit is a continuous training adherence. To avoid overexertion, knowing the individual anaerobic threshold is necessary. This applies for preventive medicine as well as for pre-habilitation to meet the proposed exercise recommendation of 150–300 min per week by the WHO. Future work implies to supervise pre-habilitation patients with the recommended regression model, as standard cycle ergometry can be performed by every general practitioner, and patients can be examined close to the place of residence. There is a need for more research in preventive medicine that focuses on developing better preventive training control methods for the general population. The main leverage point is the prevention of cardiovascular diseases, which is one of the most causes of morbidity and mortality. Furthermore, as societies are becoming older, frailty amongst the elderly population is a growing financial burden [37]. Increasing frailty goes in line with a decrease in quality of life. Regular physical activity can reduce frailty [38], and research has also shown an improvement for quality of life [39]. This challenge for health systems needed to be addressed by establishing easy access methods for training control parameters and training programs for the main population. Int. J. Environ. Res. Public Health 2023, 20, 4641 6 of 12 Although lactate performance diagnostics is a well-established method for recording performance, methodological errors must be considered. Due to constantly increasing blood lactate values and only intermittently measured lactate values by using the capillary blood of the earlobe, measurement inaccuracy must be assumed. Furthermore, certain nutritional methods (e.g., low-carb) are associated with an altered lactate curve [40]. Due to glycogen depletion, incorrect low blood lactate is measured, which can lead to a misinterpretation of the lactate curve. A large number of studies examined different threshold models, whereby an exact determination of the aerobic and anaerobic threshold is better to be regarded as an aerobic and anaerobic transition [41,42]. The earlier assumption of fixed aerobic and anaerobic thresholds soon showed individual differences in further investigations and the need to consider individual threshold methods. Despite these challenges of metabolic threshold models, the lactate determination for training recommendation was established as daily routine in contrast to other methods, in the last decades. The cardio-pulmonary exercise test (CPET), which is applied to determine VO2max and ventilatory thresholds, is also a method of assessing an individual’s physical fitness. For the collection of respiratory and metabolic parameters, this is a complex measurement, and expensive equipment with regular calibration is needed. This method is significantly more time-consuming, requires special trained nurses or sport scientists, and is therefore primarily reserved for patients with cardiac and pulmonary diseases. The RPE and the walking test are simple methods to avoid overexertion in preventive and recreational sports. Nevertheless, the application of RPE is difficult for people who are inexperienced in sports and can easily lead to overexertion or unchallenged activity. In preventive medical examination, an individual training recommendation is increasingly demanded by patients. It could be shown that the lactate accumulation shows inter-individual differences [43–45], and fixed submaximal threshold concepts (of 2 mmol/and 4 mmol/l) should not be applied for individual training recommendations. The lactate concentration in the aerobic–anaerobic tran- sition range is also dependent on muscle recruitment in different movement patterns [46,47]. Individual training recommendations should therefore be specific to the sport. Several studies have been able to prove the training effect of exercise based on the individual anaerobic threshold. The determination of the HR(IAT) during the ergometry without lactate diag- nostics can be used for recommendations of basic endurance and interval training and for prevention programs. Lactate measurement examination is an expensive, as lactate measurement equipment and special trained nurses are required, and time-consuming examination and has until now rarely been covered by health insurance companies. Depending on the individual, the test takes forty to fifty minutes, including warm-up, measuring resting heart rate and recovery time in the end. In various studies, the lactate transition range and the maxi- mum lactate steady state showed a connection with hormonal and immunological changes, which at least supports the assumption of an upper anaerobic threshold [48–50]. Therefore, the individual anaerobic threshold should be considered when making training recom- mendations for the general population, since long-term training with a disproportionate increase in lactate can lead to training non-adherence and vulnerability for infections or injuries, thus bringing the known advantages of regular physical activity [49,50]. Our regression model allows a good prediction of HR(IAT) with an RSME 8.77 bpm and a prediction of W/kg(IAT) with a deviation of 0.241 W/kg. Shen et al. examined the velocity at lactate threshold on a treadmill by using several prediction models with different heart rates [31]. As with the data in this study, age was not a significant parameter and was excluded in the regression models. However, body mass index was excluded [31], and this study only included body weight for predicting W/kg(IAT). Interestingly, women seem to show a slightly higher W/kg unless body height is considered to be negative, in which case these effects neutralize each other. Differences between the results of Shen et al. also might be attributed to different physical activity on a treadmill and a cycle ergometry [31]. Sport-specific differences for HR(IAT) and W/kg(IAT) needed to be considered [51–53], Int. J. Environ. Res. Public Health 2023, 20, 4641 7 of 12 and further research on regression models for running and rowing should be addressed in future studies. The exclusion of heart rate recovery for the prediction of W/kg(IAT) was justified by not meeting the significance level of p < 0.001. This suggests that HRR may not be a singular predictor for evaluating physical fitness [54]. As research results are inconclusive and the evidence is weak [55–57], further research should be performed in larger studies. However, HRR should be recorded as a longitudinal parameter [54], since changes in HRR showed good results in recognizing cardiopulmonary diseases [58,59]. In this context, HRR is an essential parameter, which should be monitored regularly to detect changes in autonomic function [60]. The Karvonen formula is mainly used for training control in popular sports. The Karvonen formula uses the heart rate reserve, and it requires that the maximum heart rate and resting heart rate are determined to apply the formula [61]. By multiplication with a fitness level factor (0.8 for athletes; 0.6 for recreational athletes; and 0.3 for untrained people), the heart rate at the anaerobic threshold can be calculated [61]. The Karvonen formula was applied to the examined measurement protocol results in this study. The results of using the Karvonen formula with a factor of 0.7, due to the predominantly athletic clientele of the sports medicine university outpatient clinic, are shown in Figure 2. In comparison to the measured heart rate at IAT, the scatter diagram reveals a good correlation with a higher RMSE of 11.2 bpm in comparison to the regression model of this study (RMSE 8.77 bpm). The shape of the curve indicates an overestimation of the low values and an underestimation of the high HR values at the IAT. The Karvonen formula also uses resting HR and maximum heart rate to determine HR(IAT). Both heart rates are individual values, and maximum heart rate is especially difficult to determine for the general population, especially as maximum heart rate changes with age [62–64]. Thus, an initial determination of maximum heart rate is also required for the Karvonen formula and should be acquired under medical supervision, especially for individuals > 35 years to cardiovascular adverse events. Due to the improved prediction based on a regression model determined in this study, we recommend cycle ergometry in medical supervision with the regression model identified in Figure 1. In preventive medicine, ergometry is also recommended for every sports beginner and returner over the age of 40 (for men) and over 55 (for women), according to the German guideline for preventive medical check-ups in sports. In contrast to lactate performance diagnostics, ergometry can be carried out by almost any general practitioner or as part of an occupational medical examination. However, an individual training recommendation is usually only given by sports physicians, since a respective specialization for individual training advice is missing. Due to the increasing number of cardiovascular events, obesity and an increasingly aging population, there is a health gap to reach the general population with individual training recommendations and to examine the full scope of preventive medicine. The proposed regression model differs from other studies with a significantly higher number of study protocols examined in a heterogeneous population [27,28]. In addition, further research should examine whether shorter exercise tests, such as the 6-MWT (6-min walking test), can be used for a regression model prediction of essential parameters for training control [65]. Studies in obese individuals demonstrated promising potential to assess individual respiratory threshold [65–67]. Especially obese and older subjects or individuals with other disabilities which rule out cycle ergometry might benefit [65]. Thus, a regression model for cycle ergometry with a shortened protocol should be addressed in future studies, as these shorter tests can be performed more regularly to examine training improvement and address the changed HR(IAT) after consistent training [66]. Considering the results of this retrospective study, we recommend the output of a training program with an individual training heart rate at IAT and watt range at IAT, provided after every check-up examination using cycle ergometry in medical supervision, including the recommendation of the WHO [15]. Int. J. Environ. Res. Public Health 2023, 20, 4641 8 of 12 Future Work As it is known that also children and adolescent obesity has been continuously rising during the last few years [68], there is a need to find approaches for physical activity in these age groups, since chronic diseases will start in early ages and will have a huge impact on GNP. Further studies are necessary to establish regression models for HF(IAT) for adolescents to teach them a healthy and adequate regular exercise program with a potentially better exercise adherence, since exercise adherence is one of the most encourag- ing parameters for health [20]. These exercise programs should be established in schools under supervision and with regular physical examinations. Furthermore, research has shown that pre-habilitation can be a relevant benefit prior to chemotherapy or extensive operations [69,70]. As neoadjuvant chemotherapy is associated with decreasing aerobic endurance [70,71], there is a need for easy access training therapy not only in primary but also in secondary and tertiary prevention. Measuring HF(IAT) at routine secondary and tertiary preventive examinations may improve exercise adherence; further research in these subgroups is necessary. Further goals of these examinations are an establishment of pre-habilitation offers close to home, besides the expansion of the preventive individual training recommendations for the general population. An individual training recommendation for pre-habilitation could therefore be made directly by the attending general practitioner or cardiologist. A gap in care, of mostly only a few sports medicine offers, could thus be closed. Further examinations with other ergometer types (rowing ergometer, elliptical) are planned in order to enable a conversion of HR(IAT) and different ergometer types in prevention and pre-habilitation. 5. Limitations Incorrect entries during the manual transmission of the lactate values must be consid- ered. These were minimized in advance by means of a plausibility check of the entire data set. The sample size in this study is appropriate for generating a valuable prediction in comparison to other studies [31,72]. The age and gender distribution may vary in compari- son to the general population, as the examined population includes more physically active individuals, especially in the younger age groups. Furthermore, a heart rate deviation of 8.77 bpm is not appropriate for athletes in professional sports, although a blood lactate test or cardiopulmonary exercise test (CPET) is still recommended for this clientele. This regres- sion model is suitable for cardiorespiratory endurance sports. It should be noticed that it is not applicable for resistance or interval training; individual training recommendations for these kinds of training should be considered. At the same time, ergometry offers a simple and inexpensive measuring method that can be performed in the outpatient and inpatient sector and represents a suitable procedure for popular sports and preventive medicine to monitor cardiorespiratory training. 6. Conclusions In conclusion, it is possible to derive relevant parameters for training control after a standard cycle ergometry without performing a blood lactate test by using regression models to predict HR(IAT) and W/kg(IAT) for the general population. This enables training control without blood lactate diagnostics or CPET and does achieve enormous time and financial savings for active exercise therapy as well as for preventive and rehabilitative medicine. Regular individual test repetition allows the consideration of short-term training adaption and supports continuous training progress. Author Contributions: Conceptualization, C.R. and B.W.; Methodology, C.R.; Formal analysis, C.R.; Investigation, C.R.; Writing—original draft, C.R.; Writing—review and editing, B.W.; Supervision, B.W.; Project administration, B.W. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Int. J. Environ. Res. Public Health 2023, 20, 4641 9 of 12 Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki and approved by the Institutional Review Board (or Ethics Committee) of Humboldt University (protocol code HU-KSBF_EK_2018_0004; date of approval 2018/05/09). Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: The data presented in this study are available on request from the corresponding author. The data are not publicly available due to data privacy regulations. Conflicts of Interest: The authors declare no conflict of interest regarding this publication. Abbreviations CPET cardiopulmonary exercise test HR heart rate HR (IAT) heart rate at individual anaerobic threshold HRmax maximum heart rate HRR heart rate recovery = HRmax-HR after 5 min of recovery Pmax maximum power output W/kg(IAT) Watt per kg at individual anaerobic threshold RMSE root mean square error W Watt References 1. Halle, M.; Korsten-Reck, U.; Wolfarth, B.; Berg, A. Low-grade systemic inflammation in overweight children: Impact of physical fitness. Exerc. Immunol. Rev. 2004, 10, 66–74. [PubMed] 2. Lehnert, T.; Streltchenia, P.; Konnopka, A.; Riedel-Heller, S.G.; König, H.-H. Health burden and costs of obesity and overweight in Germany: An update. Eur. J. Health Econ. 2014, 16, 957–967. [CrossRef] [PubMed] 3. Seidell, J.C.; Halberstadt, J. The global burden of obesity and the challenges of prevention. Ann. Nutr. Metab. 2015, 66 (Suppl. S2), 7–12. [CrossRef] 4. Cheng, W.; Zhang, Z.; Cheng, W.; Yang, C.; Diao, L.; Liu, W. Associations of leisure-time physical activity with cardiovascular mortality: A systematic review and meta-analysis of 44 prospective cohort studies. Eur. J. Prev. Cardiol. 2018, 25, 1864–1872. [CrossRef] 5. Lollgen, H.; Leyk, D. Prevention by physical activity. The relevance of physical fitness. Der Internist 2012, 53, 663–670. [CrossRef] 6. Einarson, T.R.; Acs, A.; Ludwig, C.; Panton, U.H. Economic Burden of Cardiovascular Disease in Type 2 Diabetes: A Systematic Review. Value Health 2018, 21, 881–890. [CrossRef] 7. Li, Q.; Blume, S.W.; Huang, J.C.; Hammer, M.; Graf, T.R. The Economic Burden of Obesity by Glycemic Stage in the United States. Pharmacoeconomics 2015, 33, 735–748. [CrossRef] [PubMed] 8. Yates, N.; Teuner, C.M.; Hunger, M.; Holle, R.; Stark, R.; Laxy, M.; Hauner, H.; Peters, A.; Wolfenstetter, S.B. The Economic Burden of Obesity in Germany: Results from the Population-Based KORA Studies. Obes. Facts 2016, 9, 397–409. [CrossRef] 9. Berry, J.D.; Pandey, A.; Gao, A.; Leonard, D.; Farzaneh-Far, R.; Ayers, C.; DeFina, L.; Willis, B. Physical fitness and risk for heart failure and coronary artery disease. Circ. Heart Fail. 2013, 6, 627–634. [CrossRef] 10. Babl, F.E.; Weiner, D.L.; Bhanji, F.; Davies, F.; Berry, K.; Barnett, P. Advanced training in pediatric emergency medicine in the United States, Canada, United Kingdom, and Australia: An international comparison and resources guide. Ann. Emerg. Med. 2005, 45, 269–275. [CrossRef] 11. Khan, K.M.; Thompson, A.M.; Blair, S.N.; Sallis, J.F.; Powell, K.E.; Bull, F.C.; Bauman, A.E. Sport and exercise as contributors to the health of nations. Lancet 2012, 380, 59–64. [CrossRef] 12. Xiang, M.; Zhang, Z.; Kuwahara, K. Impact of COVID-19 pandemic on children and adolescents’ lifestyle behavior larger than expected. Prog. Cardiovasc. Dis. 2020, 63, 531–532. [CrossRef] [PubMed] 13. Zhang, J.; Zheng, S.; Hu, Z. The Effect of Physical Exercise on Depression in College Students: The Chain Mediating Role of Self-Concept and Social Support. Front. Psychol. 2022, 13, 841160. [CrossRef] [PubMed] 14. Ji, C.; Yang, J.; Lin, L.; Chen, S. Physical Exercise Ameliorates Anxiety, Depression and Sleep Quality in College Students: Experimental Evidence from Exercise Intensity and Frequency. Behav. Sci. 2022, 12, 61. [CrossRef] 15. Bull, F.C.; Al-Ansari, S.S.; Biddle, S.; Borodulin, K.; Buman, M.P.; Cardon, G.; Carty, C.; Chaput, J.-P.; Chastin, S.; Chou, R.; et al. World Health Organization 2020 guidelines on physical activity and sedentary behaviour. Br. J. Sports Med. 2020, 54, 1451–1462. [CrossRef] 16. Löllgen, H.; Papadopoulou, T. Updated meta-analysis of prevention of cardiovascular mortality by regular physical activity. Eur. J. Prev. Cardiol. 2018, 25, 1861–1863. [CrossRef] 17. Franklin, B.A.; Cushman, M. Recent advances in preventive cardiology and lifestyle medicine: A themed series. Circulation 2011, 123, 2274–2283. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2023, 20, 4641 10 of 12 18. Ekkekakis, P.; Lind, E. Exercise does not feel the same when you are overweight: The impact of self-selected and imposed intensity on affect and exertion. Int. J. Obes. 2006, 30, 652–660. [CrossRef] [PubMed] 19. DaSilva, S.G.; Guidetti, L.; Buzzachera, C.F.; Elsangedy, H.M.; Colombo, H.; Krinski, K.; Dos Santos, S.L.C.; De Campos, W.; Baldari, C. The influence of adiposity on physiological, perceptual, and affective responses during walking at a self-selected pace. Percept. Mot. Ski. 2009, 109, 41–60. [CrossRef] 20. Janssen, I.; LeBlanc, A.G. Systematic review of the health benefits of physical activity and fitness in school-aged children and youth. Int. J. Behav. Nutr. Phys. Act. 2010, 7, 40. [CrossRef] 21. Garcia-Ortiz, L.; Recio-Rodriguez, J.I.; Agudo-Conde, C.; Patino-Alonso, M.C.; Maderuelo-Fernandez, J.-A.; Gento, I.R.; Puig, E.P.; Gonzalez-Viejo, N.; Arietaleanizbeaskoa, M.S.; Schmolling-Guinovart, Y.; et al. Long-Term Effectiveness of a Smartphone App for Improving Healthy Lifestyles in General Population in Primary Care: Randomized Controlled Trial (Evident II Study). JMIR mHealth uHealth 2018, 6, e107. [CrossRef] [PubMed] 22. Martin, L.R.; Williams, S.L.; Haskard, K.B.; DiMatteo, M.R. The challenge of patient adherence. Ther. Clin. Risk Manag. 2005, 1, 189–199. [PubMed] 23. Skinner, J.S.; McLellan, T.H. The transition from aerobic to anaerobic metabolism. Res. Q. Exerc. Sport 1980, 51, 234–248. [CrossRef] [PubMed] 24. Svedahl, K.; MacIntosh, B.R. Anaerobic threshold: The concept and methods of measurement. Can. J. Appl. Physiol. 2003, 28, 299–323. [CrossRef] [PubMed] 25. Meyer, T.; Lucía, A.; Earnest, C.P.; Kindermann, W. A conceptual framework for performance diagnosis and training prescription from submaximal gas exchange parameters–theory and application. Int. J. Sports Med. 2004, 26, S38–S48. [CrossRef] [PubMed] 26. Tomasits, J.H.P. Leistungsphysiologie; Springer: Berlin/Heidelberg, Germany, 2016. 27. Conconi, F.; Ferrari, M.; Ziglio, P.G.; Droghetti, P.; Codeca, L. Determination of the anaerobic threshold by a noninvasive field test in runners. J. Appl. Physiol. 1982, 52, 869–873. [CrossRef] 28. Ham, J.-H.; Park, H.-Y.; Kim, Y.-H.; Bae, S.-K.; Ko, B.-H.; Nam, S.-S. Development of an anaerobic threshold (HRLT, HRVT) estimation equation using the heart rate threshold (HRT) during the treadmill incremental exercise test. J. Exerc. Nutr. Biochem. 2017, 21, 43–49. [CrossRef] 29. Novais, L.D.; Silva, E.; Simões, R.P.; Sakabe, D.I.; Martins, L.E.B.; Oliveira, L.; Diniz, C.A.R.; Gallo, L., Jr.; Catai, A.M. Anaerobic Threshold by Mathematical Model in Healthy and Post-Myocardial Infarction Men. Int. J. Sports Med. 2015, 37, 112–118. [CrossRef] 30. Schmitz, B.; Niehues, H.; Thorwesten, L.; Klose, A.; Kruger, M.; Brand, S.M. Sex Differences in High-Intensity Interval Training-Are HIIT Protocols Interchangeable Between Females and Males? Front. Physiol. 2020, 11, 38. [CrossRef] 31. Shen, T.; Wen, X. Heart-rate-based prediction of velocity at lactate threshold in ordinary adults. J. Exerc. Sci. Fit. 2019, 17, 108–112. [CrossRef] 32. Garcia-Tabar, I.; Izquierdo, M.; Gorostiaga, E.M. On-field prediction vs. monitoring of aerobic capacity markers using submaximal lactate and heart rate measures. Scand. J. Med. Sci. Sports 2017, 27, 462–473. [CrossRef] 33. Stratton, E.; O’Brien, B.J.; Harvey, J.; Blitvich, J.; McNicol, A.J.; Janissen, D.; Paton, C.; Knez, W. Treadmill Velocity Best Predicts 5000-m Run Performance. Int. J. Sports Med. 2008, 30, 40–45. [CrossRef] [PubMed] 34. Roseguini, B.; Narro, F.; Oliveira, Á.R.; Ribeiro, J. Estimation of the lactate threshold from heart rate response to submaximal exercise: The pulse deficit. Int. J. Sports Med. 2006, 28, 463–469. [CrossRef] [PubMed] 35. Weiler, R.; Stamatakis, E.; Blair, S. Should health policy focus on physical activity rather than obesity? Yes. BMJ 2010, 340, c2603. [CrossRef] [PubMed] 36. Dickhuth, H.-H.; Yin, L.; Niess, A.; Röcker, K.; Mayer, F.; Heitkamp, H.-C.; Horstmann, T. Ventilatory, lactate-derived and catecholamine thresholds during incremental treadmill running: Relationship and reproducibility. Int. J. Sports Med. 1999, 20, 122–127. [CrossRef] 37. Gobbens, R.J.J.; Luijkx, K.G.; Wijnen-Sponselee, M.T.; Schols, J.M.G.A. Towards an integral conceptual model of frailty. J. Nutr. Health Aging 2009. [CrossRef] 38. Oliveira, J.S.; Pinheiro, M.B.; Fairhall, N.; Walsh, S.; Franks, T.C.; Kwok, W.; Bauman, A.; Sherrington, C. Evidence on Physical Activity and the Prevention of Frailty and Sarcopenia Among Older People: A Systematic Review to Inform the World Health Organization Physical Activity Guidelines. J. Phys. Act. Health 2020, 17, 1247–1258. [CrossRef] 39. Kojima, G.; Iliffe, S.; Jivraj, S.; Walters, K. Association between frailty and quality of life among community-dwelling older people: A systematic review and meta-analysis. J. Epidemiol. Community Health 2016, 70, 716–721. [CrossRef] 40. Faude, O.; Kindermann, W.; Meyer, T. Lactate threshold concepts: How valid are they? Sports Med. 2009, 39, 469–490. [CrossRef] 41. Kindermann, W.; Simon, G.; Keul, J. The significance of the aerobic-anaerobic transition for the determination of work load intensities during endurance training. Eur. J. Appl. Physiol. Occup. Physiol. 1979, 42, 25–34. [CrossRef] 42. McLellan, T.M.; Skinner, J.S. The use of the aerobic threshold as a basis for training. Can. J. Appl. Sport Sci. J. Can. des Sci. Appl. au Sport 1981, 6, 197–201. 43. Beneke, R.; Hütler, M.; Leithäuser, R.M. Maximal lactate-steady-state independent of performance. Med. Sci. Sports Exerc. 2000, 32, 1135–1139. [CrossRef] 44. Urhausen, A.; Coen, B.; Weiler, B.; Kindermann, W. Individual anaerobic threshold and maximum lactate steady state. Int. J. Sports Med. 1993, 14, 134–139. [CrossRef] Int. J. Environ. Res. Public Health 2023, 20, 4641 11 of 12 45. MacIntosh, B.R.; Esau, S.; Svedahl, K. The lactate minimum test for cycling: Estimation of the maximal lactate steady state. Can. J. Appl. Physiol. 2002, 27, 232–249. [CrossRef] [PubMed] 46. Beneke, R.; von Duvillard, S.P. Determination of maximal lactate steady state response in selected sports events. Med. Sci. Sports Exerc. 1996, 28, 241–246. [CrossRef] [PubMed] 47. Beneke, R.; Leithäuser, R.M.; Hütler, M. Dependence of the maximal lactate steady state on the motor pattern of exercise. Br. J. Sports Med. 2001, 35, 192–196. [CrossRef] [PubMed] 48. Lucia, A.; Sánchez, O.; Carvajal, A.; Chicharro, J.L. Analysis of the aerobic-anaerobic transition in elite cyclists during incremental exercise with the use of electromyography. Br. J. Sports Med. 1999, 33, 178–185. [CrossRef] 49. Urhausen, A.; Weiler, B.; Coen, B.; Kindermann, W. Plasma catecholamines during endurance exercise of different intensities as related to the individual anaerobic threshold. Eur. J. Appl. Physiol. Occup. Physiol. 1994, 69, 16–20. [CrossRef] 50. Gabriel, H.; Kindermann, W. The acute immune response to exercise: What does it mean? Int. J. Sports Med. 1997, 18 (Suppl. S1), S28–S45. [CrossRef] 51. Olsson, K.S.E.; Rosdahl, H.; Schantz, P. Interchangeability and optimization of heart rate methods for estimating oxygen uptake in ergometer cycling, level treadmill walking and running. BMC Med. Res. Methodol. 2022, 22, 55. [CrossRef] 52. Roecker, K.; Striegel, H.; Dickhuth, H.H. Heart-rate recommendations: Transfer between running and cycling exercise? Int. J. Sports Med. 2003, 24, 173–178. [CrossRef] 53. Larson, A.J. Variations in heart rate at blood lactate threshold due to exercise mode in elite cross-country skiers. J. Strength Cond. Res. 2006, 20, 855–860. [CrossRef] 54. Römer, C.; Wolfarth, B. Heart Rate Recovery (HRR) Is Not a Singular Predictor for Physical Fitness. Int. J. Environ. Res. Public Health 2022, 20, 792. [CrossRef] [PubMed] 55. Bosquet, L.; Gamelin, F.-X.; Berthoin, S. Reliability of Postexercise Heart Rate Recovery. Int. J. Sports Med. 2007, 29, 238–243. [CrossRef] 56. Mongin, D.; Chabert, C.; Courvoisier, D.S.; García-Romero, J.; Alvero-Cruz, J.R. Heart rate recovery to assess fitness: Comparison of different calculation methods in a large cross-sectional study. Res. Sports Med. 2021, 1–14. [CrossRef] [PubMed] 57. Daanen, H.A.; Lamberts, R.P.; Kallen, V.L.; Jin, A.; Van Meeteren, N.L. A systematic review on heart-rate recovery to monitor changes in training status in athletes. Int. J. Sports Physiol. Perform. 2012, 7, 251–260. [CrossRef] 58. Dhoble, A.; Lahr, B.D.; Allison, T.G.; Kopecky, S.L. Cardiopulmonary fitness and heart rate recovery as predictors of mortality in a referral population. J. Am. Heart Assoc. 2014, 3, e000559. [CrossRef] [PubMed] 59. Cole, C.R.; Blackstone, E.H.; Pashkow, F.J.; Snader, C.E.; Lauer, M.S. Heart-rate recovery immediately after exercise as a predictor of mortality. N. Engl. J. Med. 1999, 341, 1351–1357. [CrossRef] 60. Aztatzi-Aguilar, O.G.; Vargas-Domínguez, C.; Debray-Garcia, Y.; Ortega-Romero, M.S.; Almeda-Valdés, P.; Aguilar-Salinas, C.A.; Naranjo-Meneses, M.A.; Mena-Orozco, D.A.; Lam-Chung, C.E.; Cruz-Bautista, I.; et al. Biochemical and Hematological Relationship with the Evaluation of Autonomic Dysfunction by Heart Rate Recovery in Patients with Asthma and Type 2 Diabetes. Diagnostics 2021, 11, 2187. [CrossRef] 61. Karvonen, M.; Kentala, K.; Mustala, O. The effects of training heart rate: A longitudinal study. Ann. Med. Exp. Biol. Fenn. 1957, 35, 307–315. 62. Fairbarn, M.S.; Blackie, S.P.; McElvaney, N.G.; Wiggs, B.R.; Paré, P.D.; Pardy, R.L. Prediction of heart rate and oxygen uptake during incremental and maximal exercise in healthy adults. Chest 1994, 105, 1365–1369. [CrossRef] 63. Gellish, R.L.; Goslin, B.R.; Olson, R.E.; McDONALD, A.; Russi, G.D.; Moudgil, V.K. Longitudinal modeling of the relationship between age and maximal heart rate. Med. Sci. Sports Exerc. 2007, 39, 822–829. [CrossRef] 64. Tanaka, H.; Monahan, K.D.; Seals, D.R. Age-predicted maximal heart rate revisited. J. Am. Coll. Cardiol. 2000, 37, 153–156. [CrossRef] [PubMed] 65. Emerenziani, G.P.; Ferrari, D.; Vaccaro, M.G.; Gallotta, M.C.; Migliaccio, S.; Lenzi, A.; Baldari, C.; Guidetti, L. Prediction equation to estimate heart rate at individual ventilatory threshold in female and male obese adults. PLoS ONE 2018, 13, e0197255. [CrossRef] [PubMed] 66. Emerenziani, G.P.; Gallotta, M.C.; Migliaccio, S.; Ferrari, D.; Greco, E.A.; Saavedra, F.J.; Iazzoni, S.; Aversa, A.; Donini, L.M.; Lenzi, A.; et al. Effects of an individualized home-based unsupervised aerobic training on body composition and physiological parameters in obese adults are independent of gender. J. Endocrinol. Investig. 2017, 41, 465–473. [CrossRef] 67. Makni, E.; Moalla, W.; Trabelsi, Y.; Lac, G.; Brun, J.F.; Tabka, Z.; Elloumi, M. Six-minute walking test predicts maximal fat oxidation in obese children. Int. J. Obes. 2012, 36, 908–913. [CrossRef] [PubMed] 68. Collaboration NCDRF. Worldwide trends in body-mass index, underweight, overweight, and obesity from 1975 to 2016: A pooled analysis of 2416 population-based measurement studies in 128.9 million children, adolescents, and adults. Lancet 2017, 390, 2627–2642. [CrossRef] 69. West, M.; Loughney, L.; Lythgoe, D.; Barben, C.; Sripadam, R.; Kemp, G.; Grocott, M.; Jack, S. Effect of prehabilitation on objectively measured physical fitness after neoadjuvant treatment in preoperative rectal cancer patients: A blinded interventional pilot study. Br. J. Anaesth. 2015, 114, 244–251. [CrossRef] Int. J. Environ. Res. Public Health 2023, 20, 4641 12 of 12 70. Navidi, M.; Phillips, A.; Griffin, S.M.; Duffield, K.E.; Greystoke, A.; Sumpter, K.; Sinclair, R.C.F. Cardiopulmonary fitness before and after neoadjuvant chemotherapy in patients with oesophagogastric cancer. Br. J. Surg. 2018, 105, 900–906. [CrossRef] 71. Jack, S.; West, M.; Raw, D.; Marwood, S.; Ambler, G.; Cope, T.; Shrotri, M.; Sturgess, R.; Calverley, P.; Ottensmeier, C.; et al. The effect of neoadjuvant chemotherapy on physical fitness and survival in patients undergoing oesophagogastric cancer surgery. Eur. J. Surg. Oncol. (EJSO) 2014, 40, 1313–1320. [CrossRef] 72. Camargo Alves, J.C.; Segabinazi Peserico, C.; Nogueira, G.A.; Machado, F.A. The influence of the regression model and final speed criteria on the reliability of lactate threshold determined by the Dmax method in endurance-trained runners. Appl. Physiol. Nutr. Metab. 2016, 41, 1039–1044. [CrossRef] [PubMed] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Prediction of Relevant Training Control Parameters at Individual Anaerobic Threshold without Blood Lactate Measurement.

03-06-2023

Römer, Claudia,Wolfarth, Bernd

eng

PMC7971079

Submitted 4 June 2020 Accepted 4 January 2021 Published 15 March 2021 Corresponding author Hung-Ting Chen, [email protected] Academic editor Celine Gallagher Additional Information and Declarations can be found on page 13 DOI 10.7717/peerj.10831 Copyright 2021 Chung et al. Distributed under Creative Commons CC-BY 4.0 OPEN ACCESS Predicting maximal oxygen uptake from a 3-minute progressive knee-ups and step test Yu-Chun Chung1, Ching-Yu Huang2, Huey-June Wu3, Nai-Wen Kan1, Chin-Shan Ho4, Chi-Chang Huang4 and Hung-Ting Chen5 1 Center of General Education, Taipei Medical University, Taipei, Taiwan 2 Service Systems Technology Center, Industrial Technology Research Institute, Hsinchu, Taiwan 3 Department of Combat Sports and Chinese Martial Arts, Chinese Culture University, Taipei, Taiwan 4 Graduate Institute of Sports Science, National Taiwan Sport University, Taoyuan, Taiwan 5 Physical Education Office, Ming Chuan University, Taipei, Taiwan ABSTRACT Background. Cardiorespiratory fitness assessment is crucial for diagnosing health risks and assessing interventions. Direct measurement of maximum oxygen uptake ( ˙VO2 max) yields more objective and accurate results, but it is practical only in a laboratory setting. We therefore investigated whether a 3-min progressive knee-up and step (3MPKS) test can be used to estimate peak oxygen uptake in these settings. Method. The data of 166 healthy adult participants were analyzed. We conducted a ˙VO2 max test and a subsequent 3MPKS exercise test, in a balanced order, a week later. In a multivariate regression model, sex; age; relative ˙VO2 max; body mass index (BMI); body fat percentage (BF); resting heart rate (HR0); and heart rates at the beginning as well as at the first, second, third, and fourth minutes (denoted by HR0, HR1, HR2, HR3, and HR4, respectively) during a step test were used as predictors. Moreover, R2 and standard error of estimate (SEE) were used to evaluate the accuracy of various body composition models in predicting ˙VO2max. Results. The predicted and actual ˙VO2 max values were significantly correlated (BF% model: R2 = 0.624, SEE = 4.982; BMI model: R2 = 0.567, SEE = 5.153). The BF% model yielded more accurate predictions, and the model predictors were sex, age, BF%, HR0, 1HR3−HR0, and 1HR3−HR4. Conclusion. In our study, involving Taiwanese adults, we constructed and verified a model to predict ˙VO2 max, which indicates cardiorespiratory fitness. This model had the predictors sex, age, body composition, and heart rate changes during a step test. Our 3MPKS test has the potential to be widely used in epidemiological research to measure ˙VO2 max and other health-related parameters. Subjects Anatomy and Physiology, Cardiology, Hematology, Kinesiology, Respiratory Medicine Keywords Aerobic ability, 3-min Harvard step test, Cardiovascular function, Field tests INTRODUCTION In 2016, the American Heart Association launched a series of publications promoting the clinical evaluation of cardiorespiratory fitness (CRF) with the overall aim of improving the prevention and treatment of cardiovascular disease (CVD; Ross et al., 2016). Furthermore, How to cite this article Chung Y-C, Huang C-Y, Wu H-J, Kan N-W, Ho C-S, Huang C-C, Chen H-T. 2021. Predicting maximal oxygen uptake from a 3-minute progressive knee-ups and step test. PeerJ 9:e10831 http://doi.org/10.7717/peerj.10831 the association urged the US federal government to compile a registered CRF database (Kaminsky et al., 2013); this highlights the importance of CRF. CRF is generally defined as the integrated ability to transport oxygen from the atmosphere to the mitochondria for physical activity. Notably, CRF involves the respiratory, circulatory, and neuromuscular systems and has a clear and direct relationship with the functions of various systems. Individuals with weak CRF have an up to 70% all-cause mortality rate and 56% cardiovascular mortality rate (Kodama et al., 2009). Similarly, every 1-MET increase in athletic ability reduces all-cause mortality and cardiovascular mortality rates by 15% and 13%, respectively (Kodama et al., 2009). Numerous studies have suggested that CRF and CVD are related to all-cause mortality and cancer mortality (Blair et al., 1989; Laukkanen et al., 2004; Sui, LaMonte & Blair, 2007; Sawada et al., 2014; Sui, LaMonte & Blair, 2007). A recent meta-analysis reported CRF to be a predictor of the risk of sudden cardiac death (Jiménez-Pavón, Lavie & Blair, 2019). Therefore, CRF assessment is crucial for diagnosing health risks and assessing interventions. CRF can be measured using the respiratory data of exercising participants. Specifically, these data are used to calculate maximal oxygen uptake ( ˙VO2 max), the gold standard for CRF measurement; in the measurement, participants either run on a treadmill or use an ergometer at an exercise intensity that increases progressively until a given maximum is reached. Although submaximal exercise models and nonexercise models (without an exercise test) are alternatives for estimating ˙VO2 max in measuring CRF (Abut, Akay & George, 2016), the direct measurement of ˙VO2 max yields more objective and accurate results. However, such measurement is inconvenient because it requires expensive equipment and well-trained experimenters. In addition, participants perceive such measurement tests to be exhausting, time-consuming, and relatively risky and are thus less willing to participate. Accordingly, researchers have developed various submaximal exercise tests to indirectly estimate ˙VO2 max; moreover, retrospective studies conducted by the American Heart Association have demonstrated that CRF indicators, whether directly measured or indirectly estimated, are robust indicators of health (Ross et al., 2016). Submaximal exercise is a common method for estimating ˙VO2 max, particularly in epidemiological research and large-scale physical fitness testing that involve numerous participants. The field tests in these measurement procedures include running, shuttle running, and the step test, with the step test being the most common method for evaluating cardiovascular function (Grant, Joseph & Campagna, 1999). In particular, the YMCA step test is widely used to predict ˙VO2 max (Beutner et al., 2015). Currently, the Sports Administration of Taiwan’s Ministry of Education uses the 3-min Harvard step test for its National Physical Fitness and Cardiovascular Test. Specifically, three heart rate measurements are used to calculate the step-up index. However, previous studies have reported considerable differences in the validity of using the step test index to evaluate ˙VO2 max, with the corresponding correlation coefficient (R) being 0.35–0.94 (Buckley et al., 2004; Chang & Lin, 1995; Mazic et al., 2001; Su, Lin & Hsieh, 2006; Chang & Lin, 1995; Yoopat, Vanwonterghem & Louhevaara, 2002). Furthermore, step tests require the use of step-up boxes, and the overall test time must be at least 6 min to allow for heart rate recovery. Participants who are less physically fit or who have knee conditions may find it Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 2/16 difficult to complete the test and may also fall in the process of going up and down the stairs. A team of Japanese researchers developed a new 3-min walking test (Cao et al., 2013). Specifically, their main evaluation criteria comprised participant characteristics such as age, sex, and BMI as well as participants’ RPE during exercise. These criteria were determined to be effective predictors of ˙VO2 max, and participants thought that this method was quicker and easier. Tests of general CRF are crucial to the clinical evaluation of CVD. Additionally, the advantages and disadvantages, such as venue size, participant willingness, and the instruments, of various past field tests should be considered during the formulation of new methods, as done in the present study. Accordingly, we conducted the present study with the aim of developing a rapid, convenient, and low-risk model that can predict ˙VO2 max in Taiwanese adults. Additionally, our model accords with the principle that physical exercise ought to be progressive. We investigated the feasibility of using a 3-min progressive knee-ups and step (3MPKS) test to predict ˙VO2 max. MATERIALS AND METHODS Participants Prospective participants were excluded if they (1) had cardiovascular, pulmonary, or metabolic diseases; (2) had neurological, muscular, or skeletal disorders that affected their athletic ability; (3) had other health conditions that made them unsuited for moderate or intense exercise; or (4) were taking medications that could affect the outcome of this study. In total, among 200 participants recruited for this experiment, 166 completed the test. The data of the 166 participants were included in the analysis (age: 20–64 years; 65 men, 101 women). Among the 34 participants excluded, one participant withdrew from the experiment after experiencing suspected symptoms of arrhythmia during exercise; 11 were excluded because they failed to complete the step test within the requisite time (3 min); 12 were excluded because they could not attain the requisite step frequency and knee height for 20 consecutive seconds; nine were excluded because they had missing or improperly measured heart rate data; and one was excluded for having a ‘‘0’’ in their heart rate data. All participants signed an informed consent form after understanding their rights, the risks when participating in this study, and the purpose and method of our research. Our research plan was approved by the Institutional Review Boards (IRBs) of the Industrial Technology Research Institute and of Taipei Medical University (IRB No: N201808055). Participant characteristics are detailed in Table 1. Procedure The anthropometric and body composition measures were height, weight, and body fat percentage (BF%). BF% was measured using bioelectrical impedance analysis (InBody 720, Biospace, USA; McLester et al., 2020), and body mass index (BMI, in kg/m2) was calculated as the quotient that is weight (in kilograms) divided by the squared height (in meters). We conducted two exercise tests in a counterbalanced design. The second test was conducted exactly 1 week after the first and at the same time of the day to ensure that the participants recovered adequately from the first exercise. The participants underwent 5–10 Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 3/16 Table 1 Participant characteristics. Total Training dataset Testing dataset Sample size(n) 166 124 42 Age(years) 41.9 ± 9.6 42.2 ± 9.4 40.8 ± 10.2 Male (n) 65 44 21 Height 164.83 ± 8.35 164.33 ± 8.07 166.30 ± 9.06 Weight 65.63 ± 13.60 65.22 ± 14.08 66.85 ± 12.15 Body fat (%) 27.81 ± 7.92 27.79 ± 7.65 27.86 ± 8.75 ˙VO2 max (ml kg−1 min−1) 34.45 ± 8.69 34.06 ± 8.14 35.61 ± 10.15 HR0 86.04 ± 12.78 86.04 ± 12.99 86.02 ± 12.29 1HR3- HR0 71.00 ± 13.24 71.10 ± 13.41 70.69 ± 12.87 1HR3-HR4 14.64 ± 13.72 14.14 ± 13.94 16.65 ± 14.09 Notes. Data are presented as mean ± standard deviation. HR0, heart rate at the beginning; 1HR3-HR0, difference between third minute heart rate and beginning heart rate; 1HR3- HR4, difference between third minute and fourth minute heart rates. min of dynamic warm-up prior to both exercise tests; to mitigate extraneous influence on the results, the participants were also asked not to engage in moderate or intense exercise 48 h before both exercise tests. To measure the ˙VO2 max of the participants, we used a bicycle ergometer (839E, Monark, Varberg, Sweden) for a maximal graded exercise test. After participants sat still for 2 min, they sat on the stationary bicycle and started cycling at the speed of 70 ± 10 rpm. The participants began the exercise with a 2-min warm-up at 25 W loading, where the loading was increased by 15 W every 2 min. The testing was terminated when the participants could no longer continue the exercise due to bradypnea or fatigue, although the bicycle speed was maintained at 70 rpm. Subsequently, the participants rested for 3 min at a loading of 0 W (no resistance). Throughout the exercise testing, the participants wore a watch to monitor their heart rate and a mask to monitor their breathing. Breath-by-breath analysis was conducted on the participant data through a cardiopulmonary testing system (MetaMax 3B, Cortex, Germany). ˙VO2 max was defined as the maximum average oxygen uptake for 20 consecutive seconds. To ensure that every participant reached ˙VO2 max, we defined ˙VO2 max as being reached if two of the three following conditions were met: (1) ˙VO2 plateaus with increases in work rate; (2) the maximum respiratory exchange ratio is ≥1.10; and (3) 90% of the expected maximal heart rate, obtained by subtracting the participant’s age from 220, is reached (American College of Sports Medicine, 2009). Nearly all participants satisfied the criteria for an acceptable ˙VO2 max, with only one participant excluded from the ˙VO2 max test due to suspected symptoms of arrhythmia observed in the step test. 3MPKS test Prior to the 3MPKS test, the participants wore a sports watch with heart rate (Polar V800, USA) and stride sensors (Polar S3 BlueTooth Stride Sensor, USA). The heart rate sensor was placed at the center of each participant’s chest using a heart rate belt (Polar H10), and the step sensor was fixed on a pair of shoes, with shoelaces, to monitor their heartbeat Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 4/16 Figure 1 3MPKS heart rate model and corresponding step frequency. Full-size DOI: 10.7717/peerj.10831/fig-1 and number of steps taken. After the devices were worn, we measured the midpoint of the line connecting the anterior epicondyle to the midpoint of the sacrum. We marked the midpoint on the wall using colored tape as a reference for the height at which the knee should be lifted to when stepping. After the test started, the participants followed the appropriate rhythm and were required to lift their knee to the marked height at each step. The participants began the test at a pace of 80 spm (steps per minute), which increased by 16 spm every 30 s in six stages. The participants walked in stages 1 to 4 and had to perform stationary running in stages 5 and 6 (Fig. 1). We stopped the exercise if the participants could not achieve the requisite knee height or rhythm for 30 s. For their safety, the participants were asked to relax at a step rate of 80 spm in the first 30 s before resting in a standing position. We recorded the participants’ heart rate during the exercise, at the end of the exercise, and 1 min after the end of the exercise. Thirty-four participants were excluded because (1) their heart rate data were missing, (2) their heart rate was 0, (3) they did not maintain the requisite step frequency or knee height for 20 consecutive seconds, (4) they failed to complete the step test within the requisite duration, and (5) they were suspected of having heart arrhythmia. Potential predictor variables for the results of the 3MPKS test were based on per-second heart rate data collected during the test. The data included heart rate at the beginning as well as at the first, second, third, and fourth minutes, denoted by HR0, HR1, HR2, HR3, and HR4, respectively, and were used for subsequent analysis. Statistical analyses To construct and subsequently evaluate a model for estimating relative oxygen uptake, we divided the full sample set (n = 166) into a 75% training sample set (n = 124) and 25% test sample set through simple random sampling. We analyzed the descriptive statistics for the main parameters, for the whole sample, and for the two subsamples. Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 5/16 Development of prediction model Using Pearson correlation coefficients, we examined the relationship between the predicted and actual relative oxygen uptakes. Multiple regression analysis was used to construct a method for selecting which variables to include in the model for predicting relative oxygen uptake. Through a backward-selection regression approach, the initial model included all possible predictors, including sex (men = 1, women = 0), age, BMI, BF%, HR0, HR1, HR2, HR3, HR4, △HR0 − HR1, △HR1 − HR2, △HR2 − HR3, △HR3 − HR0, and △HR3 − HR4. Additionally, we constructed a BMI model and BF% model to predict body composition. The goodness of fit and precision of the regression equations were evaluated using the multiple coefficient of determination (R2), absolute standard error of estimate (SEE), and relative SEE (%SEE). To construct an accurate regression model, the regression assumptions were verified. We conducted a Kolmogorov–Smirnov test to examine the normality of the residuals, and we calculated the variation inflation factor (VIF) to check for multicollinearity. All statistical analyses were performed using SPSS version 20 (IBM, USA). Statistical significance was indicated by an alpha level of 0.05. RESULTS The 166 participants had an average age of 41.9 ±9.6 years (range: 22–64 years), and 40% of them were men. Their mean relative oxygen uptake was 34.45 ±8.69 mL/kg/min. The training sample and test sample did not differ significantly with respect to their parameter values (p > 0.05) Table 1. The test–retest reliability of the 3MPKS test, as evaluated in our laboratory, was excellent: the intraclass correlation coefficient (ICC) was 0.88 (95% confidence interval [CI]: 0.77– 0.94), and 60 Taiwanese adults tested 1 week apart participated in this evaluation. In general, good, moderate, and poor reliability levels are indicated by ICC values of >0.75, 0.5–0.75, and <0.5, respectively. According to the correlation matrix, ˙VO2 max had the strongest correlation with BF% among all variables (R = −0.662; training data set, n = 124). In addition, ˙VO2 max was significantly correlated with the heart rate parameters (HR0, HR2, HR3, and HR4), whose data were collected in the step test. ˙VO2 max was most and least correlated with HR4 (R = −0.442) and HR3 (R = −0.289), respectively. Despite the high correlation between ˙VO2 max and the heart rate parameters at different stages, the heart rates of the participants were expected to increase continuously from the first to third minutes of stepping, if performed properly. An individual’s heart rate typically reaches its peak immediately after exercise, and it either decreases at 1 min after exercise or does not decrease at all depending on whether the individual recovers quickly or poorly. Because heart rate is dynamic, to establish a regression model, we used combinations of heart rate parameters and adopted the difference between predicted and measured heart rate data at each stage as inputs (Table 2). The results of our other cross-validation analyses are presented in terms of CE (Constant error) values. The absolute CE values for subgroups stratified by sex and age were <1.00 Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 6/16 Table 2 Correlation between ˙VO2max and features in training dataset (n = 124). ˙VO2max Sex Age BMI BF% HR0 HR1 HR2 HR3 Sex 0.597** Age −0.342** −0.114 BMI −0.083 0.334** −0.160 BF% −0.662** −0.491** 0.109 0.448 HR0 −0.317** −0.242* −0.101 −0.058 0.227* HR1 −0.344** −0.033 −0.283* −0.039 0.274* 0.69** HR2 −0.357** −0.312** −0.093 −0.005 0.308** 0.592** 0.899** HR3 −0.289* −0.254* −0.21* 0 0.248* 0.525** 0.725** 0.8** HR4 −0.442** −0.42** −0.13 −0.063 0.334** 0.564** 0.57** 0.629** 0.702** Notes. BF%, body fat percentage. **Correlation coefficient is significant(p < 0.001). *Correlation coefficient is significant(p < 0.05). for the two models (both in training and testing data sets, n = 124 and 42). Regarding the subgroups stratified by ˙VO2 max, the CE values were negative in low-fitness, middle-fitness subgroups in training data set and low-fitness in testing data set. On the other hand, the CE values were positive in high-fitness in all two data sets (Table 3). Figures 2 and 3 present the Bland–Altman plots produced by the BF% and BMI models based on the testing data set (n = 42). As evident in the plots, the differences between the predicted and measured data were within an acceptable range. The mean error of the BF% model was −0.36 mL/kg/min (95% CI [−12.38–11.98]). For the BMI model, the mean error was 0.4 mL/kg/min (95% CI [−12.35–13.58]). In the BF% and BMI models, the errors for three and two participants, respectively, fell outside the 95% CI. We constructed a model to predict relative oxygen uptake by using multiple regression analysis. The parameters selected for the BF% model were sex, age, BF%, HR0, 1HR3 − HR0, and 1HR3 − HR4; R2 = 0.624 and SEE = 4.982 (training data set, n = 124) (Fig. 4). The parameters selected for the BMI model were sex, age, BMI, initial heart rate, 1HR3 − HR0, and 1HR3 − HR4; R2 = 0.567 and SEE = 5.153 (training data set, n = 124) (Fig. 5). We used BF% as a predictor of body composition; it is more accurate relative to BMI, which is calculated using only height and weight (Table 4). Table 4 presents the cross-validation results for the predicted residual sum of squares (PRESS) statistics (R2p = 0.64 and SEE p = 4.84), which demonstrated minimal shrinkage in the accuracy of the regression model. All regression assumptions were satisfied in our ˙VO2 max prediction models. Specifically, the Kolmogorov–Smirnov test indicated normality in the residuals (p > 0.05). No pattern was determined in the scatter plot between the residuals and predicted ˙VO2 max. Multicollinearity was absent among the predictor variables: the VIF ranges for the BF% and BMI models were 1.09–1.49 and 1.10–1.40, respectively; multicollinearity is absent if VIF ≤ 10 (O’brien, 2007). Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 7/16 Table 3 Measured versus predicted ˙VO2max constant error (CE) and standard deviations (SD) for subgroups of the training dataset and testing dataset. Subgroup n(%) BF% model(%) BMI model(kg m−2) CE SD CE SD Training set(n = 124) Sex Female 80(64.5) −0.01 3.95 0.01 4.45 Male 44(35.5) −0.02 6.23 0.01 5.99 Age <40 years 48(38.7) −0.34 4.72 −0.49 4.94 40–50 years 44(35.5) 0.21 5.35 0.22 5.24 ≥50 years 32(25.8) 0.17 4.46 0.47 4.95 ˙VO2max <29 ml/kg/min 34(27.4) −2.77 3.19 −3.13 3.66 29–38 ml/kg/min 56(45.2) −0.22 4.46 −0.22 4.60 ≥38 ml/kg/min 34(27.4) 3.09 5.18 3.51 4.75 Testing set(n = 42) Sex Female 21(50) −0.15 5.85 −0.89 5.84 Male 21(50) 0.87 6.82 0.08 7.62 Age <43 years 24(57.1) 0.67 6.22 −0.07 6.8 ≥43 years 18(42.9) −0.05 6.55 −0.85 6.79 ˙VO2max <35 ml/kg/min 22(52.4) −2.88 4.95 −3.86 5.31 ≥35 ml/kg/min 20(47.6) 3.93 5.73 3.39 6.11 Figure 2 Bland Altman plot, including limits of agreement, for predicted and measured ˙VO2 max (ml/kg/min) of BF% model by testing dataset (n = 42). Black line mean difference. Dashed line±1.96 ×SD. Full-size DOI: 10.7717/peerj.10831/fig-2 Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 8/16 Figure 3 Bland Altman plot, including limits of agreement, for predicted and measured ˙VO2 max (ml/kg/min) of BMI model by testing dataset (n = 42). Black line mean difference. Dashed line±1.96 ×SD. Full-size DOI: 10.7717/peerj.10831/fig-3 Figure 4 BF% model for testing test (n = 42). Full-size DOI: 10.7717/peerj.10831/fig-4 DISCUSSION This study developed a practical and easy-to-use model for predicting ˙VO2 max in Taiwanese people. We recruited 166 Taiwanese adults and constructed and then evaluated a prediction model. Our results suggest that age, sex, and BF% as well as heart rate during the step test are excellent predictors of ˙VO2 max. We also developed a novel 3MPKS test. Nes et al. (2011) conducted large-scale ˙VO2 max tests on 4,260 participants. They developed a nonexercise model and determined four variables (age, waist circumference, physical activity, and resting heart rate) to be excellent predictors of ˙VO2 max; for their model, R2 was 0.61 and SEE was 5.70 mL/kg/min for men, and R2 was 0.56 and SEE was 5.14 Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 9/16 Figure 5 BMI model for testing (n = 42). Full-size DOI: 10.7717/peerj.10831/fig-5 Table 4 Estimation of ˙VO2 max through multiple regression model (n = 124). BF% model (%) BMI model (kg m−2) ˙VO2max (ml kg−1 min −1) Coefficients β p value Coefficients β p value Constant 72.334 .000 82.387 .000 Sex (0=women, 1=men) 4.366 0.258 .000 9.338 0.551 .000 Age(yr) −0.261 −0.302 .000 −0.327 −0.378 .000 Body composition −0.448 −0.421 .000 −0.718 −0.346 .000 HR0 −0.134 −0.214 .001 −0.171 −0.273 .001 1HR3- HR0 −0.082 −0.136 .041 −0.099 −0.163 .017 1HR3-HR4 0.073 0.124 .048 0.081 0.139 .032 R2 0.624 0.567 SEE 4.982 5.153 SEE% 14.46 14.96 PRESS 2904.186 3107.325 SEEp 4.840 5.006 R2p 0.644 0.619 Notes. BMI, body mass index; BF%, body fat percentage; β, standardized regression weights; SEE, standard error of estimate; SEE%, SEE / mean of measured ˙VO2 max ×100.; PRESS, predicted residual error sum of squares; SEEp, PRESS standard er- ror of estimate; R2p, PRESS squared multiple correlation coefficient. mL/kg/min for women. Jackson et al. (2012) conducted a 27-year study that examined the ˙VO2 max of 11,365 people and used variables such as age, sex, BMI, waist circumference, resting heart rate, physical activity, and smoking habits to estimate CRF; for their model, R was 0.78–0.81 and SEE was 5.3–5.6 mL/kg/min. Although the nonexercise model is Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 10/16 an excellent predictor of ˙VO2 max, its SEE is generally higher than those of submaximal exercise models; compared with nonexercise models, our developed BF% model had better predictive performance and a lower standard error of estimate (R2 = 0.624 and SEE = 4.982). Abut, Akay & George (2016) reported that (1) when perceived functional ability (PFA) was used as the sole predictor of ˙VO2 max, an R value of 0.73 and a higher RMSE of 6.08 mL/kg/min could be obtained; (2) when submaximal ending speed (SM-ES) of a treadmill was used as the sole predictor, the R value increased to 0.82 and the RMSE was relatively low at 4.99 mL/kg/min; and (3) when both PFA and SM-ES were used as predictors, the R value was 0.89 and RMSE was 4.14 mL/kg/min. These findings indicate that predicted values of ˙VO2 max that are based only on participant self-reports are likely to deviate from their measured values. Although predictive performance is ostensibly improved when motion is added to the prediction model, the cost of exercise tests due to the use of this method restricts its application in large-scale tests. Several studies have developed simple models involving submaximal motion. Lee et al. (2019) investigated 568 adults and used sex, age, height, and weight and inverse recovery heart rate during a YMCA step test to predict ˙VO2 max; for their model, R was 0.78 and SEE was 4.74 mL/kg/min. The duration of their exercise test plus recovery time was only 4 min, and they used exercise-induced heart rate as a predictor; their results are similar to ours. Their study provided a simple and practical method for simultaneously estimating CRF in many Korean adults. Cao et al. (2013) used age, sex, and physical composition as well as stepping distance over a 3-min period to develop a set of prediction methods. They determined that BF% (a measure of body composition) was a better predictor than BMI (R2 = 0.83 vs. 0.80, SEE = 4.565 vs. 5.037 mL/kg/min). In contrast to our method, their method has the considerable advantages of a shorter testing time of 3 min and the fact that participants need not wear a heart rate monitor. However, their test is limited by its need for a 20-m open space. Similarly, we found that sex, age, and BF% as well as heart rate during the 3MPKS test yielded the best prediction performance (R = 0.79, SEE = 4.982 mL/kg/min). Because BMI is based on only height and weight and may not accurately represent the body characteristics of participants, BMI is a less accurate predictor than BF%. Most submaximal exercise models proposed by previous studies involve a fixed-height step test. However, the height and leg length of participants when standing may affect their physiological response in the step test (Culpepper & Francis, 1987). Relative to their European counterparts, Asian adults have shorter heights and leg lengths when standing (Stanfield et al., 2012). Therefore, differences in heart rate and oxygen consumption potentially affect the model’s prediction. The 3MPKS test employs the knee-ups and step test to measure the physical fitness and cardiopulmonary endurance of older adults (Rikli & Jones, 2001). In the test, participants must execute tasks at various knee heights based on their thigh length, and individualized exercise testing goals are provided. Moreover, most field tests involve average speed tests, such as step tests and running. In running tests specifically, if the distance is used as the capacity index but the speed or frequency of exercise is not progressively increased, participants may exercise intensely at the beginning of the test (i.e., run at a higher speed). However, due to the lack of appropriate speed Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 11/16 allocation, decremental loading occurs in participants as their physical strength decreases. The difficulty of diagnosing potential heart diseases in advance increases the risk of sudden death during running tests. To the best of our knowledge, research has not been conducted on the ethics of running tests. Most previous studies have investigated the rate of sudden death among athletes in long-distance competitions. However, cases of sudden cardiac death occur frequently worldwide during running tests, and the principle that physical activities ought to be progressive must be adhered to in physical fitness tests. Our research method used body composition and heart rate as variables. The advantages of the 3MPKS test are that it does not require a step-up box and is not subject to venue restrictions. These make the 3MPKS test accord with the principle that physical activities ought to be progressive, thus making it safer. Considering the immediacy of heart rate measurement and that of confounding factors, we used a chest-worn heart rate monitor in the experiment. Although the requirement of heart rate monitoring constitutes a disadvantage for the 3MPKS test, it is ameliorated by the prevalence of low-cost wearable devices. More comfortable than the chest-worn heart rate belt, products that combine running clothes with heart rate belts have also appeared on the market. Research has also suggested a high correlation between the heart rate measurements of various types of optical devices and chest-worn heart rate belts (Stahl et al., 2016). Therefore, when conducting a large-scale cardiorespiratory general test, the use of easily wearable optical heart rate monitors can be considered. The whole-range monitoring of heart rate can also considerably improve test safety in a field study. Notably, through whole-range monitoring, we found that one research participant was likely to have an unknown heart disease. We then terminated the experiment for the participant and recommended that the participant seek medical treatment. This example illustrates a side benefit of CRF tests. In our research model, heart rate during stepping at each stage was used as the main variable. Therefore, the test may be unsuitable for individuals who have psychological sensitivity or dysautonomia or who are taking medication. Furthermore, because our participants were adults between 20 and 64 years old, it was unclear whether our 3MKPS test is appropriate as a physical fitness and cardiorespiratory test for students (7–23 years old) and older adults (≥65 years old). Future research must include samples with greater diversity in age and ethnicity to assess whether our 3MKPS test can be applied to the wider global population. CONCLUSION This study, involving Taiwanese adults, constructed and verified a model for predicting ˙VO2 max, which is used to measure CRF. This model comprises the predictors sex, age, and body composition as well as heart rate changes during a step test. Our 3MKPS test has three advantages: it has a short testing time of 4 min, it has no venue limitations, and it does not require a step box. Furthermore, measurements can be taken for many participants simultaneously by asking them to wear a heart rate monitor and move according to a beat. Our model can also be applied to large-scale epidemiological research. In future Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 12/16 applications, the model can be combined with smartwatches or used to develop health and well-being apps, helping users to track their ˙VO2 max. Future research can further explore the correlation between various diseases and ˙VO2 max, as predicted using our simple and reliable method for measuring CRF. ADDITIONAL INFORMATION AND DECLARATIONS Funding This research was supported by Research Grants from Taipei Medical University (no. TMU105-AE1-B06) and the Sports Administration, Ministry of Education, R.O.C. for the Comprehensive Research for the Industrial Technology Research Institute’s Technology Fitness Program (no. J4653H1A20). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Grant Disclosures The following grant information was disclosed by the authors: Research Grants from Taipei Medical University: no. TMU105-AE1-B06. Sports Administration, Ministry of Education, R.O.C. for the Comprehensive Research for the Industrial Technology Research Institute’s Technology Fitness Program: no. J4653H1A20. Competing Interests The authors declare there are no competing interests. Author Contributions • Yu-Chun Chung conceived and designed the experiments, performed the experiments, analyzed the data, prepared figures and/or tables, authored or reviewed drafts of the paper, and approved the final draft. • Ching-Yu Huang conceived and designed the experiments, analyzed the data, authored or reviewed drafts of the paper, and approved the final draft. • Huey-June Wu conceived and designed the experiments, performed the experiments, authored or reviewed drafts of the paper, and approved the final draft. • Nai-Wen Kan and Chi-Chang Huang performed the experiments, authored or reviewed drafts of the paper, and approved the final draft. • Chin-Shan Ho performed the experiments, analyzed the data, authored or reviewed drafts of the paper, and approved the final draft. • Hung-Ting Chen conceived and designed the experiments, analyzed the data, prepared figures and/or tables, authored or reviewed drafts of the paper, and approved the final draft. Human Ethics The following information was supplied relating to ethical approvals (i.e., approving body and any reference numbers): Institutional Review Boards (IRBs) of the Industrial Technology Research Institute and of Taipei Medical University (IRB No: N201808055). Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 13/16 Data Availability The following information was supplied regarding data availability: Raw measurements are available in the Supplemental Files. Supplemental Information Supplemental information for this article can be found online at http://dx.doi.org/10.7717/ peerj.10831#supplemental-information. REFERENCES Abut F, Akay MF, George J. 2016. Developing new VO2 max prediction models from maximal, submaximal and questionnaire variables using support vector machines combined with feature selection. Computers in Biology and Medicine 79:182–192 DOI 10.1016/j.compbiomed.2016.10.018. American College of Sports Medicine. 2009. ACSM’s guidelines for exercise testing and prescription. Philadelphia: Lippincott Williams & Wilkins. Beutner F, Ubrich R, Zachariae S, Engel C, Sandri M, Teren A, Gielen S. 2015. Valida- tion of a brief step-test protocol for estimation of peak oxygen uptake. European Journal of Preventive Cardiology 22:503–512 DOI 10.1177/2047487314533216. Blair SN, Kohl HW, Paffenbarger RS, Clark DG, Cooper KH, Gibbons LW. 1989. Physical fitness and all-cause mortality: a prospective study of healthy men and women. Journal of the American Medical Association 262:2395–2401 DOI 10.1001/jama.262.17.2395. Buckley JP, Sim J, Eston RG, Hession R, Fox R. 2004. Reliability and validity of measures taken during the Chester step test to predict aerobic power and to prescribe aerobic exercise. British Journal of Sports Medicine 38:197–205 DOI 10.1136/bjsm.2003.005389. Cao ZB, Miyatake N, Aoyama T, Higuchi M, Tabata I. 2013. Prediction of maximal oxygen uptake from a 3-minute walk based on gender, age, and body composition. Journal of Physical Activity and Health 10:280–287 DOI 10.1123/jpah.10.2.280. Chang SC, Lin JC. 1995. The validity generalization of step test as a measure of the maximal oxygen intake. Physical Education Journal 20:351–362 DOI 10.6222/pej.0020.199512.403. Culpepper MI, Francis KT. 1987. An anatomical model to determine step height in step testing for estimating aerobic capacity. Journal of Theoretical Biology 129:1–8 DOI 10.1016/S0022-5193. Grant JA, Joseph AN, Campagna PD. 1999. The prediction of VO2max: a comparison of 7 indirect tests of aerobic power. Journal of Strength and Conditioning Research 13:346–352. Jackson AS, Sui X, O’Connor DP, Church TS, Lee DC, Artero EG, Blair SN. 2012. Longitudinal cardiorespiratory fitness algorithms for clinical settings. American Journal of Preventive Medicine 43:512–519 DOI 10.1016/j.amepre.2012.06.032. Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 14/16 Jiménez-Pavón D, Lavie CJ, Blair SN. 2019. The role of cardiorespiratory fitness on the risk of sudden cardiac death at the population level: a systematic review and meta- analysis of the available evidence. Progress in Cardiovascular Diseases 62:279–287 DOI 10.1016/j.pcad.2019.05.003. Kaminsky LA, Arena R, Beckie TM, Brubaker PH, Church TS, Forman DE, Franklin BA, Gulati M, Lavie CJ, Myers J, Patel MJ, Piña IL, Weintraub WS, Williams MA. 2013. The importance of cardiorespiratory fitness in the United States: the need for a national registry: a policy statement from the American Heart Association. Circulation 127:652–662 DOI 10.1161/CIR.0b013e31827ee100. Kodama S, Saito K, Tanaka S, Maki M, Yachi Y, Asumi M, Sugawara A, Totsuka K, Shimano H, Ohashi Y, Yamada N, Sone H. 2009. Cardiorespiratory fitness as a quantitative predictor of all-cause mortality and cardiovascular events in healthy men and women: a meta-analysis. Journal of the American Medical Association 301:2024–2035 DOI 10.1001/jama.2009.681. Laukkanen JA, Kurl S, Salonen R, Rauramaa R, Salonen JT. 2004. The predictive value of cardiorespiratory fitness for cardiovascular events in men with various risk profiles: a prospective population-based cohort study. European Heart Journal 25:1428–1437 DOI 10.1016/j.ehj.2004.06.013. Lee O, Lee S, Kang M, Mun J, Chung J. 2019. Prediction of maximal oxygen con- sumption using the Young Men’s Christian Association-step test in Korean adults. European Journal of Applied Physiology and Occupational Physiology 119:1245–1252 DOI 10.1007/s00421-019-04115-8. Mazic S, Zivotic-Vanovic M, Igracki I, Zivanic S, Velkovski S. 2001. A simple and reliable step-test for indirect evaluation of aerobic capacity. Medicinski Pregled 54:522–529 DOI 10.2478/humo-2013-000. McLester CN, Nickerson BS, Kliszczewicz BM, McLester JR. 2020. Reliability and agree- ment of various InBody body composition analyzers as compared to dual-energy X-ray absorptiometry in healthy men and women. Journal of Clinical Densitometry 23:443–450 DOI 10.1016/j.jocd.2018.10.008. Nes BM, Janszky I, Vatten LJ, Nilsen TIL, Aspenes ST, Wisløff U. 2011. Estimating VO 2peak from a nonexercise prediction model: the HUNT study, Norway. Medicine & Science in Sports & Exercise 43:2024–2030 DOI 10.1249/MSS.0b013e31821d3f6f. O’brien RM. 2007. A caution regarding rules of thumb for variance inflation factors. Quality and Quantity 41:673–690 DOI 10.1007/s11135-006-9018-6. Rikli RE, Jones CJ. 2001. Senior fitness test manual. Champaign: Human Kinetics. Ross R, Blair SN, Arena R, Church TS, Després JP, Franklin BA, Haskell WL, Kaminsky LA, Levine BD, Lavie CJ, Myers J, Niebauer J, Sallis R, Sawada SS, Sui X, Wisløff U. 2016. Importance of assessing cardiorespiratory fitness in clinical practice: a case for fitness as a clinical vital sign: a scientific statement from the American Heart Association. Circulation 134:e653–e699 DOI 10.1161/CIR.0000000000000461. Sawada SS, Lee IM, Naito H, Kakigi R, Goto S, Kanazawa M, Okamoto T, Tsukamoto K, Muto T, Tanaka H. 2014. Cardiorespiratory fitness, body mass index, and Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 15/16 cancer mortality: a cohort study of Japanese men. BMC Public Health 14:1012 DOI 10.1186/1471-2458-14-1012. Stahl SE, An HS, Dinkel DM, Noble JM, Lee JM. 2016. How accurate are the wrist- based heart rate monitors during walking and running activities? Are they accurate enough? BMJ Open Sport & Exercise Medicine 2:e000106 DOI 10.1136/bmjsem-2015-000106. Stanfield KM, Wells JC, Fewtrell MS, Frost C. 2012. Differences in body com- position between infants of South Asian and European ancestry: the London Mother and Baby study. International Journal of Epidemiology 41:1409–1418 DOI 10.1093/ije/dys139. Su MY, Lin CC, Hsieh SY. 2006. The validity and reliability of a step test with adjusted heights according to leg lengths. Sports & Exercise Research 8:87–94 DOI 10.5297/ser.200606_8(2).0007. Sui X, LaMonte MJ, Blair SN. 2007. Cardiorespiratory fitness as a predictor of nonfatal cardiovascular events in asymptomatic women and men. American Journal of Epidemiology 165:1413–1423 DOI 10.1093/aje/kwm031. Yoopat P, Vanwonterghem K, Louhevaara V. 2002. Evaluation of a step-test for assessing the cardiorespiratory capacity of workers in Thailand: a pilot study. Journal of Human Ergology 31:33–40 DOI 10.1002/car.2491. Chung et al. (2021), PeerJ, DOI 10.7717/peerj.10831 16/16

Predicting maximal oxygen uptake from a 3-minute progressive knee-ups and step test.

03-15-2021

Chung, Yu-Chun,Huang, Ching-Yu,Wu, Huey-June,Kan, Nai-Wen,Ho, Chin-Shan,Huang, Chi-Chang,Chen, Hung-Ting

eng

PMC7379642

Supplement Table 1. Internal drop-out analyses within the available database, comparing individuals with (included) and without (excluded) VO2max. Included Excluded Included Excluded Included Excluded Included Excluded Included Excluded Year n % women n % women Mean (SD) Mean (SD) Mean (SE)# Mean (SE)# Mean (SE)# Mean (SE)# % % 1995-1997 4 574 52.4 815 41.6a 40.9 (10.0) 43.3 (10.6)a 172.6 (9.3) 173.8 (9.7)a 74.8 (14.0) 78.2 (14.8)a 13.7 12.3 1998-1999 6 543 45.3 1 577 44.6 42.0 (10.3) 44.1 (10.8)a 173.9 (9.1) 174.0 (9.4) 76.1 (14.0) 77.9 (15.2)a 19.1 19.5 2000-2001 12 545 49.5 3 370 48.4 42.3 (10.7) 45.4 (11.2)a 173.3 (9.2) 173.3 (9.5) 76.0 (14.2) 79.1 (15.6)a 20.8 19.4 2002-2003 22 629 52.4 5 167 53.3 41.9 (11.2) 47.3 (11.1)a 172.8 (9.2) 172.4 (9.6)a 75.4 (14.4) 79.0 (16.4)a 19.9 17.7a 2004-2005 37 420 52.1 9 704 48.7a 42.8 (10.9) 47.5 (11.1)a 173.0 (9.2) 173.3 (9.5)a 75.9 (14.6) 80.0 (16.4)a 25.3 20.5a 2006-2007 38 519 48.6 9 703 46.0a 42.9 (11.1) 47.1 (11.4)a 173.5 (9.3) 173.6 (9.7) 77.1 (15.0) 80.5 (17.0)a 24.5 20.3a 2008-2009 43 479 46.2 9 672 47.8a 42.8 (11.3) 46.7 (12.0)a 173.9 (9.3) 173.3 (9.7)a 77.9 (15.3) 80.6 (17.9)a 25.9 22.5a 2010-2011 39 177 44.2 7 758 45.6a 42.4 (11.2) 46.0 (12.0)a 174.3 (9.4) 173.8 (9.7)a 78.7 (15.5) 82.0 (18.1)a 27.3 22.9a 2012-2013 57 246 40.8 12 135 40.6 42.1 (11.3) 46.2 (11.7)a 174.9 (9.3) 174.9 (9.7) 79.0 (15.4) 82.8 (17.9)a 31.5 26.2a 2014-2015 55 584 37.6 13 070 38.2 41.7 (11.5) 45.0 (12.3)a 175.3 (9.2) 175.2 (9.8) 79.9 (15.7) 83.2 (18.6)a 29.7 24.6a 2016-2017 36 561 36.8 8 878 37.9 41.1 (11.7) 43.9 (12.4)a 175.4 (9.3) 175.6 (9.6) 80.3 (16.0) 83.3 (18.4)a 29.5 26.5a Total 354 277 44.2 81 849 44.0 42.2 (11.2) 46.0 (11.8)a 174.2 (9.3) 174.1 (9.7)a 78.1 (15.3) 81.4 (17.4)a 26.9 22.8a a Different from individuals with VO2max during the same time period (p>0.05) using chi2-test for percentages, independent t-test for age and general linear modelling for height and weight. # Mean and SE values adjusted for age and gender SD, standard deviation SE, standard error Gender Age Height Weight >12 years of education

Decline in cardiorespiratory fitness in the Swedish working force between 1995 and 2017.

11-15-2018

Ekblom-Bak, Elin,Ekblom, Örjan,Andersson, Gunnar,Wallin, Peter,Söderling, Jonas,Hemmingsson, Erik,Ekblom, Björn

eng

PMC6766792

sensors Article A Sensor Platform for Athletes’ Training Supervision: A Proof of Concept Study Alessandro Zompanti 1,* , Anna Sabatini 1 , Marco Santonico 1, Simone Grasso 1 , Antonio Gianfelici 2, Bruno Donatucci 2, Andrea Di Castro 2 and Giorgio Pennazza 1 1 Unit of Electronics for Sensor Systems, Department of Engineering Campus Bio-Medico University of Rome, 00128 Rome, Italy; [email protected] (A.S.); [email protected] (M.S.); [email protected] (S.G.); [email protected] (G.P.) 2 Sport Medicine and Science Institute, CONI (Comitato Olimpico Nazionale Italiano), 00197 Rome, Italy; [email protected] (A.G.); [email protected] (B.D.); [email protected] (A.D.C.) * Correspondence: [email protected]; Tel.: +39-062-2541-9610 Received: 4 July 2019; Accepted: 6 September 2019; Published: 12 September 2019



 Abstract: One of the basic needs of professional athletes is the real-time and non-invasive monitoring of their activities. The use of these kind of data is necessary to develop strategies for specific tailored training in order to improve performances. The sensor system presented in this work has the aim to adopt a novel approach for the monitoring of physiological parameters, and athletes’ performances, during their training. The anaerobic threshold is herein identified with the monitoring of the lactate concentration and the respiratory parameters. The data collected by the sensor are used to build a model using a supervised method (based on the partial least squares method, PLS) to predict the values of the parameters of interest. The sensor is able to measure the lactate concentration from a sample of saliva and it can estimate a respiratory parameter, such as maximal oxygen consumption, maximal carbon dioxide production and respiratory rate from a sample of exhaled breath. The main advantages of the device are the low power; the wireless communication; and the non-invasive sampling method, which allow its use in a real context of sport practice. Keywords: athletes’ performances; low power sensor; wireless system; lactate 1. Introduction A gas sensor array for volatile fingerprinting (dubbed e-nose) and e-tongue systems (based on cyclic voltammetry analysis) are used in several applications. E-nose is applied in several fields: from the medical practice for the early detection of respiratory diseases [1], to the evaluation of food freshness [2] and the monitoring of the air quality [3]. The e-tongue systems are used for voltammetric analysis in the discrimination of the quality of olive oil [4], wine [5], water [6] and also physiological liquids [7]. Those systems have not been used, so far, for the evaluation of sport activities. In sport activities the monitoring of the physiological parameters allows one to evaluate the physical condition of the athletes. The current state of art of this research field does not follow a unique guideline, but several approaches are pursued and different theories are reported to support the effectiveness of each parameter proposed for athletes’ monitoring [8]. The accepted and available approaches can be divided into two groups: the first one based on a self-report measurement of the physical effort (e.g., Borg Scale), and the second one that consists of the objective monitoring of certain physiological parameters. In the first approach the athlete has to score his physical effort using a standard scale; the second method is performed via specific measurement instruments. These two approaches can also be jointly applied. In the application of Sensors 2019, 19, 3948; doi:10.3390/s19183948 www.mdpi.com/journal/sensors Sensors 2019, 19, 3948 2 of 9 the second approach, the anaerobic threshold is often used. It is a well-known parameter for the assessment of sportive performances, defined by Mader as the 4 millimoles per litre concentration of lactate [9]. There is also evidence about the use of other phisiological parameters, such as heart rate [10], respiratory rate [11] and other respiratory parameters for the evaluation of the physical effort. However, in spite of the lack of a standard and shared methodology, lactate concentration is the most used [12]. This fact is also evident by the availability of many different instruments for the measurement of lactate in the blood extracted by the ear lobe [12]. A key-point in the application of lactate evaluation is represented by the invasivity of the technique, which is based on blood analysis. Here, the ability of a sensor device to estimate the lactate concentration via the analysis of a saliva sample, which, of course, involves a non invasive collection procedure, is demonstrated. A step up in the field should be given by the simultaneuos application of non invasive and easy-to-use sensors able to implement the measurements of many other physiological parameters. Feasibility of this approach is granted by the design and development of dedicated electronic interfaces for sensors’ interactions as a collaborative sensor system, via low-power strategies and low-noise circuits for signal acquisition and treatment. The aims of this work are: (1) the demonstration of the feasibility of a method to estimate lactate concentration in humans by the analysis of saliva, with a proof of concenpt study showing the effectiveness of the saliva sampling procedure and the relevance of the instrumental output in the estimation of the lactate concentration; (2) the feasibility of a system approach based on more than one sensors in order to enlarge the system’s accuracy and effectiveness in the monitoring of athletes’ training. 2. Materials and Methods During the experiment, samples of saliva and of exhaled breath were collected. Saliva was analyzed using the sensor array for liquid analysis of a sensor system named BIONOTE [13]. Exhaled breath has been measured using a sensor array for volatile analysis of the BIONOTE system. Both these systems, named BIONOTE-L and BIONOTE-V respectively, are briefly described below. Further details can be found in the references. The BIONOTE-V used in this set-up is based on 20 MHz quartz microbalances, the module is composed of 7 QMBs (Quartz MicroBalance) covered by anthocyanins extracted from different plant tissues, such as blue orthesia, red cabbage and red rose [13]. The breath is stored in a Tenax cartridge [14], after collection performed via Pneumopipe® [15] and analyzed with a desorbing process run consecutively at 50, 100, 150 and 200 ◦C. Between the desorbing steps, a flow of nitrogen is distributed inside the measurement chamber. The final output for each measurement is an array of 28 numbers. These 28 numbers are the frequency shifts registered by each of the 7 QMBs for the four temperatures of desorption. More technical information about the most relevant parameters of these acoustic sensors can be found in literature [13,16,17]. The BIONOTE-L performs voltammetric analysis on the saliva samples, using 3 screen-printed electrodes made of gold (working electrode), silver (reference electrode) and platinum (counter electrode): the system has a low power, about 80 mW, 3.6 V, 22 mA and a size of about 12 cm2. These features, the order of magnitude of the power and size, allow the use of the system in wearable applications. In the case of BIONOTE-L, Data, acquired with a microcontroller, are transmitted using a BLE (Bluetooth Low Energy) module to an external device, like a smartphone or tablet. The input signal is a triangular wave ranging between −1 and 1 V with a frequency of 0.01 Hz. That signal is generated by the microcontroller and the electrode is driven by a custom electronic interface. In Figure 1 the structure of the sensor used in this study is reported [18]. When the input signal is applied to the solution, a redox reaction is induced and the output signal is measured in the working electrode. A current flows from the counter electrode to the working electrode and it is converted to a voltage value using a trans-impedance stage. The output of the sensor is an array of 500 numbers. Sensors 2019, 19, 3948 3 of 9 Sensors 2019, 19, x FOR PEER REVIEW 3 of 9 Figure 1. Device used in the experimental set-up: (a) BIONOTE-L (b) BIONOTE-V. The samples of saliva and of exhaled breath were collected during the performing of a modified Mader test. Twelve male triathlon athletes, trained in running, swimming and cycling, performed a running test on a treadmill designed for the evaluation of the athletes’ maximum effort. The study was in collaboration with the Italian National Olympic Committee (Comitato Olimpico Nazionale Italiano) Sport Medicine and Science Institute, CONI, Rome. The modified test is composed of 5 minutes running at different increasing speeds (12, 13.5, 15 and 16.5 km/h) and each step is separated from the next one by 2 minutes’ pause. During the pauses saliva samples are collected using a sampling device called Salivette® [19]. The athletes were requested to gently chew the cotton roll for one minute to stimulate saliva production and then to put it in the plastic tube. The extraction process was performed with the centrifugation of the sampler for 2 minutes at 1000 RCF and room temperature. The extracted liquid volume of 0.5 mL could be frozen to allow post measurement, or directly diluted in deionized water and analyzed: standard disposable spectrometry cuvettes, with a volume of 4 mL, were used; in order to ensure that the electrodes were correctly dipped into the liquid solution, saliva samples were diluted into the minimum required volume of 3.75 mL. Before the analysis of each sample of saliva, a white sample of deionized water was measured. The exhaled breath was collected at the beginning and at the end of the test and it required the athletes to breathe normally for 3 minutes into the Pneumopipe [16]. The cartridge was stored at 4 °C until the analysis with the BIONOTE-V (Figure 2). Output data have been used for the elaboration of the models to predict several physiological parameters (blood lactate concentration, maximal oxygen production, maximal carbon dioxide production, respiratory ratio, etc.). Both supervised and unsupervised analyses were performed, specifically PCA (principal component analysis) and PLS-DA (partial least squares discriminant analysis), using the PLS Toolbox in the MATLAB environment. PCA and PLS-DA are two data analysis techniques normally used to treat and elaborate multidimensional data-sets. They represent Figure 1. Device used in the experimental set-up: (a) BIONOTE-L (b) BIONOTE-V. The samples of saliva and of exhaled breath were collected during the performing of a modified Mader test. Twelve male triathlon athletes, trained in running, swimming and cycling, performed a running test on a treadmill designed for the evaluation of the athletes’ maximum effort. The study was in collaboration with the Italian National Olympic Committee (Comitato Olimpico Nazionale Italiano) Sport Medicine and Science Institute, CONI, Rome. The modified test is composed of 5 minutes running at different increasing speeds (12, 13.5, 15 and 16.5 km/h) and each step is separated from the next one by 2 minutes’ pause. During the pauses saliva samples are collected using a sampling device called Salivette® [19]. The athletes were requested to gently chew the cotton roll for one minute to stimulate saliva production and then to put it in the plastic tube. The extraction process was performed with the centrifugation of the sampler for 2 minutes at 1000 RCF and room temperature. The extracted liquid volume of 0.5 mL could be frozen to allow post measurement, or directly diluted in deionized water and analyzed: standard disposable spectrometry cuvettes, with a volume of 4 mL, were used; in order to ensure that the electrodes were correctly dipped into the liquid solution, saliva samples were diluted into the minimum required volume of 3.75 mL. Before the analysis of each sample of saliva, a white sample of deionized water was measured. The exhaled breath was collected at the beginning and at the end of the test and it required the athletes to breathe normally for 3 minutes into the Pneumopipe [16]. The cartridge was stored at 4 ◦C until the analysis with the BIONOTE-V (Figure 2). Output data have been used for the elaboration of the models to predict several physiological parameters (blood lactate concentration, maximal oxygen production, maximal carbon dioxide production, respiratory ratio, etc.). Both supervised and unsupervised analyses were performed, specifically PCA (principal component analysis) and PLS-DA (partial least squares discriminant analysis), using the PLS Toolbox in the MATLAB environment. PCA and PLS-DA are two data analysis Sensors 2019, 19, 3948 4 of 9 techniques normally used to treat and elaborate multidimensional data-sets. They represent an optimal solution to extract relevant information by the output of arrays of chemical sensors [20], by reducing the number of variables via a linear combination of the original parameters in the direction able to maximize information. The model calibration was performed using measurements collected during the execution of the modified Mader test. Specifically the physiological values measured are: VO2—oxygen uptake; VCO2—carbon dioxide production; Pet O2—oxygen partial pressure; Pet CO2—carbon dioxide partial pressure; FR—respiratory rate; RQ—respiratory ratio (produced CO2/O2 uptake); VT—tidal volume; Ti—duration of inspiration; and Te—duration of expiration. These parameters were real-time monitored, during the whole duration of the test, using a Quark CPET platform (COSMED srl, [21]): The Quark CPET is a state-of-the-art metabolic cart for gas exchange analysis (VO2, VCO2) either during exercise testing or resting protocols. The platform is equipped with a paramagnetic sensor for the O2, and an infrared sensor for the CO2. The athletes have to breathe into a sensor-embedded mask connected to a workstation that is able to record and evaluate data (Figure 2). Several PLS models were built analyzing the data obtained from exhaled breath samples and saliva samples. Two models were built using data obtained analyzing saliva samples with BIONOTE-L: the models were built using only the output data obtained from the BIONOTE-L; therefore, a 500-column matrix was used. The model was cross-validated using the leave-one-out method. The first model was built to predict lactate concentration in the whole range of variation for the parameter (from 0 to 10 mmol/L); the second model was built to measure the lactate in the range of 2-6 mmol/L. The same goals were achieved with two models built using a data-set obtained from the data fusion of data collected measuring both saliva samples (using BIONOTE-L) and exhaled breath samples (using BIONOTE-V), thus a 528 column matrix was used. A linear normalization was performed on the dataset, but no feature selection was performed. All the models were cross-validated using the leave-one-out method. The lactate values were measured via an earlobe blood sample collected by a doctor during the pauses. Sensors 2019, 19, x FOR PEER REVIEW 4 of 9 an optimal solution to extract relevant information by the output of arrays of chemical sensors [20], by reducing the number of variables via a linear combination of the original parameters in the direction able to maximize information. The model calibration was performed using measurements collected during the execution of the modified Mader test. Specifically the physiological values measured are: VO2—oxygen uptake; VCO2—carbon dioxide production; Pet O2—oxygen partial pressure; Pet CO2—carbon dioxide partial pressure; FR—respiratory rate; RQ—respiratory ratio (produced CO2/O2 uptake); VT—tidal volume; Ti—duration of inspiration; and Te—duration of expiration. These parameters were real-time monitored, during the whole duration of the test, using a Quark CPET platform (COSMED srl, [21]): The Quark CPET is a state-of-the-art metabolic cart for gas exchange analysis (VO2, VCO2) either during exercise testing or resting protocols. The platform is equipped with a paramagnetic sensor for the O2, and an infrared sensor for the CO2. The athletes have to breathe into a sensor-embedded mask connected to a workstation that is able to record and evaluate data (Figure 2). Several PLS models were built analyzing the data obtained from exhaled breath samples and saliva samples. Two models were built using data obtained analyzing saliva samples with BIONOTE-L: the models were built using only the output data obtained from the BIONOTE-L; therefore, a 500-column matrix was used. The model was cross-validated using the leave-one-out method. The first model was built to predict lactate concentration in the whole range of variation for the parameter (from 0 to 10 mmol/L); the second model was built to measure the lactate in the range of 2-6 mmol/L. The same goals were achieved with two models built using a data-set obtained from the data fusion of data collected measuring both saliva samples (using BIONOTE-L) and exhaled breath samples (using BIONOTE-V), thus a 528 column matrix was used. A linear normalization was performed on the dataset, but no feature selection was performed. All the models were cross-validated using the leave- one-out method. The lactate values were measured via an earlobe blood sample collected by a doctor during the pauses. Figure 2. The athlete is running on a treadmill; Respiratory parameters and heart rate are real-time monitored. During the pause between the steps, lactate from blood and saliva samples are evaluated; at the beginning and at the end of the test as well. An exhaled breath sample is also collected. 3. Results Data collected from both BIONOTE-L and BIONOTE-V were analyzed using an unsupervised method, principal component analysis (PCA), in order to evaluate the method’s effectiveness. Then, a supervised analysis, the partial least squares discriminant analysis (PLS-DA), was used to build predictive models for the determination of saliva lactate levels and respiratory parameters. During the calibration process, measured blood lactate concentrations and respiratory parameters were used to build the models. Figure 2. The athlete is running on a treadmill; Respiratory parameters and heart rate are real-time monitored. During the pause between the steps, lactate from blood and saliva samples are evaluated; at the beginning and at the end of the test as well. An exhaled breath sample is also collected. 3. Results Data collected from both BIONOTE-L and BIONOTE-V were analyzed using an unsupervised method, principal component analysis (PCA), in order to evaluate the method’s effectiveness. Then, a supervised analysis, the partial least squares discriminant analysis (PLS-DA), was used to build predictive models for the determination of saliva lactate levels and respiratory parameters. During the calibration process, measured blood lactate concentrations and respiratory parameters were used to build the models. Sensors 2019, 19, 3948 5 of 9 3.1. PCA Analysis Data obtained from BIONOTE-L, measuring saliva samples and white samples of deionized water, were analyzed using PCA. The scope of this elaboration was to check if the procedure for saliva collection and treatment is effective: is the fingerprint of a saliva sample obtained by its measurement with the BIONOTE-L different from another (similar) standard solution? If yes, as the PCA model demonstrates, it could be possible to distinguish among different saliva samples. Figure 3 shows the results: on the plane representing the score-plot of the first two principal components of the calculated model, the information is mainly contained in PC1 (90.6% of the information); two clusters can be distinguished, one for the saliva samples and the other for the white samples. Sensors 2019, 19, x FOR PEER REVIEW 5 of 9 3.1. PCA Analysis Data obtained from BIONOTE-L, measuring saliva samples and white samples of deionized water, were analyzed using PCA. The scope of this elaboration was to check if the procedure for saliva collection and treatment is effective: is the fingerprint of a saliva sample obtained by its measurement with the BIONOTE-L different from another (similar) standard solution? If yes, as the PCA model demonstrates, it could be possible to distinguish among different saliva samples. Figure 3 shows the results: on the plane representing the score-plot of the first two principal components of the calculated model, the information is mainly contained in PC1 (90.6% of the information); two clusters can be distinguished, one for the saliva samples and the other for the white samples. Figure 3. Principal component analysis (PCA) analysis of saliva and white samples are evaluated using a PCA method. Here the plot score of the first two principal component is reported. It contains almost the 97% of the explained variance. Here, two clusters can be distinguished: the blue one is composed of white samples while the orange cluster is made of saliva samples. Thus, PCA analysis shows the effectiveness of the sampling method and the ability of the BIONOTE-L to discriminate saliva samples from white samples. 3.2. PLS-DA Analysis Several PLS models were built analyzing the data obtained from exhaled breath samples and saliva samples. Two models were built using data obtained analyzing saliva samples with BIONOTE-L; their performances are reported in Table 1. The first model was built to predict lactate concentration in the whole range of variation of the parameter (from 0 to 10 mmol/L); it presents a RMSECV (root mean square in cross validation) error of 1.94 mmol/L. The second model was built to measure the lactate in the range of 2–6 mmol/L; this range is critical for the evaluation of the anaerobic threshold in order to improve the athletes’ performances. In this case, the RMSECV error decreases at 0.66 mmol/L. The score plots of the models are reported respectively in Figure 4a,b. Figure 3. Principal component analysis (PCA) analysis of saliva and white samples are evaluated using a PCA method. Here the plot score of the first two principal component is reported. It contains almost the 97% of the explained variance. Here, two clusters can be distinguished: the blue one is composed of white samples while the orange cluster is made of saliva samples. Thus, PCA analysis shows the effectiveness of the sampling method and the ability of the BIONOTE-L to discriminate saliva samples from white samples. 3.2. PLS-DA Analysis Several PLS models were built analyzing the data obtained from exhaled breath samples and saliva samples. Two models were built using data obtained analyzing saliva samples with BIONOTE-L; their performances are reported in Table 1. The first model was built to predict lactate concentration in the whole range of variation of the parameter (from 0 to 10 mmol/L); it presents a RMSECV (root mean square in cross validation) error of 1.94 mmol/L. The second model was built to measure the lactate in the range of 2–6 mmol/L; this range is critical for the evaluation of the anaerobic threshold in order to improve the athletes’ performances. In this case, the RMSECV error decreases at 0.66 mmol/L. The score plots of the models are reported respectively in Figure 4a,b. Table 1. Model parameters obtained from saliva analysis. Range Latent Variables RMSECV Lactate 0–10 mmol/L 7 1.94 mmol/L Lactate 2–6 mmol/L 35 0.66 mmol/L Sensors 2019, 19, 3948 6 of 9 Sensors 2019, 19, x FOR PEER REVIEW 6 of 9 Figure 4. Score plot of (a) lactate in the range of 0–10 mmol/L; (b) lactate in the range of 2–6 mmol/L. Table 1. Model parameters obtained from saliva analysis. Range Latent Variables RMSECV Lactate 0–10 mmol/L 7 1.94 mmol/L Lactate 2–6 mmol/L 35 0.66 mmol/L Several models were built using data obtained analyzing exhaled breath samples with BIONOTE-V; six respiratory parameters can be predicted using these models. The errors are reported in Table 2. Table 2. Model parameters obtained from exhaled breath analysis. range LVs RMSECV VCO2 3000–6000 mL/min 3 720 mL/min VO2 3000–5500 mL/min 2 894.55 mL/min Pet O2 100–125 mmHg 3 4.71 mmHg Pet CO2 30–45 mmHg 3 2.49 mmHg ReRa 0.95–1.2 [mL/min]/[mL/min] 4 0.04 [mL/min]/[mL/min] VT 2–3 L 3 0.66 L Figure 5 shows the score plot of the parameters VCO2 and VO2. Figure 5. Score plot of: (a) VCO2; (b) VO2 using data from exhaled breath. Figure 4. Score plot of (a) lactate in the range of 0–10 mmol/L; (b) lactate in the range of 2–6 mmol/L. Several models were built using data obtained analyzing exhaled breath samples with BIONOTE-V; six respiratory parameters can be predicted using these models. The errors are reported in Table 2. Table 2. Model parameters obtained from exhaled breath analysis. Range LVs RMSECV VCO2 3000–6000 mL/min 3 720 mL/min VO2 3000–5500 mL/min 2 894.55 mL/min Pet O2 100–125 mmHg 3 4.71 mmHg Pet CO2 30–45 mmHg 3 2.49 mmHg ReRa 0.95–1.2 [mL/min]/[mL/min] 4 0.04 [mL/min]/[mL/min] VT 2–3 L 3 0.66 L Figure 5 shows the score plot of the parameters VCO2 and VO2. Sensors 2019, 19, x FOR PEER REVIEW 6 of 9 Figure 4. Score plot of (a) lactate in the range of 0–10 mmol/L; (b) lactate in the range of 2–6 mmol/L. Table 1. Model parameters obtained from saliva analysis. Range Latent Variables RMSECV Lactate 0–10 mmol/L 7 1.94 mmol/L Lactate 2–6 mmol/L 35 0.66 mmol/L Several models were built using data obtained analyzing exhaled breath samples with BIONOTE-V; six respiratory parameters can be predicted using these models. The errors are reported in Table 2. Table 2. Model parameters obtained from exhaled breath analysis. range LVs RMSECV VCO2 3000–6000 mL/min 3 720 mL/min VO2 3000–5500 mL/min 2 894.55 mL/min Pet O2 100–125 mmHg 3 4.71 mmHg Pet CO2 30–45 mmHg 3 2.49 mmHg ReRa 0.95–1.2 [mL/min]/[mL/min] 4 0.04 [mL/min]/[mL/min] VT 2–3 L 3 0.66 L Figure 5 shows the score plot of the parameters VCO2 and VO2. Figure 5. Score plot of: (a) VCO2; (b) VO2 using data from exhaled breath. Figure 5. Score plot of: (a) VCO2; (b) VO2 using data from exhaled breath. The same models were built using a data-set obtained from the data fusion of data collected measuring both saliva samples (using BIONOTE-L) and exhaled breath samples (using BIONOTE-V). The data-fusion models present RMSECV errors reported in Table 3. Sensors 2019, 19, 3948 7 of 9 Table 3. Model parameters obtained using a fusion of saliva and exhaled breath data. Range LVs RMSECV VCO2 3000–6000 mL/min 2 1024 mL/min VO2 3000–5500 mL/min 2 894 mL/min Pet O2 100–125 mmHg 4 6.11 mmHg Pet CO2 30–45 mmHg 3 2.46 mmHg ReRa 0.95–1.2 [mL/min]/[mL/min] 3 0.06 [mL/min]/[mL/min] VT 2–3 L 2 0.5 L 4. Discussion The BIONOTE can be used for the detection of the anaerobic threshold and for the monitoring the athletes’ performances. The physiological parameter can be predicted using different models developed in this work. The lactate concentration error is 1.94 mmol/L and it decreases in the range of interest, 2–6 mmol/L, at 0.66 mmol/L. This value is larger than the error claimed by other commercial devices, as shown in Table 4, but its advantage is that BIONOTE is non invasive and easy-to-use, so the method used for the acquisition of the data does not need the involvement of expert medical staff. This promising result of saliva analysis also paves the way to a fruitful application for disease diagnosis, as already demonstrated with other sensors [22]. The respiratory parameters’ models obtained with the exhaled breath data show a higher error in cross validation compared to models obtained with exhaled breath and saliva data. The device and the set-up can be improved and this could give a reduction in the error in the model. Further work could be focused on developing an experimental set-up to allow direct sampling from the mouth of the athlete, or adding a functionalized layer that can improve the detection of the lactate for a specific application. By the way, it must be considered that the current method for exhaled breath collection and desorption is mediated by an adsorbing cartridge. The contribution of the exhaled breath fingerprint given by BIONOTE-V is important for athlete monitoring, but the exhaled breath sampling and measurement procedure should be further developed in order to be integrated in a same device with the BIONOTE-L. Obviously this aspect is beyond the scope of this paper. Besides, it is a useful reminder that the same wireless BLE solution used for the BIONOTE-L has already been tested for the BIONOTE-V. The device’s size and low power consumption allows the use of the device in an IoT (Internet of Things) system in which each athlete can be monitored remotely by a trainer or a doctor which can evaluate his performance, which can also be done by comparing it with respect to other sportsmen. Table 4. Comparison chart of the proposed device and five commercial portable devices [12]. Manufacturer Method Analysis time [s] Accuracy [within 2–5 mmol/L] [19] Invasiveness BIONOTE-L ESS Lab, UCBM, Italy Eletrochemical sensor 100 0.66 NO Lactate Pro2 Arkray KDK, Japan Aperometic reagent 15 0.11 YES Lactate Scout+ EKF Giagnostics, Germany Enzymatic amperometric 10 0.09 YES Nova Statsrip Xpress Nova Biomedical, USA Electrochemical biosensor 13 0.13 YES Edge Transatlenticv Science, USA Electrochemical biosensor 45 0.14 YES i-STAT Abbott Laboratories, USA Amperometric 280 0.45 YES It is worth a remark that for the real application of this method based on saliva and exhaled breath collection and measurement, two points have to be clarified in future experiments: (1) the confirmation of the method’s effectiveness for women as well (herein a male population has been tested); (2) testing Sensors 2019, 19, 3948 8 of 9 the effectiveness of this monitoring ‘action’ for supporting an improvement in the training procedures of the athletes. 5. Conclusions The BIONOTE sensor platform can be used for the detection of the anaerobic threshold and for the monitoring the athletes’ performances, measuring blood lactate concentration from saliva samples and respiratory parameters from exhaled breath samples. Even if the device shows a measurement error higher than the error claimed by other commercial devices, it has the great advantage to be non invasive and easy-to-use: so it can be used without the needing of medical staff. The contribution of exhaled breath fingerprint is important to achieve a more inclusive monitoring of the athletes’ performances: however further developments are required to allow direct sampling of exhaled breath samples, without the mediation of adsorbing cartridges. The device’s size and low power consumption allows the use of the sensor platform in an IoT (Internet of Things) system in which each athlete can be monitored remotely by a trainer or a doctor which can evaluate his performance, which can also be done by comparing it with respect to other sportsmen. Author Contributions: Conceptualization, G.P., M.S. and A.G.; methodology, M.S. and A.G.; software, A.Z.; validation, A.S., S.G., A.D.C. and A.Z.; formal analysis G.P. and M.S.; investigation, G.P., M.S. and A.G.; resources, A.D.C., B.D. and A.G.; data curation, A.S., S.G., and A.Z.; writing—original draft preparation, A.S. and A.Z.; writing—review and editing, G.P., M.S. and A.G. Funding: This research received no external funding. Conflicts of Interest: The authors declare no conflict of interest. References 1. Dragonieri, S.; Pennazza, G.; Carratu, P.; Resta, O. Electronic Nose Technology in Respiratory Diseases. Lung 2017, 195, 157–165. [CrossRef] [PubMed] 2. Loutfi, A.; Coradeschi, S.; Mani, G.K.; Shankar, P.; Rayappan, J.B.B. Electronic noses for food quality: A review. J. Food Eng. 2015, 144, 103–111. [CrossRef] 3. Zampolli, S.; Elmi, I.; Ahmed, F.; Passini, M.; Cardinali, G.C.; Nicoletti, S.; Dori, L. An electronic nose based on solid state sensor arrays for low-cost indoor air quality monitoring applications. Sens. Actuators B Chem. 2004, 101, 39–46. [CrossRef] 4. Santonico, M.; Grasso, S.; Genova, F.; Zompanti, A.; Parente, F.; Pennazza, G. Unmasking of olive oil adulteration via a multi-sensor platform. Sensors 2015, 15, 21660–21672. [CrossRef] [PubMed] 5. Gutiérrez, J.M.; Moreno-Barón, L.; Pividori, M.I.; Alegret, S.; del Valle, M. A voltammetric electronic tongue made of modified epoxy-graphite electrodes for the qualitative analysis of wine. Microchim. Acta 2010, 169, 261–268. [CrossRef] 6. Santonico, M.; Parente, F.R.; Grasso, S.; Zompanti, A.; Ferri, G.; D’Amico, A.; Pennazza, G. Investigating a single sensor ability in the characterisation of drinkable water: a pilot study. Water Environ. J. 2016, 30, 253–260. [CrossRef] 7. Grasso, S.; Santonico, M.; Bisogno, T.; Pennazza, G.; Zompanti, A.; Sabatini, A.; Maccarrone, M. An Innovative Liquid Biosensor for the Detection of Lipid Molecules Involved in Diseases of the Nervous System. Proceedings 2018, 2, 760. [CrossRef] 8. Chen, K.Y.; Janz, K.F.; Zhu, W.; Brychta, R.J. Re-defining the roles of sensors in objective physical activity monitoring. Med. Sci. Sports Exerc. 2012, 44, S13–S23. [CrossRef] [PubMed] 9. Heck, H.; Mader, A.; Hess, G.; Mücke, S.; Müller, R.; Hollmann, W. Justification of the 4-mmol/L lactate threshold. Int. J. Sports Med. 1985, 6, 117–130. [CrossRef] [PubMed] 10. Achten, J.; Jeukendrup, A.E. Heart rate monitoring. Sports Med. 2003, 33, 517–538. [CrossRef] [PubMed] 11. Nicolò, A.; Marcora, S.M.; Sacchetti, M. Respiratory frequency is strongly associated with perceived exertion during time trials of different duration. J. Sports Sci. 2016, 34, 1199–1206. [CrossRef] [PubMed] 12. Bonaventura, J.M.; Sharpe, K.; Knight, E.; Fuller, K.L.; Tanner, R.K.; Gore, C.J. Reliability and accuracy of six hand-held blood lactate analysers. J. Sports Sci. Med. 2015, 14, 203–214. [PubMed] Sensors 2019, 19, 3948 9 of 9 13. Santonico, M.; Pennazza, G.; Grasso, S.; D’Amico, A.; Bizzarri, M. Design and test of a biosensor-based multisensorial system: a proof of concept study. Sensors 2013, 13, 16625–16640. [CrossRef] 14. Tenax GR. Available online: https://www.sigmaaldrich.com/catalog/product/supelco/28272u (accessed on 11 September 2019). 15. Pennazza, G.; Santonico, M.; D’Amico, A.; Incalzi, R.A.; Petriaggi, M. Pneumopipe—Auxiliary Device for Collection and Sampling of Exhaled Air. European Patent No. 12425057.2, 25 September 2013. 16. Pennazza, G.; Santonico, M.; Incalzi, R.A.; Scarlata, S.; Chiurco, D.; Vernile, C.; D’Amico, A. Measure chain for exhaled breath collection and analysis: A novel approach suitable for frail respiratory patients. Sens. Actuators B Chem. 2014, 204, 578–587. [CrossRef] 17. Pennazza, G.; Santonico, M.; Zompanti, A.; Grasso, S.; D’Amico, A. Electronic Interface for a Gas Sensor System Based on 32 MHz QCMs: Design and Calibration. IEEE Sens. J. 2017, 18, 1419–1426. [CrossRef] 18. Pennazza, G.; Santonico, M.; Vollero, L.; Zompanti, A.; Sabatini, A.; Kumar, N.; Pini, I.; Solano, W.F.Q.; Sarro, L.; D’Amico, A. Advances in the Electronics for Cyclic Voltammetry: The Case of Gas Detection by Using Microfabricated Electrodes. Front Chem. 2018, 6, 327. [CrossRef] 19. SARSTEDT. Available online: https://www.sarstedt.com/en/products/diagnostic/salivasputum/product/51. 1534/ (accessed on 11 September 2019). 20. Natale, C.D.; Martinelli, E.; Pennazza, G.; Orsini, A.; Santonico, M. Data analysis for chemical sensor arrays. In Advances in Sensing with Security Applications; Springer: Dordrecht, The Netherlands, 2006; pp. 147–169. 21. Quark CPET: Research Grade Stationary System for Accurate and Reliable Metabolic Measurements. Available online: https://www.cosmed.com/en/products/cardio-pulmonary-exercise-test/quark-cpet (accessed on 11 September 2019). 22. Bottoni, U.; Tiriolo, R.; Pullano, S.A.; Dastoli, S.; Amoruso, G.F.; Nistico, S.P.; Fiorillo, A.S. Infrared saliva analysis of psoriatic and diabetic patients: similarities in protein components. IEEE Trans. Biomed. Eng. 2015, 63, 379–384. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

A Sensor Platform for Athletes' Training Supervision: A Proof of Concept Study.

09-12-2019

Zompanti, Alessandro,Sabatini, Anna,Santonico, Marco,Grasso, Simone,Gianfelici, Antonio,Donatucci, Bruno,Di Castro, Andrea,Pennazza, Giorgio

eng

PMC7295712

Vol.:(0123456789) 1 3 European Journal of Applied Physiology (2020) 120:1621–1628 https://doi.org/10.1007/s00421-020-04390-w ORIGINAL ARTICLE Physiological responses during simulated 16 km recumbent handcycling time trial and determinants of performance in trained handcyclists Benjamin Stone1  · Barry S. Mason1  · Ben T. Stephenson1,2  · Vicky L. Goosey‑Tolfrey1 Received: 27 January 2020 / Accepted: 6 May 2020 / Published online: 20 May 2020 © The Author(s) 2020 Abstract Purpose To characterise the physiological profiles of trained handcyclists, during recumbent handcycling, to describe the physiological responses during a 16 km time trial (TT) and to identify the determinants of this TT performance. Methods Eleven male handcyclists performed a sub-maximal and maximal incremental exercise test in their recumbent handbike, attached to a Cyclus II ergometer. A physiological profile, including peak aerobic power output (POPeak), peak rate of oxygen uptake ( ̇VO2Peak), aerobic lactate threshold (AeLT) and PO at 4 mmol L−1 (PO4), were determined. Participants also completed a 16 km simulated TT using the same experimental set-up. Determinants of TT performance were identified using stepwise multiple linear regression analysis. Results Mean values of POPeak = 252 ± 9 W, ̇VO2Peak = 3.30 ± 0.36 L min−1 (47.0 ± 6.8 mL kg−1 min−1), AeLT = 87 ± 13 W and PO4 = 154 ± 14 W were recorded. The TT was completed in 29:21 ± 0:59 min:s at an intensity equivalent to 69 ± 4% POPeak and 87 ± 5% ̇VO2Peak. POPeak (r = − 0.77, P = 0.006), PO4 (r = − 0.77, P = 0.006) and AeLT (r = − 0.68, P = 0.022) were significantly correlated with TT performance. PO4 and POPeak were identified as the best predictors of TT performance (r = 0.89, P < 0.001). Conclusion POPeak, PO4 and AeLT are important physiological TT performance determinants in trained handcyclists, dif- ferentiating between superior and inferior performance, whereas ̇VO2peak was not. The TT took place at an intensity cor- responding to 69% POPeak and 87% ̇VO2peak. Keywords Endurance performance · Paralympic · Lactate threshold · Disability sport · Spinal cord injury Abbreviations AeLT Aerobic threshold PO4 Power output at 4 mmol L−1 BLa Blood lactate concentrations HR Heart rate ̇VO2 Oxygen uptake ME% Mechanical efficiency PO Power output POPeak Peak aerobic power output HRPeak Peak heart rate ̇VO2Peak Peak rate of oxygen uptake RPE Rating of perceived exertion RER Respiratory exchange ratio TT Time trial Introduction Recumbent handcycling is an endurance sport for athletes with lower limb impairments, such as spinal cord injuries, lower limb amputations and congenital conditions (Abel et al. 2006). Since 2004 (Athens Paralympic Games), hand- cycling has been an integral part of the paracycling road programme, with 65 handcyclists contesting 13 events at the most recent 2016 Paralympic Games in Rio (Paralympics. org 2016). At national and international events (e.g. British Championships, Paralympic Games), handcyclists compete Communicated by Jean-René Lacour. * Vicky L. Goosey-Tolfrey [email protected] 1 Peter Harrison Centre for Disability Sport, School of Sport, Exercise and Health Sciences, Loughborough University, NCSEM 1.26, Loughborough University Campus, Loughborough LE11 3TU, UK 2 English Institute of Sport, Performance Centre, Loughborough University, Loughborough, UK 1622 European Journal of Applied Physiology (2020) 120:1621–1628 1 3 in individual time trials (TT) (10–30 km, lasting 20–40 min) and/or road races (40–80 km, lasting 60–150 min) with many handcycling events being scheduled on consecutive days (Zeller et al. 2017). To date, most studies examining physiological responses during handcycling have gath- ered and described heart rate (HR), capillary blood lactate concentration (BLa), power output (PO) or rate of oxygen uptake ( ̇VO2) during laboratory-based exercise tests (Janssen et al. 2001; Abel et al. 2010; de Groot et al. 2014; Kouwijzer et al. 2018b; Quittmann et al. 2018). These studies have been limited to recreational or touring bike configurations which are disparate to that of elite handcyclists. It is only recently that recumbent handcycling studies have been conducted during competition (West et al. 2015) or protocols repli- cating recumbent handcycling or competitive race intensi- ties (Fischer et al. 2015, 2020; Graham-Paulson et al. 2016; Stangier et al. 2019; Stone et al. 2019b). Physiological variables, such as peak rate of oxygen uptake ( ̇VO2Peak) and ventilatory thresholds are the best predictors of handcycling TT performance. Moreover, peak power output achieved during incremental tests to exhaustion (POPeak) has been related to performance (Jans- sen et al. 2001; Lovell et al. 2012; de Groot et al. 2014; Fischer et al. 2015). However, large variance in POPeak (129–247 W) and ̇VO2Peak (2.0–3.45 L min−1 equivalent to 26.5–42.3 mL kg−1 min−1) have been reported in the litera- ture (Janssen et al. 2001; Abel et al. 2006; Lovell et al. 2012; Fischer et al. 2015; Graham-Paulson et al. 2018; Kouwijzer et al. 2018a; Stangier et al. 2019; Stone et al. 2019a). Con- founding factors, such as training status, athlete classifica- tion and the number of years involved in the sport and indeed the protocols adopted (e.g. ramp vs. step incremental test), all contribute to these reported physiological profiles (Miller et al. 2004; Goosey-Tolfrey et al. 2008; van Drongelen et al. 2009; Lovell et al. 2012). Moreover, since the experimental designs of these studies varied considerably, using attach- able-units, touring handbikes or bespoke ergometers, this makes the findings difficult to transfer to the recumbent handbikes used in the present day. These handbikes are con- siderably lighter, but more importantly, are bespoke to the individual, with respect to crank position, crank width and backrest inclination (Stone et al. 2018). Therefore, the pur- pose of this study was to characterise the physiological pro- files of trained handcyclists, during recumbent handcycling and to identify which physiological variables are related to 16 km TT performance. Methods Participants Eleven male recumbent handcyclists (age 38 ± 10 years; body mass 72 ± 9 kg; classification 5 H3 and 6 H4; injury description 4 spinal cord lesions complete (4th–11th tho- racic vertebra), 3 spinal cord lesions incomplete (8th–9th thoracic vertebra), 3 lower limb amputees and 1 diplegic cerebral palsy) volunteered to participate. All partici- pants had competed in handcycling or paratriathlon, at a national or international level (handcycling experience 5.4 ± 5.6 years; training duration 5 ± 4 handcycling ses- sions totalling 7 ± 3 h week−1 with a self-reported distance of 178 ± 93 km week−1 and 2 ± 1 gym sessions totalling 3 ± 2 h week−1). The University’s local ethics committee approved the study. Before participation, all participants provided their written, informed consent. Experimental protocol Participants refrained from exercise in the 24 h preceding the testing. The experiment protocol was performed on two consecutive days (Fig. 1). On the first day, participants completed a sub-maximal exercise test, to determine—the aerobic threshold (AeLT), power output (PO) at 4 mmol L−1 (PO4), and maximal incremental exercise test, to assess POPeak, ̇VO2Peak, and peak heart rate (HRPeak). Ventilatory thresholds were not determined as there has been mixed results to their suitability in this sample (Kouwijzer et al. 2019; Leicht et al. 2014). Both tests were conducted in the participants’ bespoke recumbent handbike (5 ‘Carbonbike’, 4 ‘Top End’ and 2 ‘Schmicking’), which were attached to a Cyclus II ergometer (RBM electronic automation GmbH, Leipzig, Germany). Following a 10-min warm-up at a self- selected cadence and intensity, the sub-maximal test com- menced with an initial load of 20 W, increasing by 20 W every 4 min until BLa exceeded 4.0 mmol L−1 when the test was terminated. BLa was determined from 20 μL earlobe capillary blood samples, collected in the last minute of each stage and analysed using a Biosen C-Line (EFK Diagnostics GmbH, Germany). Breath-by-breath gas analysis (Cortex Metalyzer 3B, Cortex, Leipzig, Germany)—for calculation of ̇VO2, carbon dioxide production and respiratory exchange ratio (RER)—HR (Polar RS400, Kempele, Finland) and rating of perceived exertion (RPE; Borg 1998), were also collected in the last minute of each stage. AeLT was cal- culated using the log–log transformation method (Beaver et al. 1985). Moreover, this test enabled the identification of the PO corresponding to a fixed BLa of 4 mmol L−1 (PO4), by using linear interpolation methods as used in the hand- cycling literature (Stangier et al. 2019; Zeller et al. 2017). 1623 European Journal of Applied Physiology (2020) 120:1621–1628 1 3 Following 30 min’ passive rest, participants performed a maximal test to exhaustion. The starting PO was equivalent to their AeLT and was maintained for 2 min. PO was then increased by 5 W every 15 s until the athlete reached voli- tional exhaustion (failure to maintain cadence ≥ 50 rpm and an overall RPE ≥ 19) (Graham-Paulson et al. 2016). Ver- bal encouragement was provided during this test. Breath- by-breath gas analysis and HR were recorded continuously throughout the maximal test and BLa reported 2 min post- completion of the maximal test. The ergometer was set in a power control mode, ensuring the pre-set PO was controlled independently of cadence or gear selection. Cadence was self-selected in all trials as reported elsewhere (Graham- Paulson et al. 2016). On the second day, participants completed a 16-km TT in the shortest time possible. Participants selected the gear ratio to commence the TT, which could then be changed virtually by the investigators throughout the time trial, as instructed by participants. No motivation was provided dur- ing the TT and the feedback provided was PO, cadence and cumulative distance displayed on the ergometer, to maximise ecological validity. Breath-by-breath gas analysis and HR were recorded continuously throughout the TT and BLa col- lected every 4 km. Data and statistical analysis In the sub-maximal exercise test ̇VO2, HR, RER and cadence were averaged across the last minute of each stage. Mechanical efficiency (ME%), calculated as the ratio of external work to energy expended (Powers et al. 1984), in 1 min of exercise, was determined at AeLT and PO4 (RER < 1.00 for all participants). Energy expenditure was calculated from the product of ̇VO2 and RER and the standard conversion table (Péronnet and Massicotte 1991). During the maximal exercise test, the highest 30-s rolling average of ̇VO2 and HR were used to calculate ̇VO2Peak and HRPeak. A 15-s rolling average was used to calculate POPeak. In the TT, ̇VO2, HR, RER, cadence, PO, speed and ME% were averaged every km and calculated relative to ̇V O2Peak, HRPeak and POPeak, where applicable. Shapiro–Wilks tests were used to determine the distri- bution of the data. Repeated measures analysis of vari- ance was used to determine differences in ̇VO2, HR, PO and cadence across each km of the TT. Sphericity was assessed using the Mauchly’s test of sphericity. If the data were aspherical, a correction, using the Greenhouse–Geis- ser epsilon, was applied to the calculated P value (Girden 1992). If a significant difference was identified, post hoc paired t tests, with Bonferroni corrections, were applied to determine the differences in cadence or PO. Pearson’s product-moment correlation was used to ana- lyse the relationship between variables from the sub-max- imal (AeLT, ME% at AeLT, ̇VO2 at AeLT, PO4, ME% at PO4 and ̇VO2 at PO4) and maximal test [ ̇VO2Peak (absolute Fig. 1 Experimental protocol for the submaximal test, maximal test and TT with details of the collection of BLa, HR and ̇VO2 1624 European Journal of Applied Physiology (2020) 120:1621–1628 1 3 and relative), HRPeak and POPeak] to TT performance (IBM SPSS Statistics 24, Chicago, IL). Variables that were sig- nificantly correlated with 16 km TT performance were inputted into a stepwise multiple linear regression to establish the most important determinants of 16 km TT performance. Durbin–Watson test was used to indicate the independence of variables entered into the regression. Tolerance and variance inflation factors were also used to assess multicollinearity within the data. Results Outcomes of the exercise tests and TT Data from the sub-maximal exercise test, maximal exer- cise test and 16 km TT are summarised in Table 1. No significant differences in PO were identified throughout the TT (Fig. 2a). The 4-km sector split times ranged from 7:24 ± 0.14 min:s (4–8 km) to 7:17 ± 0.14 min:s (12–16 km), further indicating that, as a group, the handcyclists main- tained a relatively constant pace throughout the TT. Cadence increased throughout the TT, and during the 16th km was on average 4 rpm greater than the 12th and 13th km (P < 0.05) (Fig. 2b). ̇VO2 increased significantly during the first 2 km of the TT (P < 0.05) before plateauing during the remainder of the TT (Fig. 2c). Similarly, HR significantly increased during the first 3 km of the TT (P < 0.05) before plateauing between the 3rd km and 14th km and significantly increas- ing in the final 2 km (P < 0.001) (Fig. 2d). BLa significantly increased throughout the TT [4 km: 8.0 ± 2.3 mmol L−1; 8 km: 10.0 ± 2.6 mmol L−1; 12 km: 11.2 ± 2.7 mmol L−1; 16 km: 13.1 ± 2.7 mmol L−1 (P < 0.001)], while ME% was found to significantly decrease [0–4 km: 19.3 ± 1.2%; 4–8 km: 16.8 ± 1.2%; 8–12 km: 16.3 ± 1.3%; 12–16 km: 16.3 ± 1.5% (P < 0.001)]. The average intensity of the TT’s was equivalent to 69 ± 4% of POPeak and 87 ± 5% of ̇VO2Peak (Table 1). Determinants of TT performance The highest correlations were observed between POPeak (r = − 0.77, P = 0.006), PO4 (r = − 0.77, P = 0.006), AeLT (r = − 0.68, P = 0.022) and TT performance (Table 2). Mod- erate to low correlations were observed for all other physi- ological parameters (Table 2). The multiple linear regression indicated that PO4 was the strongest predictor accounting for 59.3% of the variance in 16 km TT performance (P = 0.006). The addition of POPeak (19.0% of overall variance) to PO4 provided a stronger prediction model, accounting for 78.3% of the variance in performance (P = 0.002). A Durbin–Wat- son value of 1.501 indicated that variables within this model were sufficiently independent and not autocorrelated. Toler- ance values of 0.741 and a variance inflation factor of 1.350 further indicate that multicollinearity was not an issue for this regression model. Discussion The current study investigated the physiological responses during a 16-km TT, in a population of national and interna- tional recumbent handcyclists. Extending the recent work of Stangier et al. (2019), this study focused on recumbent trained handcyclists yet was designed to measure the sub- maximal exercise, maximal exercise and simulated TT per- formance in their own bespoke recumbent handbikes. A key finding was that better 16 km TT performances were achieved by handcyclists with a higher POPeak, PO4 and AeLT. Conversely, ̇VO2Peak, both absolute and relative, were not correlated with TT performance. The results from the maximal exercise test revealed that the handcyclists in the current study had higher average Table 1 Physiological responses from the submaximal and maximal recumbent handcycling tests and the average response from the whole 16 km handcycling TT (n = 11) Parameter Mean ± SD Min Max Submaximal test  AeLT (W) 87 ± 13 70 108   ̇VO2 at AeLT (L min−1) 1.49 ± 0.12 1.29 1.63  ME% at AeLT (%) 16.9 ± 1.6 14.5 19.8  PO4 (W) 154 ± 14 128 173   ̇VO2 at PO4 (L min−1) 2.43 ± 0.40 1.92 3.27  ME% at PO4 (%) 17.9 ± 1.7 14.3 19.6 Maximal test   ̇VO2Peak (L min−1) 3.30 ± 0.36 2.75 4.02   ̇VO2Peak (mL kg−1 min−1) 46.98 ± 6.80 36.31 57.64  POPeak (W) 252 ± 19 229 282  BLa (mmol L−1) 10.90 ± 2.51 7.48 14.28  HRPeak (bpm) 188 ± 11 169 208  RERPeak 1.15 ± 0.07 1.06 1.26 16 km TT  Time (min:s) 29:20.7 ± 00:58.8 27:58.0 30:33.1  Velocity (km h−1) 32.7 ± 1.1 31.4 34.3  PO (W) 174 ± 15 152 190  Cadence (rpm) 94.3 ± 6.0 84.3 102.0  HR (bpm) 171 ± 12 154 194   ̇VO2peak (L min−1) 2.84 ± 0.31 2.37 3.26  BLa (mmol L−1) 10.5 ± 2.5 7.3 15.0  % ̇VO2Peak 87.07 ± 5.03 79.65 92.66  %HRPeak 92.34 ± 3.23 86.66 97.52  %POPeak 68.83 ± 3.78 64.51 74.93  ME (%) 17.1 ± 1.3 14.7 18.5 1625 European Journal of Applied Physiology (2020) 120:1621–1628 1 3 and maximal ̇VO2Peak and POPeak values than previously reported (Lovell et al. 2012; Fischer et al. 2015; Graham- Paulson et al. 2018; Stangier et al. 2019). However, the participants in the current study, with ̇VO2Peak and POPeak values of 46.9 ± 6.8 mL kg−1 min−1 and 252 ± 19 W, respec- tively, were similarly trained (averaging ≥ 7 h week−1 and # * § 110 130 150 170 190 210 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 * § ¥ # # # ¥ # # § ¥ § * 1.0 1.5 2.0 2.5 3.0 3.5 4.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 * 75 85 95 105 115 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 125 150 175 200 225 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 PO (W) Cadence (rpm) V̇ O2 (L·min-1) HR (bpm) TT Distance (km) ¥ ¥ * * * * * * * * * * * * * * * * * * * * * * * * * * # # # # # # # # § § ¥ ¥ ¥ ¥ (a) (b) (c) (d) Fig. 2 PO (a), cadence (b), ̇VO2 (c) and HR (d) responses of trained recumbent handcyclists throughout a 16-km TT (values: mean ± SD). *Significant difference to 1st km (P < 0.05), #Significant difference to 2nd km (P < 0.05), §Significant difference to 15th km (P < 0.05), ¥Sig- nificant difference to 16th km (P < 0.05) 1626 European Journal of Applied Physiology (2020) 120:1621–1628 1 3 ≥ 175 km week−1) to the participants from previous stud- ies (Lovell et al. 2012; Fischer et al. 2015; Stangier et al. 2019). Furthermore, the handcyclists that achieved the greatest ̇VO2Peak (4.02 L min−1 and 57.6 mL kg min−1) and POPeak (281 W) values averaged 5 handcycling sessions a week totalling 8 h week−1 with a self-reported distance of 170 km week−1. Therefore, the higher physiological values observed in the current study are likely to be due to advances in endurance training regimes (Zeller et al. 2017), strength training (Nevin et al. 2018) and the use of bespoke recum- bent handbikes in comparison to the modified ergometers used previously (Lovell et al. 2012; Fischer et al. 2015). Participants were likely to be more stable and comfortable in their handbike, which is perceived to be essential for the application of power in handcycling, potentially facilitating the increased ̇VO2Peak and POPeak. The trained handcyclists completed the simulated 16 km TT at a similar intensity (87 ± 5% ̇VO2Peak) to a 20 and 22 km TT, reported by Graham-Paulson et al. (2018) and Fischer et al. (2015), respectively, yet with much higher end BLa (> 6 mmol L−1). During the TT participants maintained a constant PO, similar to Fischer et al. (2015), although ME% reduced from 19.3% in the first 4 km to 16.3% in the final 4 km. Based on this data, future constant work studies rep- licating handcycling TTs, should select a PO equivalent to ~ 70% POPeak or ~ 25% greater than PO4 (using methods similar to the current study). This high exercise intensity indicates the contribution of the anaerobic pathways to sup- ply energy during a typical handcycling TT. Therefore, a high anaerobic capacity may play an important role in hand- cycling TT success, like able-bodied cycling (Støren et al. 2013), although this requires future investigation. TT performance was significantly and highly correlated with POPeak (r = − 0.77, P = 0.006), which supports previous research (Janssen et al. 2001; Lovell et al. 2012; de Groot et al. 2014; Fischer et al. 2015). The present study found, for the first time, that recumbent handcycling 16 km TT performance was strongly correlated with AeLT (r = − 0.68, P = 0.022) and PO4 (r = − 0.77, P = 0.006). Contrary to the previous handcycling literature (Janssen et al. 2001; Lovell et al. 2012; de Groot et al. 2014; Fischer et al. 2015), abso- lute ̇VO2Peak (r = − 0.06) and relative ̇VO2Peak (r = − 0.21) were not correlated with TT performance. Therefore, these results suggest that, in highly trained handcyclists with a heterogenous injury description, POPeak and blood lactate (AeLT and PO4) are better predictors of TT performance than ̇VO2Peak. In able-bodied athletes, fractional utilisa- tion of ̇VO2Peak, i.e. intensity at lactate threshold, has long been understood to better predict endurance performance in highly trained athletes whilst POpeak has consistently been shown to predict cycling performance, regardless of task duration or test profile (Bentley et al. 2001; Joyner and Coyle 2008). This has also been shown to be true when assess- ing physiological correlates to cycling performance within a paratriathlon race (Stephenson et al. 2020). As such, this finding is not unique in endurance sport but is presently con- firmed in elite handcyclists. The current study extends previous work, which was limited to handcyclists with a spinal cord injury (Lovell et al. 2012; Fischer et al. 2015; West et al. 2015; Graham- Paulson et al. 2018), by including participants with lower limb impairments and cerebral palsy, which is more repre- sentative of athletes competing at the Paralympic Games. The increased recruitable muscle mass of athletes with non- paralysed lower limbs (e.g. amputation or cerebral palsy), results in greater rates of oxygen uptake relative to athletes with complete spinal lesions (Baumgart et al. 2018). For example, in amputee athletes this is due to the capability to use lower limb musculature to brace within the handbike for greater stability and power transfer. The increased available muscle mass and the potential loss of lower limbs makes variables such as absolute ̇VO2Peak, relative ̇VO2Peak and power to weight ratio unrepresentative or misleading without scaling parameters (Goosey-Tolfrey et al. 2003). Therefore, in a population of handcyclists competing in the H3 and H4 classes [spinal lesions, lower limb amputations and cerebral palsy (UCI 2019)] variables such as POPeak and blood lactate (AeLT and PO4) are better indicators of TT performance than ̇VO2Peak. From this data, it may be recommended that testing for POPeak or PO4 is conducted with handcyclists to infer performance potential. Although assessing the latter requires specialist equipment for BLa measurement, PO4 is also commonly used for training intensity prescription. Table 2 Correlations between 16  km TT and physiological param- eters determined in the submaximal and maximal exercise tests (n = 11) * Correlation P < 0.05 Parameter Correlation coefficient (R) P Sub-maximal test  AeLT (W) − 0.68* 0.022   ̇VO2 at AeLT (L min−1) − 0.43 0.191  ME at AeLT (%) − 0.38 0.256  PO4 (W) − 0.77* 0.006   ̇VO2 at PO4 (L min−1) − 0.24 0.472  ME at PO4 (%) − 0.30 0.370 Maximal test  POPeak (W) − 0.77* 0.006  HRPeak (bpm) − 0.53 0.095   ̇VO2Peak (L min−1) − 0.06 0.868   ̇VO2Peak (mL kg−1 min−1) − 0.21 0.539 1627 European Journal of Applied Physiology (2020) 120:1621–1628 1 3 Limitations An electronically braked ergometer was used to simulate the TT. The electronic braking meant that handcyclists had to pedal continually to apply power, which differs from a road TT. The corners and gradients throughout a road TT course would allow the handcyclists to recover for short periods (0–5 s). Furthermore, as the ergometer was fixed to the floor, the influence of aerodynamic drag, steering and braking were removed. Future research should aim to collect TT data on the road, in a competition if possible, and com- pare with laboratory-based findings. Additionally, recruiting participants from the H1, H2 and H5 classes, along with a greater number within H3 (n = 5) and H4 (n = 6) would allow a comparison between classes. Finally, the inter-individual variability of physiological parameters in the study’s cohort is acknowledged. Heterogeneity is inherent in Paralympic populations due to athletes’ spectra of impairments. Further- more, variation in physiological parameters are likely largely mediated by training history and performance standard. Conclusions The current study revealed that the best predictors of 16 km TT performance, in a population of trained recumbent hand- cyclists, were PO4 and POPeak. It is suggested that PO4 and POPeak are used to infer performance level/potential rather than ̇VO2Peak within a population of H3 and H4 recumbent handcyclists (spinal cord lesion vs. lower limb amputation vs. cerebral palsy). A protocol equivalent to 70% POPeak is recommended, in future studies, to replicate 16 km TT intensity. Acknowledgements The authors would like to thank the handcyclists who participated in this study. This study was funded by the Engineer- ing and Physical Sciences Research Council (Grant no. EP/M507489/1) and supported by John Lenton, the English Institute of Sport, British Cycling and the Peter Harrison Centre for Disability Sport. Compliance with ethical standards Conflict of interest The authors report no conflict of interest. Ethical approval Approval was obtained from the ethics committee of Loughborough University. The procedures used in this study adhere to the tenets of the Declaration of Helsinki. Open Access This article is licensed under a Creative Commons Attri- bution 4.0 International License, which permits use, sharing, adapta- tion, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. References Abel T, Schneider S, Platen P, Strüder HK (2006) Performance diag- nostics in handbiking during competition. Spinal Cord 44:211– 216. https ://doi.org/10.1038/sj.sc.31018 45 Abel T, Burkett B, Schneider S, Lindschulten R, Strüder HK (2010) The exercise profile of an ultra-long handcycling race: the Sty- rkeprøven experience. Spinal Cord 48:894–898. https ://doi. org/10.1038/sc.2010.40 Baumgart JK, Brurok B, Sandbakk Ø (2018) Peak oxygen uptake in Paralympic sitting sports: a systematic literature review, meta- and pooled-data analysis. PLoS ONE 13:1–25. https ://doi. org/10.1371/journ al.pone.01929 03 Beaver W, Wasserman K, Whipp BJ (1985) Improved detection of lactate threshold during exercise using a log-log transformation. J Appl Physiol 59:1936–1940 Bentley DJ, McNaughton LR, Thompson D, Vleck VE, Batterham AM (2001) Peak power output, the lactate threshold, and time trial performance in cyclists. Med Sci Sports Exerc 33:2077–2081 Borg GAV (1998) Borg’s perceived exertion and pain scales. Human Kinetics, Champaign, IL de Groot S, Postma K, van Vliet L, Timmermans R, Valent LJ (2014) Mountain time trial in handcycling: exercise intensity and predic- tors of race time in people with spinal cord injury. Spinal Cord 52:455–461. https ://doi.org/10.1038/sc.2014.58 Fischer G, Figueiredo P, Ardigò LP (2015) Physiological performance determinants of a 22 km handbiking time trial. Int J Physiol Per- form 10:965–971. https ://doi.org/10.1123/ijspp .2014-0429 Fischer, Figueiredo, Ardigò (2020) Bioenergetics and biomechanics of handcycling at submaximal speeds in athletes with a spinal cord injury. Sports 8:16. https ://doi.org/10.3390/sport s8020 016 Girden ER (1992) ANOVA—repeated measures. Sage, Newbury Park Goosey-Tolfrey VL, Batterham AM, Tolfrey K (2003) Scaling behavior of ̇VO2peak in trained wheelchair athletes. Med Sci Sports Exerc 35:2106–2111. https ://doi.org/10.1249/01.MSS.00000 99106 .33943 .8C Goosey-Tolfrey VL, Alfano H, Fowler N (2008) The influence of crank length and cadence on mechanical efficiency in hand cycling. Eur J Appl Physiol 102:189–194. https ://doi.org/10.1007/s0042 1-007-0576-7 Graham-Paulson T, Perret C, Goosey-Tolfrey V (2016) Improvements in cycling but not handcycling 10 km time trial performance in habitual caffeine users. Nutrients 8:1–11. https ://doi.org/10.3390/ nu807 0393 Graham-Paulson T, Perret C, Goosey-Tolfrey V (2018) Case study: dose response of caffeine on 20-km handcycling time trial perfor- mance in a paratriathlete. Int J Sport Nutr Exerc Metab 28:274– 278. https ://doi.org/10.1123/ijsne m.2017-0089 Janssen TW, Dallmeijer AJ, van der Woude LH (2001) Physical capac- ity and race performance of handcycle users. J Rehabil Res Dev 38:33–40 Joyner MJ, Coyle EF (2008) Endurance exercise performance: the physiology of champions. J Physiol 586:35–44. https ://doi. org/10.1113/jphys iol.2007.14383 4 Kouwijzer I, Nooijen CF, Van Breukelen K, Janssen TW, De Groot S (2018a) Effects of push-off ability and handcycle type on hand- cycling performance in able-bodied participants. J Rehabil Med 50:563–568. https ://doi.org/10.2340/16501 977-2343 1628 European Journal of Applied Physiology (2020) 120:1621–1628 1 3 Kouwijzer I, Valent L, Osterthun R, van der Woude L, de Groot S (2018b) Peak power output in handcycling of individuals with a chronic spinal cord injury: predictive modeling, validation and reference values. Disabil Rehabil. https ://doi.org/10.1080/09638 288.2018.15010 97 Kouwijzer I, Cowan RE, Maher JL, Groot FP, Riedstra F, Valent LJ, van der Woude LH, de Groot S (2019) Interrater and intrarater reliability of ventilatory thresholds determined in individuals with spinal cord injury. Spinal Cord 57:669–678. https ://doi. org/10.1038/s4139 3-019-0262-8 Leicht CA, Griggs KE, Lavin J, Tolfrey K, Goosey-Tolfrey VL (2014) Blood lactate and ventilatory thresholds in wheelchair athletes with tetraplegia and paraplegia. Eur J Appl Physiol 114:1635– 1643. https ://doi.org/10.1007/s0042 1-014-2886-x Lovell D, Shields D, Beck B, Cuneo R, McLellan C (2012) The aero- bic performance of trained and untrained handcyclists with spi- nal cord injury. Eur J Appl Physiol 112:3431–3437. https ://doi. org/10.1007/s0042 1-012-2324-x Miller TL, Mattacola CG, Santiago MC (2004) Influence of varied, controlled distances from the crank axis on peak physiological responses during arm crank ergometry. J Exerc Physiol Online 7:61–67. https ://doi.org/10.1519/JSC.0b013 e3181 87456 4 Nevin J, Smith P, Waldron M, Patterson S, Price M, Hunt A, Blagrove R (2018) Efficacy of an 8-week concurrent strength and endurance training program on hand cycling performance. J Strength Cond Res 32:1861–1868 Paralympics.org (2016) Daily competition schedule—cycling road. https ://www.paral ympic .org/rio-2016/sched ule-resul ts/info-live- resul ts/rio-2016/eng/zz/engzz _cycli ng-road-daily -compe titio n-sched ule.htm. Accessed 13 Dec 2018 Péronnet F, Massicotte D (1991) Table of nonprotein respiratory quo- tient: an update. Can J Sport Sci 16:23–29 Powers SK, Beadle RE, Mangum M (1984) Exercise efficiency during arm ergometry: effects of speed and work rate. J Appl Physiol 56:495–499 Quittmann OJ, Abel T, Zeller S, Foitschik T, Strüder HK (2018) Lac- tate kinetics in handcycling under various exercise modalities and their relationship to performance measures in able-bodied partici- pants. Eur J Appl Physiol 118:1493–1505. https ://doi.org/10.1007/ s0042 1-018-3879-y Stangier C, Abel T, Zeller S, Quittmann OJ, Perret C, Strüder HK (2019) Comparison of different blood lactate threshold concepts for constant load performance prediction in spinal cord injured handcyclists. Front Physiol 10:1–11. https ://doi.org/10.3389/fphys .2019.01054 Stephenson BT, Shill A, Lenton J, Goosey-Tolfrey V (2020) Physi- ological correlates to in-race paratriathlon cycling performance. Int J Sports Med. https ://doi.org/10.1055/a-1103-2001 Stone B, Mason BS, Bundon A, Goosey-Tolfrey VL (2018) Elite hand- cycling: a qualitative analysis of recumbent handbike configura- tion for optimal sports performance. Ergonomics 62:449–458. https ://doi.org/10.1080/00140 139.2018.15311 49 Stone B, Mason BS, Warner MB, Goosey-Tolfrey VL (2019a) Shoulder and thorax kinematics contribute to increased power output of competitive handcyclists. Scand J Med Sci Sports 29:843–853. https ://doi.org/10.1111/sms.13402 Stone B, Mason BS, Warner MB, Goosey-Tolfrey VL (2019b) Hori- zontal crank position affects economy and upper limb kinematics of recumbent handcyclists. Med Sci Sports Exerc 51:2265–2273. https ://doi.org/10.1249/MSS.00000 00000 00206 2 Støren Ø, Ulevåg K, Larsen MH, Støa EM, Helgerud J (2013) Physio- logical determinants of the cycling time trial. J Strength Cond Res 27:2366–2373. https ://doi.org/10.1519/JSC.0b013 e3182 7f542 7 UCI (2019) Union Cycliste Internationale cycling regulations—part 16 para-cycling. http://www.uci.ch/mm/Docum ent/News/Rules andre gulat ion/16/80/73/1-GEN-20160 101-E_Engli sh.pdf. Accessed 13 Apr 2018 van Drongelen S, Maas JCC, Scheel-Sailer A, Van Der Woude LH (2009) Submaximal arm crank ergometry: effects of crank axis positioning on mechanical efficiency, physiological strain and per- ceived discomfort. J Med Eng Technol 33:151–157. https ://doi. org/10.1080/13561 82080 25656 76 West CR, Gee CM, Voss C, Hubli M, Currie KD, Schmid J, Kras- sioukov AV (2015) Cardiovascular control, autonomic function, and elite endurance performance in spinal cord injury. Scand J Med Sci Sports 25:476–485. https ://doi.org/10.1111/sms.12308 Zeller S, Abel T, Strueder HK (2017) Monitoring training load in hand- cycling: a case study. J Strength Cond Res 31:3094–3100. https :// doi.org/10.1519/JSC.00000 00000 00178 6 Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Physiological responses during simulated 16 km recumbent handcycling time trial and determinants of performance in trained handcyclists.

05-20-2020

Stone, Benjamin,Mason, Barry S,Stephenson, Ben T,Goosey-Tolfrey, Vicky L

eng

PMC10233502

Article Perceptual and Motor Skills 2023, Vol. 130(3) 1202–1220 © The Author(s) 2023 Article reuse guidelines: sagepub.com/journals-permissions DOI: 10.1177/00315125231165819 journals.sagepub.com/home/pms Mental Toughness and Resilience in Trail Runner’s Performance Nuno Gameiro1,2, Filipe Rodrigues1,2, Ra´ul Antunes1,2,3, Rui Matos1,2, Nuno Amaro1,2, Miguel Jacinto1,2, and Diogo Monteiro1,3,4 Abstract Our purpose with this study was to analyze trail runners’ psychological variables of mental toughness (MT) and resilience, and their associations with runners’ perfor- mances within a quantitative cross-sectional study. In total, we analyzed data from 307 Portuguese trail runners (60 female, 247 male), aged between 20 to 66 years (M age = 41.98; SD = 7.74). The results showed that the measurement model, including the factors of MT, resilience, and performance variables, exhibited an adequate fit to the data: χ2 = 150.01 (74); BS-p = .003; CFI= .953; TLI = .942; RMSEA = .058 90% (.045, .071) and SRMR= .042. Standardized direct effects revealed positive associations between these variables. More specifically: (a) MT was significantly associated with resilience; and (b) resilience was significantly associated with performance. The indirect regression paths showed that MT was positively associated with performance, with resilience considered a possible mediator (β = .09 IC = .010, .168; p = .02). In total, considering direct and indirect effects, the model explained 21% of performance variance among trail runners. 1ESECS – Polytechnic of Leiria, Leiria, Portugal 2Life Quality Research Centre (CIEQV), Leiria, Portugal 3Center for Innovative Care and Health Technology (ciTechCare), Polytechnic of Leiria, Leiria 2411-901, Portugal 4Research Centre in Sports Sciences, Health Sciences and Human Development (CIDESD), Vila Real, Portugal Corresponding Author: Diogo Monteiro, ESECS – Polytechnic of Leiria, Portugal; Research Centre in Sports Sciences, Health Sciences and Human Development (CIDESD), Vila Real, Portugal. Life Quality Research Centre (CIEQV), Leiria, Portugal. Email: [email protected] Keywords endurance, mental toughness, performance, resilience, trail running Introduction Exercise benefits have been well described in past literature (Spittler & Oberle, 2019) with agreement that exercise contributes favorably to the physiological, psychological, and social development of the human being (Bostancı et al., 2019). Traditional rec- reational running is a very popular form of exercise around the world (Wa´skiewicz et al., 2019) creating positive lifestyles changes (Gajardo-Burgos et al., 2021). Over recent years, there has been increasing interest in running longer distances (Spittler & Oberle, 2019), probably due to modern societal demands to overcome adversities and challenges (Goddard et al., 2019). Parallel to this tendency, there is an increase in the athlete´s participation in off road or track endurance races known as trail running (Easthope et al., 2014; Malliaropoulos et al., 2015; Suter et al., 2020), involving unsurfaced mountain trails with extensive vertical displacement and different distances (Easthope et al., 2014; Malliaropoulos et al., 2015). Trail running is an outdoor sport (Viljoen et al., 2022) that is becoming one of the most popular disciplines in endurance running (Groslambert et al., 2020; Scheer et al., 2019). The International Association of Athletics Federations has recognized trail running as a new running discipline (Perrotin et al., 2021) and there are an estimated 20 million trail runners, with increased par- ticipation in the last decade (Viljoen et al., 2022). With a strong sense of sports ethics and a sense of humility, trail running is a sport that takes place amid nature and respects the environment as demanding for both body and mind (International Trail Running Association [ITRA], 2022). Trail running races are characterized by pedestrian competitions open to everyone, with a minimum use of paved roads (20% maximum) (Malliaropoulos et al., 2015) and distances that can range from a few kilometers to more than 200 km (in a multiday marathon) (Viljoen et al., 2022). These endurance races occur in variable terrains, including over frequent significant climbs and descents (elevation gain and loss), in varied environmental conditions like cold, heat, high altitude, snow, and humidity conditions (Perrotin et al., 2021) creating high overall difficulty for a given race (ITRA, 2022). The conditions can influence both the bio- mechanical and psychological state of the runner and, therefore, the overall perfor- mance during the race (Perrotin et al., 2021). Thus, there is a need to attend to the runner’s senses, their age, and the runner’s medical condition to explore their capa- bilities and develop their physical and mental abilities (ITRA, 2022). Recently, trail running has become more accessible to non-professional athletes, despite its demands and requirements for training, work schedule and personal life sacrifices (Rochat et al., 2017). In addition, several events or changing dynamics may occur during a race (Suter et al., 2020) due to the many variable exigencies of this sport (Scheer et al., 2019), including a combination of the runner’s own physical, tactical, and psychological characteristics that can lead to success or failure (Liew et al., 2019). In trail running, as Gameiro et al. 1203 in all other sports, performance is influenced by many physical characteristics like toughness, strength, speed, agility, and by the athletes’ psychological status such as motivation, concentration, and mental ability (Namli & Demir, 2019). In this regard, toughness may be needed for correct decision making under stress-difficult conditions, staying calm under pressure, and overcoming troubles that are intertwined with per- formance; this makes mental toughness a critical competency for achieving best success (Namli & Demir, 2019). A characteristic of activity endurance is surely the demand for mental and physical reintegration following fatigue induced by the performance effort (Diotaiuti et al., 2021). Especially since performance can be challenged by unpredictable occurrences like weather conditions, mechanical failures, pain, or discomfort related to physical and mental states (Diotaiuti et al., 2021), it is imperative that athletes prepare in a com- prehensive but focused way (Scheer et al., 2019). The influence of psychological factors in endurance sports is of undeniable importance; and, while there has been growing interest in studying these factors, and their impact on performance, they have been poorly analyzed to date (M´endez-Alonso et al., 2021), leaving us with insufficient data of this kind (Scheer et al., 2019). Different types of mental resources are apt to be important for an athlete’s preparations for the exigencies and challenges of this sport (Moreira et al., 2021). Mental Toughness Mental toughness (MT) is a psychological construct with demonstrated importance in sport (Brace et al., 2020; Cooper et al., 2020; Jones & Parker, 2019) and there has been increased research and practice interest on mental toughness in sport and exercise psychology over the last two decades (Gucciardi, 2017). MT has been associated with beneficial behaviors and better sport performance outcomes (Stamatis et al., 2020), though there is still inadequate consensus regarding its conceptualization (Hardy et al., 2014). Amongst several definitions of MT, Gucciardi et al. (2015) defined it as “a personal capacity to produce consistently high levels of subjective (e.g., personal goals or strivings) or objective performance (e.g., sales, race time, GPA) despite everyday challenges and stressors as well as significant adversities.” Zeiger & Zeiger (2018) defined MT as “a state-like psychological resource that is purposeful, flexible and efficient in nature for the enactment and maintenance of goal-directed pursuits,” such that MT enables striving (e.g., effort), surviving (e.g., coping) and thriving (e.g., performing) (Jones & Parker, 2019). MT shares conceptual space and has been cor- related with various other psychological constructs (Brace et al., 2020; Jones & Parker, 2019) like optimism, pessimism, coping, youth experiences, achievement goals and sport motivation, developmental assets, and stress appraisal (Jones & Parker, 2019), as well as resilience, self-belief, and emotional intelligence (Nicholls et al., 2015). Nicholls et al. (2015) suggested that it might be the presence of other psychological constructs (rather than coping alone) that permits mentally tough individuals to dis- tinguish themselves under stressful circumstances (Nicholls et al., 2015). Evidence 1204 Perceptual and Motor Skills 130(3) suggests that MT is a multifaceted construct that supports performance excellence (Cowden et al., 2017; Shaari et al., 2020) irrespective of the type, direction, and degree of demands experienced (Cowden et al., 2017). Additionally, MT is considered central to sport performance (Anthony et al., 2016) and it is an important prerequisite for a higher sustained athletic performance (Goddard et al., 2019). MT is classified as a critical factor for success (Souter et al., 2018) because of its role in increasing a controlled adaptative response to positive and negative pressures, situations, and events (Cowden et al., 2017). The implicit association between MT and success or better performance has gained increasing attention among leading researchers, especially those who have conducted retrospective studies of elite athletes (Cowden et al., 2017). Despite the promising potential of developing MT, there is no evidence to date for any specific approach to training this attribute. Nevertheless, Stamatis et al. (2020) sug- gested that due to sport-specific differences in MT, interventions to enhance MTshould consider the cultural and contextual attributes of each sport. Additionally, some in- vestigators have begun to quantify the predictive role of MT for competitive (Cowden, 2016) and noncompetitive (Gucciardi et al., 2016) performance indicators. These results have not been discussed in detail and there is still a scarcity of empirical research regarding this conceptual association between MT and athletic performance. More specifically, it is still unclear whether MT is noticeable in athletes with better per- formance, higher achievement, or successful outcomes, or whether MT is more apt to be evident in association with non-performance factors like resilience, self-belief, and emotional intelligence (Cowden et al., 2017). Cowden et al. (2017) highlighted the importance of conducting statistically based studies that more accurately control the relationship between levels of MT on performance outcomes. Stamatis et al. (2020) highlighted the importance of conducting studies with objective indicators of athletic performance to provide evidence for MT interventions. Resilience As previously noted, another construct that is frequently mentioned alongside MT is resilience (Liew et al., 2019). Resilience in sport has aroused interest due to athletes’ needs to use and optimize a range of mental qualities to protect them from or to overcome stressors, adversities, and failures (Galli & Gonzalez, 2015; Sarkar & Fletcher, 2014). Sport is an excellent context in which to study resilience for cop- ing with unexpected adversities like serious injuries, or stressors of a psychosocial nature (e.g., losing a match, maladaptive interactions with coach), a physiological nature (e.g., high training loads) or a non-typical circumstance (e.g., pandemic situ- ations) (den Hartigh et al., 2022). Additionally, athletes submit themselves continu- ously to evaluative environments with high consequences (e.g., winning or losing) (Galli & Gonzalez, 2015). Thus, to distinguish psychological resilience from other forms of resilience, Fletcher and Sarkar (2012) first defined resilience as “the role of mental processes and behavior in promoting personal assets and protecting an indi- vidual from the potential negative effect of stressors.” Additionally, resilience has been Gameiro et al. 1205 seen as the individual’s capacity to recognize personal limits, and to accept and go further to face difficulties with optimism (Diotaiuti et al., 2021). Roebuck et al. (2020) conceptualized resilience as a “high-order trait that reflects the ability of a person to maintain normal psychological functioning in the setting of a stressor.” There still exists controversy between the definition and the concept of resilience in research and sport practice, with some investigators noting continued difficulties operationalizing and measuring resilience, both in this context and in non-sport settings (Galli & Gonzalez, 2015). Resilience has been studied, with multidisciplinary interest, as a dynamic process with a personalized perspective (den Hartigh et al., 2022), leading sport scientists to adopt one of two possible approaches (Galli & Gonzalez, 2015). On one hand, they have examined the psychosocial factors that predict performance following an initial failure on the same task, seeing resilience as a coping behavior characterized by performing successfully after an initial failure or trying after-the-fact to identify how to enhance resilience/performance (Galli & Gonzalez, 2015). From the other per- spective, resilience has been investigated by attempting to understand the thoughts, beliefs, emotions, and behaviors of athletes who demonstrate a capacity to overcome adversity in sport (Galli & Gonzalez, 2015). Nonetheless, the athlete’s personal qualities of resilience (Galli & Gonzalez, 2015) namely positivity, determination, competitiveness, commitment, maturity, persistence, passion for the sport (Sarkar & Fletcher, 2014) as well as social and environmental contexts appear to have an im- portant role (Galli & Gonzalez, 2015). In qualitative studies, researchers have focused on psychological elements that protect athletes against stressors; they have emphasized positive personality, motivation, confidence, focus and perceived social support as main protective factors (Sarkar & Fletcher, 2014). Additionally, positive personality traits, and, more specifically, adaptative perfectionism, optimism and competitiveness have been linked with dealing with stressors (den Hartigh et al., 2022). Consequently, studying this dynamic process of bouncing back to normal functioning following adversity and noting specifically how long this takes have been seen as keys to un- derstanding how to prevent performance depletion and contend with psychological or physical stresses (den Hartigh et al., 2022). Utilizing biopsychosocial data, it is possible to determine warning signs of a loss of resiliency (den Hartigh et al., 2022). Current Integrative Research As noted above, resilience can interact with other psychological constructs in sport like hardiness, coping, MT, and post traumatic growth (Galli & Gonzalez, 2015). Thus, Gucciardi et al. (2008) and Nicholls et al. (2015) focused on the relationship between resilience and MT in sport and, more recently, Moreira et al. (2021) showed that, while MT incorporates characteristics of resilience and hardiness, it also enables one to thrive in situations where there are positive effects and perceived positive pressure. Resilience is very similar to MT (Cowden et al., 2016) and both constructs have very often been cited together (Liew et al., 2019; Gucciardi et al., 2009). Like MT, conceptualization, operationalization, and measurement of resilience have not yet generated a consensus 1206 Perceptual and Motor Skills 130(3) (Cowden et al., 2016). While they share similar conceptual space, their relationships to each other have not been explicitly clarified (Nicholls et al., 2015). Nonetheless, it is important to clarify some dissimilarities. Clough et al. (2002) proposed that confidence, a component of MT, is the distinguishing factor between both constructs. Anthony et al. (2020) also underlined the importance of resilience. For instance, Anthony et al. (2020) showed that resilience could act as an emergent outcome, both in terms of individual and group systems, that allows one to bounce back quickly to homeostasis following adversity; whereas MT is only related to the psychological capacity of individuals or resources, acting as a protective factor. Aryanto and Larasati (2020) pointed to the fact that resilient individuals control their behavior by remaining focused, despite iden- tifying and controlling negative influences, while MT individuals can reject outside negative effects to the point that they are unaware of them. MT could be applied to positive circumstances, representing a group of personal attributes that impact the way in which adversity, challenges and goals are surrounded and assessed (Cowden et al., 2016). On the other hand, resilience is mostly associated with negative contexts, including possession of and/or the presence of protective and vulnerability factors, such that resilience influences the relationship between risk and positive adaptation and may influence and be influenced by important attributes outside of the self (e.g., perceived social support) (Cowden et al., 2016). Nonetheless, Cowden et al. (2016) emphasized the studies conducted by Gerber et al. (2013) in which MT was seen as “a resilience resource or protective factor that moderated the association between risk and adaption levels to facilitate positive outcomes.” In line with the inherent growth of interest in sports generally, the growing interest in trail running races by sport professionals has contributed to increased attention to the characteristics that athletes might develop through training and competitions to better their race performances (M´endez-Alonso et al., 2021). In this way, owing to the limitations of physical training, the possibilities and seeming limitlessness of psy- chological training has become a newer crucial focus (Zeiger & Zeiger, 2018). Given the promising results of MT (M´endez-Alonso et al., 2021; Guszkowska & Wojcik, 2021) and resilience (Diotaiuti et al., 2021; Galli & Gonzalez, 2015) constructs in endurance performance sports, endurance runners have begun to try to raise their levels of tenacity, determination, and tolerance of negative affect (e.g., resilience traits) (Diotaiuti et al., 2021). Nonetheless, quantitative studies analyzing these constructs through validated measurement instruments in endurance sports like trail running are still scarce. In the present study, our objective was to analyze these psychological variables among trail runners and to study the association between these variables and athletic performance as measured by the International Trail Running Association Index (ITRA Index). More specifically, we studied the relationship between MT and resilience and athletic performance, hoping to contribute a quantitative study that would lend a better understanding of these two constructs, outlining their similarities and differences. Gameiro et al. 1207 Method Participants A total of 307 Portuguese trail runners (60 female, 247 male), aged between 20 to 66 years (M age = 41.98; SD = 7.74) participated in this study Data were collected in accordance with the Helsinki Declaration World Medical Association (2013); and the Ethics Committee of the Polytechnic of Leiria gave its approval for its implementation (CE/IPLEIRIA/26/2021). Potential participants were contacted though the Portuguese Trail Running Association platform. Additionally, social network pages, forums, and individual teams were contacted. A Google form was sent to all potential participants of this study. To be included in this study, potential participants needed to obtain a valid ITRA profile. This means that athletes needed to have finished at least one certified ITRA race in the previous 36 months. Before data collection, potential participants were informed about the main objectives of the study, and the estimated time to complete the assessment battery (approximately 12 minutes). Before completing the questionnaires, participants had to complete a check box, ensuring that they understood the aims of the study, and that they consented to participate. Participants were thanked for their contribution, but no compensation was provided. Measures Mental Toughness. In this study, we used the Sport Mental Toughness Questionnaire (SMTQ) developed by Sheard et al. (2009) in a Portuguese version by Fonseca (2012). The SMTQ was established to ascertain the athlete´s mental toughness levels. This questionnaire is comprised of 14-items that are answered on a four-point Likert-type scale, ranging from “not at all true” [1] to “very true” [4]. The items are grouped into three factors: Confidence – six items (e.g., “I interpret threats as positive opportuni- ties.”); Constancy – four items (e.g., “I give up in difficult situations.”) and Control – four items (e.g., “I am overcome by self-doubt.”). Previous studies supported the validity and reliability of this measure among athletes (Sheard et al., 2009; Zeiger & Zeiger, 2018). For the present study, we utilized the three factor constructs (Confidence, Constancy, and Control) representing mental toughness (Miçoo˘gulları, 2017). Resilience. The 10 item Connor-Davidson Resilience Scale (CD-RISC-10) developed by Connor and Davidson (2003) is available in a Portuguese version adapted by Almeida et al. (2020). The CD-RISC-10 measures resilience in the general population, and it has been tested among athletes with adequate psychometric properties (Galli & Gonzalez, 2015). This scale is comprised of 10 items, that are answered on a scale from 0 (not true at all) to 4 (true nearly all the time). The items are grouped into a single factor representing the level of the respondent’s resilience. Several studies (e.g., Nartova- Bochaver et al., 2021) have supported validity and reliability of this measure in different countries. 1208 Perceptual and Motor Skills 130(3) Performance. The International Trail Running Association (ITRA) Performance Index is a tool for ranking athletes based on their sport performance level, and it has been used to compare the speed of trail runners around the world (M´endez-Alonso et al., 2021). The scale has a maximum of 1000 points corresponding to the best possible perfor- mance (world record performance for that distance), considering the athlete’s finish time and specific race characteristics, namely distance, elevation gain/loss and average altitude (ITRA, 2022). This scale utilizes an indirect normative comparison method (based on the statistical analysis of a database of more than 5.3 million individual results), with the technical characteristics of the terrain also considered (ITRA, 2022). A general performance index is calculated from the weighted mean of the five best race results over the previous 36 months (permitting reliable statistical calculations and giving the possibility of an injured athlete continuing to appear), regardless of the distance of each race (ITRA, 2022). By finishing a certified ITRA race (from a minimum of 2 km to more than 190 km), the result will appear in the ITRA Performance Index (ITRA, 2022). The number of races finished by participants in this study varied between 1 to 69 (M = 15.94; SD = 13.01). In addition, the average distance (in ki- lometers) made in certified trail races was 971.12 and the average distance of positive ascents was 46.45 km, while the average distance of negative ascents was 45.90 km. Statistical Analysis Descriptive statistics included means (and standard deviations) and bivariate corre- lations for studied variables. In addition, a two-step maximum likelihood approach (ML) (Kline, 2016) in AMOS 27.0 was performed. A confirmatory factor analysis (CFA) first tested the psychometric properties of the measurement model, including its convergent and discriminant validity and composite reliability (Hair et al., 2019). Convergent validity was assessed via average variance extracted (AVE), considering values higher than or equal .50 as adequate (Fornell & Larcker, 1981). Discriminant validity was estimated through the square correlations between factors, and it was considered adjusted when the square correlations were below the AVE of each factor (Hair et al., 2019). Additionally, the internal consistency of each of the latent variables under study was calculated, from the composite reliability (Raykov, 1997), assuming as a cut-off value for adequacy coefficients, ≥.70 (Hair et al., 2019; Raykov, 1997). Next, a structural model was established to test the hypothesis. The model´s fit for both the measurement model and the structural model was observed through the traditional goodness-of-fit indexes. Specifically, we used the Comparative Fit Index (CFI) and Tucker-Lewis Index (TLI) and the absolutes of the Standardized Root Mean Residual (SRMR) and Root Mean Square Error of Approximation (RMSEA) with a confidence interval (CI 90%), as recommended by several authors (Kline, 2016; Hair et al., 2019; Byrne, 2016; Marsh et al., 2004) and with the following adopted cut-off values: CFI and TLI ≥ .90; RMSEA and SRMR ≤ .08 (Kline, 2016; Hair et al., 2019; Byrne, 2016; Marsh et al., 2004). Standardized direct and indirect effects on the dependent variable were also analyzed. The significance of direct and indirect effects was analyzed using a Gameiro et al. 1209 bootstrap resampling procedure (1000 bootstrap samples), through a 95% CI. The indirect effect was considered significant (≤0.05) if the 95% CI did not include zero (Williams & Mackinnon, 2008). We chose to consider confidence intervals rather than the probability of significance (p-value) due to recent evidence of mediation without a significant relationship between variables (Hayes, 2018). Results Preliminary Analysis The Full Information Robust Maximum Likelihood (FIML) was used to handle the small amount of missing data at the item level (missing at random = 4%) as proposed by Enders (2010). Additionally, no outliers (univariate and multivariate) were identified. Item-level descriptive statistics indicated no deviations from univariate normality because skewness and kurtosis assumptions of the data distribution were comprised between
2 and +2 and
7 and +7, respectively (Hair et al., 2019). Mardia’s coefficient for multivariate kurtosis (47.83) exceeded expected values (5.0) for the assumption of multivariate normality (Byrne, 2016). Therefore, the Bollen-Stine bootstrap on 2000 samples was employed for subse- quent analysis (Nevitt & Hancock, 2001). Finally, the collinearity diagnosis was checked using variance inflation factor (VIF) and tolerance tests and these results revealed values between 1.56 to 1.88 for VIF and 0.38 to 0.77 for the tolerance test, demonstrating acceptable conditions for regression analysis (Kline, 2016; Hair et al, 2019). Then, descriptive statistics and bivariate correlations were calculated for all variables under analysis. Descriptive statistics showed that the participants presented scores above the midpoint for the measures of MT and resilience. Looking at bivariate correlations, positive and significant associations were found between all variables under analysis, specifically including these observations: (a) MT was positively associated with both resilience and performance; and (b) resilience was positively associated with per- formance. The measurement model including the factors MT, resilience, and perfor- mance variables, exhibited an adequate fit to the data: χ2 = 150.01 (74); BS-p = .003; CFI = .953; TLI = .942; RMSEA = .058 90% (.045, .071) and SRMR = .042, since CFI and TLI were above and SRMR and RMSEAwere below the previous reported cut-off values. As seen by the CR coefficients, each factor showed scores above the cutoff (>.70), revealing adequate internal consistency. Based on the results of the measurement model and reliability analysis, convergent and discriminant validity were calculated. Con- vergent validity was achieved, since the AVE scores were above the acceptable cut-off values, as seen in Table 1. According to the squared correlations and AVE scores, all factors demonstrated adequate discriminant validity since the squared correlations of each latent variable were lower than the AVE scores in each latent variable. The results provide preliminary support to conduct SEM analysis and examine the direct between 1210 Perceptual and Motor Skills 130(3) mental toughness and resilience with performance and indirect effect between mental toughness and performance via resilience. The results from the SEM analysis showed that the structural model provided an acceptable fit to the data, with χ2 = 150.01 (75); BS-p = .003; CFI = .954; TLI = .944; RMSEA = .057 90% (.044, .070) and SRMR = .042 since CFI and TLI were above and SRMR and RMSEA were below the previous reported cut-off values. Standardized direct effects revealed positive associations (see Figure 1) between variables. Spe- cifically, MT displayed a significant association with resilience; and resilience was significantly associated with performance. The indirect regression paths showed that MT was positively associated with performance, with resilience a possible mediator (β = .09 IC = .010, .168; p = .02) in this relationship. In total, considering direct and indirect effects the model explained 21% of the performance variance in trail runners. Discussion In this study we aimed to analyze the associations across mental toughness, resilience, and athletic performance in Portuguese trail runners. Overall, our hypothesis was confirmed and will be discussed below in the context of existing literature. The positive association found between MT and resilience was in line with previous research in which Nicholls et al. (2015) found significant associations between MT and resilience, concluding that “mentally tough athletes are able to excel under pressure.” MT, in the presence of other psychological constructs, can distinguish performance (not just coping) under stressful circumstances. Similarly, Gucciardi et al. (2008) found, through qualitative investigation, that “mentally tough athletes are resilient,” and Cowden et al. (2016) found a strong positive association in competitive South African Tennis players between MT and resilience, further affirming the conceptual similarities between the two constructs. Our standardized direct effects analysis also revealed a positive association between resilience and performance, in line with Hosseini and Besharat’s (2010) finding that resilience predicted athletes’ sporting achievement, psychological well-being and distress. Table 1. Descriptive Statistics, Bivariate Correlation, Convergent and Discriminant Validity, and Composite Reliability of the Participants Responses to MT and Resilience Measures and Their Trail Running Performance. M SD 1 2 3 AVE CR 1. MT 3.09 .48 1 — .67 .82 2. Resilience 3.12 .54 .62** 1 — .56 .84 3. Performance 650.91 543.87 .23** .13** 1 — — Note. MT = mental toughness; M = mean; SD = standard deviation; AVE = average variance extracted; CR = composite reliability. Gameiro et al. 1211 Our indirect regression analysis revealed a positive association between MT and athletic performance, with resilience considered as a possible mediator. This result empathizes the close relationship between these two constructs and gives new input to the study of psychological constructs in athletic performance. Despite the close re- lationship between resilience and MT, resilience retains its uniqueness, including the conditions of positive adaptation and adversity that Galli and Gonzalez (2015) de- scribed. The decisive role of resilience in facing severe adversities that can occur outside of the sport context (Cowden et al., 2016) seems to give resilience a mediation role for “the enactment and maintenance of goal-directed pursuits” as Zeiger and Zeiger (2018) suggested. Notwithstanding limited knowledge of the conceptual association between MT and athletic performance, due to scarce other empirical results to date, our results are in line with reviews by Cowden et al., (2017) and Guszkowska and Wojcik (2021) who reported a positive correlation between MT and sporting performance across different sports, regardless of the participants’ age, gender, or skill levels. This result emphasizes that MT is a multifaceted construct that is a central prerequisite to excellent sport performance (Cowden et al., 2017). Our results are also in line with M´endez-Alonso et al. (2021) who found that MTand resilience are psychological predictors of success in ultra-trail runners (i.e., there were better classification and race times in athletes with higher values of MT and resilience). M´endez-Alonso et al. (2021) also highlighted that ultra-trail showed higher values of MT and resilience than either athletes in other sports or sedentary individuals, and this was also previously reported by Galli and Gonzalez (2015). Furthermore, past studies found higher levels of psychological constructs among endurance athletes, especially MT (Aryanto & Larasati, 2020) and resilience (Roebuck et al., 2020), raising the question of whether these attributes are intrinsic characteristics or consequences of training and/or competing. M´endez-Alonso et al. (2021) stated that each race works as a means of training these psychological factors, increasing an athlete’s MTand resilience. This training aspect is probably due to the unpredictable conditions and specific characteristics of each trail running race. This is a characteristic that sports profes- sionals should consider in assessments of MT and resilience at pre-testing, during a race, and post- race. Although, Brace et al. (2020) demonstrated high levels of MT in a sample of elite level ultra-marathon runners, they did not find performance effects during a race among those with higher MT values, possibly suggesting that other Figure 1. Standardized Direct Effects. 1212 Perceptual and Motor Skills 130(3) factors can influence performance more. These results reinforce the important role of possible indirect associations between psychological constructs in performance var- iance. Endurance sports demand mental and physical integration owing to the impact of fatigue prompted by the high effort sport performance requires (Diotaiuti et al., 2021). Specifically, in trail running, it is necessary to make a permanent adjustment to different conditions (e.g., elevation and climate make it imperative that the athlete control various pace, nutrition, posture, loneliness, and fatigue). However, these athletes present a more effective way of contending with unpleasant situations and perform more effectively, acting with self-awareness of their own effort and fatigue (Guszkowska & Wojcik, 2021). This fact makes trail running different from others sports or even from other kinds of running races, adding importance to the use of permanent psychological control adaptations to reach sports goals, highlighting the general importance of mental characteristics and the primordial role of psychological preparation (MT and resilience training) in sports (Guszkowska & Wojcik, 2021). Performance in these races is multifactorial, with many factors involved (Diotaiuti et al., 2021), and our results show the importance of just these two constructs, while other psychological social, and physiological factors may explain remaining variance in athletic performance. Notwithstanding these other unknowns, the complexity of trail running makes it an ideal target for future sports research, and for endurance sports. Our findings show that MT, mediated by resilience, explained 21% of the total performance variance in trail runners, providing a new perspective of the possible importance to intervention training in psychological constructs for these kind of en- durance efforts. Sports professionals should be aware that mental training should be an integrant part of a holistic psychosocial program (Fletcher & Sarkar, 2016). Limitations and Directions for Future Research Although previous studies analyzed the association between MT, resilience, and performance, this is the first study to analyze simultaneously MT, resilience, and performance in trail runners. While the present study contributed new insights into these associated psychological constructs, some limitations should be addressed. First, we used a cross-sectional design that precludes us from determining causal rela- tionships between these variables. Experimental studies are needed to examine the effects of mental toughness and resilience on athletic performance. Second, our data were limited to a Portuguese trail running sample and may not generalize well to athletes from other cultures and/or sport contexts. Third we collected these data only in one moment, namely at the middle of the competitive calendar, and there may be variations in these results if data were collected at other times. Future investigators should analyze MT and resilience during a competitive race calendar, taking into consideration the possible changes that may occur during a season. Additionally, the association between these and other constructs normally related, should be studied directly and indirectly, using not only cross-sectional but experi- mental designs. To obtain greater generalizability, future studies might be applied Gameiro et al. 1213 toward participants in other sports and cultures. Finally, investigations of whether or how these psychological constructs might be trained and enhanced should be part of future research. Conclusion There is a growing appreciation for the importance that psychological preparation should assume in endurance sports and, specifically, in trail running, characterized by unpredictable and stressful conditions that makes variance in performance excellence multifactorial. These athletes must make permanent adjustments to different conditions higher performance has been associated with higher values of such psychological constructs as MT and resilience. Our findings showed that these constructs explained 21% of trail runners’ performance variance when considering their direct and indirect effects. This, highlights the close relationship between these two constructs and their joint influence on performance variation, contributing to a holistic view of athletic performance. Psychological training in endurance sports practice, especially including MT and resilience training, would seem advantageous obtaining better performances. These training programs should consider the cultural and contextual attributes of each sport and social and the athletes’ environmental context. Declaration of Conflicts of Interest The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. Funding The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This project was supported by national funds through the Portuguese Foundation for Science and Technology, I.P., under the project UID04045/2020 Data Availability The datasets presented in this article are not readily available because the data are under confidentiality requirements. Requests to access the datasets should be directed to Diogo Monteiro ([email protected]) Institutional Review Board Statement The study was conducted in accordance with the Helsinki Declaration. Ethical approval was obtained by the Ethical Committee of the Polytechnic of Leiria before data collection (reference number: (CE/IPLEIRIA/26/2021). ORCID iDs Raul Antunes  https://orcid.org/0000-0002-5485-9430 1214 Perceptual and Motor Skills 130(3) Rui Matos  https://orcid.org/0000-0002-2034-0585 Diogo Monteiro  https://orcid.org/0000-0002-7179-6814 References Almeida, M. H., Dias, S., Xavier, M., & Torgal, J. (2020). Validação Exploratória e Confirmatória da Escala de Resiliˆencia Connor-Davidson (CD-RISC-10) numa Amostra de Inscritos em Centros de Emprego. Acta Medica Portuguesa, 33(2), 124–132. https://doi.org/10.20344/ amp.12243 Anthony, D. R., Gordon, S., & Gucciardi, D. F. (2020). A qualitative exploration of mentally tough behaviour in Australian football. Journal of Sports Sciences, 38(3), 308–319. https:// doi.org/10.1080/02640414.2019.1698002 Anthony, D. R., Gucciardi, D. F., & Gordon, S. (2016). A meta-study of qualitative research on mental toughness development. International Review of Sport and Exercise Psychology, 9(1), 160–190. https://doi.org/10.1080/1750984X.2016.1146787 Aryanto, D., & Larasati, A. (2020). Factors influencing mental toughness. In 5th ASEAN conference on psychology, counselling and humanities (ACPCH 2019), Gelugor, Malaysia, Atlantis Press, 395: 307-309. Bostancı, ¨O., Karaduman, E., & Mayda, M. (2019). Investigation of self confidence levels in elite extreme athletes. Physical Education of Students, 23(3), 106–111. https://doi.org/10.15561/ 20755279.2019.0301 Brace, A. W., George, K., & Lovell, G. P. (2020). Mental toughness and self-efficacy of elite ultra-marathon runners. PloS One, 15(11), Article e0241284. https://doi.org/10.1371/ journal.pone.0241284 Byrne, B. M. (2016). Strucutural equation modeling with AMOS: Basic concepts applications and programming (2nd ed.). Routledge. https://doi.org/10.4324/9780203805534 Clough, P. J., Earle, K., & Sewell, D. (2002) Mental Toughness: The Concept and Its Mea- surement. In I. Cockerill (Ed.), Solutions in Sport Psychology (pp. 32–43). London: Thomson. Connor, K. M., & Davidson, J. R. T. (2003). Development of a new resilience scale: The connor- davidson resilience scale (CD-RISC). Depression and Anxiety, 18(2), 76–82. https://doi.org/ 10.1002/da.10113 Cooper, K. B., Wilson, M. R., & Jones, M. I. (2020). A 3000-mile tour of mental toughness: An autoethnographic exploration of mental toughness intra-individual variability in endurance sport. International Journal of Sport and Exercise Psychology, 18(5), 607–621. https://doi. org/10.1080/1612197X.2018.1549583 Cowden, R. G. (2016). Competitive performance correlates of mental toughness in Tennis: A preliminary analysis. Perceptual and Motor Skills, 123(1), 341–360. https://doi.org/10. 1177/0031512516659902 Cowden, R. G., Meyer-Weitz, A., & Oppong Asante, K. (2016). Mental toughness in competitive Tennis: Relationships with resilience and stress. Frontiers in Psychology, 7, Article 320. https://doi.org/10.3389/fpsyg.2016.00320 Gameiro et al. 1215 Cowden, R. (2017). Mental toughness and success in sport: A review and prospect. The Open Sports Sciences Journal, 10(1):1–14. https://doi.org/10.2174/1875399X01710010001 Den Hartigh, R. J. R., Meerhoff, L. R. A., Van Yperen, N. W., Neumann, N. D., Brauers, J. J., Frencken, W. G. P., Emerencia, A., Hill, Y., Platvoet, S., Atzmueller, M., Lemmink, K. A. P. M., & Brink, M. S. (2022). Resilience in sports: A multidisciplinary, dynamic, and personalized perspective. International Review of Sport and Exercise Psychology. Ad- vanced online publication. https://doi.org/10.1080/1750984X.2022.2039749 Diotaiuti, P., Corrado, S., Mancone, S., & Falese, L. (2020). Resilience in the endurance runner: The role of self-regulatory modes and basic psychological needs. Frontiers in Psychology, 11, Article, 558287. https://doi.org/10.3389/fpsyg.2020.558287 Easthope, C. S., Nosaka, K., Caillaud, C., Vercruyssen, F., Louis, J., & Brisswalter, J. (2014). Reproducibility of performance and fatigue in trail running. Journal of Science and Medicine in Sport, 17(2), 207–211. https://doi.org/10.1016/j.jsams.2013.03.009 Enders, C. (2010). Applied missing data analysis. The Guilford Press. Fletcher, D., & Sarkar, M. (2012). A grounded theory of psychological resilience in Olympic champions. Psychology of Sport and Exercise, 13(5), 669–678. https://doi.org/10.1016/j. psychsport.2012.04.007 Fletcher, D., & Sarkar, M. (2016). Mental fortitude training: An evidence-based approach to developing psychological resilience for sustained success. Journal of Sport Psychology in Action, 7(3), 135–157. https://doi.org/10.1080/21520704.2016.1255496 Fonseca, C. (2012). Estudo da validade do question´ario de robustez mental no desporto. [Dissertação de Mestrado]. Universidade T´ecnica de Lisboa. Faculdade de Motricidade Humana. Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. Journal of Marketing Research, 18(1), 39–50. https://doi. org/10.2307/3151312 Gajardo-Burgos, R., Monrroy-Uarac, M., Barr´ıa-Pailaquil´en, R. M., Norambuena-Noches, Y., van Rensburg, D. C. J., Bascour-Sandoval, C., & Besomi, M. (2021). Frequency of injury and illness in the final 4 weeks before a trail running competition. International Journal of Environmental Research and Public Health, 18(10), Article 5431. https://doi.org/10.3390/ ijerph18105431 Galli, N., & Gonzalez, S. P. (2015). Psychological resilience in sport: A review of the literature and implications for research and practice. International Journal of Sport and Exercise Psychology, 13(3), 243–257. https://doi.org/10.1080/1612197X.2014.946947 Gerber, M., Kalak, N., Lemola, S., Clough, P. J., Perry, J. L., Pühse, U., Elliot, C., Holsboer- Trachsler, E., & Brand, S. (2013). Are adolescents with high mental toughness levels more resilient against stress? Stress and Health: Journal of the International Society for the Investigation of Stress, 29(2), 164–171. https://doi.org/10.1002/smi.2447 Goddard, K., Roberts, C. M., Anderson, L., Woodford, L., & Byron-Daniel, J. (2019). Mental toughness and associated personality characteristics of marathon des sables athletes. Frontiers in Psychology, 10, Article 2259. https://doi.org/10.3389/fpsyg.2019.02259 Groslambert, A., Baron, B., Ouvrard, T., Desmoulins, L., Lacroix, E., Gimenez, P., Grosprˆetre, S., & Grappe, F. (2020). Influencing factors of pacing variations and performance in a 44- 1216 Perceptual and Motor Skills 130(3) kilometer mountain trail race. Advances in Physical Education, 10(02), 81–96. https://doi. org/10.4236/ape.2020.102008 Gucciardi, D. F. (2017). Mental toughness: Progress and prospects. Current Opinion in Psy- chology, 16, 17–23. https://doi.org/10.1016/j.copsyc.2017.03.010 Gucciardi, D. F., Gordon, S., & Dimmock, J. A. (2008). Towards an understanding of mental toughness in Australian football. Journal of Applied Sport Psychology, 20(3), 261–281. https://doi.org/10.1080/10413200801998556 Gucciardi, D. F., Gordon, S., & Dimmock, J. A. (2009). Development and preliminary validation of a mental toughness inventory for Australian football. Psychology of Sport and Exercise, 10(1), 201–209. https://doi.org/10.1016/j.psychsport.2008.07.011 Gucciardi, D. F., Hanton, S., Gordon, S., Mallett, C. J., & Temby, P. (2015). The concept of mental toughness: Tests of dimensionality, nomological network, and traitness. Journal of Personality, 83(1), 26–44. https://doi.org/10.1111/jopy.12079 Gucciardi, D. F., Peeling, P., Ducker, K. J., & Dawson, B. (2016). When the going gets tough: Mental toughness and its relationship with behavioural perseverance. Journal of Science and Medicine in Sport, 19(1), 81–86. https://doi.org/10.1016/j.jsams.2014.12.005 Guszkowska, M., & Wojcik, K. (2021). Effect of mental toughness on sporting performance: Review of studies. Baltic Journal of Health and Physical Activity, Suppl 1(2), 1–12. https:// doi.org/10.29359/bjhpa.2021.suppl.2.01 Hair, J., Black, W., Babin, B., & Anderson, R. (2019). Multivariate data analysis (8th ed.). Pearson Educational. Hardy, L., Bell, J., & Beattie, S. (2014). A neuropsychological model of mentally tough behavior. Journal of Personality, 82(1), 69–81. https://doi.org/10.1111/jopy.12034 Hayes, A. F. (2018). Introduction to mediation, moderation, and conditional process analysis: A regression-based approach (2nd ed.). Guilford Press. Hosseini, S. A., & Besharat, M. A. (2010). Relation of resilience whit sport achievement and mental health in a sample of athletes. Procedia - Social and Behavioral Sciences, 5, 633–638. https://doi.org/10.1016/j.sbspro.2010.07.156 ITRA – International Trail Running Association. Discover Trail Running. (Accessed 4th January 2022),https://itra.run/About/DiscoverTrailRunning Jones, M. I., & Parker, J. K. (2019). An analysis of the size and direction of the association between mental toughness and olympic distance personal best triathlon times. Journal of Sport and Health Science, 8(1), 71–76. https://doi.org/10.1016/j.jshs.2017.05.005 Kline, R. B. (2016). Principles and practice of structural equation modeling (3rd ed.). Guilford Press. Liew, G. C., Kuan, G., Chin, N. S., & Hashim, H. A. (2019). Mental toughness in sport. German Journal of Exercise and Sport Research, 49(4), 381–394. https://doi.org/10.1007/s12662- 019-00603-3 Malliaropoulos, N., Mertyri, D., & Tsaklis, P. (2015). Prevalence of injury in ultra trail running. Human Movement, 16(2), 55–59. https://doi.org/10.1515/humo-2015-0026 Marsh, H. W., Hau, K.-T., & Wen, Z. (2004). In search of golden rules: Comment on hypothesis- testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu Gameiro et al. 1217 and Bentler’s (1999) findings. Structural Equation Modeling: A Multidisciplinary Journal, 11(3), 320–341. https://doi.org/10.1207/s15328007sem1103_2 M´endez-Alonso, D., Prieto-Saborit, J. A., Bahamonde, J. R., & Jim´enez-Arber´as, E. (2021). Influence of psychological factors on the success of the ultra-trail runner. International Journal of Environmental Research and Public Health, 18(5), 2704. https://doi.org/10. 3390/ijerph18052704 Miçoo˘gulları, B. (2017) The sports mental toughness questionnaire (SMTQ): a psychometric evaluation of the Turkish version, 11(2), 90–100. Studia Sportiva. https://doi.org/10.5817/ StS2017-2-9 Moreira, C. R., Codonhato, R., & Fiorese, L. (2021). Transcultural adaptation and psychometric proprieties of the mental toughness inventory for Brazilian athletes. Frontiers in Psychology, 12, Article 663382. https://doi.org/10.3389/fpsyg.2021.663382 Namli, S., & Demir, G. T. (2019). Is mental toughness in elite athletes a predictor of moral disengagement in sports? Journal of Education and Learning, 8(6), 56. https://doi.org/10. 5539/jel.v8n6p56 Nartova-Bochaver, S., Korneev, A., & Bochaver, K. (2021). Validation of the 10-item connor- davidson resilience scale: The case of Russian youth. Frontiers in Psychiatry, 12, Article 611026. https://doi.org/10.3389/fpsyt.2021.611026 Nevitt, J., & Hancock, G. R. (2001). Performance of bootstrapping approaches to model test statistics and parameter standard error estimation in structural equation modeling. Structural Equation Modeling: A Multidisciplinary Journal, 8(3), 353–377. https://doi.org/10.1207/ S15328007SEM0803_2 Nicholls, A. R., Perry, J. L., Jones, L., Sanctuary, C., Carson, F., & Clough, P. J. (2015). The mediating role of mental toughness in sport. The Journal of Sports Medicine and Physical Fitness, 55(7–8), 824–834. https://www.minervamedica.it/en/journals/sports-med- physical-fitness/article.php?cod=R40Y2015N07A0824# Perrotin, N., Gardan, N., Lesprillier, A., Le Goff, C., Seigneur, J. M., Abdi, E., Sanudo, B., & Taiar, R. (2021). Biomechanics of trail running performance: Quantification of spatio- temporal parameters by using low cost sensors in ecological conditions. Applied Sciences, 11(5), Article 2093. https://doi.org/10.3390/app11052093 Raykov, T. (1997). Estimation of composite reliability for congeneric measures. Applied Psy- chological Measurement, 21(2), 173–184. https://doi.org/10.1177/01466216970212006 Rochat, N., Hauw, D., Antonini Philippe, R., Crettaz von Roten, F., & Seifert, L. (2017). Comparison of vitality states of finishers and withdrawers in trail running: An enactive and phenomenological perspective. PloS One, 12(3), Article e0173667. https://doi.org/10.1371/ journal.pone.0173667 Roebuck, G. S., Urquhart, D. M., Che, X., Knox, L., Fitzgerald, P. B., Cicuttini, F. M., Lee, S., Segrave, R., & Fitzgibbon, B. M. (2020). Psychological characteristics associated with ultra-marathon running: An exploratory self-report and psychophysiological study. Aus- tralian Journal of Psychology, 72(3), 235–247. https://doi.org/10.1111/ajpy.12287 Sarkar, M., & Fletcher, D. (2014). Psychological resilience in sport performers: A review of stressors and protective factors. Journal of Sports Sciences, 32(15), 1419–1434. https://doi. org/10.1080/02640414.2014.901551 1218 Perceptual and Motor Skills 130(3) Scheer, V., Vieluf, S., Janssen, T. I., & Heitkamp, H. C. (2019). Predicting competition per- formance in short trail running races with lactate thresholds. Journal of Human Kinetics, 69, 159–167. https://doi.org/10.2478/hukin-2019-0092 Shaari, J. S., Hooi, L. B., Radzi, J. A., Siswantoyo, Sumartininggsih, S., & Zainal Abidin, N. E. (2020). Mental toughness measurement for psychological skills training intervention, translation and adaptation. International Journal of Mechanical and Production Engi- neering Research and Development, 10(3), 1231–1242. https://doi.org/10.24247/ ijmperdjun2020107 Sheard, M., Golby, J., & van Wersch, A. (2009). Progress toward construct validation of the sports mental toughness questionnaire (SMTQ). European Journal of Psychological As- sessment, 25(3), 186–193. https://doi.org/10.1027/1015-5759.25.3.186 Souter, G., Lewis, R., & Serrant, L. (2018). Men, mental health and elite sport: A narrative review. Sports Medicine - Open, 4(1), 57. https://doi.org/10.1186/s40798-018-0175-7 Spittler, J., & Oberle, L. (2019). Current trends in ultramarathon running. Current Sports Medicine Reports, 18(11), 387–393. https://doi.org/10.1249/JSR.0000000000000654 Stamatis, A., Grandjean, P., Morgan, G., Padgett, R. N., Cowden, R., & Koutakis, P. (2020). Developing and training mental toughness in sport: A systematic review and meta-analysis of observational studies and pre-test and post-test experiments. BMJ Open Sport and Exercise Medicine, 6(1), Article e000747. https://doi.org/10.1136/bmjsem-2020-000747 Suter, D., Sousa, C. V., Hill, L., Scheer, V., Nikolaidis, P. T., & Knechtle, B. (2020). Even pacing is associated with faster finishing times in ultramarathon distance trail running-the “Ultra- Trail du Mont Blanc” 2008-2019. International Journal of Environmental Research and Public Health, 17(19), 7074. https://doi.org/10.3390/ijerph17197074 Viljoen, C., Janse van Rensburg, D. C. C., van Mechelen, W., Verhagen, E., Silva, B., Scheer, V., Besomi, M., Gajardo-Burgos, R., Matos, S., Schoeman, M., Jansen van Rensburg, A., van Dyk, N., Scheepers, S., & Botha, T. (2022). Trail running injury risk factors: A living systematic review. British Journal of Sports Medicine, 56(10), 577–587. https://doi.org/10. 1136/bjsports-2021-104858 Wa´skiewicz, Z., Nikolaidis, P. T., Chalabaev, A., Rosemann, T., & Knechtle, B. (2019). Mo- tivation in ultra-marathon runners. Psychology Research and Behavior Management, 12, 31–37. https://doi.org/10.2147/PRBM.S189061 Williams, J., & Mackinnon, D. P. (2008). Resampling and distribution of the product methods for testing indirect effects in complex models. Structural Equation Modeling: A Multidisci- plinary Journal, 15(1), 23–51. https://doi.org/10.1080/10705510701758166 World Medical Association (2013). World medical association declaration of Helsinki: Ethical principles for medical research involving human subjects. JAMA, 310(20), 2191–2194. https://doi.org/10.1001/jama.2013.281053 Zeiger, J. S., & Zeiger, R. S. (2018). Mental toughness latent profiles in endurance athletes. PloS One, 13(2), Article e0193071. https://doi.org/10.1371/journal.pone.0193071 Gameiro et al. 1219 Author Biographies Nuno Gameiro holds a master’s degree in Exercise Prescription and Health Promotion and is a researcher at ESECS (Polytechnique of Leiria) and CIEQV. He has a degree in Physiotherapy, with specialities in Sport and Elite Sport. His research focuses on Sport and Exercise Psychology and Performance. Other areas of research interest include Biomechanics and Sports injuries. Filipe Rodrigues is an adjunct lecturer at the ESECS (Polytechnique of Leiria) and researcher at the CIEQV. His research focuses on motivational and cognitive theories to understand health-related behavior change in diverse domains. Other areas of research interest include health psychology, interpersonal behaviors, coaching andneuropsychology. Nuno Amaro holds a PhD in Sport Sciences and is an adjunct professor at the ESECS -Polytechnique of Leiria and is an integrated member of the Life Quality Research Centre (CIEQV), Polytechnic of Leiria. He has a degree and master’s degree in Sport Sciences. His research activity focuses, mostly, on Strength & Conditioning and swimming performance; Motricity for Children and Tripela. Raul Antunes holds a PhD in Sports Science and is a professor at the ESECS -Polytechnique of Leiria and is an integrated member of the Life Quality Research Centre (CIEQV), Polytechnic of Leiria. He has a degree and master’s degree in Sport. His research activity focuses, mostly, on the sport and exercise psychology, specifically the analysis of the de- terminants of physical activity and its consequences on quality of life and well-being (emotional and cognitive component). Rui Matos holds a PhD in Human Movement Sciences and has a tenure position as a Professor at the Polytechnic of Leiria. He is the vice-Coordinator of the Quality of Life Research Centre. He has a degree in Physical Education Teaching and a master’s degree in children Motor Development. His academic/research field is linked to Motor Development, Motor Competence, and their relationship with quality of life. Miguel Jacinto is a PhD student in Sport Science, Adapted Physical Activity area at the Faculty of Sport Sciences and Physical Education-University of Coimbra. Is also a research scholarship student at the Leiria branch of the Life Quality Research Center and invited professor at ESECS-polytechnic of Leiria. His research relates the variables of physical exercise, people with disabilities and quality of life. Diogo Monteiro holds a PhD in Sports Science and is a professor at the ESECS -Poly- technique of Leiria and is an integrated member of the Research Center in Sport, Health and Human Development. He has a degree in Sport and Exercise Psychology and master’s degree in Sport and Exercise Psychology. His academic/research field is linked to motivational determinants in sport and exercise and behavioral change, with a special focus on sedentary behavior, physical activity, healthy lifestyles, well-being, exercise adherence, sport dropout, and persistence. 1220 Perceptual and Motor Skills 130(3)

Mental Toughness and Resilience in Trail Runner's Performance.

03-24-2023

Gameiro, Nuno,Rodrigues, Filipe,Antunes, Raúl,Matos, Rui,Amaro, Nuno,Jacinto, Miguel,Monteiro, Diogo

eng

PMC5474287

Research Article Developing a Low-Cost Force Treadmill via Dynamic Modeling Chih-Yuan Hong,1 Lan-Yuen Guo,2 Rong Song,3 Mark L. Nagurka,4 Jia-Li Sung,1 and Chen-Wen Yen1,5 1Department of Mechanical and Electromechanical Engineering, National Sun Yat-sen University, Kaohsiung, Taiwan 2Department of Sports Medicine, Kaohsiung Medical University, Kaohsiung, Taiwan 3School of Engineering, Sun Yat-sen University, Guangzhou, China 4Department of Mechanical Engineering, Marquette University, Milwaukee, WI, USA 5Department of Physical Therapy, Kaohsiung Medical University, Kaohsiung, Taiwan Correspondence should be addressed to Chen-Wen Yen; [email protected] Received 2 January 2017; Accepted 2 April 2017; Published 4 June 2017 Academic Editor: Emiliano Schena Copyright © 2017 Chih-Yuan Hong et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. By incorporating force transducers into treadmills, force platform-instrumented treadmills (commonly called force treadmills) can collect large amounts of gait data and enable the ground reaction force (GRF) to be calculated. However, the high cost of force treadmills has limited their adoption. This paper proposes a low-cost force treadmill system with force sensors installed underneath a standard exercise treadmill. It identifies and compensates for the force transmission dynamics from the actual GRF applied on the treadmill track surface to the force transmitted to the force sensors underneath the treadmill body. This study also proposes a testing procedure to assess the GRF measurement accuracy of force treadmills. Using this procedure in estimating the GRF of “walk-on-the-spot motion,” it was found that the total harmonic distortion of the tested force treadmill system was about 1.69%, demonstrating the effectiveness of the approach. 1. Introduction In walking and running, the inertial force acting on the human body is equal to the sum of the ground reaction force (GRF) exerted by the ground on the foot and the gravita- tional force of the body weight. Many important gait param- eters can be derived from the GRF. These include temporal features such as the time instants of heel strike and toe off and the time durations of stance and swing phases as well as the step frequency. As a result, GRF can provide important information about gait behavior. GRF data have been used to investigate gait symmetry [1], calculate leg stiffness [2], quantify impacts [3], under- stand propulsion and braking [4], compute muscle forces, joint forces and moments [5, 6], explain running economy [7, 8], and top running speeds [9]. GRF data have also been used to assess the effects of health-related conditions that can influence gait. These conditions include knee replacement [10], hip arthroplasty [11], aging effect [12], knee arthrosis [13], Parkinson’s disease [14], peripheral arterial disease [15, 16], patellofemoral pain syndrome [17], osteoarthritis [18], cerebral palsy [19], multiple sclerosis [20], lower extremity muscle fatigue [21], stroke [22, 23], weighted walking [24], and hemiplegia [25]. To measure GRF during gait, most previous studies have relied on a force platform-embedded walkway. The most common configuration of a force platform consists of a metal plate mounted on load cells that give an electri- cal output proportional to the force applied to the plate. Typically, only a few steps of gait data are collected in each experimental trial. The necessity of proper foot placement on the force platform also complicates the experimental process. In addition, intentional behavior is likely to change the GRF and alter the gait pattern. This problem is par- ticularly pronounced in testing individuals who exhibit gait difficulty. It is very difficult to perform constant Hindawi Journal of Healthcare Engineering Volume 2017, Article ID 9875471, 9 pages https://doi.org/10.1155/2017/9875471 speed walking or running studies using floor-mounted force platforms. Since it has been found that the differences between treadmill and overground locomotion are small [26–28] and can be negligible after only a few minutes of treadmill- walking practice [29], treadmills have been employed exten- sively to study gait. To enhance the utility of treadmills, force platform-instrumented treadmills (commonly called force treadmills) have been used to quickly and continuously collect large amounts of GRF data during gait. These force treadmills offer several advantages over conventional walkway-based measurement systems. First, force treadmills reduce the time and space requirements substantially. Second, with a tread- mill, controlling the speed of locomotion becomes a straight- forward task. Third, body weight support modules can be added to the treadmills to ensure safety. Fourth, it is easier to integrate complementary measurement devices (such as elec- tromyographic systems and oxygen consumption-measuring instruments) in the treadmill design in comparison to using a walkway-based system. Treadmill training is frequently prescribed as a treatment option for patients with gait abnormalities. By using a force treadmill to quantitatively analyze gait patterns and detect gait abnormalities, medical therapists can adjust the intensity of treadmill training on an individual basis. In addition, pre- vious studies have shown that the feedback of auditory, vibrotactile, and visual gait information can alter or improve gait features such as walking speed [30, 31], gait coordination [32], trunk sway [33], stride length [31], hip mechanics [34], cadence [31], step length symmetry [35], knee movement [36], gait cycle length [37], duration of gait [37], and swing phase speed [37]. With the capability of generating many important gait features, force treadmills represent an ideal platform for implementing such biofeedback systems. Based on the location of the force transducers, force treadmills can be divided into two categories: direct measure- ment force treadmills (DMFTs) and indirect measurement force treadmills (IMFTs). By incorporating force platforms internally, DMFTs can measure GRF directly without con- sidering the structural dynamics of the treadmill body [38–42]. Typically, DMFTs were built by installing force platforms under the track surface of the treadmill. This con- ceptually simple setup, however, requires complex mechan- ical design and a tedious assembly and calibration process in order to prevent erroneous force components generated by the moving parts (the motor and mechanism) of the treadmill [39, 42–44]. In contrast, by mounting the treadmill on top of force transducers, IMFTs simplify the mechanical design of force treadmills [45–49]. The friction forces generated by the mov- ing components (such as belt, motor, and rollers) of the treadmill become internal forces and are not measured by the force sensors attached externally to the treadmill frame. The tradeoff of such a simplified design is the potential infi- delity of the GRF measurements. Unless the treadmill frame can be made rigid, forces transmitted to the force transducers of the IMFT are generally not the same as the actual GRF applied to the treadmill track surface. To resolve this prob- lem, current IMFTs are designed to possess a very high natural frequency to prevent the GRF from exciting the dynamics of the treadmill structure. This high structural natural frequency specification can only be achieved when the treadmill body is light and rigid. As a result, one needs to use low-density, high-stiffness materials in a specially designed mechanical structure for the treadmill frame. These requirements inevitably increase the manufacturing cost. The other reason for the high price of current force treadmill systems is that, due to their special design require- ments, these treadmills are typically custom made or manu- factured in very small quantities. In comparison, standard exercise treadmills are mass produced and, as such, much more affordable. Considering the utility of force treadmills and the fact that their high cost has limited their adoption, the goal of this study is to introduce a systematic approach to convert a stan- dard exercise treadmill into a force treadmill via a straightfor- ward system identification method. A distinct feature of the proposed approach is that it relaxes the high structural natu- ral frequency requirement for the treadmill frame. As a result, the construction cost of the force treadmills can be reduced considerably. This work also proposes an experi- mental procedure to assess the GRF measurement accuracy of force treadmills. 2. Methods 2.1. The Dynamic Modeling Method. This subsection iden- tifies the dynamic specifications that need to be satisfied by conventional IMFTs. Furthermore, it introduces the basic idea of the proposed approach by addressing the problems caused by such specifications. Denoting the force applied to the IMFT track surface as x t and the force transmitted to the force transducers placed under the IMFT body as y t , this study assumes that the GRF transmission dynamics of transmitting the force from x t to y t can be modeled as a linear time-invariant single-input single-output (SISO) system. In particular, with x t as the input and y t as the output, the GRF transmission dynamics of the IMFT are represented by the following frequency-domain transfer function H f : H f = Y f X f , 1 where f denotes the frequency (Hz) and X f and Y f rep- resent the Fourier transforms of x t and y t , respectively. Since an IMFT can only measure y t , to ensure that the actual GRF signal x t can be approximated closely by y t , conventional IMFTs were designed to behave like a distor- tionless transmission system in the low-frequency range. An SISO system is a distortionless transmission system if it satisfies the following condition: y t = kx t − td , 2 where t is the time variable, k is an arbitrary constant, and td is the time delay of this distortionless transmission system. Therefore, the transmission is considered to be distortionless if the input and the output have identical wave shapes with a 2 Journal of Healthcare Engineering proportionality constant k. A delayed output that retains the input waveform is also considered distortionless. These spec- ifications of distortionless transmission can be converted into the frequency domain by taking the Fourier transform of (2) which yields Y f = kX f e−j2πf td 3 Therefore, the corresponding amplitude response is H f = k, 4 and the phase response is ∠H f = −td2πf 5 Hence, a distortionless transmission system must have a constant amplitude response and a phase response that declines linearly with frequency f . By modeling the GRF transmission dynamics of an IMFT as a linear time- invariant second-order system with natural frequency f n and a damping ratio ξ, its amplitude and phase responses can be expressed, respectively, as [50] H u = 1 1 − u2 2 + 4ξ2u2 6 and ∠H u = −tan−1 2ξu 1 − u2 , 7 where the dimensionless frequency variable u = f /f n. If f is much smaller than f n, the amplitude and phase responses of this standard second-order system can be approximated by H u ≈ 1 8 and ∠H u ≈ − 2ξu 9 Therefore, a linear time-invariant second-order system behaves like a distortionless transmission system when f << f n. This is the reason why the structural natural fre- quency of a conventional IMFT needs to be considerably higher than the bandwidth of the GRF signal. Experimental studies have found that, on average, 99% of the vertical direction GRF signal power was contained under 12.75 Hz when walking at a comfortable speed [12]. Never- theless, human GRF contain frequency components as high as 60 Hz for walking [51] and 100 Hz for running [52]. To quantitatively demonstrate the importance of high structural natural frequency of the treadmill structure, we assume that the natural frequency f n to be 45 Hz (Kram et al. [45] indi- cated that the vertical direction structural natural frequencies of the six force treadmills that they reviewed are all lower than 45 Hz). With f =12.75 Hz, the corresponding dimen- sionless frequency u is 12.75/45 ≈0.283. By using (6) with u= 0.283 and ξ= 0, it can be shown that H u ≈ 1 087 which represents an 8.7% deviation from the desired specification of H u = 1. To reduce such a deviation, two more recently developed IMFTs increase their structural natural frequen- cies to 160 Hz [45] and 219 Hz [46], respectively. At f =12.75 Hz, the corresponding H u improves to 1.006 and 1.003, respectively. When modeled as a linear time-invariant second order system, it is well known that the structural natural frequency f n of the treadmill can be determined from f n = 1 2π k m , 10 where k (N/m) is the stiffness and m (kg) is the mass of the treadmill. Clearly, f n can be increased by reducing the weight of the treadmill. This is the reason why previous force tread- mills often removed parts such as side handrails, front rails, and the control panel to make the treadmill lighter. However, these changes also degraded the functionality and safety of the treadmill system. The natural frequency f n can also be increased by using higher strength materials to increase the stiffness. The lightweight and high strength material require- ments inevitably increase the cost of the treadmill. To relax the high natural frequency requirement for the IMFT structure, this study tries to compensate for the effect of the GRF transmission dynamics of the treadmill by identi- fying its transfer function model. In particular, by applying an excitation force x t to the treadmill track surface and measuring the resulting x t and y t , we can identify the transfer function from x t to y t from (1). Using the inverse dynamic model of the identified transfer function, we can then estimate the actual GRF from xc t = F−1 ̂H −1 f Y f , 11 where ̂H f represents the identified transfer function of the treadmill GRF transmission dynamics. In the remaining parts of the manuscript, x t , y t , and xc t will be referred to as the actual, the uncompensated, and the compensated GRF signals, respectively. The experimental procedure for implementing the proposed approach will be described in the following subsection. 2.2. The Experimental Procedure. Figure 1 illustrates the con- figuration of the experimental system which consists of two subsystems, namely, a force treadmill and a force platform. The treadmill (7355, Fit Plus, Taiwan) has bed dimensions of 1.5 m length and 70 cm width. The speed control system provides a range from 0 to 22 km/hr with a minimum incre- ment of 0.1 km/hr. The weight of the treadmill is 150 kg. To convert this standard exercise treadmill into a force treadmill, this study installed four load cells (Sensolink SLP-1 with maximum capacity of 100 kg) into the legs that support the treadmill body. The four circles shown in Figure 2 specify the location of the force transducers. As shown in Figure 1, to measure the actual GRF signal x t , a force platform is placed at the center of the treadmill track surface. Similar to a commercially available force plat- form, the force platform built here is a rectangular plate with force transducers located at its four corners. The force 3 Journal of Healthcare Engineering treadmill and the force platform employed in this study use the same load cell unit. The size of the platform is 40 cm by 40 cm. We have carefully compared the measurements obtained by this force platform and a commercial force platform (Kistler 9286AA) to verify comparable repeatability and accuracy. After amplification, analog voltage signals obtained by the four load cells of the force treadmill are converted to dig- ital signals via a four channels, 24 bit DAQ (data acquisition) card (NI 9234). The voltages generated by the load cells of the force platform are also processed by an independent but identical set of voltage amplifiers and a DAQ card. The digitized force signals were sent to a PC using a USB chassis (NI cDAQ-9174) and low pass filtered by a distortionless phase 20th-order Butterworth filter with a cutoff frequency of 150 Hz. The sampling frequency was set to 1024 Hz. The experimental system used the graphical programming environment NI LabVIEW (National Instrument, Austin, TX, USA) for performing system control, signal processing, and graphical user interface (GUI) functions. The experimental work consists of two phases. The first phase identifies the GRF transmission dynamics of the tread- mill by finding its transfer function model. A dead blow ham- mer with a nonmarring head was used to strike the center of the force platform. By measuring the resulting x t and y t with the force platform and force treadmill, respectively, the transfer function of the GRF transmission dynamics was determined from (1). The second phase of the exper- imental work was to assess the accuracy of the estimated GRF signals. The test input signals were produced by asking ten male subjects (age 24.20 ±3.29 years, weight 73.09 ±15.42 N) to walk “on the spot” for 20 seconds when standing on the force platform which was placed on the center of the treadmill track surface. The fidelity of the estimated GRF was evaluated quantitatively by its total harmonic distortion (THD), defined as THD = ∫ T 0 x t − x t 2 dt ∫ T 0 x2 t dt , 12 where ̂x t and x t represent the estimated and actual GRF signals, respectively. In this study, the duration for each of the walk-on-the-spot tests was T = 20s. Note that, by defining the distorted signal as the difference between ̂x t and x(t), the THD represents the ratio of the energy of the GRF esti- mation error signal to the energy of the actual GRF signal. 3. Results and Discussions Figure 3 plots the amplitude spectrum of the identified trans- fer function of the treadmill GRF transmission dynamics obtained in the first phase of the experimental study. As shown in Figure 3, the amplitude response of the identified transfer function is very different from that of a distortionless transmission system. Specifically, its amplitude response is relatively flat only in the low-frequency range of 0 to 5 Hz and becomes highly oscillatory in the higher frequency region. This clearly reveals the importance of compensating for the effect of GRF transmission dynamics to improve GRF measurement accuracy for an IMFT. To demonstrate the efficacy of the proposed approach, based on the data obtained in the second phase of the Treadmill system Amplifcation circuits Data acquisition Force platform Voltage amplifer Voltage amplifer Treadmill load cells DAQ USB chassis DAQ PC Figure 1: Configuration of the experimental system. Rear 750 mm Right Lef 1500 mm 2100 mm 880 mm Front 1650 mm Track surface Motor & transmission system Figure 2: Top view of the treadmill surface. 4 Journal of Healthcare Engineering experimental study, the THDs were computed by using the uncompensated and the compensated GRF signals as the estimated GRF signal. The resulting THDs for the 20 partic- ipants of the walk-on-the-spot experiment are plotted in Figure 4. As shown in Figure 4, the THDs obtained by the compensated GRF are considerably smaller than the THDs of the uncompensated GRF. In particular, for the uncompen- sated GRF, the mean and standard deviation of the THDs are 9.64% and 6.3%, respectively. In comparison, by using the compensated GRF as the estimated GRF, the proposed approach reduces the mean of the THDs to 1.69% and the standard deviation of the THDs to 1.38%. Such improve- ments can also be observed from Figure 5 that plots the time responses of the actual, the compensated, and the uncom- pensated GRF signals for a typical 2 s period of the walk- on-the-spot experiment. As shown by Figure 5, the time responses of the actual and compensated GRF signals are relatively close. In contrast, the uncompensated GRF signal tends to oscillate around the actual GRF signal and often overshoots the actual GRF signal, particularly at the sharp corners of the actual GRF time response profile. To compare the efficacy of the proposed approach to the conventional IMFT design, the IMFT was modeled as a second-order linear system whose frequency spectra can be represented by (6) and (7). With the actual GRF signal of the walk-on-the-spot experiment as the input and the corre- sponding output of the second order linear system of (6) and (7) as the estimated GRF, the THDs can be determined for the IMFT mathematical model. The resulting mean THDs of the twenty participants are plotted in Figure 6 as a function of f n for ξ =0.01, 0.05, and 0.1. As expected, THD decreases with the increasing f n. Since Kram et al. [45] indicated that the vertical direction natural frequencies of the six force treadmills that they reviewed are all lower than 45 Hz, we first inspect the THDs for ωn= 45 Hz. Based on the results of Figure 6, when ωn = 45Hz, the THDs are 7.07%, 3.59%, and 2.07% for ξ =0.05, 0.1, and 0.2, respectively. Note that the solid line of Figure 6 corresponds to the mean THD obtained by the proposed approach which is 1.69%. In order to reduce THD to be smaller than 1.69%, the natural frequency has to increase to 88Hz, 65 Hz, and 60 Hz for ξ =0.01, 0.05, and 0.1, respectively. As shown in Figure 6, THDs vary from person to per- son. Although accurate prediction of individually dependent THDs does not seem possible, it is still valuable to under- stand factors that can influence the accuracy of the esti- mated GRF. Considering the potential influences of noise on the system identification process in the high-frequency range, it is hypothesized that the THD is positively corre- lated with the bandwidth of the GRF signals. Due to unavailability of the actual GRF signal during the normal treadmill operations, this study investigates the association between the bandwidth of the compensated GRF signal and its THD. In particular, by specifying the 98% bandwidth as the portion of the signal spectrum in the frequency domain which contains 98% of the signal energy, Figure 7 depicts the scatter diagram of THD versus 98% bandwidth of the compensated GRF signal. With a p value of 1.41 ×10−7, the value of the corresponding correlation coefficient is 0.891. Such a strong correlation demonstrates that the inaccuracy of the compensated GRF signal increases with its bandwidth. To the best of our knowledge, such an association between GRF frequency content and the GRF measurement accuracy has never been studied systematically. Such knowledge can help us estimate the degree of inaccu- racy of the GRF measurements in dealing with GRF signals with different frequency contents. The experimental results presented in this work demon- strate the feasibility of the proposed approach. However, the success of the approach relies on the linear system assumption of (1). For a poorly constructed treadmill, this assumption of linearity may not be valid. It is also possible that the GRF transmission dynamics of the treadmill are too complex to be compensated accurately. Therefore, choos- ing a treadmill with a relatively solid structure should be an important consideration in implementing the proposed approach. Since increasing the structural natural frequency of an IMFT tends to increase its cost and an IMFT with poor rigid- ity may be too difficult to be compensated accurately, a pos- sible compromise between cost and performance of an IMFT is to build a relatively rigid but inexpensive IMFT with a less than ideal structural natural frequency and then improve its GRF measurement accuracy with the proposed approach. A possible future work is to systematically study the tradeoffs between the cost and accuracy for such a hybrid hardware- software force treadmill design. For the existing IMFTs, the proposed approach can be used to examine the frequency responses of their GRF transmission dynamics. This can help us better understand the dynamic behaviors of the existing IMFTs since their frequency responses have rarely been investigated systematically. The proposed approach can also be applied to quantify the accuracy of the existing IMFTs by computing the distortions of their GRF signals. If neces- sary, the proposed approach can also be used to improve their GRF measurement accuracy by compensating the effect of the GRF transmission dynamics. 0 50 100 150 ‒60 ‒50 ‒40 ‒30 ‒20 ‒10 0 10 20 30 Frequency (Hz) Amplitude (dB) Figure 3: The amplitude spectrum of the identified transfer function. 5 Journal of Healthcare Engineering 4. Conclusion The goal of this work is to reduce the cost and extend the applicability of force platform-instrumented treadmills (force treadmills). By identifying the influences of treadmill structural dynamics on ground reaction force (GRF) mea- surements and by installing force transducers underneath the treadmill body, a standard exercise treadmill can be con- verted to a force treadmill. A previous work showed that treadmill structures need to be highly rigid in order to ensure that the resultant force measured by these force sensors closely approximates the actual GRF applied to the track sur- face. The high cost for building such treadmills has limited their adoption. To relax the requirement of high structural rigidity, the proposed approach adopts a system identification approach to model the GRF transmission dynamics from the treadmill track surface to the force sensors underneath the treadmill. Actual GRF Uncompensated GRF Compensated GRF 700 710 720 730 740 750 760 770 780 790 800 GRF signals (N) 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 Time (s) Figure 5: GRF signal time responses of a typical 2-second period for the walk-on-the-spot experiment. 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 0 5 10 15 20 25 30 Participants Harmonic distortion (%) Compensated GRF Uncompensated GRF Figure 4: Total harmonic distortions of the estimated GRF signals for the walk-on-the-spot experiment. 6 Journal of Healthcare Engineering By using the inverse dynamic model of the identified transfer function, the approach can be used to estimate the actual GRF by compensating for the effect of the GRF transmission dynamics of the treadmill. In addition to developing a compensation method to enhance the GRF measurement accuracy, this work intro- duces an experimental procedure to assess the accuracy of the estimated GRF signals. As shown by the test results obtained from the walk-on-the-spot experiment, the mean total harmonic distortion of the estimated GRF signals is only 1.69%. This study also found that the inaccuracy of the esti- mated GRF signal increases with its bandwidth. In addition to converting standard exercise treadmills to force treadmills, the proposed approach can be used to assess and improve the GRF measurement accuracy of existing force treadmills. Conflicts of Interest The authors declare that they have no competing interest. Acknowledgments This research was partly supported by the Ministry of Science and Technology in Taiwan, under Grant MOST 104-2221-E-110-010. 0 1 2 3 4 5 6 7 5 10 15 20 25 30 35 Harmonic distortion (%) Bandwidth (Hz) Figure 7: The scatter diagram of the total harmonic distortion and the 98% bandwidth of the compensated GRF. fn (Hz) 30 40 50 60 80 90 100 0.1 1 1.69 10 100 Harmonic distortion (%) Proposed approach 70
훇: 0.1
훇: 0.05
훇: 0.01 Figure 6: The total harmonic distortion of a second-order IMFT model for the walk-on-the-spot experiment. 7 Journal of Healthcare Engineering References [1] G. Giakas and V. Baltzopoulos, “Time and frequency domain analysis of ground reaction forces during walking: an investi- gation of variability and symmetry,” Gait & Posture, vol. 5, no. 3, pp. 189–197, 1997. [2] D. J. Dutto and G. A. Smith, “Changes in spring-mass charac- teristics during treadmill running to exhaustion,” Medicine and Science in Sports and Exercise, vol. 34, no. 8, pp. 1324– 1331, 2002. [3] J. S. Gottschall and R. Kram, “Ground reaction forces during downhill and uphill running,” Journal of Biomechanics, vol. 38, no. 3, pp. 445–452, 2005. [4] E. J. Goldberg, S. A. Kautz, and R. R. Neptune, “Can treadmill walking be used to assess propulsion generation?” Journal of Biomechanics, vol. 41, no. 8, pp. 1805–1808, 2008. [5] C. L. Vaughan, B. L. Davis, and J. C. O’connor, Dynamics of Human Gait, Human Kinetics Publishers, Champaign, IL, USA, 1992. [6] A. Erdemir, S. McLean, W. Herzog, and A. J. van den Bogert, “Model-based estimation of muscle forces exerted during movements,” Clinical Biomechanics, vol. 22, no. 2, pp. 131– 154, 2007. [7] G. D. Heise and P. E. Martin, “Are variations in running economy in humans associated with ground reaction force characteristics?” European Journal of Applied Physiology, vol. 84, no. 5, pp. 438–442, 2001. [8] S. Wright and P. G. Weyand, “The application of ground force explains the energetic cost of running backward and forward,” The Journal of Experimental Biology, vol. 204, no. 10, pp. 1805–1815, 2001. [9] P. G. Weyand, D. B. Sternlight, M. J. Bellizzi, and S. Wright, “Faster top running speeds are achieved with greater ground forces not more rapid leg movements,” Journal of Applied Physiology, vol. 89, no. 5, pp. 1991–1999, 2000. [10] F. Verdini, T. Leo, S. Fioretti, M. G. Benedetti, F. Catani, and S. Giannini, “Analysis of ground reaction forces by means of wavelet transform,” Clinical Biomechanics, vol. 15, no. 8, pp. 607–610, 2000. [11] J. L. McCrory, S. C. White, and R. M. Lifeso, “Vertical ground reaction forces: objective measures of gait following hip arthroplasty,” Gait & Posture, vol. 14, no. 2, pp. 104–109, 2001. [12] N. Stergiou, G. Giakas, J. B. Byrne, and V. Pomeroy, “Frequency domain characteristics of ground reaction forces during walking of young and elderly females,” Clinical Biome- chanics, vol. 17, no. 8, pp. 615–617, 2002. [13] H. Gök, S. Ergin, and G. Yavuzer, “Kinetic and kinematic char- acteristics of gait in patients with medial knee arthrosis,” Acta Orthopaedica Scandinavica, vol. 73, no. 6, pp. 647–652, 2002. [14] R. Bartsch, M. Plotnik, J. W. Kantelhardt, S. Havlin, N. Giladi, and J. M. Hausdorff, “Fluctuation and synchronization of gait intervals and gait force profiles distinguish stages of Parkin- son's disease,” Physica a, vol. 383, no. 2, pp. 455–465, 2007. [15] M. M. Scott-Pandorf, N. Stergiou, J. M. Johanning, L. Robin- son, T. G. Lynch, and I. I. Pipinos, “Peripheral arterial disease affects ground reaction forces during walking,” Journal of Vascular Surgery, vol. 46, no. 3, pp. 491–499, 2007. [16] D. McGrath, T. N. Judkins, I. I. Pipinos, J. M. Johanning, and S. A. Myers, “Peripheral arterial disease affects the fre- quency response of ground reaction forces during walking,” Clinical Biomechanics, vol. 27, no. 10, pp. 1058–1063, 2012. [17] P. Levinger and W. Gilleard, “Tibia and rearfoot motion and ground reaction forces in subjects with patellofemoral pain syndrome during walking,” Gait & Posture, vol. 25, no. 1, pp. 2–8, 2007. [18] S. P. Moustakidis, J. B. Theocharis, and G. Giakas, “A fuzzy decision tree-based SVM classifier for assessing oste- oarthritis severity using ground reaction force measure- ments,” Medical Engineering & Physics, vol. 32, no. 10, pp. 1145–1160, 2010. [19] S. E. Williams, S. Gibbs, C. B. Meadows, and R. J. Abboud, “Classification of the reduced vertical component of the ground reaction force in late stance in cerebral palsy gait,” Gait & Posture, vol. 34, no. 3, pp. 370–373, 2011. [20] S. R. Wurdeman, J. M. Huisinga, M. Filipi, and N. Stergiou, “Multiple sclerosis affects the frequency content in the vertical ground reaction forces during walking,” Clinical Biomechan- ics, vol. 26, no. 2, pp. 207–212, 2011. [21] A. A. Zadpoor and A. A. Nikooyan, “The effects of lower extremity muscle fatigue on the vertical ground reaction force: a meta-analysis,” Proceedings of the Institution of Mechanical Engineers. Part H, vol. 226, no. 8, pp. 579–588, 2012. [22] C. Y. Chen, P. W. Hong, C. L. Chen et al., “Ground reaction force patterns in stroke patients with various degrees of motor recovery determined by plantar dynamic analysis,” Chang Gung Medical Journal, vol. 30, no. 1, pp. 62–72, 2007. [23] E. Szczerbik, M. Krawczyk, and M. Syczewska, “Ground reac- tion force analysed with correlation coefficient matrix in group of stroke patients,” Acta of Bioengineering and Biomechanics, vol. 16, no. 2, pp. 3–9, 2014. [24] C. R. James, L. T. Atkins, H. S. Yang, J. S. Dufek, and B. T. Bates, “Kinematic and ground reaction force accommodation during weighted walking,” Human Movement Science, vol. 44, pp. 327–337, 2015. [25] H. D. Kim, J. G. Kim, D. M. Jeon et al., “Analysis of vertical ground reaction force variables using foot scans in hemiplegic patients,” Annals of Rehabilitation Medicine, vol. 39, no. 3, pp. 409–415, 2015. [26] S. C. White, H. J. Yack, C. A. Tucker, and H. Y. Lin, “Comparison of vertical ground reaction forces during overground and treadmill walking,” Medicine and Science in Sports and Exercise, vol. 30, no. 10, pp. 1537–1542, 1998. [27] P. O. Riley, G. Paolini, U. Della Croce, K. W. Paylo, and D. C. Kerrigan, “A kinematic and kinetic comparison of overground and treadmill walking in healthy subjects,” Gait & Posture, vol. 26, no. 1, pp. 17–24, 2007. [28] B. Kluitenberg, S. W. Bredeweg, S. Zijlstra, W. Zijlstra, and I. Buist, “Comparison of vertical ground reaction forces during overground and treadmill running. A validation study,” BMC Musculoskeletal Disorders, vol. 13, no. 1, p. 235, 2012. [29] M. Van de Putte, N. Hagemeister, N. St-Onge, G. Parent, and J. A. de Guise, “Habituation to treadmill walking,” Bio-Medical Materials and Engineering, vol. 16, no. 1, pp. 43–52, 2006. [30] I. Lim, E. van Wegen, C. de Goede et al., “Effects of external rhythmical cueing on gait in patients with Parkinson's disease: a systematic review,” Clinical Rehabilitation, vol. 19, no. 7, pp. 695–713, 2005. [31] A. J. Espay, Y. Baram, A. K. Dwivedi et al., “At-home train- ing with closed-loop augmented-reality cueing device for improving gait in patients with Parkinson disease,” Journal of Rehabilitation Research and Development, vol. 47, no. 6, pp. 573–581, 2010. 8 Journal of Healthcare Engineering [32] M. Roerdink, C. J. Lamoth, G. Kwakkel, P. C. van Wieringen, and P. J. Beek, “Gait coordination after stroke: benefits of acoustically paced treadmill walking,” Physical Therapy, vol. 87, no. 8, pp. 1009–1022, 2007. [33] L. J. Janssen, L. L. Verhoeff, C. G. Horlings, and J. H. Allum, “Directional effects of biofeedback on trunk sway during gait tasks in healthy young subjects,” Gait & Posture, vol. 29, no. 4, pp. 575–581, 2009. [34] B. Noehren, J. Scholz, and I. Davis, “The effect of real-time gait retraining on hip kinematics, pain and function in subjects with patellofemoral pain syndrome,” British Journal of Sports Medicine, vol. 45, no. 9, pp. 691–696, 2011. [35] S. J. Kim and H. I. Krebs, “Effects of implicit visual feedback distortion on human gait,” Experimental Brain Research, vol. 218, no. 3, pp. 495–502, 2012. [36] D. Schliessmann, C. Schuld, M. Schneiders et al., “Feasibility of visual instrumented movement feedback therapy in indi- viduals with motor incomplete spinal cord injury walking on a treadmill,” Frontiers in Human Neuroscience, vol. 8, p. 416, 2014. [37] M. Drużbicki, A. Guzik, G. Przysada, A. Kwolek, and A. Brzozowska-Magoń, “Efficacy of gait training using a tread- mill with and without visual biofeedback in patients after stroke: a randomized study,” Journal of Rehabilitation Med- icine, vol. 47, no. 5, pp. 419–425, 2015. [38] E. C. Jansen, D. Vittas, S. Hellberg, and J. Hansen, “Normal gait of young and old men and women. Ground reaction force measurement on a treadmill,” Acta Orthopaedica Scandina- vica, vol. 53, no. 2, pp. 193–196, 1982. [39] R. Kram and A. J. Powell, “A treadmill-mounted force platform,” Journal of Applied Physiology, vol. 67, no. 4, pp. 1692–1698, 1989. [40] J. B. Dingwell, B. L. Davis, and D. M. Frazier, “Use of an instrumented treadmill for real-time gait symmetry evaluation and feedback in normal and trans-tibial amputee subjects,” Prosthetics and Orthotics International, vol. 20, no. 2, pp. 101–110, 1996. [41] E. R. Draper, “A treadmill-based system for measuring symmetry of gait,” Medical Engineering & Physics, vol. 22, no. 3, pp. 215–222, 2000. [42] L. B. Bagesteiro, D. Gould, and D. J. Ewins, “A vertical ground reaction force-measuring treadmill for the analysis of pros- thetic limbs,” Brazilian Journal of Biomedical Engineering, vol. 27, pp. 3–77, 2011. [43] S. H. Collins, P. G. Adamczyk, D. P. Ferris, and A. D. Kuo, “A simple method for calibrating force plates and force treadmills using an instrumented pole,” Gait & Posture, vol. 29, no. 1, pp. 59–64, 2009. [44] L. H. Sloot, H. Houdijk, and J. Harlaar, “A comprehensive protocol to test instrumented treadmills,” Medical Engineering & Physics, vol. 37, no. 6, pp. 610–616, 2015. [45] R. Kram, T. M. Griffin, J. M. Donelan, and Y. H. Chang, “Force treadmill for measuring vertical and horizontal ground reac- tion forces,” Journal of Applied Physiology, vol. 85, no. 2, pp. 764–769, 1998. [46] A. Belli, P. Bui, A. Berger, A. Geyssant, and J. R. Lacour, “A treadmill ergometer for three-dimensional ground reaction forces measurement during walking,” Journal of Biomechanics, vol. 34, no. 1, pp. 105–112, 2001. [47] F. Dierick, M. Penta, D. Renaut, and C. Detrembleur, “A force measuring treadmill in clinical gait analysis,” Gait & Posture, vol. 20, no. 3, pp. 299–303, 2004. [48] G. J. Verkerke, A. L. Hof, W. Zijlstra, W. Ament, and G. Rakhorst, “Determining the center of pressure during walking and running using an instrumented treadmill,” Journal of Biomechanics, vol. 38, no. 9, pp. 1881–1885, 2005. [49] G. Paolini, U. Della Croce, P. O. Riley, F. K. Newton, and K. D. Casey, “Testing of a tri-instrumented-treadmill unit for kinetic analysis of locomotion tasks in static and dynamic loading conditions,” Medical Engineering & Physics, vol. 29, no. 3, pp. 404–411, 2007. [50] R. C. Dorf and R. H. Bishop, Modern Control Systems, Prentice-Hall, Englewood Cliffs, NJ, USA, 12th edition, 2010. [51] G. F. Harris, K. R. Acharya, and R. A. Bachschmidt, “Investiga- tion of spectral content from discrete plantar areas during adult gait: an expansion of rehabilitation technology,” IEEE Transactions on Rehabilitation Engineering, vol. 4, no. 4, pp. 360–374, 1996. [52] M. A. Lafortune, M. J. Lake, and E. Hennig, “Transfer function between tibial acceleration and ground reaction force,” Journal of Biomechanics, vol. 28, no. 1, pp. 113–117, 1995. 9 Journal of Healthcare Engineering

Developing a Low-Cost Force Treadmill via Dynamic Modeling.

06-04-2017

Hong, Chih-Yuan,Guo, Lan-Yuen,Song, Rong,Nagurka, Mark L,Sung, Jia-Li,Yen, Chen-Wen

eng

PMC3289163

  © 2010 by Sports Medicine Research Center, Tehran University of Medical Sciences, All rights reserved.  ORIGINAL ARTICLE 41 A Regression Equation for the Estimation of Maximum Oxygen Uptake in Nepalese Adult Females Pinaki Chatterjee*1,2, PhD; Alok K Banerjee2, PhD; Paulomi Das3,4, PhD; Parimal Debnath4, PhD 1. Department of Physiology, SR College of Dental Sciences and Research, Faridabad, India 2. Department of Physical Education, Kalyani University, Kalyani, West Bengal, India 3. Department of Physiology, Nepalgunj Medical College, Chisapani, Banke, Nepal 4. Department of Physical Education, Jadavpur University, Kolkata, India * Corresponding Author; Address: Reader & HOD, Department of Physiology, S. R. College of Dental Sciences and Research, Faridabad, India E-mail: [email protected] Received: Jun 09, 2009 Final Revision: Jul 21, 2009 Accepted: Oct 21, 2009 Key Words: Cardiovascular fitness; VO2max; Beep test; Indirect measurement; Sedentary Abstract Purpose: Validity of the 20-meter multi stage shuttle run test (20-m MST) has not been studied in Nepalese population. The purpose of this study was to validate the applicability of the 20-m MST in Nepalese adult females. Methods: Forty female college students (age range, 20.42 ~24.75 years) from different colleges of Nepal were recruited for the study. Direct estimation of VO2 max comprised treadmill exercise followed by expired gas analysis by scholander micro-gas analyzer whereas VO2 max was indirectly predicted by the 20-m MST. Results: The difference between the mean (±SD) VO2 max values of direct measurement (VO2 max=32.78 +/-2.88 ml/kg/min) and the 20-m MST (SPVO2 max = 32.53+/-3.36 ml/kg/min) was statistically insignificant (P>0.1). Highly significant correlation (r=0.94, P<0.01) existed between the maximal speed of the 20-m MST and VO2 max. Limits of agreement analysis also suggest that the 20-m MST can be applied for the studied population. Conclusion: The results of limits of agreement analysis suggest that the application of the present form of the 20-m MST may be justified in the studied population. However, for better prediction of VO2 max, a new equation has been computed based on the present data to be used for female college students of Nepal. Asian Journal of Sports Medicine, Vol 1 (No 1), March 2010, Pages: 41-45 INTRODUCTION irect measurement of maximum oxygen uptake (VO2 max) is recognized as the best single index of aerobic fitness[1]. But the test of the direct measurement of cardiorespiratory endurance (VO2 max) itself is difficult, exhausting and often hazardous to perform regardless the type of ergometer used [2]. Since the direct testing procedure is rather complicated on larger populations, several indirect running and walking field tests have been developed. Scientists often calculate VO2 max with indirect protocols[3]. It has been stated that equations for predicting VO2 max indirectly using field tests are very sensitive to populations tested on. Therefore, before applying any D     Chatterjee P, et al Vol 1, No 1, March 2010 42 indirect protocol for prediction of VO2 max, the validity of the test should be established in a particular population. The 20-meter multistage shuttle run test (20-m MST); [4,5], popularly known as Beep test, is often used worldwide for measurement of aerobic capacity [6,7,8,9,10]. But in Nepal, the scientists have not yet used this test. Cooper et al,[11] studied the repeatability and criterion related validity of the 20-m multistage fitness test as a predictor of maximal oxygen uptake in active young men. Suminski et al,[12] established the validity of the 20-m MST for measuring aerobic fitness of Hispanic youth of 10 to 12 years of age. Chatterjee et al, [13,14] studied the validity of 20-m MST in junior Taekwondo players and male university students of India. However, the validity and suitability of this test have not been studied in any Nepalese population until now. Nepal is the neighboring country of India, but a point to be noted here is that there are racial differences as well as differences in habitual activities and that the people of Nepal live at high altitudes. A recent study suggests that gender-distinctive equations provide more accurate prediction of VO2 max from 20-m MST [15]. For this reason, only female adults were recruited as subjects in the mentioned study and not males. Keeping in view all these facts, the present study was undertaken with an objective to assess the applicability of the 20-m MST to predict VO2 max in female college students of Nepal. METHODS AND SUBJECTS Subjects: 40 female college students from different colleges of Nepal were volunteered for the study. The subjects had the mean age of 22.04 yr., height of 157.41 cm, and weight of 49.83 kg. The experimental protocol was fully explained to the participants and they underwent familiarization trial of the beep test few days before the actual test. They had a light breakfast 2-3 hours before the test and refrained from any energetic physical activity for that period. The participants had no history of any major disease and did not follow any physical-conditioning program, except from some recreational sports. The tests were demonstrated to the subjects before actual administration and they agreed to sign a statement of informed consent. All institutional policies concerning the human subjects in research were followed. The tests for all the subjects were done in the morning so that diurnal variation can be avoided, if there was any. Experimental Design: Maximum oxygen consumption of each subject was determined by both indirect and direct methods at an interval of 4 days by random sequencing. Indirect one in the half of the subjects followed the direct method whereas indirect one was followed by direct method in the other half of the subjects. This was done so to avoid any possibility of bias. Subjects were asked to take complete rest at least for half an hour prior to the exercise, so that pulmonary ventilation and pulse rate might come down to a steady state [16]. Prediction of maximum Oxygen uptake capacity by the 20-m MST: Subjects started running back and forth a 20-m course and must touch the 20-m line. The initial speed was 8.5 km/hr. The speed got progressively faster (0.5 km/hr every minute), in accordance with a pace dictated by a sound signal on an audiotape. Several shuttle runs made up each stage. The subjects were instructed to keep pace with the signal for as long as possible. When the subjects could no longer follow the pace, the last stage announced was used to predict the maximal oxygen uptake using the equation of Leger et al.[5]. The equation: Y= -27.4+6.0X, Where Y= VO2 max (ml/kg/min) & X= Maximal shuttle run speed (km/hr) Direct measurement of maximum oxygen uptake capacity: The subjects walked on a treadmill to warm up at a speed of 4 km/hr at a 4.5 inclination for five minutes [17]. Running at a constant speed of 7 km/hr for a maximum duration of 5 min followed this. The inclination gradient was increased successively from 4.5 until the subject was unable to continue the task. In no case did it exceed 7.5 inclinations. The criteria to reach maximum state were exhaustion and withdrawal from running within the scheduled 5-min time period, when the heart rate reached the predicted maximum heart rate and when a further increase of inclination did not bring about any significant rise in oxygen uptake[16].     Vol 1, No 1, March 2010 Validity of 20-m MST 43 Low resistance high velocity Collin’s Triple “J type” plastic valve was used for the collection of gas by open circuit method[16]. The valve was connected with the Douglas bag (150-liter) and the expired gas was collected in the second minute of the exhausting final workload if signs of severe exhaustion supervened. No gas collection was made in the first minute of the workload. The expired gas measured in a wet gasometer (Toshniwal, Germany CAT No. C G 05.10) and the aliquots of gas samples were analyzed in a Scholander micro gas analysis apparatus following the standard procedure [18]. Statistical Analyses: The aired t-test, Pearson’s product moment correlation, linear regression statistics and Bland and Altman approach for limit of agreement were adopted for statistical analyses of the data. Statistical package for Social Sciences (SPSS) MS windows Release 11.5 was used for statistical analyses. To determine validity of the results, repeatability was investigated where 22 subjects performed the test (20-m MST) twice. The results showed non-significant bias between the two applications of the 20-m MST (mean of the difference +/- standard deviation of the difference = -0.13±1.8 ml/kg/min; t = -0.32; P=0.7 with 95% limits of agreement). RESULTS Means and standard deviations of physical characteristics, predicted VO2 max (SPVO2 max) by 20-m MST and directly measured VO2 max of the participants are presented in the table 1. No significant variation was observed (P>0.1) between the values of directly measured and predicted VO2 max. The mean difference between VO2 max and SPVO2 max was 0.27 ml/kg/min with 95% confidence interval of -0.11 to 0.66 ml/kg/min. This indicates that 20-m MST predicted the maximum oxygen uptake capacity between -0.11 to 0.66 ml/kg/min. The standard error when predicting the VO2 max from shuttle run test was 0.53. Analysis of data by Bland and Altman[19] method of approach for limits of agreement between SPVO2 max and VO2 max reveals that limits of agreement are –2.15 to 2.69 (Fig 1). These parameters are small enough for the 20-m MST to be used confidently in place of the direct method. Limits of agreement analysis suggest that application of the present form of the 20-m MST should be justified for the studied population. Highly significant correlation (r=0.94, P<0.01) existed between the maximal speed of the 20-m MST and VO2 max.   DISCUSSION The following equation, derived on the basis of present data will better predict the aerobic fitness in female college students of Nepal: Y= - 15.207 + 4.806 X Where Y= VO2 max (ml/kg/min) and Table 1: Physical parameters, predicted and measured VO2 max of the subjects (N=40) Parameter Minimum Maximum Mean Std. Deviation Age (yr.) 20.42 24.75 22.04 1.14 Height (cm) 154.10 160.30 157.41 1.79 Weight (kg) 42.50 57.20 49.83 4.21 VO2 max‡ (ml/kg/min) 26.90 38.00 32.78 2.88 SPVO2 max* (ml/kg/min) 26.60 38.60 32.53 3.36 Speed (km/hr) 9.00 11.00 9.99 0.56 ‡VO2 max: maximum oxygen uptake / * SPVO2 max: predicted VO2 max      Chatterjee P, et al Vol 1, No 1, March 2010 44 -3 -2 -1 0 1 2 3 0 10 20 30 40 50 Average of VO2 max (ml/kg/min) obtained from two methods Difference between VO2 max and SPVO2 max (ml/kg/min) Mean +2SD = 2.69, Mean =0.27, Mean- 2SD =-2.15 Fig. 1: Plotting of difference between VO2 max values against their means (Bland and Altman method of approach) X= Maximal shuttle run speed (km/hr) Using the above new equation the limits of agreement between directly measured VO2 max and predicted VO2 max from the 20-m MST (SPVO2 max) are -2.01 to 2.03. The result suggests that better limits of agreement exist between the two methods when this newly developed equation is used for prediction of VO2 max from the 20-m MST. Therefore, from the present observations it is concluded that the 20-m MST is recommended as a valid method to evaluate aerobic fitness in terms of VO2 max among female adults (age 20.42~24.75 yr.) of Nepal. A recent study has indicated that there are sport- specific differences when predicting VO2 max results yielded from the MST [20]. In another recent study by Cetin et al. on Taekwondo athletes, the authors concluded that VO2 max can be predicted from shuttle run test scores, but not as indicated with the test package. In order to obtain the true scores, one must apply a regression equation[21]. Studies by Chatterjee et al. on two different population of India also suggested separate regression equations for prediction of VO2 max in a particular population[13,14]. In our present study too, it is found that 20-m MST can be used in the studied population, but for better prediction a new regression equation has been derived. CONCLUSION The regression equation developed on the basis of present data is recommended to be used for the population studied. This is likely to be the most useful method when a large number of subjects are to be evaluated without the help of a well-equipped laboratory, with fewer expenses and within a short period of time. In a country like Nepal where laboratory facilities for direct evaluation of aerobic fitness is scanty, this method may be of great importance. Efforts should be taken to validate the applicability of 20-m MST in different Nepalese population including various sports disciplines.     Vol 1, No 1, March 2010 Validity of 20-m MST 45 ACKNOWLEDGMENTS The authors thank to all the subjects of this study and the authorities of their respective colleges for their all- out cooperation during the study. Conflict of interests: None declared REFERENCES 1. Astrand PO, Rodahl K, Dahl H, Stromme S. Evaluation of physical performance on the basis of tests. In: Textbook of Work Physiology: Physiological Bases of Exercise. 4th ed. New York: McGraw Hill. 2003; P.273. 2. Fox EL. A simple, accurate technique for predicting maximal aerobic power. J Appl Physiol. 1973;35:914-6. 3. Das SK, Bhattacharya G. A comparison of cardiorespiratory fitness in non-athletes and athletes of eastern India. Indian J Physiol Allied Sci. 1995;49:16-23. 4. Leger LA, Mercier D, Gadoury C, et al. The multistage 20 metre shuttle run test for aerobic fitness. J Sports Sci. 1988; 6:93-101. 5. Leger L, Gadoury C. Validity of the 20-m shuttle run test with 1 minute Stages to predict VO2 Max in Adults. Canadian J Sports Sci. 1989;14:21-6. 6. Wong TW, Yu TS, Wang XR, Robinson P. Predicted maximal oxygen uptake in normal Hong Kong Chinese schoolchildren and those with respiratory diseases. Pediatr Pulmonol. 2001;31:126-32. 7. Mota J, Guerra S, Leandro C, et al. Association of maturation, sex, and body fat in cardiorespiratory fitness. Am J Hum Biol. 2002;14:707-12. 8. Guerra S, Ribeiro JC, Costa R, et al. Relationship between cardiorespiratory fitness, body composition and blood pressure in school children. J Sports Med Physical Fitness. 2002;42:207-13. 9. Vicente-Rodriguez G, Jimenez-Ramirez J, Ara I, et al. Enhanced bone mass and physical fitness in prepubescent footballers. Bone. 2003;33:853-9. 10. Vicente-Rodriguez G, Dorado C, Perez-Gomez J, et al. Enhanced bone mass and physical fitness in young female handball players. Bone. 2004;35:1208-15. 11. Cooper SM, Baker JS, Tong RJ, et al. The repeatability and criterion related validity of the 20-m multistage fitness test as a predictor of maximal oxygen uptake in active young men. Br J Sports Med. 2005;39(4):19. 12. Suminski RR, Ryan ND, Poston CS, Jackson AS. Measuring aerobic fitness of Hispanic youth 10 to 12 years of age. Int J Sports Med. 2004;25(1):61-7. 13. Chatterjee P, Banerjee AK, Majumdar P, et al. Validity of the 20-m Multi Stage Shuttle Run Test for the Prediction of VO2max in Junior Taekwondo Players of Indial. Int J Appl Sports Sci. 2006;18(1):1-7. 14. Chatterjee P, Banerjee AK, Das P, et al. A regression equation for the estimation of maximum oxygen uptake in Indian male university students. Int J Applied Sports Sci. 2008;20:1-9. 15. Stickland MK, Petersen SR, Bouffard M. Prediction of maximal aerobic power from the 20-m multi-stage shuttle run test. Canadian J Appl Physiol. 2003;28:272-82. 16. Chatterjee S, Chakravarti B. Comparative study of maximum aerobic capacity by three ergometries in untrained college women. Jpn J Physiol. 1986;36:151-62. 17. Slonim NB, Gillespie DG, Harold WH. Peak oxygen uptake of healthy young man as determined by a treadmill method. J Appl Physiol. 1957;10:401-4. 18. Consolazio C F, Johnson RE, Pekora LJ. Analysis of Gas Samples. In: Samuel J, Hans field S, editors. Physiological Measurements of Metabolic Function in Man. 2nd ed. New York: McGraw Hill. 1963; Pp:507-10. 19. Bland JM, Altman DG. Statistical method for assessing agreement between two methods of clinical measurements. Lancet. 1986;1(8476):307-310. 20. Cetin C, Karatosun H, Baydar ML, et al. A regression equation to predict true maximal oxygen consumption of taekwondo athletes using a field-test. Saudi Med J. 2005;26:848-50. 21. Gibson A, Broomhead S, Lambert MI, et al. Prediction of maximal oxygen uptake from a 20-m shuttle run as measured directly in runners and squash players. J Sports Sci. 1998;16:331-5.

A regression equation for the estimation of maximum oxygen uptake in nepalese adult females.

[]

Chatterjee, Pinaki,Banerjee, Alok K,Das, Paulomi,Debnath, Parimal

eng

PMC10397271

1 Vol.:(0123456789) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports Elderly female ultra‑marathoners reduced the gap to male ultra‑marathoners in Swiss running races Beat Knechtle 1,2*, Anja Witthöft 3, David Valero 4, Mabliny Thuany 5, Pantelis T. Nikolaidis 6, Volker Scheer 4, Pedro Forte 7,8,9 & Katja Weiss 2 Recent studies showed that female runners reduced the performance gap to male runners in endurance running with increasing age and race distance. However, the investigated samples were generally small. To investigate this further, the present study examined sex differences by age across various race distances (5, 10 km, half‑marathon, marathon, and ultra‑marathon) using a large dataset of over 1,100,000 race records from Switzerland over two decades (1999–2019). The study explored performance and participation disparities between male and female runners by employing diverse methods, such as descriptive statistics, histograms, scatter and line plots, correlations, and a predictive machine learning model. The results showed that female runners were more prevalent in shorter races (5, 10 km, half‑marathon) and outnumbered male runners in 5 km races. However, as the race distance increased, the male‑to‑female ratio declined. Notably, the performance gap between sexes reduced with age until 70 years, after which it varied depending on the race distance. Among participants over 75 years old, ultra‑marathon running exhibited the smallest sex difference in performance. Elderly female ultra‑marathoners (75 years and older) displayed a performance difference of less than 4% compared to male ultra‑marathoners, which may be attributed to the presence of highly selected outstanding female performers. Running is high popular1 and is organized in different event formats such as track running2,3, road running4 and trail running (off-road running)5. Road-based running races are held over different distances, such as 5  km6, 10  km4, half-marathon7, marathon8 and ultra-marathon9. Recently, the interest in sports science to study female athletes profiles has grown (i.e., physiology, professionalism, and contextual factors that affect performance)10,11, with the aspect of sex difference in running garnering high-interest12–17. Twenty years ago, it was assumed that longer running distances were associated with higher sex differences. This might have been confounded by the reduced number of female runners in longer running distances15. It was also assumed that a sex difference of ~ 11–12% would be unchanged independent of the distance15 and would not change over years17,18. Today, the sex difference is still higher in longer running distances compared to shorter distances16, and the sex difference of different sports disciplines remained stable at ~ 10%12. It highlighted the need for researching the sex differ- ences in different running distances. Recent studies showed that elderly female ultra-marathoners reduced the gap to male ultra-marathoners of the same age13,14. Age seems to be of higher importance than the length of a race. It has been shown that female runners reduced the gap to male runners with increasing age, not with the increasing length of a race14. It is important to understand the contextual and environmental factors that may explain sex differences in running competitions. Modality popularity is dependent on the number of participants and competitive level. Therefore, females reducing the performance gap to male runners may improve participation in the sports modality. Upon that, it is important to understand the performance differences between sex (i.e., running speed), the ratio of OPEN 1Medbase St. Gallen am Vadianplatz, Vadianstrasse 26, 9001 St. Gallen, Switzerland. 2Institute of Primary Care, University of Zurich, Zurich, Switzerland. 3Kinderspital St. Gallen, St. Gallen, Switzerland. 4Ultra Sports Science Foundation, Pierre-Benite, France. 5Faculty of Sports, University of Porto, Porto, Portugal. 6School of Health and Caring Sciences, University of West Attica, Athens, Greece. 7CI-ISCE, Higher Institute of Educational Sciences of the Douro, Penafiel, Portugal. 8Instituto Politécnico de Bragança, Bragança, Portugal. 9Research Center in Sports, Health and Human Development, Covilhã, Portugal. *email: [email protected] 2 Vol:.(1234567890) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ participants between sexes and the evolution of the participant number over time to explain the sex differences in running. This first step may allow forecasting strategies to reduce sex differences in running. Studies investigating the sex difference in endurance running performance analyzed rather small samples and/or single distances19. The present study investigated the sex difference in performance in running races of 5, 10 km, half-marathon, marathon, and ultra-marathon with race records of two decades of a single country with a sample size of more than one million race records. Based upon recent findings regarding the reduction of the sex difference in longer running distances and older age groups, we hypothesized to confirm recent findings. Methods Ethical approval. This study was approved by the Institutional Review Board of Kanton St. Gallen, Switzer- land, with a waiver of the requirement for informed consent of the participants as the study involved the analysis of publicly available data (EKSG 01/06/2010). The study was conducted in accordance with recognized ethical standards according to the Declaration of Helsinki adopted in 1964 and revised in 2013. Data set. Athletes’ data from all 5, 10 km, half-marathon, marathon and ultra-marathon races held in Swit- zerland between 1999 and 2019 were collected from different sources such as “swiss-running” (www. swiss- runni ng. ch), “runme” (www. runme. ch/ de/ laufk alend er/ schwe iz), “datasport” (www. datas port. com/ de) and “DUV” (https:// stati stik.d- u-v. org/ calen dar. php). For all race distances, road-based races and trail runs were included. For ultra-marathons, any running races longer than the official marathon distance of 42.195 km and longer than 6 h were combined20. Data preparation. For each race, data from successful participants, including name and surname, sex, age, year of birth, distance and race time, event name and terrain type, were obtained from the websites and recorded in an EXCEL file. The average running speed (km/h) was calculated from the race distance and time. While the continuous age variable was available, race records were also classified (and later aggregated) by age group according to the official 5-year age group intervals. Records from runners younger than 18 were discarded. The junior category (18 years) includes only runners aged 18 and 19, while runners older than 75 were considered 75+ years. Race records were also aggregated by year for male and female runners separately. The male-to-female ratio was calculated by dividing the number of male records by the number of female records each year. Statistical analysis. Data normality was assessed by plotting histograms of the race speed in each race distance and for each sex. Descriptive statistics is then presented through the mean and standard deviation. For each race distance and age group, the male-to-female ratio was calculated by dividing the number of male records by the number of female records in each age group. The percentage difference in average speed was calculated as 100 * (male speed − female speed)/male speed. Pearson correlations were calculated between the male-to-female ratio, the percentage difference of speed and the age, where the male-to-female ratio and per- centage of speed difference were calculated for each year of age (18 through to 100). Statistical significance was assessed through the calculation of p-values, where the threshold set at p < 0.05. A machine learning (ML) predictive model was built and evaluated through the R2 and MAE metrics, with further analysis through SHAP values and feature importance/interaction analysis. All data processing, analysis and visualization were done in a Google Colab notebook using Python and statistical/machine learning free packages such as numpy, pandas, matplotlib, seaborn, statsmodels, scipy.stats, sklearn, catboost, shap. Results A total of 1,149,182 race records from 419,042 runners competing in 243 race events were considered. Table 1 presents the number of runners by distance and sex. Female runners were more numerous than male runners in the 5-km run. The male-to-female ratio increased with increasing race distance. Figure 1 summarizes the trend in the number of runners over the years by distance and sex, along with the male-to-female ratio. The male-to-female ratio generally decreased in all race distances over the years, caused by a more progressive increase in the number of female runners (Fig. 1, pink lines) compared to the number of male runners (Fig. 1, blue lines). The number of male marathoners in Swiss races has steadily decreased since 2005. Figure 2 shows the male-to-female ratio by age group and race distance. There is a general growing trend in the ratio from age 25 years until age group 70 years. Thereafter, the ratio increased further in 5 km and marathon, remained flat in 10 km, but decreased in half-marathon and ultra-marathon. Table 1. Number of male and female runners by race distance and male-to-female ratio. Race distance Male Female Total Male-to-female ratio 5 km 44,866 56,954 101,820 0.78 10 km 204,648 125,469 330,117 1.63 Half-marathon 268,486 117,275 385,761 2.28 Marathon 187,778 43,777 231,555 4.28 Ultra-marathon 83,040 16,889 99,929 4.91 3 Vol.:(0123456789) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ Figure 3 presents the histograms of the average running speed by distance and sex. In general, male runners were running faster than female runners in all race distances. The fastest average running speed was achieved at 10 km for both female and male runners. According to descriptive statistics, female runners participated more in the shorter distances and less in marathons and ultra-marathons, but the performance by sex did not differ so much in the longer distances. Figure 4 presents the sex difference in percent difference by age for the five-race distances. According to descriptive statistics, there is a decreasing trend from the age 20 years to age 70 years in the 5, 10 km, and half Figure 1. The trend in the number of runners and male-to-female ratio over the years by distance and sex. 4 Vol:.(1234567890) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ marathon distances. This decreasing trend starts at 45 years in marathon distances and at age 50 years in ultra- marathons. In 5 km and marathon distances, the sex difference increases after age 65 years, while in 10 km, half-marathon and ultra-marathon, the sex difference continues decreasing beyond this age. Table 2 presents the associations between the male-to-female ratio versus the % difference in running speed, between the male-to-female ratio and age group, and between the % difference in running speed and age group. According to descriptive statistics, the male-to-female ratio increased with age, with some exceptions in the youngest and the oldest age. Furthermore, there were always more male than female runners (except for 5 km), and males were always faster than females, where the differences in running speed declined with increasing age, except in the marathon distance. In 5 km, we found a significant and positive association (r = 0.57, p < 0.05) between the male-to-female ratio and the age and a significant and negative association (r = − 0.43, p < 0.05) between the difference in running speed and age. In 10 km, similar significant correlations can be observed, with slightly higher values. In the half-marathon, we found a significant and negative association (r = − 0.45, p < 0.05) between the male-to-female ratio and the percent difference in running speed and a significant and positive correlation (r = 0.67, p < 0.05) between the M/W ratio and the age. In the marathon, there was a significant and positive association (r = 0.28, p = 0.03) between the male-to-female ratio and the percent difference in running speed and a significant and positive association (r = 0.73, p < 0.05) between the male-to-female ratio and the age. Last, a significant and negative association (r = − 0.42, p = 0.0004) between the percent difference in running speed and age was found in ultra-marathon running. Predictive model. To explore potential non-linear relationships between the variables of interest, an ML tree-ensemble/gradient boosting model was built and evaluated. The model uses the Cat Boost Regressor algo- rithm with 200 learners to predict the average race speed (km/h) from the runner´s age and sex, the distance (km) and the terrain type (flat/trail). The model was trained with over 860 K race records or 75% of the full sample of 1,149,182, and later evaluated over the remaining 287 K records (25%), achieving predictive accuracy scores of R2 = 0.53, MAE = 1.32 km/h. The model features relative importance were also computed (Terrain 44%, Distance 32%, Sex 18%, Age 6%) along with the SHAP values and feature interactions. Predictive model key indicators Sample size 1,149,182 CatBoost model 200 trees MAE (km/h) 1.32 R2 0.53 Figure 2. The male-to-female ratio across age groups for all race distances. 5 Vol.:(0123456789) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ Features relative importances Terrain_type 44.13 Distance_km 32.06 Runner_sex 17.66 Runner_age 6.14 Figure 3. Histograms of average running speed by distance and sex. 6 Vol:.(1234567890) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ SHAP summary plots. Figure 5 presents the SHAP summary plots of individual prediction dots relative to reference value zero. They show how the model distributes its predictions for different values of each feature. Terrain and Distance make the broadest contributions to the model output and are rated as the model’s most important features. Flat races (Terrain_type = 0, blue dots) score on the positive side of the x-axis (higher speeds), while trail races (Terrain_type = 1, red strip of dots) get reduced speeds. A similar pattern can be observed in the distance chart. Longer distances (red dots, but also violet ones—distance is continuous) can be seen increasingly towards the left (negative side of the x-axis), whilst low values of the distance predictor accumulate in the positive side. The sex distinction is clear, with males (’1’, red dots) accumulating almost completely on the right side of the chart. Finally, the age predictor, also a continuous variable, shows a similar pattern to the first two: lower ages (blue points) obtain the model’s best predictions of speed, and from there the x-axis turns darker blue, purple, and red towards the left as the age increases. SHAP dependence plots. Figure 6 shows the SHAP dependence plots for the age of the runners, race distance, sex, and terrain. Each row represents the SHAP values for one predictor while its interactions with the other tree predictors are shown in each column. Regarding age, the first chart shows the blue dots corresponding to female, and the red dots to male, and shows an interesting trend changing gradually between 40 and 50 years. The model gives females better predictions than males from the age of 50 and up to 60 or 65 years. The second chart is largely dominated by blue (Terrain 0, flat races) and just shows the performance decline with age. The third chart has some interest, given that some red dots (long-distance races) obtain better speeds in age ranges (between 30 and 50 years and between 60 and 80 years). A possible explanation is the specialization of runners. For the race distance predictor, the first chart shows a broad dispersion (blue and red dots overlapping) across the distance axis, indicating a small interaction of sex with the distance. An interesting chunk of blue points Figure 4. % speed difference by sex for all race distances. Table 2. Correlations between the male-to-female ratio, % difference in running speed, and age (r = correlation coefficient, p = p value). 5 km 10 km Half-Marathon Marathon Ultra-marathon r p r p r p r p r p M/W ratio vs % running speed difference 0.04 0.73 − 0.01 0.92 − 0.45 0.0001 0.28 0.03 0.025 0.84 M/W ratio vs age 0.57 2.97e−07 0.66 6.3e−10 0.67 6.9e−10 0.73 2.8e−11 0.18 0.16 % running speed difference vs age − 0.43 0.0002 − 0.55 7.1e−07 − 0.11 0.35 0.39 0.001 −0.42 0.0004 7 Vol.:(0123456789) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ (Runner_sex 0, female) can be seen from the 150 km distance, indicating the model gives a significant number of higher speed predictions to female than to male runners in these long distances. The second chart shows that blue (flat run) dominates high-speed predictions for very short distances, but then red is above blue from approx. 50 km and up to 150 km. The third chart (distance/age interaction) shows equally a significant dispersion, where runners of all ages get the full range of predictions across all distances. For sex, the first chart shows red dots (trail race predictions) closer between male and female runners, with the larger differences being between the flat races (blue points). The second chart shows that female performance (Sex 0, left bar) is more distant to male performance in the early ages (blue points) or, in other words, the model predicts the performance gap decreases with age. A similar pattern can be seen in the third chart, where blue points (performance in short-distance races) present the largest difference between male and female runners, progressively coming closer through violets and red colors (mid and old ages). For terrain, the first chart shows that in trail races male runners obtain the lowest model predictions (red strip at the bottom of the right bar). The standard deviation of the data in the second chart is relatively high, indicating significant dispersion. The third and last chart shows that mid distances get higher speed predictions than high and short distances in both flat and trail races. This is more noticeable in trail terrain (ID 1). Discussion This study investigated the sex difference in running performance in a sample size of more than one million race records with the hypothesis to confirm the recent finding of reducing the sex difference in longer running distances and older age groups. The main important findings were (i) female runners participated more in the shorter race distances and less often in marathons and ultra-marathons, (ii) female runners were more numerous than male runners in the 5-km races, (iii) the male-to-female ratio increased with increasing race distance, (iv) the male-to-female ratio decreased in all race distances over the years, (v) the number of male marathoners con- tinuously decreased since 2005, (vi) male runners were running faster than female runners in all distances, (vii) the fastest average running speed was in the 10 km events, and (viii) female runners reduced the performance gap with male runners gradually from age group 20–24 years to age group 65–69 years in 5, 10 km and half-marathon, but not so much in marathon and ultra-marathon, with differences thereafter for the different race distances. Smallest sex difference in ultra‑marathon running. The smallest sex difference was found in the longest race distance (ultra-marathon) and oldest age group (75+), supporting the hypothesis. Female runners reduced the gap to male runners in ultra-marathon by increasing the age group from 50 to 54 years. The male- to-female ratio decreased after age group 65–69 years in ultra-marathon, while it increased in marathons. The male-to-female ratio increased with age in most race distances but not in ultra-marathon. The decrease in sex Figure 5. SHAP summary plots. 8 Vol:.(1234567890) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ difference and the increase in male-to-female ratio with increasing age was due to a higher number of faster and more competitive female runners in older age groups. This pattern was consistent across most race distances, except for an outlier in the oldest age group in marathons. The correlations between running speed difference, male-to-female ratio, and age group were similar across all race distances, with the few female runners in older age groups and ultra-marathon likely being exceptional performers. A study of ultra-marathoners in timed and multi-day events found that peak performance age increased with longer race durations and more race finishes, indicating that successful ultra-marathoners improved with age and experience21. A study investigating the effect of age and years of running experience in runners aged from 20 to 80 years showed that the number of years of running had a positive effect on running economy22. Master Figure 6. SHAP dependence plots for age of the runners, race distance, sex, and terrain (from top to down). 9 Vol.:(0123456789) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ ultra-marathoners improved their performance in long ultra-marathons such as ‘Badwater’ and ‘Spartathlon’23. It was also shown by Van den Berghe et al. that a low-loading and high-load-bearing tolerance running style could be advantageous for completing endurance running events for long-term runners24. A study comparing 50- and 100-mile races showed that female runners reduced the gap to male runners with increasing age, where the sex difference was smaller in 100 miles compared to 50 miles13. Another study investigating data from the American Master Road Running Records from 5 km to 144 h showed that female runners reduced the gap to male runners with increasing age, not with increasing length or duration of the performance14. The sex difference in performance can be due to different reasons such as biological differences (e.g., muscle mass, body fat, body size, muscle strength, limb length)25–28, physiological differences (e.g., aerobic capacity, running economy, fatigue resistance, substrate efficiency, energetic demands)29–31, participation9,32, experience and decision making33,34, motivation35, sociocultural36, and psychological differences3,37. Elderly female runners may have advantages over elderly male runners due to their ability to store and use elastic energy more efficiently, as studies have shown27. Female runners have a higher proportional area of type I muscle fibers and are better at using fatty acids during prolonged exercise and preserving carbohydrates25,38. This may lead to a more even pacing and less fatigue than male runners25,38. Considering physiological differences, it has been assumed that female ultra-marathoners have better fatigue resistance compared to equally trained male ultra-marathoners who are faster in the marathon than the female ultra-marathoners30. Since males have a higher aerobic capacity and a higher skeletal muscle mass than females, the distance running gap will not narrow between females and males28. Finally, decreasing testosterone levels due to aging also compromises male physical performance39. However, female runners seem more prone to support fatigue, so the aging effects in performance may be lower30. This may also support the reduced gap in sex differences with aging. Participation of females in ultra-marathon is an important aspect since the percentage of female runners is generally low in these races9,40. Lower participation of female runners compared to male runners overestimates the decline in age-related performance, especially in very old females41. A small sex difference in ultra-marathon running is more likely due to a low number of participants than an outstanding physiology42. An analysis of 20 ultra-marathons from 45 to 160 km showed that the sex difference in running was lower in the longer distances and the largest when fewer female and male runners were in a race43. Another aspect is competitiveness3,35,44–46. It is well-known that males are more engaged in direct competi- tion than females3 due to their higher competitiveness47. However, another study refuted that females were less competitive than males48. A study investigating 10-km races showed a significant annual decrease in the male-to-female ratio of finishers, with increasingly more female runners finishing in the sub-hour range4. Fur- thermore, it has been reported that females prefer smaller competitions49, which is the case in ultra-marathon running with lower numbers of participants. The sex difference in performance can also be due to the motivat- ing factors to participate in competitions. Between 1975 and 2013, master athletes improved their performance, where the magnitude of improvement was higher in the older age groups leading to gradual closing to younger athletes50. It has been reported that the motivation to enter an athletic competition is based on social conditions and predisposition3. A study of marathoners aged 20–79 found that sex differences in running speed increased with age, primarily attributed to the lower number of females than males51. In marathon running, however, a successful finisher can achieve a similar race performance from 20 to 55  years52, which would not be possible in ultra-marathon running. The sex difference in marathon versus ultra-marathon in the 75+ age group may be attributed to the declining number of male marathoners after 2005. Conversely, the number of female maratho- ners in the Venice Marathon increased from 2003 to 2019, providing a counterpoint53. Differences in the trend of sex difference by race distance. There was a general trend of decreas- ing sex difference from the 20–24 age group to the 65–69 age group, except for 5 km and the marathon, which increased after this age. In contrast, the sex difference continuously decreased with increasing age in 10 km, half-marathon, and ultra-marathon races. The male-to-female ratio by age group and distance may explain these trends, as observed in an analysis of races held in Oslo from 2008 to 2018. Female runners comprised a higher percentage of finishers in the 10 km race, but fewer in the half-marathon and marathon, and the male-to-female ratio was lowest in the 10 km and highest in the marathon54. Another explanation could be the age itself. The age of ~ 65–70 years is also important regarding the age- related performance decline55,56. In age group athletes, performance declines curvilinear from the age of 35 years until the age of ~ 65–70  years57. McClelland and Weyand16 recently noted a sex difference of ~ 12% for running distances from 800 m to 10 km, attributed to differences in energy supply and demands. Accordingly, Jobe et al.58 observed that males were 9–13% faster than females in all running events of the United States Olympic trials (from 100 m to marathon). Considering the physiological mechanisms underpinning the decrease of sex difference with the increasing race distance, an explanation might be the different taxing of the energy trans- fer systems in races varying for distance58. As the race distance increases, there is a larger contribution of the aerobic processes and a smaller anaerobic mechanism59. Thus, the existence of smaller sex differences in aerobic than in anaerobic capacity might relate to the decreased sex difference in increasing distances60. This is also in accordance with the latest study by Le Mat et al.61, although the race distances analyzed by the authors exceeded the ultra-marathon distances analyzed in this paper (45–260 km) a clear trend of decreasing sex difference with increasing race distance was shown. Limitations, strengths, and practical applications. A limitation of the present study was that it ana- lyzed race data from a single country; thus, considering the differences in performance and participation trends among countries, the findings should be generalized with caution to other countries. On the other hand, the strength was the large dataset that allowed drawing safe conclusions about the variation of sex difference by race 10 Vol:.(1234567890) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ distance and age group. Another potential limitation is the reliance on descriptive statistics in the result interpre- tation. The finding has practical applications for scientists and professionals working with long-distance runners to set optimal training goals for female athletes. Specifically, the main goals in training optimization considered should be body composition, sex differences in performance and performance differences throughout the aging process. Conclusion In summary, elderly female ultra-marathoners (age group 75+) show the smallest performance difference from male ultra-marathoners compared to other running distances from 5 km to a marathon. This is probably due to ‘highly selected’ female ultra-marathoners who perform exceptionally well. Future studies might investigate the experience and motivation of elderly female ultra-marathoners. Data availability For this study, we have included official results and split times from “swiss-running” (www. swiss- runni ng. ch), “runme” (www. runme. ch/ de/ laufk alend er/ schwe iz), “datasport” (www. datas port. com/ de) and “DUV” (https:// stati stik.d- u-v. org/ calen dar. php). The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Received: 4 March 2023; Accepted: 29 July 2023 References 1. Scheerder, J., Breedveld, K. & Borgers, J. Who is doing a run with the running boom? In Running across Europe: The Rise and Size of One of the Largest Sport Markets (eds Scheerder, J. et al.) 1–27 (Palgrave Macmillan UK, 2015). 2. Ganse, B., Kleerekoper, A., Knobe, M., Hildebrand, F. & Degens, H. Longitudinal trends in master track and field performance throughout the aging process: 83,209 results from Sweden in 16 athletics disciplines. Geroscience 42, 1609–1620. https:// doi. org/ 10. 1007/ s11357- 020- 00275-0 (2020). 3. Deaner, R. O., Addona, V. & Mead, M. P. U. S. masters track participation reveals a stable sex difference in competitiveness. Evol. Psychol. 12, 848–877 (2014). 4. Cushman, D. M., Markert, M. & Rho, M. Performance trends in large 10-km road running races in the United States. J. Strength Cond. Res. 28, 892–901. https:// doi. org/ 10. 1519/ jsc. 00000 00000 000249 (2014). 5. Scheer, V. et al. Defining off-road running: A position statement from the ultra sports science foundation. Int. J. Sports Med. 41, 275–284. https:// doi. org/ 10. 1055/a- 1096- 0980 (2020). 6. Deaner, R. O. & Lowen, A. Males and females pace differently in high school cross-country races. J. Strength Cond. Res. 30, 2991–2997. https:// doi. org/ 10. 1519/ jsc. 00000 00000 001407 (2016). 7. Yang, S. et al. Gender and age differences in performance of over 70,000 Chinese finishers in the half- and full-marathon events. Int. J. Environ. Res. Public Health https:// doi. org/ 10. 3390/ ijerp h1913 7802 (2022). 8. Reusser, M. et al. Increased participation and decreased performance in recreational master athletes in “Berlin Marathon” 1974– 2019. Front. Physiol. https:// doi. org/ 10. 3389/ fphys. 2021. 631237 (2021). 9. Thuany, M. et al. Trends in participation, sex differences and age of peak performance in time-limited ultramarathon events: A secular analysis. Medicina https:// doi. org/ 10. 3390/ medic ina58 030366 (2022). 10. Emmonds, S., Heyward, O. & Jones, B. The challenge of applying and undertaking research in female sport. Sports Med. Open 5, 51. https:// doi. org/ 10. 1186/ s40798- 019- 0224-x (2019). 11. Elliott-Sale, K. J. et al. Methodological considerations for studies in sport and exercise science with women as participants: A work- ing guide for standards of practice for research on women. Sports Med. 51, 843–861. https:// doi. org/ 10. 1007/ s40279- 021- 01435-8 (2021). 12. Thibault, V. et al. Women and men in sport performance: The gender gap has not evolved since 1983. J. Sports Sci. Med. 9, 214–223 (2010). 13. Waldvogel, K. J., Nikolaidis, P. T., Di Gangi, S., Rosemann, T. & Knechtle, B. Women reduce the performance difference to men with increasing age in ultra-marathon running. Int. J. Environ. Res. Public Health https:// doi. org/ 10. 3390/ ijerp h1613 2377 (2019). 14. Sousa, C. V., da Silva Aguiar, S., Rosemann, T., Nikolaidis, P. T. & Knechtle, B. American masters road running records-the per- formance gap between female and male age group runners from 5 km to 6 days running. Int. J. Environ. Res. Public Health https:// doi. org/ 10. 3390/ ijerp h1613 2310 (2019). 15. Coast, J. R., Blevins, J. S. & Wilson, B. A. Do gender differences in running performance disappear with distance?. Can. J. Appl. Physiol. 29, 139–145. https:// doi. org/ 10. 1139/ h04- 010 (2004). 16. McClelland, E. L. & Weyand, P. G. Sex differences in human running performance: Smaller gaps at shorter distances?. J. Appl. Physiol. 133, 876–885. https:// doi. org/ 10. 1152/ jappl physi ol. 00359. 2022 (1985). 17. Millard-Stafford, M., Swanson, A. E. & Wittbrodt, M. T. Nature versus nurture: Have performance gaps between men and women reached an asymptote?. Int. J. Sports Physiol. Perform. 13, 530–535. https:// doi. org/ 10. 1123/ ijspp. 2017- 0866 (2018). 18. Sparling, P. B., O’Donnell, E. M. & Snow, T. K. The gender difference in distance running performance has plateaued: An analysis of world rankings from 1980 to 1996. Med. Sci. Sports Exerc. 30, 1725–1729. https:// doi. org/ 10. 1097/ 00005 768- 19981 2000- 00011 (1998). 19. Costello, J. T., Bieuzen, F. & Bleakley, C. M. Where are all the female participants in sports and exercise medicine research?. Eur. J. Sport Sci. 14, 847–851. https:// doi. org/ 10. 1080/ 17461 391. 2014. 911354 (2014). 20. Zaryski, C. & Smith, D. J. Training principles and issues for ultra-endurance athletes. Curr. Sports Med. Rep. 4, 165–170 (2005). 21. Knechtle, B., Valeri, F., Zingg, M. A., Rosemann, T. & Rüst, C. A. What is the age for the fastest ultra-marathon performance in time-limited races from 6 h to 10 days?. Age https:// doi. org/ 10. 1007/ s11357- 014- 9715-3 (2014). 22. Dos Anjos Souza, V. R. et al. Running economy in long-distance runners is positively affected by running experience and negatively by aging. Physiol. Behav. 258, 114032. https:// doi. org/ 10. 1016/j. physb eh. 2022. 114032 (2023). 23. Zingg, M. A., Knechtle, B., Rüst, C. A., Rosemann, T. & Lepers, R. Analysis of participation and performance in athletes by age group in ultramarathons of more than 200 km in length. Int. J. Gen. Med. 6, 209–220. https:// doi. org/ 10. 2147/ ijgm. S43454 (2013). 24. Van den Berghe, P., Breine, B., Haeck, E. & De Clercq, D. One hundred marathons in 100 days: Unique biomechanical signature and the evolution of force characteristics and bone density. J. Sport Health Sci. 11, 347–357. https:// doi. org/ 10. 1016/j. jshs. 2021. 03. 009 (2022). 25. Sandbakk, Ø., Solli, G. S. & Holmberg, H. C. Sex differences in world-record performance: The influence of sport discipline and competition duration. Int. J. Sports Physiol. Perform. 13, 2–8. https:// doi. org/ 10. 1123/ ijspp. 2017- 0196 (2018). 11 Vol.:(0123456789) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ 26. Wells, C. L. & Plowman, S. A. Sexual differences in athletic performance: Biological or behavioral?. Phys. Sportsmed 11, 52–63. https:// doi. org/ 10. 1080/ 00913 847. 1983. 11708 602 (1983). 27. Lindle, R. S. et al. Age and gender comparisons of muscle strength in 654 women and men aged 20–93 yr. J. Appl. Physiol. 1997(83), 1581–1587. https:// doi. org/ 10. 1152/ jappl. 1997. 83.5. 1581 (1985). 28. Cheuvront, S. N., Carter, R., Deruisseau, K. C. & Moffatt, R. J. Running performance differences between men and women: An update. Sports Med. 35, 1017–1024. https:// doi. org/ 10. 2165/ 00007 256- 20053 5120- 00002 (2005). 29. Bassett, A. J. et al. The biology of sex and sport. JBJS Rev. 8, e0140. https:// doi. org/ 10. 2106/ jbjs. Rvw. 19. 00140 (2020). 30. Bam, J., Noakes, T. D., Juritz, J. & Dennis, S. C. Could women outrun men in ultramarathon races?. Med. Sci. Sports Exerc. 29, 244–247. https:// doi. org/ 10. 1097/ 00005 768- 19970 2000- 00013 (1997). 31. Tiller, N. B. et al. Do sex differences in physiology confer a female advantage in ultra-endurance sport?. Sports Med. 51, 895–915. https:// doi. org/ 10. 1007/ s40279- 020- 01417-2 (2021). 32. Baker, A. B. & Tang, Y. Q. Aging performance for masters records in athletics, swimming, rowing, cycling, triathlon, and weightlift- ing. Exp. Aging Res. 36, 453–477. https:// doi. org/ 10. 1080/ 03610 73x. 2010. 507433 (2010). 33. Deaner, R. O., Carter, R. E., Joyner, M. J. & Hunter, S. K. Men are more likely than women to slow in the marathon. Med. Sci. Sports Exerc. 47, 607–616. https:// doi. org/ 10. 1249/ mss. 00000 00000 000432 (2015). 34. Deaner, R. O., Addona, V., Carter, R. E., Joyner, M. J. & Hunter, S. K. Fast men slow more than fast women in a 10 kilometer road race. PeerJ 4, e2235. https:// doi. org/ 10. 7717/ peerj. 2235 (2016). 35. Deaner, R. O. Distance running as an ideal domain for showing a sex difference in competitiveness. Arch. Sex Behav. 42, 413–428. https:// doi. org/ 10. 1007/ s10508- 012- 9965-z (2013). 36. Hallam, L. & Amorim, F. Expanding the gap: An updated look into sex differences in running performance. Front. Physiol. 12, 804149–804149. https:// doi. org/ 10. 3389/ fphys. 2021. 804149 (2022). 37. Deaner, R. O., Addona, V. & Hanley, B. Risk taking runners slow more in the marathon. Front. Psychol. 10, 1–12. https:// doi. org/ 10. 3389/ fpsyg. 2019. 00333 (2019). 38. Besson, T. et al. Sex differences in endurance running. Sports Med. https:// doi. org/ 10. 1007/ s40279- 022- 01651-w (2022). 39. Schaap, L. A. et al. Low testosterone levels and decline in physical performance and muscle strength in older men: Findings from two prospective cohort studies. Clin. Endocrinol. 68, 42–50. https:// doi. org/ 10. 1111/j. 1365- 2265. 2007. 02997.x (2008). 40. Thuany, M. et al. A macro to micro analysis to understand performance in 100-mile ultra-marathons worldwide. Sci. Rep. 13, 1415. https:// doi. org/ 10. 1038/ s41598- 023- 28398-2 (2023). 41. Senefeld, J. W. & Hunter, S. K. Are masters athletic performances predictive of human aging in men and women?. Mov. Sport Sci. Sci Mot. 200, 5–12. https:// doi. org/ 10. 1051/ sm/ 20190 18 (2019). 42. Deaner, R. O. Physiology does not explain all sex differences in running performance. Med. Sci. Sports Exerc. 45, 146–147. https:// doi. org/ 10. 1249/ MSS. 0b013 e3182 690110 (2013). 43. Senefeld, J., Smith, C. & Hunter, S. K. Sex differences in participation, performance, and age of ultramarathon runners. Int. J. Sports Physiol. Perform. 11, 635–642. https:// doi. org/ 10. 1123/ ijspp. 2015- 0418 (2016). 44. Deaner, R. O. & Mitchell, D. More men run relatively fast in U.S. road races, 1981–2006: A stable sex difference in non-elite run- ners. Evol. Psychol. 9, 600–621 (2011). 45. Flory, J., Gneezy, U., Leonard, K. & List, J. Gender, age, and competition: A disappearing gap?. J. Econ. Behav. Organ. 150, 256–276 (2018). 46. Carpenter, J., Frank, R. & Huet-Vaughn, E. Gender differences in interpersonal and intrapersonal competitive behavior. J. Behav. Exp. Econ. 77, 170–176. https:// doi. org/ 10. 1016/j. socec. 2018. 10. 003 (2018). 47. Deaner, R. O., Lowen, A., Rogers, W. & Saksa, E. Does the sex difference in competitiveness decrease in selective sub-populations? A test with intercollegiate distance runners. PeerJ 3, e884. https:// doi. org/ 10. 7717/ peerj. 884 (2015). 48. Wieland, A. & Sarin, R. Domain specificity of sex differences in competition. J. Econ. Behav. Organ. 83, 151–157. https:// doi. org/ 10. 1016/j. jebo. 2011. 06. 019 (2012). 49. Hanek, K. J., Garcia, S. M. & Tor, A. Gender and competitive preferences: The role of competition size. J. Appl. Psychol. 101, 1122–1133. https:// doi. org/ 10. 1037/ apl00 00112 (2016). 50. Akkari, A., Machin, D. & Tanaka, H. Greater progression of athletic performance in older Masters athletes. Age Ageing 44, 683–686. https:// doi. org/ 10. 1093/ ageing/ afv023 (2015). 51. Hunter, S. K. & Stevens, A. A. Sex differences in marathon running with advanced age: Physiology or participation?. Med. Sci. Sports Exerc. 45, 148–156. https:// doi. org/ 10. 1249/ MSS. 0b013 e3182 6900f6 (2013). 52. Leyk, D. et al. Performance, training and lifestyle parameters of marathon runners aged 20–80 years: Results of the PACE-study. Int. J. Sports Med. 30, 360–365. https:// doi. org/ 10. 1055/s- 0028- 11059 35 (2009). 53. Albertin, G. et al. “Venice marathon”: Participation of female Master Athletes shows a constant increase from 2003 to 2019. Eur. J. Transl. Myol. https:// doi. org/ 10. 4081/ ejtm. 2021. 10266 (2021). 54. Nikolaidis, P., Cuk, I., Clemente-Suárez, V., Villiger, E. & Knechtle, B. Number of finishers and performance of age group women and men in long-distance running: Comparison among 10km, half-marathon and marathon races in Oslo. Res. Sports Med. 29, 56–66. https:// doi. org/ 10. 1080/ 15438 627. 2020. 17267 45 (2021). 55. Ganse, B., Drey, M., Hildebrand, F., Knobe, M. & Degens, H. Performance declines are accelerated in the oldest-old track and field athletes 80 to 94 years of age. Rejuvenation Res. 24, 20–27. https:// doi. org/ 10. 1089/ rej. 2020. 2337 (2021). 56. Ransdell, L. B., Vener, J. & Huberty, J. Masters athletes: An analysis of running, swimming and cycling performance by age and gender. J. Exerc. Sci. Fit. 7, S61–S73. https:// doi. org/ 10. 1016/ S1728- 869X(09) 60024-1 (2009). 57. Reaburn, P. & Dascombe, B. Endurance performance in masters athletes. Eur. Rev. Aging Phys. Activity 5, 31–42. https:// doi. org/ 10. 1007/ s11556- 008- 0029-2 (2008). 58. Jobe, T. K., Shaffer, H. N., Doci, C. L. & Gries, K. J. Sex differences in performance and depth of field in the United States olympic trials. J. Strength Cond. Res. 36, 3122–3129. https:// doi. org/ 10. 1519/ jsc. 00000 00000 004295 (2022). 59. Trappe, S. Marathon runners: How do they age?. Sports Med. 37, 302–305. https:// doi. org/ 10. 2165/ 00007 256- 20073 7040- 00008 (2007). 60. Galloway, M. T., Kadoko, R. & Jokl, P. Effect of aging on male and female master athletes’ performance in strength versus endur- ance activities. Am. J. Orthop. 31, 93–98 (2002). 61. Le Mat, F. et al. Running endurance in women compared to men: Retrospective analysis of matched real-world big data. Sports Med. 53, 917–926. https:// doi. org/ 10. 1007/ s40279- 023- 01813-4 (2023). Author contributions B.K. and K.W. drafted the manuscript, D.V. performed the statistical analysis and prepared methods and results, A.W. obtained the data, M.T., P.N., V.S., and P.F. helped in drafting the final version. All authors read and approved the final manuscript. Competing interests The authors declare no competing interests. 12 Vol:.(1234567890) Scientific Reports | (2023) 13:12521 | https://doi.org/10.1038/s41598-023-39690-6 www.nature.com/scientificreports/ Additional information Correspondence and requests for materials should be addressed to B.K. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. © The Author(s) 2023

Elderly female ultra-marathoners reduced the gap to male ultra-marathoners in Swiss running races.

08-02-2023

Knechtle, Beat,Witthöft, Anja,Valero, David,Thuany, Mabliny,Nikolaidis, Pantelis T,Scheer, Volker,Forte, Pedro,Weiss, Katja

eng

PMC8412947

Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 1 of 16 Muscle- specific economy of force generation and efficiency of work production during human running Sebastian Bohm1,2*, Falk Mersmann1,2, Alessandro Santuz1,2, Arno Schroll1,2, Adamantios Arampatzis1,2 1Humboldt- Universität zu Berlin, Department of Training and Movement Sciences, Berlin, Germany; 2Berlin School of Movement Science, Humboldt- Universität zu Berlin, Berlin, Germany Abstract Human running features a spring- like interaction of body and ground, enabled by elastic tendons that store mechanical energy and facilitate muscle operating conditions to minimize the metabolic cost. By experimentally assessing the operating conditions of two important muscles for running, the soleus and vastus lateralis, we investigated physiological mechanisms of muscle work production and muscle force generation. We found that the soleus continuously shortened throughout the stance phase, operating as work generator under conditions that are considered optimal for work production: high force- length potential and high enthalpy efficiency. The vastus lateralis promoted tendon energy storage and contracted nearly isometrically close to optimal length, resulting in a high force- length- velocity potential beneficial for economical force generation. The favorable operating conditions of both muscles were a result of an effective length and velocity- decoupling of fascicles and muscle- tendon unit, mostly due to tendon compliance and, in the soleus, marginally by fascicle rotation. Introduction During locomotion, muscles generate force and perform work in order to support and accelerate the body, and the activity of the lower- limb muscles accounts for most of the metabolic energy cost needed to walk or run (Kram and Taylor, 1990; Kram, 2000; Dickinson et al., 2000). Running is char- acterized by a spring- like interaction of the body with the ground, indicating a significant conversion of the body’s kinetic and potential energy to strain energy - via the elongation of elastic elements, mainly tendons - that can be recovered in the propulsive second half of the stance phase (Dickinson et al., 2000; Roberts and Azizi, 2011; Cavagna et al., 1964). In addition, the elasticity of tendons influences the operating conditions of the muscles, which in turn are associated with their metabolic cost (Roberts, 2002). For a given muscle force, the metabolic cost depends on the muscle’s oper- ating force- length and force- velocity potential (Bohm et al., 2019; Bohm et al., 2018; Nikolaidou et al., 2017) (fraction of maximum force according to the force- length [Gordon et al., 1966] and force- velocity [Hill, 1938] curves) because it determines the number of recruited muscle fibers and thus the active muscle volume (Roberts, 2002). This means that quasi isometric contractions close the optimum of the force- length curve, that is, with a high force- length- velocity potential, are theoreti- cally most economical for generating a given force. During steady- state running, however, the human system does not perfectly conserve all the mechanical energy in each stride. Therefore, muscular work by active muscle shortening is needed to maintain the running movement, yet it increases the meta- bolic cost a) due to the reduced force- velocity potential, which will increase the active muscle volume for a given force (Roberts and Azizi, 2011), and b) due to the higher metabolic energy consumption RESEARCH ARTICLE *For correspondence: sebastian. bohm@ hu- berlin. de Competing interest: The authors declare that no competing interests exist. Funding: See page 13 Received: 03 February 2021 Preprinted: 19 February 2021 Accepted: 06 August 2021 Published: 02 September 2021 Reviewing Editor: Carlos Isales, Medical College of Georgia at Augusta University, United States Copyright Bohm et al. This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 2 of 16 of each fiber when actively shortening (Smith et al., 2005; He et al., 2000). The active shorting range and velocity of a muscle during movements can be reduced by its tendon and, thus, an important benefit of tendon elasticity is a reduction in the metabolic cost of running. At the muscle level, it has been shown that the triceps surae muscle group produces muscular work/energy during the stance phase of steady- state running (Lai et al., 2014). The soleus is the largest muscle in this group (Albracht et al., 2008) and does work by active shortening throughout the entire stance phase (Bohm et al., 2019; Bohm et al., 2021). In the first part of the stance phase, the performed muscular work is stored in the Achilles tendon as elastic strain energy. During the later propulsion phase, the tendon strain energy recoil contributes to the muscular energy produc- tion, suggesting an energy amplification behavior (Roberts and Azizi, 2011) within the triceps surae muscle- tendon unit (MTU) during running. On the contrary, the vastus lateralis muscle, as the largest muscle of the quadriceps femoris muscle group (Mersmann et al., 2015), operates nearly isometri- cally despite a lengthening- shortening behavior of the vastus lateralis MTU (Bohm et al., 2018; Monte et al., 2020). The almost isometric contraction suggests a negligible mechanical work production by the vastus lateralis during running and a spring- like energy exchange between body and vastus later- alis MTU, which promotes energy conservation (Dickinson et al., 2000; Roberts and Azizi, 2011). The triceps surae and the quadriceps muscle group are considered to be crucial for running perfor- mance (Arampatzis et al., 2006; Hamner and Delp, 2013). The quadriceps femoris decelerates and supports the body early in stance while the triceps surae accounts for the propulsion later in the stance phase (Dorn et al., 2012; Santuz et al., 2020; Hamner and Delp, 2013). The soleus and vastus later- alis, as the largest muscles of both muscle groups, show marked differences in their morphological and architectural properties with shorter fascicles and higher pennation angles in the soleus (Bohm et al., 2019; Maganaris et al., 1998) compared to vastus lateralis (Bohm et al., 2018; Marzilger et al., 2018). Because of the long fascicles of the vastus lateralis, a unit of force generated by this muscle is metabolically more expensive (Biewener and Roberts, 2000) compared to the soleus. Our previous findings (Bohm et al., 2018) suggest that the vastus lateralis operates at a high force- length- velocity potential during running, which would indicate a fascicle contraction condition that could minimize the energetic cost of muscle force generation. The soleus muscle instead operates as a muscular work generator through active shortening, though close to the optimum of the force- length curve. Operating with increasing shortening velocity decreases the force- velocity potential according to the force- velocity relationship (Bohm et al., 2019; Bohm et al., 2021) and may increase the ener- getic cost of muscle force generation, marking a trade- off between mechanical work production and metabolic expenses. The enthalpy efficiency (Barclay, 2015) (or mechanical efficiency; Hill, 1939; Hill, 1964) quantifies the fraction of chemical energy from ATP hydrolysis that is converted into mechanical work and depends on the shortening velocity, with a steep increase at low shortening velocities up to a maximum at around 20% of the maximum shortening velocity (Vmax) and a decrease thereafter (Hill, 1939; Barclay et al., 1993; Hill, 1964). Previous findings suggest that the soleus fascicles continu- ously shorten at a moderate velocity during the stance phase of running (Bohm et al., 2019), covering a range that corresponds to a high efficiency. Therefore, the soleus muscle may operate at fascicle conditions that would be beneficial for economical work/energy production. The muscle fascicle behavior is strongly influenced by the decoupling of the fascicles from the MTU excursions due to tendon elasticity and fascicle rotation (Azizi et al., 2008; Alexander, 1991; Zuur- bier and Huijing, 1992; Wakeling et al., 2011). The previously reported decoupling of the soleus muscle indicates that tendon elasticity and fascicle rotation affect the operating fascicle length and velocity during running (Bohm et al., 2019; Werkhausen et al., 2019); however, their integration in the regulation of the efficiency- fascicle velocity dependency is unclear. Regarding the vastus lateralis muscle, it was suggested that proximal muscles like the knee extensors feature less compliant tendons compared to the distal triceps surae muscles, thus limiting the decoupling between fascicles and MTU (Farris and Sawicki, 2012; Biewener and Daley, 2007; Biewener, 2016). However, in our previous study, we found significantly smaller vastus lateralis fascicle length changes compared to the vastus lateralis MTU (Bohm et al., 2018), indicating an important decoupling within the vastus lateralis MTU due to tendon elasticity. The purpose of this study was to assess the soleus and the vastus lateralis fascicle behavior with regard to the operating force- length- velocity potential and enthalpy efficiency to investigate physi- ological mechanisms for muscle work production and muscle force generation during running. The  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 3 of 16 soleus muscle actively shortens during the stance phase at moderate velocities, which may match the plateau of the enthalpy efficiency- velocity curve, and operates close to the optimum of the force- length curve. Therefore, we hypothesized that the soleus muscle as a work generator operates at a high force- length potential and a high enthalpy efficiency, minimizing the metabolic cost of work production. On the other hand, the vastus lateralis muscle that promotes energy conservation seems to operate at a favorable length and almost isometrically. Thus, we hypothesized a high force- length and a high force- velocity potential that would reduce the metabolic energy cost of muscle force generation. In order to investigate the regulation of the efficiency and force potentials, we further quantified the length and velocity- decoupling of the fascicles from the MTU as well as the electro- myographic (EMG) activity. Because of experimental constrains, the two muscles were measured in two groups, respectively. Results There were no significant differences in the anthropometric characteristics between groups (age p=0.369, height p=0.536, body mass p=0.057). The experimentally assessed optimal fascicle length for force generation (L0) of the soleus was on average 41.3 ± 5.2 mm and significantly shorter than L0 of the vastus lateralis with 94.0 ± 11.6 mm (p<0.001). The forces that corresponded to L0 of soleus and vastus lateralis (Fmax) were 2887 ± 724 N and 4990 ± 914 N (p<0.001), respectively. Furthermore, the assessed Vmax was 279 ± 35 mm/s for the soleus, significantly lower than the Vmax of the vastus lateralis with 1082 ± 133 mm/s (p<0.001). The stance and swing times during running were 304 ± 23 ms and 439 ± 26 ms for the soleus group and 290 ± 22 ms and 448 ± 30 ms for the vastus lateralis group (p=0.075, p=0.369). The EMG compar- ison showed that the soleus was active throughout the entire stance phase of running while the vastus lateralis was mainly active in the first part of the stance and with an earlier peak of activation (soleus 41 ± 5% of the stance phase, vastus lateralis 35 ± 4% of the stance phase, p<0.001, Figure 1). During the stance phase, the MTU of both muscles showed a lengthening- shortening behavior, but the vastus lateralis MTU started to shorten earlier (soleus 59 ± 2% of the stance phase, vastus lateralis 50 ± 2% of the stance phase, p<0.001, Figure 1). The soleus and the vastus lateralis fascicle length were clearly decoupled from the MTU length with smaller operating length ranges throughout the whole stance phase (Figure 1). The soleus fascicles operated at a length close to L0 at touchdown and then shortened continuously until the foot lift- off (0.994–0.752 L/L0, Figure 1). The operating length of the vastus lateralis fascicles remained closely above L0 over the entire stance phase and was on average significantly longer compared to the soleus fascicles (soleus 0.899 ± 0.104 L/L0, vastus lateralis 1.054 ± 0.082 L/L0, p<0.001, Figure 1). The stance phase- averaged force- length potential of both muscles was high and not significantly different (p=0.689, Figure 2). The average pennation angle of the soleus was significantly greater than that of the vastus lateralis (soleus 23.9 ± 5.1°, vastus lateralis 13.3 ± 1.8°, p<0.001) and increased continuously throughout stance, whereas it remained almost unchanged in the vastus lateralis (Figure 1). The average operating velocity of the soleus fascicles was significantly higher compared to the vastus lateralis (soleus 0.799 ± 0.260 L0/s, vastus lateralis 0.084 ± 0.258 L0/s, p<0.001), which showed an almost isometric contraction throughout stance. Consequently, the force- velocity potential (p<0.001) and thus the overall force- length- velocity potential (p<0.001) of the soleus was significantly lower compared to the vastus lateralis during the stance phase (Figure 2). However, the higher short- ening velocity of the soleus was close to the optimum for maximum enthalpy efficiency, leading to a significantly higher enthalpy efficiency over the stance phase in comparison to the vastus lateralis (p<0.001, Figure 3). The fascicle, muscle belly, and MTU length changes throughout stance as well as the resulting velocity decoupling coefficients (DC) are illustrated in Figure 4 for both muscles, where DCTendon quan- tifies the decoupling due to tendon compliance, DCBelly due to fascicle rotation, and DCMTU the overall decoupling of MTU and fascicles. There was a clear length and velocity- decoupling of MTU and belly due to tendon compliance in both muscles (Figure 4). The statistical parametric mapping (SPM) anal- ysis revealed a significantly lower DCTendon of the soleus compared to the vastus lateralis between 4% and 8% of the stance phase (p=0.032) since decoupling started later for the soleus. Between 20% and 57% of the stance phase (p<0.001) and between 65%  of the stance phase until lift- off, the soleus DCTendon was significantly higher than vastus lateralis (p<0.001, Figure 4). The DCTendon averaged over  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 4 of 16 the stance phase of the soleus was also significantly greater (p<0.001, Table 1). Furthermore, the velocity- decoupling of muscle belly and fascicles due to fascicle rotation progressively increased in the second part of the stance phase for the soleus but was negligible for the vastus lateralis (Figure 4). The soleus DCBelly was significantly higher from 33%  of the stance phase until lift- off compared to the vastus lateralis as shown by the SPM analysis (p<0.001, Figure 4) but also when averaged over the entire stance phase (p<0.001, Table 1). DCBelly was markedly lower than DCTendon, indicating that the tendon took over the majority of the overall decoupling in both muscles (Figure 4). Accordingly and similarly to DCTendon, the SPM analysis for the overall decoupling of MTU and fascicles showed that 0 20 40 60 80 100 0.6 0.8 1.0 1.2 1.4 Fascicle length Norm (L/L 0) 0 20 40 60 80 100 5.0 6.0 7.0 8.0 9.0 MTU length Norm (L/L 0) 2.5 3.0 3.5 4.0 4.5 0 20 40 60 80 100 0 10 20 30 40 Pennation angle (°) 0 20 40 60 80 100 Stance phase (%) 0.0 0.2 0.4 0.6 0.8 1.0 EMG Norm SOL MTU length Norm (L/L0) Fascicle length Norm (L/L 0) Pennation angle (°) EMG Norm VL MTU length Norm (L/L0) A D C B Stance phase (%) SOL VL SOL VL SOL VL SOL VL Stance phase (%) Stance phase (%) Stance phase (%) Figure 1. Soleus (SOL, n = 19) and vastus lateralis (VL, n = 14) muscle- tendon unit (MTU) length (A) and muscle fascicle length (normalized to optimal fascicle length L0, B), pennation angle (C), and electromyographic (EMG) activity (normalized to a maximum voluntary isometric contraction, D) during the stance phase of running (mean ± SD). The online version of this article includes the following source data for figure 1: Source data 1. Numerical data represented in the graph 1.  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 5 of 16 DCMTU of the soleus was significantly lower between 4% and 8% of the stance phase (p=0.032) and significantly higher from 20% to 57% of the stance phase and from 65%  of the stance phase until lift- off compared to the vastus lateralis (p<0.001, Figure 4). The stance phase- averaged DCMTU of the soleus was significantly greater compared to the vastus lateralis as well (p<0.001, Table 1). Discussion We mapped the operating length and velocity of the soleus and the vastus lateralis fascicles during running onto the individual force- length, force- velocity, and enthalpy efficiency- velocity curves in order to investigate physiological mechanisms for muscle force generation and muscle work production SOL VL Force-length-velocity potential * * 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0 Force-velocity potential Force-length potential A B C Figure 2. Soleus (SOL, n = 19) and vastus lateralis (VL, n = 14) force- length potential (A), force- velocity potential (B), and overall force- length- velocity potential (C) averaged over the stance phase of running. *Significant difference between muscles (p<0.05). The online version of this article includes the following source data for figure 2: Source data 1. Numerical data represented in the graph 2. 0.0 0.2 0.4 0.6 0.8 1.0 Fascicle velocity (V/Vmax) -0.1 0.0 0.1 0.2 0.3 0.4 0.5 Enthalpy efficiency * SOL VL Figure 3. Soleus (SOL, n = 19) and vastus lateralis (VL, n = 14) enthalpy efficiency (mean ± SD) averaged over the stance phase of running onto the enthalpy efficiency- fascicle velocity relationship (dashed line). *Significant difference between muscles (p<0.05). The online version of this article includes the following source data for figure 3: Source data 1. Numerical data represented in the graph 3.  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 6 of 16 Figure 4. Soleus (SOL, n = 19, top row) and vastus lateralis (VL, n = 14, mid row) muscle- tendon unit (MTU) vs. belly length changes (left), belly vs. fascicle length changes (mid), and MTU vs. fascicle length changes (right) over the stance phase of running with respect to the length at touchdown (0%  stance phase).  Differences between curves illustrate the length- decoupling due to tendon compliance, fascicle rotation, and the overall decoupling, respectively. The bottom row shows the resulting velocity- decoupling coefficients (DCs) as the absolute velocity differences between fascicles, belly, and MTU normalized to the maximum shorting velocity (see Materials and methods). Intervals of stance with a significant difference between both muscles are illustrated as hatched areas (p<0.05). The online version of this article includes the following source data for figure 4: Source data 1. Numerical data represented in the graph 4. Table 1. Average tendon (DCTendon), belly (DCBelly), and muscle- tendon unit (DCMTU) decoupling coefficients for the soleus (SOL) and vastus lateralis (VL) muscles during the stance phase of running (mean ± SD). SOL (n = 19) VL (n = 14) DCTendon (V/Vmax) 0.567 ± 0.128 0.180 ± 0.053* DCBelly (V/Vmax) 0.016 ± 0.008 0.003 ± 0.002* DCMTU (V/Vmax) 0.574 ± 0.127 0.179 ± 0.014* *Statistically significant difference between the two muscles (p<0.05).  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 7 of 16 in the two muscles. The soleus continuously shortened throughout the stance phase and produced muscular work at a shortening velocity close to the enthalpy efficiency optimum. Vastus lateralis oper- ated with smaller length changes, almost isometrically, resulting in a high force- velocity potential, which is beneficial for economic force generation. Both muscles operated close to L0, that is, at a high force- length potential. Tendon compliance was responsible for the majority of the overall decoupling of MTU and fascicles in both muscles, enabling favorable conditions for muscle force or muscle work production. Only in the soleus muscle did fascicle rotation contribute to the overall decoupling, indi- cating an additional, yet comparatively minor effect on the fascicle dynamics during locomotion. The triceps surae and quadriceps muscle groups are the main contributors for locomotion and thus responsible for a great portion of the metabolic energy cost of running (Farris and Sawicki, 2012; Fletcher and MacIntosh, 2015; Uchida et al., 2016; Hamner and Delp, 2013). While the quadriceps mainly decelerates and supports body mass in the early stance phase, the triceps surae contributes to the acceleration of the center of mass during the second part of the stance phase (Dorn et al., 2012; Hamner and Delp, 2013). The soleus is the largest muscle of the triceps surae (Albracht et al., 2008) and the vastus lateralis of the quadriceps (Mersmann et al., 2015) and thus both muscles are important for the running movement. We found that the soleus actively shortened throughout the entire stance phase, indicating continuous work/energy production. The average velocity at which the soleus shortened was very close to the optimal velocity for maximal enthalpy efficiency. Enthalpy effi- ciency quantifies the fraction of chemical energy from ATP hydrolysis that is converted into mechanical muscular work (Hill, 1964; Barclay, 2015) with a peak at around 20%  of Vmax (Hill, 1939; Barclay et al., 1993). Consequently, the mechanical work performed by the soleus muscle, being essential during running (Arampatzis et al., 1999; Stefanyshyn and Nigg, 1998; Hamner and Delp, 2013; Lai et al., 2015) and high enough in magnitude to significantly influence the overall metabolic energy cost of locomotion (Bohm et al., 2019; Sawicki et al., 2020; Beck et al., 2019), was generated at a high enthalpy efficiency (94%  of maximum efficiency). Considering that also the soleus force- length potential was close to the maximum (0.92) and that a high potential may decrease the active muscle volume for a given muscle force (Beck et al., 2019; Biewener and Roberts, 2000; Fletcher and MacIntosh, 2017), our results provide evidence of a cumulative contribution of two different mech- anisms (high force- length potential and high enthalpy efficiency) to an advantageous muscular work production of the soleus during running. The vastus lateralis was mainly active in the first part of the stance phase and its fascicles operated with very small length changes, that is, almost isometrically, confirming earlier reports (Bohm et al., 2018; Monte et al., 2020). This indicates that the vastus lateralis dissipates and/or produces negligible amounts of mechanical energy during running, yet generating force for the deceleration and support of the body mass. The observed decoupling of the vastus lateralis MTU and fascicles showed that the deceleration of the body mass in the early stance phase was not a result of an energy dissipation by the contractile element (active stretch) but rather an energy absorption by the tendinous tissue. Tendons feature low damping characteristics, resulting in a hysteresis of only 10%  (Pollock and Shadwick, 1994; Bennett et al., 1986), and, therefore, the main part of the absorbed energy of the body’s deceleration is expected to be stored as elastic tendon strain energy, which is then returned later in the second part of the stance phase. The high force- length (0.93) and force- velocity (0.96) potential of the vastus lateralis muscle throughout stance indi- cates an energy exchange within the vastus lateralis MTU under almost optimal conditions for muscle force generation during running. Operating at high potentials reduces the active muscle volume for a given force (Biewener and Roberts, 2000; Fletcher and MacIntosh, 2017) and thus the metabolic energy cost of muscle force generation. By actively shortening the soleus delivered energy during the entire stance phase to the skeleton, providing the main muscular work required for running. On the other side, the contractile elements of the vastus lateralis muscle did not contribute to the required muscular work and operated in concert with the elastic tendon in favor of energy storage (Roberts and Azizi, 2011). Our findings showed that, although the human body interacts with the ground in a spring- like manner during steady- state running to store mechanical energy (Dickinson et al., 2000; Roberts and Azizi, 2011), there are indeed muscles that operate as work generators, like the soleus, and others that promote energy conservations, like the vastus lateralis. Further, our results indicate that the fascicle operating length and velocity of the soleus muscle, the main work generator, is optimized for high enthalpy efficiency, while of the vastus lateralis muscle, which promote energy conservation, for a high potential of force  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 8 of 16 generation. The consequence of the active shortening of the soleus muscle for work production is a decrease of the force- velocity potential during the stance phase, which may increase the active muscle volume and shortening- related cost (Hill, 1938; Fenn, 1924; Smith et al., 2005; He et al., 2000). However, the soleus muscle features shorter fascicles (L0 = 41 mm) compared to the vastus lateralis muscle (L0 = 94 mm), and, for this reason, a given force generated by the soleus is energetically less expensive (Biewener and Roberts, 2000). The specific morphology of the soleus muscle certainly compensates for the reductions of the force- velocity potential and provides advantages for its function as work generator during submaximal steady- state running. Furthermore, operating around the ‘sweet spot’ of the shortening velocity for high enthalpy efficiency facilitates the economical muscular work production, while either a too high or a too low shortening velocity would be disadvantageous. The advantageous operating conditions specific for soleus and vastus lateralis during submaximal running shown here for a moderate speed of 2.5 m/s seem to persist at faster running speeds as well. This is indicated by recent evidence of a comparable muscle operating length and velocity of the soleus (Lai et al., 2015) and vastus lateralis (Monte et al., 2020) over a broad range of running speeds, respec- tively. In addition, the operating behavior of both muscles seems to reflect their respective muscle group. The gastrocnemius muscles, as the second largest plantar flexors, have been shown to operate at a length similar to soleus and only at slightly higher velocities (Lai et al., 2018), suggesting efficient work production too. For the other monoarticular knee extensors, vastus medialis and intermedius, the resting fascicle length is about similar to the vastus lateralis (Ward et al., 2009), and, since they share the same single patellar tendon, we also do not expect that those muscles operate substantially different (Arnold et al., 2013), that is, likewise at a high force potential. The almost optimal conditions for muscular work production and muscle force generation of the soleus and vastus lateralis were a result of an effective decoupling between MTU and fascicle length that was regulated by an appropriate muscle activation. For the soleus, the activation level increased in the first part of stance phase, contracting the muscle while the MTU increased in length. This activation pattern not only prevented the muscle to be stretched but also induced continuous short- ening around the plateau of the force- length curve at a high enthalpy efficiency. The respective high DCTendon further indicates that a part of the body’s mechanical energy was stored as strain energy in the Achilles tendon in addition to the generated work by fascicle shortening. During MTU shortening (propulsion phase), the soleus EMG activity decreased and the tendon recoiled, enabling the high shortening velocities of the MTU while maintaining the fascicle operating conditions close to the effi- ciency optimum. The simultaneous release of the stored strain energy from the tendon further added to the ongoing muscle work production, that is, energy amplification. The vastus lateralis muscle showed higher levels of activation during the initial part of the stance phase and earlier deactivation than soleus. The timing and level of activation regulated the decoupling within the vastus lateralis MTU during the body mass deceleration in a magnitude that the lengthening and shorting of the MTU was fully accomplished by the tendinous tissue. Consequently, the vastus lateralis fascicles operated at a high force- length- velocity potential and the body’s energy was stored within the MTU. Although being substantial for soleus and vastus lateralis, the SPM analysis revealed higher values of DCTendon for soleus during the major part of the stance phase (average value for soleus 0.57 V/Vmax and vastus lateralis 0.18 V/Vmax), indicating a greater decoupling within the soleus MTU compared to the vastus lateralis MTU. In the soleus muscle, fascicle rotation (changes in pennation angle) had an additional effect on the overall decoupling between MTU and fascicles. The results showed an increase in DCBelly in the second part of the stance phase where the soleus belly velocity was high during the MTU short- ening. However, the decoupling by the fascicle rotation was considerably smaller compared to the tendon decoupling. Over the stance phase, belly and tendon decoupling were 1.6% Vmax and 57% Vmax and during the MTU shortening phase 2.6% Vmax and 72% Vmax, respectively, suggesting a rather minor functional role of fascicle rotation during submaximal running. In the vastus lateralis, fascicle rotation was virtually absent and consequently DCBelly values showed no relevant decoupling effect at all. Note that because of the extensive experimental protocol for each muscle it was not possible to measure soleus and vastus lateralis in the same participants. However, both groups are a represen- tative sample and no significant differences were found in anthropometrics and relevant gait param- eters. Furthermore, for the determination of the vastus lateralis force- length curve, the muscle was not isolated, hence the curve also includes the contribution of the vastus medialis and intermedius. The underlying assumption for this approach is that the force- length curves of these three synergistic  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 9 of 16 knee extensors are comparable, which is supported by the study of Herzog et al. (Herzog et al., 1990). Besides, it is currently not possible to measure enthalpy efficiency directly during running. Instead, we used an experimentally determined efficiency- velocity curve reported by Hill (Hill, 1964) and confirmed by others (Barclay et al., 2010) to relate the measured operating fascicle velocities of both muscles to the enthalpy efficiency. We were also not able to directly measure Vmax of both muscles and despite using a biologically funded value, its choice affects the force- velocity potential and enthalpy efficiency. However, when conducting a sensitivity analysis by substantially reducing or increasing Vmax by 30%, the force- velocity potential of vastus lateralis only changed for Vmax -30% and Vmax +30 % from 0.96 to 0.94 and 0.98 and of soleus from 0.63 to 0.52 and 0.65 while the enthalpy effi- ciency changed from 0.082 to 0.081 and 0.016 for vastus lateralis and from 0.425 to 0.439 and 0.403 for soleus, respectively, without an impact on the significance of the differences between muscles (potential p<0.001, efficiency p<0.001). These results support the robustness of our primary outcomes and strengthen our conclusions. In conclusion, the present study demonstrated that during the stance phase of steady- state running, when the human body interacts with the environment in a spring- like manner, the soleus muscle acts as work generator and the vastus lateralis muscle as energy conservator. Furthermore, our findings provide evidence that the soleus operates under conditions optimal for muscular work production (i.e., high force- length potential and high enthalpy efficiency) and the vastus lateralis under conditions optimal for muscle force generation (i.e., high force- length and high force- velocity potential). Materials and methods Participants and experimental design Thirty- three physically active adults, accustomed to regular running on a recreational basis (i.e., no competitive runners), were included in the present investigation. None of the participants reported any history of neuromuscular or skeletal impairments in the 6 months prior to the recordings. The ethics committee of the university approved the study (EA2/076/15), and the participants gave written informed consent in accordance with the Declaration of Helsinki. From the right leg, either the soleus (n = 19, 29 ± 6 years, 177 ± 9 cm, 69 ± 9 kg, seven females) or vastus lateralis (n = 14, age 28 ± 4 years, height 179 ± 7 cm, body mass 75 ± 8 kg, three females) muscle fascicle length, fascicle pennation angle, and EMG activity were recorded during running on a treadmill at 2.5  m/s. Corresponding MTU lengths were calculated from the kinematic data and individually measured tendon lever arms. We further assessed the soleus and vastus lateralis force- fascicle length and force- fascicle velocity relationship to calculate the force- length and force- velocity potential of the soleus and the vastus lateralis muscle fascicles during running. The operating fascicle velocity was additionally mapped on the enthalpy efficiency- velocity relationship to assess the enthalpy efficiency of both muscles. The contribution of the decoupling of the fascicle length and velocity from the MTU to the operating force potential and enthalpy efficiency at the level of tendon and muscle belly during running was examined for both muscles as well. All data for one participant were collected on the same day and sensors (EMG, ultrasound, reflective markers) remained attached between the different parts of the experiment. Joint kinematics, fascicle behavior, and electromyographic activity during running After a familiarization phase, a 4 min running trial on a treadmill (soleus: h/p cosmos mercury, Isny, Germany; vastus lateralis: Daum electronic, ergo_run premium8, Fürth, Germany) was performed and kinematics of the right leg were captured by a Vicon motion capture system (version 1.8, Vicon Motion Systems, Oxford, UK, 250  Hz) using an anatomically referenced reflective marker setup (greater trochanter, lateral femoral epicondyle and malleolus, fifth metatarsal, and tuber calcanei). The kinematic data were used to determine the touchdown of the foot and the toe- off as consecutive minima in knee joint angle over time (Fellin et al., 2010). Furthermore, the kinematics of the ankle and knee joint served to calculate the MTU length change of the soleus and vastus lateralis during running as the product of ankle joint angle changes and Achilles tendon lever arm as well as knee joint angle changes and patellar tendon lever arm (Lutz and Rome, 1996), respectively. We used the ultrasound- based tendon- excursion method for the Achilles tendon lever arm determination (An  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 10 of 16 et al., 1984). The patellar tendon lever arm was measured using magnetic resonance imaging in fully extended knee joint position and calculated as a function of the knee joint angle change using the data by Herzog and Read (Herzog and Read, 1993; for a detailed description of both tendon lever arm measurements, see Bohm et al., 2019; Bohm et al., 2018; Bohm et al., 2021). The initial soleus and vastus lateralis MTU length was calculated based on the regression equation provided by Hawkins and Hull (Hawkins and Hull, 1990) at neutral ankle joint angle for the soleus MTU and at touchdown for the vastus lateralis MTU. During the running trial, ultrasound images of either the soleus or vastus lateralis muscle fascicles were recorded synchronously to the kinematic data (soleus: Aloka Prosound Alpha 7, Hitachi, Tokyo, Japan, 6 cm linear array probe, UST- 5713T, 13.3 MHz, 146 Hz; vastus lateralis: My Lab60, Esaote, Genova, Italy, 10 cm linear array probe LA923, 10 MHz, 43 Hz). The ultrasound probe was mounted over the medial aspect of the soleus muscle belly or on the vastus lateralis muscle belly (≈50%  of femur length) using a custom anti- skid neoprene- plastic cast. The fascicle length was post- processed from the ultrasound images using a self- developed semi- automatic tracking algorithm (Marzilger et al., 2018) that calculated a representative reference fascicle on the basis of multiple muscle fascicle portions identified from the entire displayed muscle (for details, see Bohm et al., 2018; Marzilger et al., 2018; Figure 5). Visual inspection of each image was conducted and correc- tions were made if necessary. At least nine steps were analyzed for each participant and then aver- aged (Bohm et al., 2018; Giannakou et al., 2011). The pennation angle was calculated as the angle between the deeper aponeurosis and the reference fascicle (Figure 5). The length changes of the muscle belly of soleus and vastus lateralis were calculated as the differences of consecutive products of fascicle length and the respective cosine of the pennation angle (Fukunaga et al., 2001). Note rF Fascicle length (mm) 5000 4000 3000 2000 1000 0 Force (N) 20 40 60 80 A upper apon. deeper apon. 1 cm F θ upper aponeurosis deeper aponeurosis θ 0 2000 4000 6000 8000 Force (N) Fascicle length (mm) 60 70 80 90 100 B F 1 cm Figure 5. Experimental setup for the determination of the soleus (A) and vastus lateralis (B) force- fascicle length relationship. Maximum isometric plantar flexions (MVC) at eight different joint angles were performed on a dynamometer. During the MVCs, ultrasound images of the soleus and vastus lateralis were recorded and a representative muscle fascicle length (F) was calculated based on multiple fascicle portions (short dashed lines). Accordingly, an individual force- fascicle length relationship for the soleus and vastus lateralis muscle was derived from the MVCs (squares) by means of a second- order polynomial fit (dashed line, bottom graphs, MVCs and curves of one representative participant).  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 11 of 16 that this does not give the length of the entire soleus or vastus lateralis muscle belly but rather the projection of the instant fascicle length onto the plane of the MTU, which can be used to calculate the changes of the belly length (Bohm et al., 2019). The velocities of fascicles, belly, and MTU were calculated as the first derivative of the lengths over time. Surface EMG of the vastus lateralis and the soleus were measured by means of a wireless EMG system (Myon m320RX, Myon AG, Baar, Switzerland, 1000 Hz). A fourth- order high- pass Butterworth filter with 50 Hz cut- off frequency, a full- wave rectification, and then a low- pass filter with 20 Hz cut- off frequency were applied to the raw EMG data. The EMG activity was averaged over the same steps that were analyzed for the soleus parameters and for the vastus lateralis over 10 running steps. EMG values were then normalized for each participant to the maximum obtained during a individual maximum voluntary contraction. Assessment of the force-length, force-velocity, and enthalpy efficiency- velocity relationship To determine the soleus and the vastus lateralis force- length relationship, eight maximum voluntary plantar flexion or knee extension contractions (MVCs) in different joint angles were performed with the right leg on an isokinetic dynamometer (Biodex Medical, Syst. 3, Inc, Shirley, NY), following a standardized warm- up (Bohm et al., 2019; Bohm et al., 2018; Nikolaidou et al., 2017; Figure 5). For the plantar flexion MVCs, the participants were placed in prone position with the knee in fixed flexed position (~120°) to restrict the contribution of the bi- articular m. gastrocnemius to the plantar flexion moment (Hof and van den Berg, 1977) and the joint angles were set in a randomized equally distrib- uted order ranging from 10° plantar flexion to the individual maximum dorsiflexion angle. Regarding the knee extensions, participants were seated with a hip joint angle of 85° to reduce the contribution of the bi- articular m. rectus femoris (Herzog et al., 1990), while the knee joint angle ranged between 20° to 90° knee joint angle (0° = knee extended) in randomly ordered 10° intervals. The resultant moments at the ankle and knee joint were calculated under consideration of the effects of gravita- tional and passive moments and any misalignment between joint axis and dynamometer axis using an established inverse dynamics approach (Arampatzis et al., 2005; Arampatzis et al., 2004). The required kinematic data were recorded during the MVCs based on anatomically referenced reflective markers (medial and lateral malleoli and epicondyle, calcaneal tuberosity, second metatarsal, and greater trochanter) by a Vicon motion capture system (250 Hz). Furthermore, the contribution of the antagonistic moment produced by tibialis anterior during the plantar flexion MVCs or by the hamstring muscles during the knee extension MVCs was taken into account by means of an EMG- based method according to Mademli et al. (Mademli et  al., 2004), considering the force- length dependency of the antagonists (Bohm et al., 2021). The force applied to the Achilles or patellar tendon during the plantar flexion or knee extension MVCs was calculated as quotient of the joint moment and individual tendon lever arm, respectively. The soleus or the vastus lateralis fascicle behavior during the MVCs was synchronously captured by ultrasonography and fascicle length was determined using the same methodology described above (Figure 5). Accordingly, an individual force- fascicle length relationship was calculated for soleus or vastus lateralis by means of a second- order polynomial fit and Fmax and L0 was derived, respectively (Figure 5). The force- velocity relationship of the soleus and the vastus lateralis muscle was further assessed using the classical Hill equation (Hill, 1938) and the muscle- specific Vmax and constants of arel and brel. For Vmax, we took values of human soleus and vastus lateralis type 1 and 2 fibers measured in vitro at 15 °C reported by Luden et al. (Luden et al., 2008). The values were then adjusted (Ranatunga, 1984) for physiological temperature conditions (37 °C) and an average fiber type distribution of the human soleus (type 1 fibers: 81%, type 2: 19%) and vastus lateralis muscle (type 1 fibers: 37%, type 2: 63%) reported in literature (Johnson et al., 1973; Luden et al., 2008; Edgerton et al., 1975; Larsson and Moss, 1993) was the basis to derive a representative value of Vmax. For the soleus muscle under the in vivo condition, Vmax was calculated as 6.77 L0/s and for the vastus lateralis as 11.51 L0/s. For L0, we then referred to the individually measured optimal fascicle length (described above, Figure 5). The constant arel was calculated as 0.1 + 0.4 FT, where FT is the fast twitch fiber type percentage, which then equals to 0.175 for the soleus and 0.351 for the vastus lateralis (Winters and Stark, 1985; Winters and Stark, 1988). The product of arel and Vmax gives the constant brel as 1.182 for the soleus and 4.042 for the vastus lateralis (Umberger et al., 2003). Based on the assessed force- length and  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 12 of 16 force- velocity relationships, we calculated the individual force- length and force- velocity potential of both muscles as a function of the fascicle operating length and velocity during the stance phase of running. The product of both potentials then gives the overall force- length- velocity potential. Furthermore, we determined the enthalpy efficiency- velocity relationship for the soleus and the vastus lateralis muscle fascicles in order to calculate the enthalpy efficiency of both muscles as a function of the fascicle operating velocity during running. For this purpose, we used the experimental efficiency values provided by the paper of Hill, 1964 in Table 1 for a/P0 = 0.25 (Hill, 1964). By means of the classical Hill equation (Hill, 1938), we then transposed the original efficiency values that were presented as a function of relative load (relative to maximum tension) to shortening velocity (normal- ized to Vmax). The values of enthalpy efficiency and shortening velocity were then fitted using a cubic spline, giving the right- skewed parabolic- shaped curve with a peak efficiency of 0.45 at a velocity of 0.18 V/Vmax. The resulting function was then used to calculate the enthalpy efficiency of the soleus and the vastus lateralis during running based on the average value of the fascicle velocity over stance, accordingly. Assessment of decoupling within the MTU To quantify the decoupling of fascicle, belly, and MTU velocities over the time course of stance, we calculated a decoupling coefficient to account for the tendon compliance (DCTendon, equation 1), fascicle rotation (DCBelly, equation 2), as well as for the overall decoupling of MTU and fascicle veloci- ties that includes both components (DCMTU, equation 3). DCTendon ( t ) = VMTU ( t ) − VBelly ( t ) /Vmax (1) DCBelly ( t ) = |VBelly ( t ) − VFascicle ( t ) |/Vmax (2) DCMTU ( t ) = |VMTU ( t ) − VFascicle ( t ) |/Vmax (3) where V(t) is the velocity at each percentage of the stance phase (i.e. t = 0, 1, …, 100% stance). We introduced these new decoupling coefficients because previously suggested decoupling ratios (i.e., tendon gearing = VMTU/VBelly, belly gearing [or architectural gear ratio] = VBelly/VFascicle, MTU gearing = VMTU/VFascicle; Azizi et al., 2008; Wakeling et al., 2011) may feature limitations for the application under in vivo conditions, that is, considering that muscle belly and fascicle velocities may be very close to or even zero during functional tasks as walking and running (Bohm et al., 2019; Bohm et al., 2018), which results in non- physiological gear ratios. Statistics A t- test for independent samples was used to test for group differences in anthropometric character- istics, temporal gait parameters, and differences between the soleus and the vastus lateralis fascicle belly, MTU, and EMG parameters. The Mann–Whitney U test was applied in case the assumption of normal distribution, tested by the Kolmogorov–Smirnov test with Lilliefors correction, was not met. The level of significance was set to α = 0.05, and the statistical analyses were performed using SPSS (IBM Corp., version 22, NY). Furthermore, SPM (independent samples t- test, α = 0.05) was used to test for differences between the DCTendon, DCBelly, and DCMTU of the soleus and the vastus lateralis throughout the stance phase of running. SPM was conducted using the software package spm1D (version 0.4, http://www. spm1d. org; Pataky, 2012). Acknowledgements Funding for this research was supplied by the German Federal Institute of Sport Science (grant no. ZMVI14- 070604/17- 18). The magnetic resonance image acquisition was funded by the foundation Stiftung Oskar- Helene- Heim. We further acknowledge support by the German Research Foundation (DFG) and the Open Access Publication Fund of Humboldt- Universität zu Berlin.  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 13 of 16 Additional information Funding Funder Grant reference number Author German Federal Institute of Sport Science ZMVI14-070604/17-18 Adamantios Arampatzis Stiftung Oskar Helene Heim Sebastian Bohm The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication. Author contributions Sebastian Bohm, Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Project administration, Visualization, Writing - original draft, Writing – review and editing; Falk Mers- mann, Data curation, Methodology, Writing – review and editing; Alessandro Santuz, Methodology, Writing – review and editing; Arno Schroll, Methodology, Software, Writing – review and editing; Adamantios Arampatzis, Conceptualization, Funding acquisition, Methodology, Supervision, Writing - original draft, Writing – review and editing Author ORCIDs Sebastian Bohm http:// orcid. org/ 0000- 0002- 5720- 3672 Falk Mersmann http:// orcid. org/ 0000- 0001- 7180- 7109 Alessandro Santuz http:// orcid. org/ 0000- 0002- 6577- 5101 Adamantios Arampatzis http:// orcid. org/ 0000- 0002- 4985- 0335 Ethics Human subjects: The ethics committee of the Humboldt- Universität zu Berlin approved the study and the participants gave written informed consent in accordance with the Declaration of Helsinki. Decision letter and Author response Decision letter https:// doi. org/ 10. 7554/ eLife. 67182. sa1 Author response https:// doi. org/ 10. 7554/ eLife. 67182. sa2 Additional files Supplementary files • Transparent reporting form Data availability The final processed data can be found at: https:// doi. org/ 10. 6084/ m9. figshare. 14046749. The following dataset was generated: Author(s) Year Dataset title Dataset URL Database and Identifier Bohm S, Mersmann F, Santuz A, Schroll A, Arampatzis A 2021 Data_Muscle- specific economy of force generation and efficiency of work production during human running https:// doi. org/ 10. 6084/ m9. figshare. 14046749 figshare, 10.6084/ m9.figshare.14046749 References Albracht K, Arampatzis A, Baltzopoulos V. 2008. Assessment of muscle volume and physiological cross- sectional area of the human triceps Surae muscle in vivo. Journal of Biomechanics 41: 2211–2218. DOI: https:// doi. org/ 10. 1016/ j. jbiomech. 2008. 04. 020, PMID: 18555257 Alexander RM. 1991. Energy- saving mechanisms in walking and running. The Journal of Experimental Biology 160: 55–69 PMID: 1960518.  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 14 of 16 An KN, Takahashi K, Harrigan TP, Chao EY. 1984. Determination of muscle orientations and moment arms. Journal of Biomechanical Engineering 106: 280–282. DOI: https:// doi. org/ 10. 1115/ 1. 3138494, PMID: 6492774 Arampatzis A, Brüggemann GP, Metzler V. 1999. The effect of speed on leg stiffness and joint kinetics in human running. Journal of Biomechanics 32: 1349–1353. DOI: https:// doi. org/ 10. 1016/ S0021- 9290( 99) 00133- 5, PMID: 10569714 Arampatzis A, Karamanidis K, De Monte G, Stafilidis S, Morey- Klapsing G, Brüggemann GP. 2004. Differences between measured and resultant joint moments during voluntary and artificially elicited isometric knee extension contractions. Clinical Biomechanics 19: 277–283. DOI: https:// doi. org/ 10. 1016/ j. clinbiomech. 2003. 11. 011 Arampatzis A, Morey- Klapsing G, Karamanidis K, DeMonte G, Stafilidis S, Brüggemann GP. 2005. Differences between measured and resultant joint moments during isometric contractions at the ankle joint. Journal of Biomechanics 38: 885–892. DOI: https:// doi. org/ 10. 1016/ j. jbiomech. 2004. 04. 027, PMID: 15713310 Arampatzis A, De Monte G, Karamanidis K, Morey- Klapsing G, Stafilidis S, Brüggemann G- P. 2006. Influence of the muscle- tendon unit’s mechanical and morphological properties on running economy. The Journal of Experimental Biology 209: 3345–3357. DOI: https:// doi. org/ 10. 1242/ jeb. 02340, PMID: 16916971 Arnold EM, Hamner SR, Seth A, Millard M, Delp SL. 2013. How muscle fiber lengths and velocities affect muscle force generation as humans walk and run at different speeds. The Journal of Experimental Biology 216: 2150–2160. DOI: https:// doi. org/ 10. 1242/ jeb. 075697, PMID: 23470656 Azizi E, Brainerd EL, Roberts TJ. 2008. Variable gearing in pennate muscles. PNAS 105: 1745–1750. DOI: https:// doi. org/ 10. 1073/ pnas. 0709212105, PMID: 18230734 Barclay CJ, Constable JK, Gibbs CL. 1993. Energetics of fast- and slow- twitch muscles of the mouse. The Journal of Physiology 472: 61–80. DOI: https:// doi. org/ 10. 1113/ jphysiol. 1993. sp019937, PMID: 8145164 Barclay CJ, Woledge RC, Curtin NA. 2010. Is the efficiency of mammalian (mouse) skeletal muscle temperature dependent? The Journal of Physiology 588: 3819–3831. DOI: https:// doi. org/ 10. 1113/ jphysiol. 2010. 192799, PMID: 20679354 Barclay CJ. 2015. Energetics of contraction. Comprehensive Physiology 5: 961–995. DOI: https:// doi. org/ 10. 1002/ cphy. c140038, PMID: 25880520 Beck ON, Punith LK, Nuckols RW, Sawicki GS. 2019. Exoskeletons improve locomotion economy by reducing active muscle volume. Exercise and Sport Sciences Reviews 47: 237–245. DOI: https:// doi. org/ 10. 1249/ JES. 0000000000000204, PMID: 31436749 Bennett MB, Ker RF, Imery NJ, Alexander1 RMcN. 1986. Mechanical properties of various mammalian tendons. Journal of Zoology 209: 537–548. DOI: https:// doi. org/ 10. 1111/ j. 1469- 7998. 1986. tb03609.x Biewener AA, Roberts TJ. 2000. Muscle and tendon contributions to force, work, and elastic energy savings: A comparative perspective. Exercise and Sport Sciences Reviews 28: 99–107 PMID: 10916700. Biewener AA, Daley MA. 2007. Unsteady locomotion: Integrating muscle function with whole body dynamics and neuromuscular control. The Journal of Experimental Biology 210: 2949–2960. DOI: https:// doi. org/ 10. 1242/ jeb. 005801, PMID: 17704070 Biewener AA. 2016. Locomotion as an emergent property of muscle contractile dynamics. The Journal of Experimental Biology 219: 285–294. DOI: https:// doi. org/ 10. 1242/ jeb. 123935, PMID: 26792341 Bohm S, Marzilger R, Mersmann F, Santuz A, Arampatzis A. 2018. Operating length and velocity of human vastus lateralis muscle during walking and running. Scientific Reports 8: 5066. DOI: https:// doi. org/ 10. 1038/ s41598- 018- 23376- 5, PMID: 29567999 Bohm S, Mersmann F, Santuz A, Arampatzis A. 2019. The force- length- velocity potential of the human soleus muscle is related to the energetic cost of running. Proceedings. Biological Sciences 286: 20192560. DOI: https:// doi. org/ 10. 1098/ rspb. 2019. 2560, PMID: 31847774 Bohm S, Mersmann F, Santuz A, Arampatzis A. 2021. Enthalpy efficiency of the soleus muscle contributes to improvements in running economy. Proceedings. Biological Sciences 288: 20202784. DOI: https:// doi. org/ 10. 1098/ rspb. 2020. 2784, PMID: 33499791 Cavagna GA, Saibene FP, Margaria R. 1964. Mechanical work in running. Journal of Applied Physiology 19: 249–256. DOI: https:// doi. org/ 10. 1152/ jappl. 1964. 19. 2. 249, PMID: 14155290 Dickinson MH, Farley CT, Full RJ, Koehl MA, Kram R, Lehman S. 2000. How animals move: An integrative view. Science 288: 100–106. DOI: https:// doi. org/ 10. 1126/ science. 288. 5463. 100, PMID: 10753108 Dorn TW, Schache AG, Pandy MG. 2012. Muscular strategy shift in human running: Dependence of running speed on hip and ankle muscle performance. The Journal of Experimental Biology 215: 1944–1956. DOI: https:// doi. org/ 10. 1242/ jeb. 064527, PMID: 22573774 Edgerton VR, Smith JL, Simpson DR. 1975. Muscle fibre type populations of human leg muscles. Histochem. J 7: 259–266. DOI: https:// doi. org/ 10. 1007/ BF01003594, PMID: 123895 Farris DJ, Sawicki GS. 2012. The mechanics and energetics of human walking and running: a joint level perspective. J. R. Soc. Interface 9: 110–118. DOI: https:// doi. org/ 10. 1098/ rsif. 2011. 0182, PMID: 21613286 Fellin RE, Rose WC, Royer TD, Davis IS. 2010. Comparison of methods for kinematic identification of footstrike and toe- off during overground and treadmill running. Journal of Science and Medicine in Sport 13: 646–650. DOI: https:// doi. org/ 10. 1016/ j. jsams. 2010. 03. 006, PMID: 20478742 Fenn WO. 1924. The relation between the work performed and the energy liberated in muscular contraction. The Journal of Physiology 58: 373–395. DOI: https:// doi. org/ 10. 1113/ jphysiol. 1924. sp002141, PMID: 16993634 Fletcher JR, MacIntosh BR. 2015. Achilles tendon strain energy in distance running: Consider the muscle energy cost. Journal of Applied Physiology 118: 193–199. DOI: https:// doi. org/ 10. 1152/ japplphysiol. 00732. 2014, PMID: 25593218  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 15 of 16 Fletcher JR, MacIntosh BR. 2017. Running economy from a muscle energetics perspective. Frontiers in Physiology 8: 433. DOI: https:// doi. org/ 10. 3389/ fphys. 2017. 00433, PMID: 28690549 Fukunaga T, Kubo K, Kawakami Y, Fukashiro S, Kanehisa H, Maganaris CN. 2001. In vivo behaviour of human muscle tendon during walking. Proceedings. Biological Sciences 268: 229–233. DOI: https:// doi. org/ 10. 1098/ rspb. 2000. 1361, PMID: 11217891 Giannakou E, Aggeloussis N, Arampatzis A. 2011. Reproducibility of gastrocnemius medialis muscle architecture during treadmill running. Journal of Electromyography and Kinesiology 21: 1081–1086. DOI: https:// doi. org/ 10. 1016/ j. jelekin. 2011. 06. 004 Gordon AM, Huxley AF, Julian FJ. 1966. The variation in isometric tension with sarcomere length in vertebrate muscle fibres. The Journal of Physiology 184: 170–192. DOI: https:// doi. org/ 10. 1113/ jphysiol. 1966. sp007909, PMID: 5921536 Hamner SR, Delp SL. 2013. Muscle contributions to fore- aft and vertical body mass center accelerations over a range of running speeds. Journal of Biomechanics 46: 780–787. DOI: https:// doi. org/ 10. 1016/ j. jbiomech. 2012. 11. 024, PMID: 23246045 Hawkins D, Hull ML. 1990. A method for determining lower extremity muscle- tendon lengths during flexion/ extension movements. Journal of Biomechanics 23: 487–494. DOI: https:// doi. org/ 10. 1016/ 0021- 9290( 90) 90304- L, PMID: 2373721 He ZH, Bottinelli R, Pellegrino MA, Ferenczi MA, Reggiani C. 2000. ATP consumption and efficiency of human single muscle fibers with different myosin isoform composition. Biophysical Journal 79: 945–961. DOI: https:// doi. org/ 10. 1016/ S0006- 3495( 00) 76349- 1, PMID: 10920025 Herzog W, Abrahamse SK, ter Keurs HE. 1990. Theoretical determination of force- length relations of intact human skeletal muscles using the cross- bridge model. Pflugers Archiv 416: 113–119. DOI: https:// doi. org/ 10. 1007/ BF00370231, PMID: 2352828 Herzog W, Read LJ. 1993. Lines of action and moment arms of the major force- carrying structures crossing the human knee joint. Journal of Anatomy 182: 213–230 PMID: 8376196. Hill AV. 1938. The heat of shortening and the dynamic constants of muscle. Proceedings of the Royal Society of London. Series B - Biological Sciences 126: 136–195. DOI: https:// doi. org/ 10. 1098/ rspb. 1938. 0050 Hill AV. 1939. The mechanical efficiency of frog’s muscle. Proceedings of the Royal Society of London. Series B - Biological Sciences 127: 434–451. DOI: https:// doi. org/ 10. 1098/ rspb. 1939. 0033 Hill AV. 1964. The efficiency of mechanical power development during muscular shortening and its relation to load. Proceedings of the Royal Society of London. Series B, Biological Sciences 159: 319–324. DOI: https:// doi. org/ 10. 1098/ rspb. 1964. 0005, PMID: 14114167 Hof AL, van den Berg Jw. 1977. Linearity between the weighted sum of the EMGs of the human triceps surae and the total torque. J. Biomech 10: 529–539. DOI: https:// doi. org/ 10. 1016/ 0021- 9290( 77) 90033- 1, PMID: 914876 Johnson MA, Polgar J, Weightman D, Appleton D. 1973. Data on the distribution of fibre types in thirty- six human muscles. Journal of the Neurological Sciences 18: 111–129. DOI: https:// doi. org/ 10. 1016/ 0022- 510x( 73) 90023-3 Kram R, Taylor CR. 1990. Energetics of running: a new perspective. Nature 346: 265–267. DOI: https:// doi. org/ 10. 1038/ 346265a0, PMID: 2374590 Kram R. 2000. Muscular force or work: what determines the metabolic energy cost of running? Exercise and Sport Sciences Reviews 28: 138–143 PMID: 10916707. Lai A, Schache AG, Lin YC, Pandy MG. 2014. Tendon elastic strain energy in the human ankle plantar- flexors and its role with increased running speed. The Journal of Experimental Biology 217: 3159–3168. DOI: https:// doi. org/ 10. 1242/ jeb. 100826, PMID: 24948642 Lai A, Lichtwark GA, Schache AG, Lin YC, Brown NAT, Pandy MG. 2015. In vivo behavior of the human soleus muscle with increasing walking and running speeds. Journal of Applied Physiology 118: 1266–1275. DOI: https:// doi. org/ 10. 1152/ japplphysiol. 00128. 2015 Lai AKM, Lichtwark GA, Schache AG, Pandy MG. 2018. Differences in in vivo muscle fascicle and tendinous tissue behavior between the ankle plantarflexors during running. Scandinavian Journal of Medicine & Science in Sports 28: 1828–1836. DOI: https:// doi. org/ 10. 1111/ sms. 13089 Larsson L, Moss RL. 1993. Maximum velocity of shortening in relation to myosin isoform composition in single fibres from human skeletal muscles. J. Physiol 472: 595–614. DOI: https:// doi. org/ 10. 1113/ jphysiol. 1993. sp019964, PMID: 8145163 Luden N, Minchev K, Hayes E, Louis E, Trappe T, Trappe S. 2008. Human vastus lateralis and soleus muscles display divergent cellular contractile properties. American Journal of Physiology. Regulatory, Integrative and Comparative Physiology 295: R1593–R1598. DOI: https:// doi. org/ 10. 1152/ ajpregu. 90564. 2008, PMID: 18815206 Lutz GJ, Rome LC. 1996. Muscle function during jumping in frogs I Sarcomere length change, EMG pattern, and jumping performance. Am. J. Physiol 271: C563-C570. DOI: https:// doi. org/ 10. 1152/ ajpcell. 1996. 271. 2. C563, PMID: 8769996 Mademli L, Arampatzis A, Morey- Klapsing G, Brüggemann GP. 2004. Effect of ankle joint position and electrode placement on the estimation of the antagonistic moment during maximal plantarflexion. J. Electromyogr. Kinesiol 14: 591–597. DOI: https:// doi. org/ 10. 1016/ j. jelekin. 2004. 03. 006, PMID: 15301777 Maganaris CN, Baltzopoulos V, Sargeant AJ. 1998. In vivo measurements of the triceps surae complex architecture in man: implications for muscle function. J. Physiol 512: 603–614. DOI: https:// doi. org/ 10. 1111/ j. 1469- 7793. 1998. 603be. x, PMID: 9763648  Research article Physics of Living Systems Bohm et al. eLife 2021;10:e67182. DOI: https:// doi. org/ 10. 7554/ eLife. 67182 16 of 16 Marzilger R, Legerlotz K, Panteli C, Bohm S, Arampatzis A. 2018. Reliability of a semi- automated algorithm for the vastus lateralis muscle architecture measurement based on ultrasound images. Eur. J. Appl. Physiol 118: 291–301. DOI: https:// doi. org/ 10. 1007/ s00421- 017- 3769- 8, PMID: 29214464 Mersmann F, Bohm S, Schroll A, Boeth H, Duda G, Arampatzis A. 2015. Muscle shape consistency and muscle volume prediction of thigh muscles. Scandinavian Journal of Medicine & Science in Sports 25: e208-e213. DOI: https:// doi. org/ 10. 1111/ sms. 12285, PMID: 24975992 Monte A, Baltzopoulos V, Maganaris CN, Zamparo P. 2020. Gastrocnemius medialis and vastus lateralis in vivo muscle- tendon behavior during running at increasing speeds. Scandinavian Journal of Medicine & Science in Sports 30: 1163–1176. DOI: https:// doi. org/ 10. 1111/ sms. 13662, PMID: 32227378 Nikolaidou ME, Marzilger R, Bohm S, Mersmann F, Arampatzis A. 2017. Operating length and velocity of human M. Vastus lateralis fascicles during vertical jumping. Royal Society Open Science 4: 170185. DOI: https:// doi. org/ 10. 1098/ rsos. 170185, PMID: 28573027 Pataky TC. 2012. One- dimensional statistical parametric mapping in Python. Comput. Methods Biomech. Biomed. Engin 15: 295–301. DOI: https:// doi. org/ 10. 1080/ 10255842. 2010. 527837, PMID: 21756121 Pollock CM, Shadwick RE. 1994. Relationship between body mass and biomechanical properties of limb tendons in adult mammals. Am. J. Physiol 266: R1016. DOI: https:// doi. org/ 10. 1152/ ajpregu. 1994. 266. 3. R1016 Ranatunga KW. 1984. The force- velocity relation of rat fast- and slow- twitch muscles examined at different temperatures. J. Physiol 351: 517–529. DOI: https:// doi. org/ 10. 1113/ jphysiol. 1984. sp015260, PMID: 6747875 Roberts TJ. 2002. The integrated function of muscles and tendons during locomotion. Comp. Biochem. Physiol. A. Mol. Integr. Physiol 133: 1087–1099. DOI: https:// doi. org/ 10. 1016/ S1095- 6433( 02) 00244- 1, PMID: 12485693 Roberts TJ, Azizi E. 2011. Flexible mechanisms: the diverse roles of biological springs in vertebrate movement. J. Exp. Biol 214: 353–361. DOI: https:// doi. org/ 10. 1242/ jeb. 038588, PMID: 21228194 Santuz A, Ekizos A, Kunimasa Y, Kijima K, Ishikawa M, Arampatzis A. 2020. Lower complexity of motor primitives ensures robust control of high- speed human locomotion. Heliyon 6: e05377. DOI: https:// doi. org/ 10. 1016/ j. heliyon. 2020. e05377, PMID: 33163662 Sawicki GS, Beck ON, Kang I, Young AJ. 2020. The exoskeleton expansion: improving walking and running economy. J. NeuroEngineering Rehabil 17: 25. DOI: https:// doi. org/ 10. 1186/ s12984- 020- 00663- 9, PMID: 32075669 Smith NP, Barclay CJ, Loiselle DS. 2005. The efficiency of muscle contraction. Prog. Biophys. Mol. Biol 88: 1–58. DOI: https:// doi. org/ 10. 1016/ j. pbiomolbio. 2003. 11. 014, PMID: 15561300 Stefanyshyn DJ, Nigg BM. 1998. Contribution of the lower extremity joints to mechanical energy in running vertical jumps and running long jumps. Journal of Sports Sciences 16: 177–186. DOI: https:// doi. org/ 10. 1080/ 026404198366885, PMID: 9531006 Uchida TK, Hicks JL, Dembia CL, Delp SL. 2016. Stretching Your Energetic Budget: How Tendon Compliance Affects the Metabolic Cost of Running. PLOS ONE 11: e0150378. DOI: https:// doi. org/ 10. 1371/ journal. pone. 0150378, PMID: 26930416 Umberger BR, Gerritsen KGM, Martin PE. 2003. A model of human muscle energy expenditure. Computer Methods in Biomechanics and Biomedical Engineering 6: 99–111. DOI: https:// doi. org/ 10. 1080/ 1025584031000091678, PMID: 12745424 Wakeling JM, Blake OM, Wong I, Rana M, Lee SSM. 2011. Movement mechanics as a determinate of muscle structure, recruitment and coordination. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences 366: 1554–1564. DOI: https:// doi. org/ 10. 1098/ rstb. 2010. 0294, PMID: 21502126 Ward SR, Eng CM, Smallwood LH, Lieber RL. 2009. Are current measurements of lower extremity muscle architecture accurate. Clin. Orthop 467: 1074–1082. DOI: https:// doi. org/ 10. 1007/ s11999- 008- 0594- 8, PMID: 18972175 Werkhausen A, Cronin NJ, Albracht K, Paulsen G, Larsen AV, Bojsen- Møller J, Seynnes OR. 2019. Training- induced increase in Achilles tendon stiffness affects tendon strain pattern during running. PeerJ 7: e6764. DOI: https:// doi. org/ 10. 7717/ peerj. 6764, PMID: 31086731 Winters JM, Stark L. 1985. Analysis of fundamental human movement patterns through the use of in- depth antagonistic muscle models. IEEE Transactions on Biomedical Engineering 32: 826–839. DOI: https:// doi. org/ 10. 1109/ TBME. 1985. 325498 Winters JM, Stark L. 1988. Estimated mechanical properties of synergistic muscles involved in movements of a variety of human joints. Journal of Biomechanics 21: 1027–1041. DOI: https:// doi. org/ 10. 1016/ 0021- 9290( 88) 90249- 7, PMID: 2577949 Zuurbier CJ, Huijing PA. 1992. Influence of muscle geometry on shortening speed of fibre, aponeurosis and muscle. Journal of Biomechanics 25: 1017–1026. DOI: https:// doi. org/ 10. 1016/ 0021- 9290( 92) 90037- 2, PMID: 1517262

Muscle-specific economy of force generation and efficiency of work production during human running.

09-02-2021

Bohm, Sebastian,Mersmann, Falk,Santuz, Alessandro,Schroll, Arno,Arampatzis, Adamantios

eng

PMC7573630

1 Vol.:(0123456789) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports Efficient trajectory optimization for curved running using a 3D musculoskeletal model with implicit dynamics Marlies Nitschke1*, Eva Dorschky1, Dieter Heinrich2, Heiko Schlarb3, Bjoern M. Eskofier1, Anne D. Koelewijn1,4 & Antonie J. van den Bogert4 Trajectory optimization with musculoskeletal models can be used to reconstruct measured movements and to predict changes in movements in response to environmental changes. It enables an exhaustive analysis of joint angles, joint moments, ground reaction forces, and muscle forces, among others. However, its application is still limited to simplified problems in two dimensional space or straight motions. The simulation of movements with directional changes, e.g. curved running, requires detailed three dimensional models which lead to a high-dimensional solution space. We extended a full-body three dimensional musculoskeletal model to be specialized for running with directional changes. Model dynamics were implemented implicitly and trajectory optimization problems were solved with direct collocation to enable efficient computation. Standing, straight running, and curved running were simulated starting from a random initial guess to confirm the capabilities of our model and approach: efficacy, tracking and predictive power. Altogether the simulations required 1 h 17 min and corresponded well to the reference data. The prediction of curved running using straight running as tracking data revealed the necessity of avoiding interpenetration of body segments. In summary, the proposed formulation is able to efficiently predict a new motion task while preserving dynamic consistency. Hence, labor-intensive and thus costly experimental studies could be replaced by simulations for movement analysis and virtual product design. In recent years, interest in musculoskeletal simulation to reconstruct and predict human movements has been growing1. Motion reconstruction based on captured data yields insight into further variables of interest, e.g. joint moments or muscle forces2–7. Furthermore, simulations can be applied to predict changes of kinematics as well as joint and muscle function in response to interventions or environmental changes. They can support decisions in orthopaedic surgeries8,9 and the design of prostheses10,11, exoskeletons12, or shoes13. Therefore, predictive simulations can replace time-consuming and expensive prototyping and experimental studies. Commonly, biomechanical parameters are computed in a consecutive approach using inverse kinematics (IK), inverse dynamics (ID), and static optimization. This results in inconsistencies between kinematics and kinetics14. Additionally, each time step is analyzed separately in a discrete set of optimization problems rather than solving one optimization problem over time. Since either the motion or the forces are not simulated but prescribed, these methods cannot be applied to predict novel movements. These limitations can be overcome by solving an open-loop optimal control problem, also known as trajectory optimization. Movements are reconstructed or predicted by obtaining state and control trajectories from one single constrained non-linear optimization problem7. Alternatively, human motion can be simulated using a neuromuscular model with a controller based on reflexes15,16 or central pattern generators17. Because of their relevance to daily activities and sports, it is important to also simulate movements with changes in direction additional to straight walking and running. It is also important that such movements can be optimized with respect to performance and/or injury risk, without having to collect human motion data from those movements. This will expand the use of existing data and avoid potentially risky human experiments. OPEN 1Machine Learning and Data Analytics Lab, Department of Computer Science, Friedrich-Alexander-Universität Erlangen-Nürnberg (FAU), Erlangen, Germany. 2Department of Sport Science, University of Innsbruck, Innsbruck, Austria. 3adidas AG, Herzogenaurach, Germany. 4Department of Mechanical Engineering, Cleveland State University, Cleveland, USA. *email: [email protected] 2 Vol:.(1234567890) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ Potential applications are in knee injury prevention18, shoe performance in curved running19 or 3D controllers for exoskeletons and active prostheses20. Simulations of gait including turning were simulated with a reflex- based controller16, but reflex loops limit the space of possible control inputs. Trajectory optimizations find open-loop control trajectories without limiting the inputs, but have only been performed in two dimensional (2D) space2,7,11,13,21–23 or were restricted to forward motions24–27. A three dimensional (3D) model is needed to simulate movements with directional changes leading to a high-dimensional solution space of the optimal control problem. Therefore, effective numerical methods are necessary to put 3D optimal control simulations into application. Specifically, we need the capability to find solutions from an initial guess that is far from the solution, the solution should be found in a reasonable amount of time, and the solutions should exactly satisfy task requirements, such as a specified running speed and change of direction. The latter is of critical importance for sports applications. Often, trajectory optimization problems were solved using an explicit formulation of the multibody dynamics leading to large computational costs25–30. The reconstruction of one cycle of walking or running took for example around 2 h for a 3D model with 21 degrees of freedom (DOFs) and 66 muscle tendon units (MTUs)25 using direct collocation with OpenSim’s explicit implementation of the dynamics5,31. This seems to be inefficient when taking into account that the initial guess was close to the simulation results since it was generated from human motion data of the same movement task25. An implicit formulation of the model dynamics was proposed by Van den Bogert et al.7 for a 2D model to improve the numerical conditioning of the optimal control problem and thus reduced computational cost. They used direct collocation to solve tracking as well as predictive optimal control simulations. This approach was successfully applied to analyze loading asymmetry in transtibial amputee gait11 and the effect of midsole materi- als of shoes13. Recently, Falisse et al.24 used implicit dynamics to develop a framework for rapid simulations of a 3D model. They generated predictive simulations of walking and running with a full-body 3D model with 29 DOFs and 92 MTUs32 on average in 36 min. However, until now no movements with directional changes in 3D space were simulated. The purpose of our work was to further extend the current state-of-the-art in trajectory optimization for musculoskeletal models by computational efficient simulations of movements with directional changes. To this end, we created a complex full-body 3D musculoskeletal model called “running model for motions in all direc- tions”, short “runMaD”, adapted from Hamner et al.32. To reduce computational cost, dynamics were formulated implicitly, and derivatives were formulated analytically. We demonstrated the efficacy, the tracking capabilities, as well as the predictive power of the proposed trajectory optimization with implicitly formulated dynamics and direct collocation using three simulations: prediction of static standing, tracking of straight running, and prediction of curved running. Methods In the following, we describe the developed musculoskeletal model, the general trajectory optimization approach and how we generated simulations of standing, straight running, and curved running. 3D musculoskeletal model. The proposed musculoskeletal model “runMaD” is a full-body 3D model with 33 DOFs, operated using 92 MTUs in the trunk and legs and 10 torque actuators in the arms (see Fig. 1). This model was adapted from the OpenSim model created by Hamner et al.32. The order of rotations in the pelvis was changed33 and the subtalar and metatarsophalangeal (mtp) joints were unlocked to simulate movements with directional changes. Ranges of motion were enlarged for knee flexion and pronation/supination angle at the elbow to fit the recorded motion. Muscular and segmental properties were taken from Hamner’s model. We refer to section S1 in the Supplementary Information for a detailed description of all model adaptations. All muscles were modeled as three element Hill-type muscles with a contractile element (CE) with contrac- tion and activation dynamics, a parallel elastic element (PEE), and a series elastic element (SEE) (Fig. 2)34. The dynamics were described implicitly for each muscle with respect to the activation a and the state variable s , which was the projection of the CE length lCE on the muscle line of action7. The variable s was used instead of lCE to avoid singularities with respect to the pennation angle φ7. Muscle-tendon lengths were described as polynomial functions of joint angles, which were fitted using the muscle moment arm data of Hamner et al.32. In accordance with Falisse et al.24, polynomial functions were chosen since they have well-defined derivatives. The model’s state vector was defined by x = ( q ˙q s a )T , where q contained the DOFs, ˙q the derivatives of the DOFs, s the CE length state of all muscles, and a the activation state of all muscles. The control vector was defined as u = ( ne m )T , where ne denoted the neural excitation of all muscles and m the arm actuation torque divided by 10 Nm for each of the DOFs in the arm (Eq. S15 in Supplementary Information). Dynamics of the musculoskeletal system were combined with a penetration-based ground contact model to describe the full dynamics of the model implicitly as function of the states x , the state derivatives ˙x , and the controls u: Details on the system dynamics are given in section S2 in the Supplementary Information. Trajectory optimization. Optimal control problems were formulated to generate movement simulations. The goal was to find a state trajectory x(t) , a control trajectory u(t) , and the duration of the simulated movement Tsim such that the objective function J(x(t), u(t), Tsim) was minimized with respect to the following constraints: (1) f(x(t), ˙x(t), u(t)) = 0. (2) f(x(t), ˙x(t), u(t)) = 0 (dynamic equilibrium) 3 Vol.:(0123456789) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ The periodicity constraint ensured that the state of the model at the end of the gait cycle, Tsim , was equal to the state at the beginning rotated in the horizontal plane with Rper and shifted by horizontal translation tper. The objective was defined as weighted sum of a tracking term, a muscular effort term, a torque term, and a regularization term: (3) xL(t) ≤ x(t) ≤ xU(t) (bounds on states) (4) uL(t) ≤ u(t) ≤ uU(t) (bounds on controls) (5) x(Tsim) = Rper x(0) + tper (periodicity of states) (6) u(Tsim) = u(0) (periodicity of controls) (7) J(x(t), u(t), Tsim) = WTrack Jtrack + Wmus Jmus + Wtor Jtor + Wreg Jreg. acromial elbow ulna radius hand torso pelvis wrist radioulnar humerus femur tibia calcaneus talus toes hip knee ankle subtalar metatarso- phalangeal y G x G Figure 1. Musculoskeletal model “runMaD” with segments in black, joints in blue, ground contact points in pink, and the global coordinate system in green. The musculoskeletal model was visualized using OpenSim 4.0 (https ://opens im.stanf ord.edu). PEE SEE s CE ϕ lCE lMTU Figure 2. Hill-type muscle model34 with contractile element (CE), parallel elastic element (PEE), series elastic element (SEE), muscle-tendon length lMTU , length of the CE lCE , and pennation angle φ . The state variable s represents the projection of lCE on the muscle line of action. 4 Vol:.(1234567890) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ For tracking, the squared difference between simulated data ysim and the corresponding mean measured data µymeas of multiple gait cycles was minimized for all time points t: The squared difference was normalized by the variance of measured data σ 2 ymeas to make it dimensionless. The variance σ 2 ymeas was adapted to be at least 10 % of the mean of the variance to avoid division by small numbers. This is for example necessary for the variance of the ground reaction force (GRF) during swing phase. Furthermore, the terms were normalized by the duration of the simulation Tsim and weighted with WVar,i for each signal vari- able i. Volume-weighted and cubed neural excitation ne was minimized for each of the Nmus muscles to reduce muscular effort and to solve the muscle redundancy problem: where the ratio of muscle volume was computed with the maximum isometric force FISO and the optimal length of the CE lCE,opt . The muscle effort was divided by the cubic norm of horizontal translation speeds  vx vz 3 to compensate for different running speeds. It was shown previously that muscle activation is linear to movement speed35,36. Besides the control of the muscles, the torque controls mi actuating the arms were minimized: Finally, a small regularization term was added to enhance convergence by minimizing the derivatives of the states x and controls u: We used regularization for both states and controls since we found that this yields the lowest number of iterations without losing simulation accuracy. Simulations. We performed three simulations: prediction of static standing, tracking of straight running, and prediction of curved running. In the following, we describe the data acquisition, the optimal control prob- lems of the three simulations, and the details of implementation and solution process. An overview of the pipe- line is given in Fig. 3. Experimental data. We recorded straight and curved running of a male subject (92 kg, 1.95 m) using 42 reflec- tive markers, 16 infrared cameras (Vicon MX, Oxford, UK), and two force plates (Kistler Instruments Corp, Winterthur, CH) for tracking and as reference. The sampling frequency was set to 200 Hz and 1000 Hz, respec- tively. Straight running was performed at a speed of vx = 4.0 ms−1 and vz = 0 . Curved running was performed in a circle with radius r = 3.7 m at a norm horizontal speed of  vx vz  = 2.7 ms−1 . The API of OpenSim 4.05 was called within MATLAB (Mathwork, Natick, MA, USA) to scale the generic model using marker trajecto- ries in neutral pose with arms besides the body (N-pose) and to compute IK and ID. For ID, joint angles were filtered within OpenSim with a 3rd order dual-pass low-pass Butterworth filter with a cut-off frequency set to 15 Hz37. The GRFs were filtered with the same filter to avoid artifacts in the computed joint moments38. After processing, single gait cycles were extracted from right to right heel strike using the minimum of the right heel marker. The mean and standard deviation (SD) of 12 gait cycles were computed for straight and curved running after linearly interpolating to the number of samples of the shortest cycles. The subject gave informed consent prior to participation. The study was approved by the ethical committee of the Friedrich-Alexander-Universität Erlangen-Nürnberg (Re.-No. 106_13 B). All methods were carried out in accordance with relevant guidelines and regulations. Standing. The goal of the prediction of static standing was to find a neutral pose of the model in equilibrium without data tracking. As this simulation was independent of time, ˙x was set to 0 and no periodicity constraints (Eqs. 5 and 6) were used. The weights of the objective terms were chosen empirically for all simulations. For standing, Wmus = Wtor = 1 was set. All bounds of x and u are provided in Table S2 in the Supplementary Infor- mation. The simulation was solved 50 times with different random initial guesses for states x and controls u to reduce the likelihood of ending up in a local minimum. The result with the lowest objective was chosen as solution. (8) Jtrack = 1 Tsim Tsim  0 Ntrack  i=1 WVar,i ysim,i(t) − µymeas,i(t) σymeas,i(t) 2 dt. (9) Jmus = 1 Tsim Nmus  vx vz 3 Tsim  0 Nmus  i=1      FISO,i lCE,opt,i Nmus  i=1 FISO,i lCE,opt,i ne,i(t)3      dt, (10) Jtor = 1 Tsim Ntor Tsim  0 Ntor  i=1 mi(t)2 dt. (11) Jreg = 1 Tsim (Nstates + Ncontrols) Tsim  0 Nstates  i=1 ˙xi(t)2 + Ncontrols  i=1 ˙ui(t)2  dt. 5 Vol.:(0123456789) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ Straight running. Straight running was reconstructed by tracking straight running data. All joint angles, the global orientation of the pelvis and the GRFs of both feet in all directions were tracked (Eq. 8) similar to our approach in 2D7. A weighted arithmetic mean was used to balance the influence between joint angle tracking and GRF tracking independently of the number of signals: WVar,i = Wang/GRF NAng Wang + NGRF WGRF with Wang = 1 and WGRF = 5 and the numbers of tracked signals NAng and NGRF . The running speed used for weighting in the muscular effort term (Eq. 9) was computed from the tracking data. The weights of the objective terms were set to WTrack = 1 , Wmus = 103 , and Wtor = 1 such that after optimization the weighted objectives of tracking and effort were of same scale. The weight of the regularization term was small. In the straight running simulation, the periodicity of the gait cycle was ensured by allowing only translation in the horizontal plane with prescribed running speeds (Eq. 5). Hence, Rper in Eq. (5) was the identity matrix and only a translation for the global pelvis position was applied: Additionally, the states x and controls u were limited by lower and upper bounds (Eqs. 3 and 4). To define a global start position of the motion, the pelvis position at the first node was fixed to qpel_tx[0] = qpel_tz[0] = 0 . The standing solution was used as initial guess. Curved running. Curved running was predicted by tracking straight running data and constraining the model to run in a circle. Only joint angles and vertical GRFs were tracked to allow a circular motion (Eq. 8). The norm horizontal speed was obtained from the reference data of curved running to weight the muscular effort term (Eq. 9). In contrast to straight running, Wmus and Wtor were increased by factor 10 to allow more deviation from the tracking data. With help of the periodicity constraint (Eq. 5), we ensured that the model ran counterclockwise in a circle around the y-axis so that the left leg was on the inside. The circle was centered at  qpel_tx qpel_tz  =  0 0  with central angle θ . All entries in the state vector x were constrained to be equal for t = 0 and t = Tsim except for the global pelvis position and rotation: The norm horizontal speed and the radius of the measured curved running were used to obtain the central angle: To define a global start position of the motion, the pelvis position at the first node was fixed to qpel_tx[0] = −r and qpel_tz[0] = 0 . The straight running solution was used as initial guess. (12)   qpel_tx(Tsim) qpel_ty(Tsim) qpel_tz(Tsim)   =   qpel_tx(0) qpel_ty(0) qpel_tz(0)   +  vx Tsim 0 vz Tsim  . (13)    qpel_tx(Tsim) qpel_ty(Tsim) qpel_tz(Tsim) qpel_rot(Tsim)    =    cos(θ) 0 sin(θ) 0 0 1 0 0 −sin(θ) 0 cos(θ) 0 0 0 0 1       qpel_tx(0) qpel_ty(0) qpel_tz(0) qpel_rot(0)    +    0 0 0 θ    . (14) θ = 2 arcsin  vx vz  Tsim 2 r  . Figure 3. Processing pipeline. Work presented in this paper is highlighted in blue and OpenSim applications in red. The unscaled and scaled model, the experimental data, as well as the simulation results are provided in OpenSim file formats in the electronic supplementary material and at https ://simtk .org/proje cts/runma d. 6 Vol:.(1234567890) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ Implementation and solution process. The implicit multibody dynamics of the skeletal model were derived using Autolev (Symbolic Dynamics Inc., Sunnyvale, CA, USA) including Jacobian matrices generated by sym- bolic differentiation. Muscle dynamics and the ground contact model were implemented in C. All dynamic equations were compiled as MEX-functions in MATLAB. The objectives and the task constraints as well as their analytic derivatives were coded in MATLAB. The optimal control problems were solved for the scaled musculoskeletal model using direct collocation and backward Euler discretization. One collocation node was used for the static standing simulation. 50 collocation nodes were chosen for the running simulations since we found in preliminary simulations that 50, 100 and 200 collocation nodes yielded similar results. Running data was linearly interpolated to 50 samples. The non-linear optimization problems were solved with IPOPT 3.12.339. Settings were adapted to terminate the optimization at a tolerance of 10−5, a constraint violation tolerance of 10−3, and a complementary tolerance of 10−3. All optimiza- tions were run on one core of a workstation with a 3.2 GHz Xeon E5-1660v4 processor. Results In total, 50 predictive simulations of static standing using different random initial guesses, one tracking simula- tion of straight running, and one predictive simulation of curved running were solved. CPU times and iterations required for optimization are summarized in Table 1. The entire process from generating standing from random initial guesses to the prediction of curved running took 1 h 17 min 28 s. The tracking of straight running using standing as initial guess required more iterations than the prediction of curved running using straight running as initial guess since standing is quite a different task than running. The results corresponded to upright standing and natural running motion (Fig. 4 and videos in the electronic supplementary material). Joint angles and GRFs of straight running were close to the reference data of straight running, since the difference was generally less than one SD (Fig. 5). However, simulated knee flexion was smaller during the swing phase compared to the reference data. The difference between the right and left subtalar angle was well represented in the simulation of straight running even though subtalar and mtp angles deviated from the reference data more than one SD. The differences in GRFs were larger than one SD between 20% and 40% of the gait cycle. The estimated muscle activation patterns of the 18 largest muscles were similar to the electromyo- graphy (EMG) measurements reported by Cappellini et al.40 for straight running (Fig. 5). Predicted joint angles and GRFs of curved running matched the reference data of curved running but were not as similar as for the tracking simulation (Fig. 6). In particular, the ranges of motion were underestimated in the hip and knee flexion, knee flexion was smaller during stance, and maximum vertical GRFs were overestimated. The pelvis rotation and the horizontal GRFs cannot be compared directly for curved running since the global frames of simulation and reference were not aligned but rotated around the vertical axis. Muscle activations for curved running were similar to straight running but of less amplitude and less symmetric (Fig. 6). The model Table 1. Solver performance for the simulations of standing, straight running, and curved running. For standing, the CPU times and number of iterations are listed for the chosen result which had the lowest objective of the 50 simulations. Additionally, the mean and sum for all 50 standing simulations is given. CPU time in mm:ss Iterations Standing Result 00:10 355 Mean 00:14 513 Sum 11:26 25,669 Straight running 46:22 1,065 Curved running 19:40 715 Figure 4. Stick figures showing the simulation process. From left to right: Initial guess used for standing, result of standing, result of straight running, and result of curved running. For visualization, every fifth and every 25th node was plotted for straight and curved running, respectively, and the curved running result was extended to fill a whole circle. 7 Vol.:(0123456789) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ Figure 5. Joint angles, GRFs, and muscle activations of the 18 largest muscles of the straight running simulation at 4.0 ms−1. GRFs are scaled to body weight (BW) and muscle activations are normalized to the peak activation of straight running. The degrees of freedom (DOFs) and muscles are named according to their definition in the model file runMaD.osim. An overview of all muscles is provided in Table S1 in the Supplementary Information. Black, red, and blue solid lines indicate the simulated variables of the torso, the right side, and left side, respectively. For GRFs and joint angles, shaded areas show mean ± standard deviation (SD) of the measured gait cycles of straight running. For muscles, shaded red areas show mean electromyography (EMG) data of the right side for running at 3.3 ms−1 reported by Cappellini et al.40 normalized to the maximum of the simulated activations of each muscle. 8 Vol:.(1234567890) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ and data of the simulated and measured movements is provided in OpenSim file formats in the electronic sup- plementary material and at https ://simtk .org/proje cts/runma d. Discussion In this work, we presented a 3D full-body musculoskeletal model adapted to running with directional changes and an implicit formulation of its dynamics for efficient trajectory optimization using direct collocation. We gen- erated a predictive simulation of standing, a tracking simulation of straight running, and a predictive simulation with directional change, more precisely curved running, to demonstrate the efficacy, the tracking capabilities, and the predictive power of the approach. Computational efficiencies must be compared with caution due to influences of computational power, imple- mentation, the initial guess, and the choice of the numerical solution method, e.g. direct collocation7,25,26, multiple shooting41,42, and simulated annealing27. Nevertheless, it can be concluded that the CPU times of approximately 46 min and 20 min for straight and curved running, respectively, were small compared to simulations with explicit dynamics25,26. We were able to solve tracking simulations of straight running faster than Lin and Pandy25 and Lin et al.26, while starting from an initial guess that was constructed without IK, static optimization, or computed muscle control (CMC). Furthermore, our model “runMaD” had a higher complexity resulting in approximately 20 % more unknowns, i.e. states and controls, per collocation node. The CPU times were similar to Falisse et al.24, who simulated straight running, confirming the advantages of implicit formulation of model dynamics. Falisse et al.24 used algorithmic derivatives whereas we implemented them analytically. Future studies are needed to compare the efficiency of algorithmic and analytical derivatives in optimal control simulations. Previously, it was shown that using algorithmic derivatives is faster than finite differences which are used by OpenSim43. Machine learning approaches were recently investigated in the field of computer graphics to speed up musculoskeletal simulation44,45. Jiang et al.44 learned a mapping from muscle-actuation space to joint-actuation space which would have to be retrained if model parameters are changing. Lee et al.45 used a two-stage deep reinforcement learning approach to simulate a full-body 3D musculoskeletal model. However, muscle activation was not included in the reward of the trajectory mimicking which might lead to non-optimal use of the muscles. In all three simulations, kinematics were natural and matched the reference data. However, smaller knee flexion during the swing phase resulted in a lower foot clearance in both running simulations (Figs. 5 and 6). This movement pattern might be more energy efficient since increased foot clearance requires more effort. Subtalar and mtp angles deviated from the reference data and seemed to compensate each other’s error (Figs. 5 and 6). The reference data might be erroneous since the subject wore shoes and the foot segments are small which increases sensitivity to errors in marker positions and soft tissue artifacts. Additionally, simplifications in the foot model likely caused inaccuracies in the simulation. Although our proposed model is allowing subtalar and mtp motion in contrast to the model of Hamner et al.32, our foot model does also not reflect the fine foot structures. For a detailed analysis of foot motion, a finite element model of the foot, like developed by Akrami et al.46, could be incorporated into the musculoskeletal model. Instead of data tracking, movements could be reconstructed by constraining the movement path25,26,42. Nevertheless, the bounds cannot be exceeded and are difficult to choose especially if a change in motion should be predicted. In the predictive simulation of curved running, interpenetration of the legs occurred to reduce muscular effort (see video in electronic supplementary material). As a result of this, kinematics deviated slightly from reference data, i.e. ranges of motion were underestimated in the hip and knee flexion (Fig. 6). Interpenetration could be avoided by prescribing a minimal distance between joint origins24 which would however require prior knowledge about the motion path. Alternatively, a constraint or an error term could be added to prevent intersection of segments similar to what was done in computer graphics47. In both cases, the use of a stochastic environment48 would be beneficial to avoid segments moving close to each other. Similarly, the predictive simulation was lacking knee flexion during stance since it does not account for uncertainties while minimizing effort24,49. Even though GRFs of straight running were tracked well, impact of the initial contact was not distinct due to a fast progression towards the forefoot (Fig. 5). This was caused by the relatively coarse sampling with 50 nodes. Additionally, the simple foot and contact model might have influenced especially the impact. For curved running, maximum vertical GRFs were overestimated compared to the reference data (Fig. 6) probably due to the interpenetration of the legs. Nevertheless, the outer leg, i.e. the right leg, showed a higher maximum vertical GRF compared to the inner leg in agreement with the reference data. Joint moments were smoother in the simulation compared to the ID since the effort term has a smoothing effect (Figs. S1 and S2 in the Supplementary Information). ID depends on filtering of joint kinematics and GRFs38 whereas trajectory optimization benefits from a physics-based filtering. The mtp moments cannot be compared to the reference until late stance phase since GRFs were applied to the calcaneus in ID whereas we simulated ground contact at the calcaneus and toe segment. However, the mismatch of simulated and measured subtalar and mtp angles, contributed to the deviation in joint moments. In contrast to static optimization, muscle activations and controls were computed while accounting for muscle and tendon dynamics. This is especially important for the analysis of fast movements, like running50. The simulated muscle activations for straight running followed mainly the patterns of EMG measurements40 (Fig. 5) despite a higher velocity (4.0 ms−1 vs. approximately 3.3 ms−1). In comparison to straight running, peak muscle activations were smaller for curved running since running velocity was smaller (4.0 ms−1 vs. 2.7 ms−1) (Fig. 6). Furthermore, muscle activations were less symmetric due to the asymmetric movement task. We were able to predict curved running at 2.7 ms−1 while using tracking data of straight running at a different velocity of 4.0 ms−1. This confirms the predictive power of the trajectory optimization, because the target and tracking velocity did not have to match. 9 Vol.:(0123456789) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ Figure 6. Joint angles, GRFs, and muscle activations of the 18 largest muscles of the curved running simulation at 2.7 ms−1. GRFs are scaled to body weight (BW) and muscle activations are normalized to the peak activation of straight running. The degrees of freedom (DOFs) and muscles are named according to their definition in the model file runMaD.osim. An overview of all muscles is provided in Table S1 in the Supplementary Information. Black, red, and blue solid lines indicate the simulated variables of the torso, the right side, and left side, respectively. Shaded areas show mean ± standard deviation (SD) of the measured gait cycles of curved running. 10 Vol:.(1234567890) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ Three general limitations of trajectory optimization have to be mentioned. First, it cannot be ensured that the global optimum of the optimal control problem was found. In the standing simulation, we used multiple initial guesses to minimize the risk of ending in a local optimum. Second, reported kinematics and kinetics were affected by the objective function and thus by the choice of weights of the objective terms. GRFs were weighted more in comparison to joint angles since the GRFs were the only tracked signals containing information about the forces within the body. For higher tracking weight WTrack in the straight running simulation, GRFs and joint angles were tracked better. However, this came at the cost of non-smooth activation signals. The signals contained alternating phases of activation and deactivation to allow the joint angles to closely match the data. When the same weights of the straight running simulation were used for the curved running simulation, i.e. when effort weight was decreased, the result was closer to the tracking data resulting in higher knee flexion, higher knee moment and higher activation in knee extensors during stance. However, the other variables were predicted worse. Weights in the objective function could be obtained from data using inverse optimal control instead of selecting them empirically41,51. Third, it is not yet known which energy measure is minimized in human walking. Several studies proposed that metabolic energy is minimized52–54, while others hypothesized that it is more likely that muscular effort, which is related to activation and thus neural excitation, is minimized in human gait22,49,55. Since the actual movement objective is unknown, we included tracking. The simulation is still predictive because a new motion task was simulated based on another one. A different option is to manually tune weightings of energy measures to predict walking and running24. However, instead of data, this requires expert input. In conclusion, we presented a comprehensive 3D full-body musculoskeletal model modified for biomechani- cal analysis of running with directional change, i.e. curved running. Model dynamics were formulated implicitly resulting in computational efficient simulations. The efficiency makes large scale inverse optimal control stud- ies or sensitivity studies actually feasible. Furthermore, virtual product design11,13 would considerably benefit. Predicted kinematics and kinetics confirmed the predictive power of the proposed approach and were very promising but limited by the fact that the true objective of human motion is still unknown. For this reason, this work might be an important step towards efficient and biomechanical accurate predictive simulations of move- ments including directional changes. Received: 30 April 2020; Accepted: 21 September 2020 References 1. Ezati, M., Ghannadi, B. & McPhee, J. A review of simulation methods for human movement dynamics with emphasis on gait. Multibody Syst. Dyn. 47, 265–292 (2019). 2. Dorschky, E., Nitschke, M., Seifer, A.-K., Van den Bogert, A. J. & Eskofier, B. M. Estimation of gait kinematics and kinetics from inertial sensor data using optimal control of musculoskeletal models. J. Biomech. 95, 109278 (2019). 3. Erdemir, A., McLean, S., Herzog, W. & Van den Bogert, A. J. Model-based estimation of muscle forces exerted during movements. Clin. Biomech. 22, 131–154 (2007). 4. Pandy, M. G. Computer modeling and simulation of human movement. Annu. Rev. Biomed. Eng. 3, 245–273 (2001). 5. Seth, A. et al. Opensim: Simulating musculoskeletal dynamics and neuromuscular control to study human and animal movement. PLoS Comput. Biol. 14, 1–20 (2018). 6. Valero-Cuevas, F. J., Hoffmann, H., Kurse, M. U., Kutch, J. J. & Theodorou, E. A. Computational models for neuromuscular func- tion. IEEE Rev. Biomed. Eng. 2, 110–135 (2009). 7. Van den Bogert, A. J., Blana, D. & Heinrich, D. Implicit methods for efficient musculoskeletal simulation and optimal control. Procedia IUTAM 2, 297–316 (2011). 8. Arnold, A. S., Liu, M. Q., Schwartz, M. H., Ounpuu, S. & Delp, S. The role of estimating muscle-tendon lengths and velocities of the hamstrings in the evaluation and treatment of crouch gait. Gait Posture 23, 273–281 (2006). 9. Goldberg, S. R., Õunpuu, S. & Delp, S. L. The importance of swing-phase initial conditions in stiff-knee gait. J. Biomech. 36, 1111–1116 (2003). 10. Fey, N. P., Klute, G. K. & Neptune, R. R. Optimization of prosthetic foot stiffness to reduce metabolic cost and intact knee loading during below-knee amputee walking: A theoretical study. J. Biomech. Eng. 134, 1–10 (2012). 11. Koelewijn, A. D. & Van den Bogert, A. J. Joint contact forces can be reduced by improving joint moment symmetry in below-knee amputee gait simulations. Gait Posture 49, 219–225 (2016). 12. Harant, M., Sreenivasa, M., Millard, M., Šarabon, N. & Mombaur, K. Parameter optimization for passive spinal exoskeletons based on experimental data and optimal control. In 2017 IEEE-RAS 17th International Conference on Humanoid Robotics (Humanoids), 535–540 (2017). 13. Dorschky, E., Krüger, D., Kurfess, N., Schlarb, H., Wartzack, S., Eskofier, B. M. & Van den Bogert, A. J. Optimal control simulation predicts effects of midsole materials on energy cost of running. Comput. Methods Biomech. Biomed. Eng. 22, 869–879 (2019). 14. Fluit, R., Andersen, M., Kolk, S., Verdonschot, N. & Koopman, H. Prediction of ground reaction forces and moments during various activities of daily living. J. Biomech. 47, 2321–2329 (2014). 15. Geyer, H. & Herr, H. A muscle-reflex model that encodes principles of legged mechanics produces human walking dynamics and muscle activities. IEEE Trans. Neural Syst. Rehabil. Eng. 18, 263–273 (2010). 16. Song, S. & Geyer, H. A neural circuitry that emphasizes spinal feedback generates diverse behaviours of human locomotion. J. Physiol. 593, 3493–3511 (2015). 17. Kim, Y., Tagawa, Y., Obinata, G. & Hase, K. Robust control of CPG-based 3D neuromusculoskeletal walking model. Biol. Cybern. 105, 269–282 (2011). 18. McLean, S. G., Huang, X. & Van den Bogert, A. J. Investigating isolated neuromuscular control contributions to non-contact anterior cruciate ligament injury risk via computer simulation methods. Clin. Biomech. 23, 926–936 (2008). 19. Taboga, P. & Kram, R. Modelling the effect of curves on distance running performance. PeerJ 7, e8222 (2019). 20. Song, S. & Geyer, H. Regulating speed and generating large speed transitions in a neuromuscular human walking model. In 2012 IEEE International Conference on Robotics and Automation, 511–516 (2012). 21. Miller, R. H. & Hamill, J. Computer simulation of the effects of shoe cushioning on internal and external loading during running impacts. Comput. Methods Biomech. Biomed. Eng. 12, 481–490 (2009). 22. Miller, R. H., Umberger, B. R., Hamill, J. & Caldwell, G. E. Evaluation of the minimum energy hypothesis and other potential optimality criteria for human running. Proc. R. Soc. B Biol. Sci. 279, 1498–1505 (2012). 11 Vol.:(0123456789) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ 23. Miller, R. H. & Hamill, J. Optimal footfall patterns for cost minimization in running. J. Biomech. 48, 2858–2864 (2015). 24. Falisse, A., Serrancoli, G., Dembia, C. L., Gillis, J., Jonkers, I., De Groote, F. Rapid predictive simulations with complex muscu- loskeletal models suggest that diverse healthy and pathological human gaits can emerge from similar control strategies. J. R. Soc. Interface 16, 20190402 (2019). 25. Lin, Y.-C. & Pandy, M. G. Three-dimensional data-tracking dynamic optimization simulations of human locomotion generated by direct collocation. J. Biomech. 59, 1–8 (2017). 26. Lin, Y.-C., Walter, J. P. & Pandy, M. G. Predictive simulations of neuromuscular coordination and joint-contact loading in human gait. Ann. Biomed. Eng. 46(8), 1216–1227 (2018). 27. Miller, R. H. A comparison of muscle energy models for simulating human walking in three dimensions. J. Biomech. 47, 1373–1381 (2014). 28. Lee, L.-F. & Umberger, B. R. Generating optimal control simulations of musculoskeletal movement using opensim and matlab. PeerJ 4, e1638 (2016). 29. Mordatch, I., Mishra, N., Eppner, C. & Abbeel, P. Combining model-based policy search with online model learning for control of physical humanoids. In 2016 IEEE International Conference on Robotics and Automation (ICRA), vol. 2016-June, 242–248 (IEEE, 2016). 30. Porsa, S., Lin, Y.-C. & Pandy, M. G. Direct methods for predicting movement biomechanics based upon optimal control theory with implementation in opensim. Ann. Biomed. Eng. 44, 2542–2557 (2016). 31. Delp, S. L. et al. Opensim: open-source software to create and analyze dynamic simulations of movement. IEEE Trans. Biomed. Eng. 54, 1940–1950 (2007). 32. Hamner, S., Seth, A. & Delp, S. L. Muscle contributions to propulsion and support during running. J. Biomech. 43, 2709–2716 (2010). 33. Baker, R. Pelvic angles: a mathematically rigorous definition which is consistent with a conventional clinical understanding of the terms. Gait Posture 13, 1–6 (2001). 34. Hill, A. V. The heat of shortening and the dynamic constants of muscle. Proc. R. Soc. Lond. B Biol. Sci. 126, 136–195 (1938). 35. Kyröläinen, H., Komi, P. V. & Belli, A. Changes in muscle activity patterns and kinetics with increasing running speed. J. Strength Cond. Res. 13, 400 (1999). 36. Neptune, R. R., Sasaki, K. & Kautz, S. A. The effect of walking speed on muscle function and mechanical energetics. Gait Posture 28, 135–143 (2008). 37. Van den Bogert, A. J. & de Koning, J. On optimal filtering for inverse dynamics analysis. In Proceedings of the IXth Biennial Confer- ence of the Canadian Society for Biomechanics., 214–215 (Vancouver, British Columbia, 1996). 38. Derrick, T. R. et al. ISB recommendations on the reporting of intersegmental forces and moments during human motion analysis. J. Biomech. 99, 109533 (2020). 39. Wächter, A. & Biegler, L. T. On the implementation of an interior-point filter line-search algorithm for large-scale nonlinear programming. Math. Program. 106, 25–57 (2006). 40. Cappellini, G., Ivanenko, Y. P., Poppele, R. E. & Lacquaniti, F. Motor patterns in human walking and running. J. Neurophysiol. 95, 3426–3437 (2006). 41. Clever, D., Malin Schemschat, R., Felis, M. L. & Mombaur, K. Inverse optimal control based identification of optimality criteria in whole-body human walking on level ground. In 2016 6th IEEE International Conference on Biomedical Robotics and Biomechatronics (BioRob), 1192–1199 (IEEE, 2016). 42. Felis, M. L. & Mombaur, K. Synthesis of full-body 3-D human gait using optimal control methods. In 2016 IEEE International Conference on Robotics and Automation (ICRA), vol. 2016-June, 1560–1566 (IEEE, 2016). 43. Falisse, A., Serrancolí, G., Dembia, C. L., Gillis, J. & De Groote, F. Algorithmic differentiation improves the computational efficiency of opensim-based trajectory optimization of human movement. PLoS ONE 14, e0217730 (2019). 44. Jiang, Y., Van Wouwe, T., De Groote, F. & Liu, C. K. Synthesis of biologically realistic human motion using joint torque actuation. ACM Trans. Graph. 38, 1–12 (2019). 45. Lee, S., Lee, K., Park, M. & Lee, J. Scalable muscle-actuated human simulation and control. ACM Trans. Graph. 38(4), 1–13 (2019). 46. Akrami, M. et al. Subject-specific finite element modelling of the human foot complex during walking: sensitivity analysis of material properties, boundary and loading conditions. Biomech. Model. Mechanobiol. 17, 559–576 (2018). 47. Bogo, F. et al. Keep it SMPL: Automatic estimation of 3D human pose and shape from a single image. In Computer Vision—ECCV 2016: 14th European Conference, Amsterdam, The Netherlands, October 11-14, 2016, Proceedings, Part V, 561–578 (Springer Inter- national Publishing, 2016). 48. Koelewijn, A. D. & Van den Bogert, A. J. A solution method for predictive simulations in a stochastic environment. J. Biomech. 104, 109759. https ://doi.org/10.1016/j.jbiom ech.2020.10975 9 (2020). 49. Ackermann, M. & Van den Bogert, A. J. Optimality principles for model-based prediction of human gait. J. Biomech. 43, 1055–1060 (2010). 50. De Groote, F., Kinney, A. L., Rao, A. V. & Fregly, B. J. Evaluation of direct collocation optimal control problem formulations for solving the muscle redundancy problem. Ann. Biomed. Eng. 44, 2922–2936 (2016). 51. Nguyen, V. Q., Johnson, R. T., Sup, F. C. & Umberger, B. R. Bilevel optimization for cost function determination in dynamic simulation of human gait. IEEE Trans. Neural Syst. Rehabil. Eng. 27, 1426–1435 (2019). 52. Gordon, K. E., Ferris, D. P. & Kuo, A. D. Metabolic and mechanical energy costs of reducing vertical center of mass movement during gait. Arch. Phys. Med. Rehabil. 90, 136–144 (2009). 53. Ralston, H. J. Energy-speed relation and optimal speed during level walking. Internationale Zeitschrift für Angewandte Physiologie Einschliesslich Arbeitsphysiologie 17, 277–283 (1958). 54. Zarrugh, M. Y., Todd, F. N. & Ralston, H. J. Optimization of energy expenditure during level walking. Eur. J. Appl. Physiol. 33, 293–306 (1974). 55. McDonald, K. A., Hieronymi, A., Cusumano, J. P. & Rubenson, J. Optimization in human walking: Decoupling whole-body ener- getics and local muscle effort. In ISB/ASB Conference 2019 (2019). Acknowledgements This work was supported by adidas AG (M.N., A.D.K., A.J.v.d.B.), the Bavarian Ministry of Economic Affairs, Regional Development and Energy within the Embedded Systems Initiative (E.D.), the German Research Foun- dation within the framework of the Heisenberg professorship programme under Grant ES 434/8-1 (B.M.E.), the national science foundation under Grant No. 1344954 (A.D.K., A.J.v.d.B.), and a Graduate Scholarship from the Parker-Hannifin Corporation (A.D.K.). Author contributions A.J.v.d.B. had the original idea of using an implicit formulation of the model dynamics. M.N., E.D., A.D.K, and A.J.v.d.B. developed and implemented the model, which was supported by D.H., H.S., and B.M.E. The data 12 Vol:.(1234567890) Scientific Reports | (2020) 10:17655 | https://doi.org/10.1038/s41598-020-73856-w www.nature.com/scientificreports/ was recorded by H.S. The simulations and the manuscript were created by M.N. All authors contributed to the interpretation of the results and to the editing of the manuscript. Funding Open Access funding enabled and organized by Projekt DEAL. Competing interests The authors declare no competing interests. Additional information Supplementary information is available for this paper at https ://doi.org/10.1038/s4159 8-020-73856 -w. Correspondence and requests for materials should be addressed to M.N. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. © The Author(s) 2020

Efficient trajectory optimization for curved running using a 3D musculoskeletal model with implicit dynamics.

10-19-2020

Nitschke, Marlies,Dorschky, Eva,Heinrich, Dieter,Schlarb, Heiko,Eskofier, Bjoern M,Koelewijn, Anne D,van den Bogert, Antonie J

eng

PMC6888451

International Journal of Environmental Research and Public Health Article Functional Laterality of the Lower Limbs Accompanying Special Exercises in the Context of Hurdling Janusz Iskra 1, Ryszard Marcinów 1, Bo˙zena Wojciechowska-Maszkowska 1,* and Mitsuo Otsuka 2 1 Faculty of Physical Education and Physiotherapy, Department of Sport University of Technology in Opole, 45-758 Opole, Poland; [email protected] (J.I.); [email protected] (R.M.) 2 Faculty of Sport and Health Science, Ritsumeikan University, 1-1-1 Nojihigashi, Kusatsu 525–8577, Shiga, Japan; [email protected] * Correspondence: [email protected] Received: 10 October 2019; Accepted: 6 November 2019; Published: 7 November 2019



 Abstract: Background: The purpose of this study was to investigate the lateralization of the lead leg during special exercises and the relationship with athletic performance throughout a hurdling session. Methods: Thirty-eight physical education students participated in the study. A novel three-part “OSI” test (walking over hurdles arranged in a circle, spiral, and straight line) was performed, and various hurdle practices (jogging and running) were selected as research tools. The lead leg selected by the participants was taken into consideration, and the relationship between the chosen lead leg and athletic performance in the five tests was established. Results: The lateralization of the lead leg changed depending on the shape of the running course. The results of further analysis showed (i) no correlation between the use of the right leg as the lead leg in three tests conducted at a marching pace, and (ii) a significant positive correlation between tests performed at the marching and running paces. Conclusion: Hurdlers flexibly change the dominant leading leg depending on the shape of the running course. The results of this research could prove helpful in the training of athletes for hurdling competitions, especially young runners in 400-m hurdles involving straight and corner tracks. Keywords: hurdle run; functional asymmetry; hurdle exercises; teaching hurdles 1. Introduction The human body, considered in relation to the sagittal plane, is two-sided. Asymmetry can be discussed in terms of morphological and functional aspects. Functional asymmetry is associated with the dominance of one of the cerebral hemispheres and, as a result, the dominance of one of the upper or lower limbs. Quantitative differences, such as the results of performance and jumping tests, are related to the concept of dynamic asymmetry. The expression of functional asymmetry—the selection and dominant use of one of the limbs—is called lateralization. We assessed the degree of lateralization by means of classic human dynamic measurements—for instance, by comparing dynamic characteristics in regard to the right and left limbs [1,2]. This study applied a mixture of observation methods used in sport sciences, among other fields [3,4]. The dominant role of a selected lower limb is often determined by the specificity of a given discipline. Such phenomena can be observed in activities such as martial arts, football, and selected athletic competitions such as jumps or hurdles [5–7]. The analysis of the laterality of the lower limbs in conditions associated with sports competitions also applies to aspects not directly related to sports, such as horse racing [8]. Int. J. Environ. Res. Public Health 2019, 16, 4355; doi:10.3390/ijerph16224355 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2019, 16, 4355 2 of 10 Hurdles are a type of athletic competition in which the cyclical nature of the race is mixed with acyclicality because the athletes need to clear 10 hurdles. The asymmetric nature of clearing hurdles involves the problem of whether the hurdler’s right or left leg should initially lead. The “lead leg” is strictly defined for 100/110-m hurdles (only the left or right leg) but can be alternated in the 400-m hurdle race. The rules regarding hurdles in 100/110-m races include the requirement to clear 10 hurdles. Competitors with advanced technical skills try to clear hurdles in a three-step rhythm and consequently attack every hurdle with the same selected lead leg. The three-step rhythm between hurdles presents a considerable challenge for less-advanced athletes. Because less-advanced athletes try to maintain a rhythm involving four steps, they clear obstacles by alternating between the right and left leg [9–11]. Only a few competitors in the history of the 400-m hurdle competition (at the highest, world-class level) have been able to complete the race by attacking the hurdles with only a one-sided lead leg [12]. Most competitors alternate their lead leg during the race, which is often associated with the need to decrease their pace in subsequent stages of the sprint [13–15]. Hurdling around a curve in 400-m hurdles seems to be particularly difficult. This challenge is related to the athletes leaning their weight sideways into the center of the curve to counteract the centrifugal force [16]. When acquiring hurdling skills, training to clear hurdles using either leg is important [17,18]. Similar to jumping competitions (advancing using the left or right leg), some people use a specific lead leg for hurdle clearance in hurdles [2], and some practice this movement during classes with special exercises that are performed at a marching or jogging pace. The importance of including such exercises in the teaching process has been confirmed in scientific reports [10]. 1.1. Objective of the Study The objective of the present study was to compare the lateralization of the lead leg during special exercises performed at a marching or jogging pace with that of the lead leg during hurdle clearance at standard distances for a specific level of competition at a maximal pace. 1.2. Research Questions This study aimed to answer three research questions: 1. Are there differences between the lateralization of the lead leg selected in three tests that involve clearing hurdles at a marching pace, taking into account different directions of exercise practice? 2. Are there differences in terms of the lead-leg lateralization of the subjects among five tests that involve clearing hurdles at marching, jogging, and maximal paces? 3. Are there correlations between the functional asymmetries of the selection of the right or left leg as the lead leg during hurdle clearance at marching, jogging, and maximal paces? 1.3. Hypotheses The following hypotheses were proposed: 1. There are no differences in the lateralization of the lead leg selected by participants in three tests involving marches over hurdles. 2. There are no differences in the lead leg selected by participants in all forms of hurdle exercises. 2. Materials and Methods 2.1. Participants The study involved a group of students with a specialty in motor skill training at the Faculty of Physical Education and Physiotherapy at the Opole University of Technology (Table 1). The group comprised 12 women (age: 23.23 ± 2.07 years; body weight: 57.47 ± 3.57 kg; body height: 1.66 ± 3.24 m) and 26 men (age: 24.24 ± 2.11 years; body weight: 75.81 ± 5.11 kg; body height: 181.79 ± 4.45 m). None of the participants were professional hurdlers; all were students of physical education. We did not Int. J. Environ. Res. Public Health 2019, 16, 4355 3 of 10 divide the group into males and females, as many authors have shown that both males and females encounter the same problem of choosing a so-called hurdle stride pattern “rhythm” [16,19,20]. The tests were carried out in an athletic sports hall on a tartan track. The first test started in the late morning (i.e., from 11:00 a.m.), and subsequent tests were performed on the same day at half-hour intervals. The subjects voluntarily agreed to participate in the research and were informed of the purpose of the study. In this group, there were no professional athletes (hurdlers). There were no reasons to exclude any students from participating in this study. The whole teaching session was typical of physical education for students. Table 1. The arrangement of hurdles in the test performed at jogging and running paces. Test Distance Between Hurdles (m) a b c d e f g Test 4 4.0 5.0 6.0 7.0 8.0 9.0 10.0 Test 5 F 12 (A) 7.5 M 13 (A) 8.2 (A): approach. 2.2. Study Procedure In our study, we used an observation method that included a mixed method of analysis [3]. After observing many hurdling exercises, we counted left and right movements and converted the values into percentages, which were subjected to statistical analysis. The laterality of the lead leg was assessed using the five trials (specialist tests) described below. 2.2.1. “OSI” Test: A Novel Test Developed in this Study The subjects had to clear hurdles (76 cm for women, 91 cm for men) arranged at uneven intervals and along various trajectories: in a straight line, along the circumference of a circle, and along S-shaped curves. The distances between hurdles (from 2.5 to 5.5 m, every 0.5 m) were determined randomly for each test (Figure 1). The subjects stood on the starting line with their legs in a step position. The subjects did not receive any advice regarding the starting position or technique for clearing the hurdles. While the subject performed the tests, the selection of the lead leg (left or right) was recorded for each hurdle clearance. Test no. 1: March along a circle, "O". The subject started the test at the start/finish line. At the signal of the coach, the subject began to march around the circumference of the circle in a counterclockwise direction. An additional final hurdle was placed at the start after the subject began the lap. The radius of the circle and the distances between hurdles are specified in Figure 1. Test no. 2: March along a curved track, "S". The subject started the test at the beginning of the curve, which curved in a counterclockwise direction. In the second part of the curve, the direction of the march shifted to clockwise. The radius of the curve and distances between hurdles are shown in Figure 1. Test no. 3: March along a straight track, "I". The distances between the hurdles are displayed in Figure 1. The distance in a maximal hurdle run is standard for students of physical education. There was a main difference between hurdle clearance on a straight track and that in a run around the track. When teaching the 400-m hurdle run, we applied exercises (mainly at a marching pace) on various movement tracks (circle slalom). For this distance, we used an “anticlockwise” direction for the run with eight variants (eight tracks with different curvatures). Int. J. Environ. Res. Public Health 2019, 16, 4355 4 of 10 Int. J. Environ. Res. Public Health 2019, 16, x FOR PEER REVIEW 4 of 10 Figure 1. Hurdle arrangement (a square denotes a hurdle) and the intervals between hurdles for tests performed in a circle, on a curve, and on a straight track. There was a main difference between hurdle clearance on a straight track and that in a run around the track. When teaching the 400-m hurdle run, we applied exercises (mainly at a marching pace) on various movement tracks (circle slalom). For this distance, we used an “anticlockwise” direction for the run with eight variants (eight tracks with different curvatures). 2.2.2. Two Tests Involving Hurdle Clearance at Jogging and Sprinting Paces Test no. 4: This test involved a light jog covering a distance of 50 m and the clearing of hurdles with a height of 76 or 91 cm (for women and men, respectively). The subject started the test at the start/finish line (legs in a step position). When the coach provided a signal, the subject began clearing the hurdles, attacking them with the leg of their choice. While the subject performed the trial, the selection of the lead leg (left or right) was recorded for each hurdle that was cleared. The subject did not receive any advice regarding the starting position or technique for clearing the hurdles. The distances between the hurdles were set randomly (from 4 to 10 m). The velocity of this exercise (test) was determined as “medium run/jog”. The time of the test was 14.98 ± 1.48 s (ranging from 12.54 to 17.98 s) (Table 1). Test no. 5: This test involved sprinting (at maximal velocity) for 60 m and clearing hurdles with a height of 76 or 91 cm (for women and men, respectively). The subject started the test at the start/finish line (legs in a step position). At the signal of the coach, the subjects began to run, clearing the hurdles with the leg of their choice. The subject did not receive any advice regarding the starting position or the best manner for clearing the hurdles. While the subject performed the test, the lead leg (left or right) was recorded for each hurdle that was cleared. The distances between the hurdles were set to 7.50 m for women and 8.20 m for men (in both cases, these distances are equal to approximately 4.5 times the average body height of the subjects participating in the study) following the study by Iskra and Mynarski [10]. The approach (i.e., the distance from the start to the first hurdle) was 12 or 13 m. This run was the final run conducted for students after the hurdle course. In the protocol in this test, the time of the run was 10.85 ± 0.94 s (from 9.20 to 12.98 s). Figure 1. Hurdle arrangement (a square denotes a hurdle) and the intervals between hurdles for tests performed in a circle, on a curve, and on a straight track. 2.2.2. Two Tests Involving Hurdle Clearance at Jogging and Sprinting Paces Test no. 4: This test involved a light jog covering a distance of 50 m and the clearing of hurdles with a height of 76 or 91 cm (for women and men, respectively). The subject started the test at the start/finish line (legs in a step position). When the coach provided a signal, the subject began clearing the hurdles, attacking them with the leg of their choice. While the subject performed the trial, the selection of the lead leg (left or right) was recorded for each hurdle that was cleared. The subject did not receive any advice regarding the starting position or technique for clearing the hurdles. The distances between the hurdles were set randomly (from 4 to 10 m). The velocity of this exercise (test) was determined as “medium run/jog”. The time of the test was 14.98 ± 1.48 s (ranging from 12.54 to 17.98 s) (Table 1). Test no. 5: This test involved sprinting (at maximal velocity) for 60 m and clearing hurdles with a height of 76 or 91 cm (for women and men, respectively). The subject started the test at the start/finish line (legs in a step position). At the signal of the coach, the subjects began to run, clearing the hurdles with the leg of their choice. The subject did not receive any advice regarding the starting position or the best manner for clearing the hurdles. While the subject performed the test, the lead leg (left or right) was recorded for each hurdle that was cleared. The distances between the hurdles were set to 7.50 m for women and 8.20 m for men (in both cases, these distances are equal to approximately 4.5 times the average body height of the subjects participating in the study) following the study by Iskra and Mynarski [10]. The approach (i.e., the distance from the start to the first hurdle) was 12 or 13 m. This run was the final run conducted for students after the hurdle course. In the protocol in this test, the time of the run was 10.85 ± 0.94 s (from 9.20 to 12.98 s). 2.3. Statistical Analysis The results were analyzed using Statistica 13.1 software (TIBCO Sofware Inc., Tulsa, OK, USA). The significance level was set to p ≤ 0.05. The normality of the distribution was assessed by the Shapiro–Wilk test. Student’s t-test was used to analyze the differences in the selection of the left or Int. J. Environ. Res. Public Health 2019, 16, 4355 5 of 10 right leg in the trials involving hurdle clearance along a curve and along a straight line, whereas the differences in the results based on the arrangement of hurdles (circle, curve, and straight line) were assessed by analysis of variance (ANOVA). The analysis of the dependencies between the leg selection patterns in the trials performed at a marching pace and those performed at jogging and running paces were performed using Spearman’s Rank correlation coefficient. The Cohen effect sizes in all statistical tests were determined using G*Power 3.1 (Heinrich-Heine-Universität Düsseldorf, Düsseldorf, Germany). Tests for correlation and regression analyses [21]: d = 0–0.1 was considered as no effect, 0.2–0.4 as small effect, 0.5–0.7 as intermediate effect and ≥0.8 as large effect. Similarly, partial η2 effect size: 0.01–0.06 small, 0.06–0.14 medium and ≥0.14 large effect. 3. Results In this study, we used an observation method that included a mixed method of analysis [3]. After observing many hurdle exercises, we counted left and right movements and converted the values into percentages, which were then subjected to statistical analysis. Table 2 shows the mean results for the lead leg selected by subjects in tests in which they cleared hurdles at a marching pace. In all tests, the subjects had a greater tendency to use the right leg to clear the hurdles, with a large difference between the maximum and minimum results. The analysis of the results demonstrates that there were even cases in which subjects cleared all the hurdles with the right leg. Such cases were not recorded for the left leg. Table 2. Lead-leg functional asymmetry in tests involving hurdle clearance at a marching pace (data in %) and the results of tests performed in a jog and a run. Test No. Type Of Test Lead Leg Mean (x) Stand. Dev. (SD) Min. Max. Skewness Kurtosis 1 March in a circle “O” L 3.11 1.20 0 6 0.28 1.22 R 3.89 1 7 −0.28 2 March around a bend “S” L 2.53 1.75 0 6 0.11 −0.44 R 4.47 1 7 −0.11 3 March in a straight “I” L 2.63 1.60 0 6 −0.11 −0.24 R 4.37 1 7 0.11 Total OSI L 8.26 3.64 0 18 0.47 1.47 R 12.74 3 21 −0.47 4 Hurdle jog L 2.71 1.59 0 6 −0.17 −0.66 R 4.29 1 7 0.17 5 Hurdle run L 2.47 2.50 0 7 0.81 −0.49 R 4.53 0 7 −0.81 Table 2 contains a summary of the mean results for lead leg selection in the tests in which the subjects cleared the hurdles at jogging and running paces. Similar to the tests in which the subjects cleared the hurdles at a marching pace, a preference for the right leg was observed in the tests performed at jogging and running paces. The results of test nos. 4 and 5 also demonstrate cases in which the subjects cleared all hurdles with the right leg. Cases in which the subject cleared all hurdles with the left leg were only observed in test no. 5 (i.e., the test that was closest to the conditions of an actual competition). Table 3 presents a comparative analysis of the mean results of the tests performed at a marching pace. The analysis shows statistically significant differences between the selection of the left and right leg as the lead leg in trials involving hurdle clearance around a curve and along a straight line. In these tests, the subjects more frequently used their right leg as the lead leg. In the test in which participants performed the hurdling test in a circle at a marching pace, the difference was not statistically significant. The mean totals in all three trials performed by subjects at a marching pace revealed statistically significant differences in the choice of lead leg. The majority of the subjects cleared the hurdles with the right leg more frequently than they used the left leg. Such differences were not observed in the hurdling test performed in a circle at a marching pace. Int. J. Environ. Res. Public Health 2019, 16, 4355 6 of 10 Table 3. Student’s t-test results for the selection of the right and left lead legs during tests at a marching pace and at jogging and running paces (data in %). Test No. Type Of Test Left Leg, L Right Leg, R Stand. Dev. (SD) t p d 1 March in a circle “O” 44.36 55.64 17.19 −2.02 0.051 0.57 2 March around a bend “S” 36.09 63.91 25.02 −3.42 0.002 * 2.97 3 March in a straight line “I” 37.59 62.41 22.87 −3.34 0.002 * 7.67 Total OSI 39.35 60.65 17.35 −3.78 0.000 * 6.14 4 Hurdle jog 38.75 61.28 22.75 −3.06 0.004 * 0.99 5 Hurdle run 35.34 64.66 35.73 −2.52 0.016 * 0.96 * p ≤ 0.05; d: Cohen effect size. A comparative analysis of the mean results of the test performed at the jogging and running paces (Table 3) demonstrated that the differences in the selection of the left or right leg were statistically significant. Similar to the tests performed at a marching pace, in the trials at the jogging and running paces, the subjects more frequently selected the right leg as the lead leg. The selection of the right leg was preferred in four of the five cases. However, an additional question remained: In attempts to clear hurdles at a marching pace, does the arrangement of the hurdles (circle, curve, or straight line) affect the mean frequency at which the right leg is selected? The ANOVA did not reveal such differences (Table 4). The mean results were rescored on the basis of the attempts to clear hurdles with the right leg (i.e., for various arrangements of the hurdles), and these results proved to be similar and were not significantly different. This collection of data proves that the trajectory of the track affected the selection of the limb used for hurdle clearance. Table 4. Differences in the results of the tests involving hurdle clearance using the right leg for various hurdle arrangements (data in %). Hurdle Arrangement Layout “O” Layout “S” Layout “I” ANOVA F p partial η2 Lead leg R (x/SD) 55.64 (±17.18) 63.91 (±25.02) 62.41 (±22.87) 1.53 0.22 (NS) 0.05 NS: lack of statistical significance; partial η2-effect size. Table 5 contains the statistical analysis results for the relation between the lead leg selection in the tests performed at a marching pace (test nos. 1, 2, and 3) and the lead leg selection in the tests performed at jogging and sprinting paces (test nos. 4 and 5). Interestingly, the results did not show a correlation based on the hurdle arrangement in the test performed at a marching pace. The results of all trials that were performed at a marching pace were only correlated with the final result of other tests performed at a marching pace. The results of the remaining tests were not significantly correlated with each other. The results of the lead leg selection in the tests performed on curved tracks (circle, spiral) did not correlate with the results of the hurdle clearance tests performed at jogging and sprinting paces, whereas the result of the lead leg selection in the straight line test was significantly correlated with the result of the lead leg selection in the tests performed at a jogging pace (0.64 for p ≤ 0.05). The results of the selection of the lead leg in the “OSI” test correlated with the lead leg selection results in the test performed at a jogging pace (0.58 for p ≤ 0.05) and did not affect the selection of the lead leg in the test performed at a sprinting pace (0.47 for p ≤ 0.05). A significant level of dependence was also observed between the selection of the lead leg during the hurdle clearance in the tests performed at jogging and sprinting paces (0.58 for p ≤ 0.05). Thus, significant dependencies can be established for exercises performed at similar speeds (march–jog and jog–run). Significant speed differences (march–run) resulted in the subjects varying the manner in which they cleared the hurdles. Int. J. Environ. Res. Public Health 2019, 16, 4355 7 of 10 Table 5. Spearman’s Rank correlation of the trial results for hurdle clearance with the right leg at marching, jogging, and running paces. Test March “O” March “S” March “I” Total “OSI” Jog Run March “O” March “S” 0.18 March “I” 0.26 0.46 * Total “OSI” 0.50 * 0.80 * 0.81 * Jog 0.11 0.45 0.64 * 0.58 * Run 0.30 0.33 * 0.44 * 0.47 * 0.58 * * p ≤ 0.05. 4. Discussion The problem of lateralization in physical education and sports is important in various activities. Analyses of the functional dominance of the limbs, especially the lower limbs, have shown four types of laterality: lateralization combined with right leg dominance (type I) and left limb dominance (type II), right-footedness, and left-footedness [22]. The dominance of a particular leg can be assessed in various ways, from questionnaires to specific physical tests [23]. The term “specific” in this study also applies to hurdles. In this study, we did not account for the potential influence of motor preparation on the choice of lead leg and, more broadly, the choice of stride pattern. This problem has been addressed in many previous studies [24]. Speed, endurance, and strength are known to be connected to the strategy used in the 400-m hurdle run, but this is a separate problem for another study. Hurdle competitions include two athletic events—the 100/110 m and the 400 m—that differ in terms of the specialized effort required (focused on speed and speed-endurance, respectively). Analysis of the results of all tests demonstrated that the subjects more frequently used the right leg to clear the hurdles in the test performed at a running pace (x = 4.53 m/s), whereas the selection of the right leg as the lead leg was least common in the hurdle test performed in a circle at a marching pace (x = 3.89 m/s). Similarly, the preference of the right leg as the lead leg is common in the 100/110-m hurdle race, with over 60% of competitors participating in the most prestigious athletic events preferring to use their right leg as their lead leg [16]. The analysis of the results showed that the subjects participating in the test cleared hurdles using the right leg as the lead leg more frequently during trials performed on straight track sections. However, this finding differs from the results of the hurdle test performed in a circle. The direction of the running movement followed the direction of the race in athletic competition conditions (i.e., counterclockwise). This may indicate that some of the subjects who normally used their right leg as their lead leg in the other tests on a circular track decided to select the left leg for the test involving a counterclockwise run. This observation supports the results reported in earlier studies [16,19], which demonstrated that subjects find it easier to clear hurdles with the inner left leg (as the body adapts to counteract the centrifugal force) during hurdle clearance on a curve. In fact, Starosta and K˛edziora [25] found that the 400-m hurdle race specifically causes the lead leg to change, regardless of the leg that is normally preferred. In test no. 5 (hurdle clearance at the maximal running pace), the distance between hurdles was set to approximately 4.5 times the mean body height, which is the distance used for beginners. At such distances, novice athletes should (in conditions similar to competition conditions) be able to clear hurdles in a three-step rhythm without attacking each hurdle with the same leg [11,26,27]. However, among the 38 subjects in the study, only 21 cleared all hurdles with the same leg (i.e., taking an odd number of steps between hurdles), and the other subjects changed their lead leg for the task of hurdle clearance. These results could be attributable to issues that extend beyond lateralization. As has been Int. J. Environ. Res. Public Health 2019, 16, 4355 8 of 10 repeatedly stated, the selection of the lead leg in hurdles can also be determined by the level of motor skills (including speed, glycolytic endurance, and leg strength) [28]. The shorter-distance race is performed on only a straight track, as opposed to longer tracks that include straight sections, curves, and arched parts; this difference leads to variation in technical skills between hurdlers [19,29]. Analysis of the correlation for the marching-pace tests confirms this conclusion. The results of the marching-pace test performed on a straight track correlated with the results of the test involving jogging and running along a straight track, but such a correlation could not be established between the lead leg selection with hurdles arranged on a straight track and that with hurdles arranged around curves (with different hurdle arrangements and moving either clockwise or counterclockwise), as shown in Table 5. The alternating selection of the lead leg due to the arrangement of hurdles has often been attributed to fatigue related to anaerobic exercise [28,30]. The results of this study suggest an association with the direction/track followed during a hurdle test. Hurdling exercises performed at a marching pace demonstrate that the technical exercise of hurdlers plays a significant role [19,20,29]. Research has shown that this approach to the development of technical skills is closely related to hurdle performance (both in the form of jogging and running). Therefore, performing hurdle exercises at a marching pace can have a considerable impact on the progression of an athlete’s hurdling technique. In this study, we used an observation method that included a mixed method of analysis [3]. After observing many hurdle exercises, we counted left and right movements and converted the values into percentages, which were then subjected to statistical analysis. High clearance at high (jogging) and maximal velocity (hurdle sprinting and running) was more difficult than special hurdle exercises at a marching pace. This difficulty is likely the reason that students chose the right leg to “jump” over hurdles. They considered (contrary to standard teaching for all hurdlers–from beginners to professional) this decision to be not only effective but also safe. In previous studies, we did not find relationships between right (dominant) handedness and lead leg preference in a hurdle run [31]. The research was limited by a group of physical education students who did not specialize in professional hurdling, so our results cannot be generalized to professional athletes. 5. Conclusions The hypotheses in this study were confirmed only partially. 1. From all tests involving hurdle clearance (marching, jogging, and running), the study demonstrates the dominance of the right leg. Therefore, the left leg can be considered the dominant take-off leg. 2. The lack of a correlation between the leg selection at the marching pace and that at the running pace demonstrates that the specific approach followed during hurdle clearance depends on the profile of the track. 3. The results of tests involving hurdle clearance at various speeds (marching, jogging, and running) demonstrate a correlation only between tests conducted at similar speeds (march–jog and jog–run). Such similarities between tests were not observed when hurdling clearance was performed at considerably different speeds (march–run). 4. The results of this research could prove helpful in the teaching hurdling running to young athletes, especially over longer distances, taking into account races following a curve (i.e., after a distance of 150–300 m in hurdling races). Author Contributions: Conceptualization: J.I.; data curation: R.M.; formal analysis: R.M.; funding acquisition: B.W.-M.; investigation: J.I.; methodology: J.I. and M.O.; supervision: M.O.; writing—original draft: B.W.-M.; writing—review & editing: J.I. and B.W.-M. Funding: This research received no external funding. Conflicts of Interest: Authors declare that there are no conflicts of interest. Int. J. Environ. Res. Public Health 2019, 16, 4355 9 of 10 References 1. Wola´nski, N. Human biological development. In Rozwój Biologiczny Człowieka; PWN: Warszawa, Poland, 2005; pp. 53–56. (In Polish) 2. Hewit, J.; Cronin, J.; Hume, P. Multidirectional leg asymmetry assessment in sport. Strength Cond. J. 2012, 34, 82–86. [CrossRef] 3. Anguera, M.T.; Camerino, O.; Castaner, M.; Sanchea-Algarra, P.; Onwuegbuzie, A.J. The specificity of observational studies in physical activity and sports sciences: Moving forward in mixed methods research and proposals for achieving quantitative and qualitative symmetry. Front. Psychol. 2017, 8, 2196. [CrossRef] [PubMed] 4. Pic, M. Performance and Home advantage in handball. J. Hum. Kinet. 2018, 63, 61–71. [CrossRef] [PubMed] 5. Stokłosa, H. Functional body asymmetry in experienced weight lifters and wrestlers. Biol. Sport 1994, 1, 65–67. 6. Eikenberry, A.; McAuliffe, J.; Welsh, T.N.; Zerpa, C.; McPherson, M.; Newhouse, I. Starting with the “right” foot minimizes sprint start time. Acta Psychol. 2007, 127, 495–500. [CrossRef] 7. Hoffman, J.R.; Ratamess, N.A.; Klatt, M.; Faigenbaum, A.D.; Kang, J. Do bilateral power deficits influence direction-specific movement patterns? Res. Sports Med. 2007, 15, 125–132. [CrossRef] 8. Williams, D.E.; Norris, B.J. Laterality in stride pattern preferences in race horses. Anim. Behav. 2007, 74, 941–950. [CrossRef] 9. Hay, L.; Schoebel, P. Spatio-temporal invariants in hurdle racing patterns. Hum. Mov. Sci. 1990, 9, 37–54. [CrossRef] 10. Iskra, J.; Mynarski, W. The influence somatic and motor fitness on hurdle results by untrained boys aged 11–15. J. Hum. Kinet. 2000, 4, 111–132. 11. Otsuka, M.; Ito, M.; Ito, A. Analysis of hurdle running at various interhurdle distances in an elementary school PE class. Int. J. Sport Health Sci. 2010, 8, 35–42. [CrossRef] 12. Quercetani, R.L. Athletics: A History of Modern Track and Field Athletics (1860–2000): Men and Women; ESP: Milan, Italy, 2000. 13. Ditroilo, M.; Marini, M. Analysis of the race distribution for male 400 m hurdlers competing at the 2000 Sydney Olympic Games. New Stud. Athl. 2001, 3, 15–30. 14. Quinn, M.D. External effects in the 400-m hurdles race. J. Appl. Biomech. 2010, 26, 171–179. [CrossRef] [PubMed] 15. Otsuka, M.; Isaka, T. Intra-athlete and inter-group comparisons: Running pace and step characteristics of elite athletes in the 400-m hurdles. PLoS ONE 2019, 14, e0204185. [CrossRef] [PubMed] 16. Iskra, J.; Cóh, M. Biomechanical Studies on Running the 400 M Hurdles. Hum. Mov. 2011, 12, 315–323. [CrossRef] 17. Thompson, P. The Official IAAF Guide to Teaching Athletics; Warners Midlands Plc: Bourne, UK, 2009; pp. 39–57. 18. Gasilewski, J.; Iskra, J. Efekty nauczania biegu przez płotki w aspekcie treningu motorycznego i technicznego. Antropomotoryka 2011, 53, 49–54. 19. McFarlane, B. The Science of Hurdling and Speed; Athletics Canada: Ottawa, ON, Canada, 2000; pp. 106–109. 20. Arnold, M. Hurdling; British Athletic Federation: Birmingham, UK, 1992; pp. 23–26. 21. Faul, F.; Erdfelder, E.; Buchner, A.; Lang, A.G. Statistical power analyses using G*Power 3.1: Tests for correlation and regression analyses. Behav. Res. Methods 2009, 41, 1149–1160. [CrossRef] [PubMed] 22. Gabbard, C.; Hart, S. Foot laterality in children, adolescents and adults. Laterality 1996, 1, 199–2005. [CrossRef] 23. Chapman, J.; Chapman, L.; Allen, J. The measurement of foot preference. Neuropsychologia 1987, 25, 577–584. [CrossRef] 24. Ozaki, Y.; Ueda, T.; Fukuda, T.; Inai, T.; Kido, E.; Narisako, D. Regulation of stride length Turing the approach run in the 400-m hurdles. J. Hum. Kinet. 2019, 69, 59–63. [CrossRef] 25. Starosta, W.; K˛edziora, R. Methods of clearing hurdles by the best male and female hurdlers from the perspective of laterality. In Problemy Badawcze w Lekkoatletyce; Kowalski, P., Migasiewicz, J., Eds.; AWF: Wrocław, Poland, 1995; pp. 65–71. (In Polish) 26. Iskra, J.; Bacik, B.; Król, H. The effect of specific exercises on changes in hurdle technique. In Current Research in Motor Control; Raczek, J., Wa´skiewicz, Z., Juras, G., Eds.; Akademia Wychowania Fizycznego: Katowice, Poland, 2000; pp. 104–110. 27. Hagen, R.; Trebels, A.H. Hurdenlauf—Ein individuelles Rhythmusproblem. Sportpadagogik 1998, 4, 46–49. Int. J. Environ. Res. Public Health 2019, 16, 4355 10 of 10 28. Zauhal, H.; Jabbour, G.; Jacob, C.; Duvigneau, D.; Botcazou, M.; Ben Abderrahaman, A.; Prioux, J.; Moussa, E. Anaerobic and aerobic energy system contribution to 400-m flat and 400-m hurdles track running. J. Strength Cond. Res. 2010, 24, 2309–2315. [CrossRef] [PubMed] 29. Iskra, J. 400 m hurdlerstraining. In Trening Płotkarzy na 400 m; Akademia Wychowania Fizycznego: Katowice, Poland, 2014; pp. 153–201. (In Polish) 30. Ward-Smith, A.J. A mathematical analysis of the bioenergetics of hurdling. J. Sports Sci. 1997, 15, 517–526. [CrossRef] [PubMed] 31. Hyjek, J. Czynniki Warunkuj ˛ace Rytm w Biegu Przez Plotki Osób o Ró˙znym Poziomie Zaawansowania Sportowego; Academy of Physical Education: Katowice, Poland, 2012. (In Polish) © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Functional Laterality of the Lower Limbs Accompanying Special Exercises in the Context of Hurdling.

11-07-2019

Iskra, Janusz,Marcinów, Ryszard,Wojciechowska-Maszkowska, Bożena,Otsuka, Mitsuo

eng

PMC7665126

International Journal of Environmental Research and Public Health Review Predictive Performance Models in Long-Distance Runners: A Narrative Review José Ramón Alvero-Cruz 1 , Elvis A. Carnero 2,3 , Manuel Avelino Giráldez García 4 , Fernando Alacid 5 , Lorena Correas-Gómez 6 , Thomas Rosemann 7 , Pantelis T. Nikolaidis 8,* and Beat Knechtle 7 1 Faculty of Medicine, University of Málaga, Andalucía TECH, 29071 Málaga, Spain; [email protected] 2 Translational Research Institute for Metabolism and Diabetes, Florida Hospital Sanford, Orlando, FL 32804, USA; [email protected] 3 Sanford Burnham Prebys Medical Discovery Institute, La Jolla, CA 92037, USA 4 Faculty of Sports Science and Physical Education, University of A Coruña, 15179 Oleiros, Spain; [email protected] 5 Department of Education, Health Research Centre, University of Almería, 04120 Almería, Spain; [email protected] 6 Faculty of Education Sciences, University of Málaga, Andalucía TECH, 29071 Málaga, Spain; [email protected] 7 Institute of Primary Care, University of Zurich, 8006 Zurich, Switzerland; [email protected] (T.R.); [email protected] (B.K.) 8 School of Health and Caring Sciences, University of West Attica, 12243 Athens, Greece * Correspondence: [email protected]; Tel.: +30-6977-8202-98 Received: 11 October 2020; Accepted: 6 November 2020; Published: 9 November 2020



 Abstract: Physiological variables such as maximal oxygen uptake (VO2max), velocity at maximal oxygen uptake (vVO2max), running economy (RE) and changes in lactate levels are considered the main factors determining performance in long-distance races. The aim of this review was to present the mathematical models available in the literature to estimate performance in the 5000 m, 10,000 m, half-marathon and marathon events. Eighty-eight articles were identified, selections were made based on the inclusion criteria and the full text of the articles were obtained. The articles were reviewed and categorized according to demographic, anthropometric, exercise physiology and field test variables were also included by athletic specialty. A total of 58 studies were included, from 1983 to the present, distributed in the following categories: 12 in the 5000 m, 13 in the 10,000 m, 12 in the half-marathon and 21 in the marathon. A total of 136 independent variables associated with performance in long-distance races were considered, 43.4% of which pertained to variables derived from the evaluation of aerobic metabolism, 26.5% to variables associated with training load and 20.6% to anthropometric variables, body composition and somatotype components. The most closely associated variables in the prediction models for the half and full marathon specialties were the variables obtained from the laboratory tests (VO2max, vVO2max), training variables (training pace, training load) and anthropometric variables (fat mass, skinfolds). A large gap exists in predicting time in long-distance races, based on field tests. Physiological effort assessments are almost exclusive to shorter specialties (5000 m and 10,000 m). The predictor variables of the half-marathon are mainly anthropometric, but with moderate coefficients of determination. The variables of note in the marathon category are fundamentally those associated with training and those derived from physiological evaluation and anthropometric parameters. Keywords: prediction equations; performance; long-distance runners Int. J. Environ. Res. Public Health 2020, 17, 8289; doi:10.3390/ijerph17218289 www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2020, 17, 8289 2 of 23 1. Introduction The great popularity of long-distance running has seen an unprecedented increase in the last 10 years. This has generated, in coaches and athletes, a great interest in the development of performance prediction models based on linear regression equations, with the aim of helping many athletes in their preparation for competitions. These predictions are based on a combination of physiological, anthropometric, nutritional and training factors (modifying frequency, volume and intensity), most obtained in exercise physiology laboratories, through variables related to training load [1,2]. Performance in long-distance disciplines can be defined as the final time or race time, and its understanding is important both for designing training programs and for determining scheduled training and race pace. However, accurate knowledge is frequently difficult to obtain, especially in long-distance races, as it would involve high training loads, which can, at times, indicate poor race planning in inexperienced runners who normally use polarized training methods [3]. This and other factors associated with the control of training, result in predictive models being recognized and useful for coaches or professional runners. The physiological adaptations produced by training in amateur runners are well understood and are generally those performed at submaximal intensities with continuous training strategies [4]. In high-level athletes, these improvements are seen particularly with tempo runs and short-interval training, as methods to improve performance [5]. Therefore, transferring the results and conclusions obtained from amateur athletes to high-level athletes is not advisable [6]. Performance in endurance running is influenced by a variety of factors, both anthropometric and training. Morphological (somatotype components) and anthropometric characteristics such as skinfolds, body fat percentage, circumferences, lower limb length, weight, height and body mass index appear to influence performance. Accordingly, certain characteristics have a better relationship between energy expenditure and performance [7,8]. There are numerous studies on physiological factors in the literature on performance prediction in long-distance runners. Classically, maximal oxygen uptake (VO2max), running economy (RE) and anaerobic threshold (AT) stand out as the main variables that have been used to predict performance in long-distance races [9,10], but a large gap exists in the field of performance prediction based on field tests. The aim of this narrative review was to undertake a descriptive, analytical and detailed analysis of the determinants and predictive ability of anthropometric, physiological (laboratory test), training and combined variables, as well as field assessments (field tests), to estimate performance in specialties of long-distance races (5000 m, 10,000 m, half–marathon and marathon). 2. Materials and Methods This document is classified as a narrative review and was carried out under a framework of assigning key attributes based on Search, Appraisal, Synthesis and Analysis (SALSA) [11]. Accordingly, the search was exhaustive. The synthesis is a tabular exposition of the data and the analysis may be chronological, conceptual or thematic [11]. In general terms, this narrative review presents all the known published works that include runners of different levels: all of these in different types of runner (amateur, moderately trained, highly trained, high-level and elite) with the common denominator that they are generally trained both in length of time and number of weekly sessions. Also included are all studies that found associations between anthropometric and physiological parameters and performance in the middle-distance (5000 m and 10,000 m) and long-distance (half-marathon and marathon) events. 2.1. Search The abstracts of original English articles registered in the Pubmed, SciELO (Scientific Electronic Library On line), ScienceDirect and SportDiscus databases were reviewed. The terms entered in the Int. J. Environ. Res. Public Health 2020, 17, 8289 3 of 23 search engines were as follows: “runners”, “long distance runners”, “performance”, “performance prediction”, “anthropometric”, “physiological determinants”, “performance determinants”, “5000 m”, “10,000 m”, “half-marathon” and “marathon”, as well as the combinations of all of them, depending on the specialty examined. 2.2. Selection Criteria The selection criteria were all relevant articles, as well as books and monographs. The first evaluation consisted of reading the abstract and the full text of the selected studies, followed by an analysis of the results. 2.3. Exclusion Criteria Case studies, duplicate articles and abstracts without clear and sufficient information were excluded. 3. Results The flow chart (Figure 1) shows the final selection of 58 articles, with 12 articles identified for the 5000 m modality, 13 for the 10,000 m, 12 for the half-marathon and 21 for the marathon. Int. J. Environ. Res. Public Health 2020 16, x 3 of 21 m”, “10,000 m”, “half-marathon” and “marathon”, as well as the combinations of all of them, depending on the specialty examined. 2.2. Selection Criteria The selection criteria were all relevant articles, as well as books and monographs. The first evaluation consisted of reading the abstract and the full text of the selected studies, followed by an analysis of the results. 2.3. Exclusion Criteria Case studies, duplicate articles and abstracts without clear and sufficient information were excluded. 3. Results The flow chart (Figure 1) shows the final selection of 58 articles, with 12 articles identified for the 5000 m modality, 13 for the 10,000 m, 12 for the half-marathon and 21 for the marathon. Figure 1. Diagram of study search and selection process. In Table 1, the variables are grouped as demographic, laboratory assessments, field test, training, anthropometric and others. Table 1. Partial and total figures for performance prediction variables in long-distance specialties. Long-Distance Specialties Variables 5000 m 10,000 m HM M Total % of Total Demographic 4 1 1 1 7 5.1 Aerobic Metabolism 26 14 3 16 59 43.4 Training 1 5 2 28 36 26.5 Anthropometry 2 5 16 5 28 20.6 Field test 0 1 2 0 3 1.47 Others 0 1 0 3 4 2.94 Figure 1. Diagram of study search and selection process. In Table 1, the variables are grouped as demographic, laboratory assessments, field test, training, anthropometric and others. Table 1. Partial and total figures for performance prediction variables in long-distance specialties. Long-Distance Specialties Variables 5000 m 10,000 m HM M Total % of Total Demographic 4 1 1 1 7 5.1 Aerobic Metabolism 26 14 3 16 59 43.4 Training 1 5 2 28 36 26.5 Anthropometry 2 5 16 5 28 20.6 Field test 0 1 2 0 3 1.47 Others 0 1 0 3 4 2.94 Subtotals/Total 33 27 24 51 137 100 HM: Half-marathon, M: Marathon. Int. J. Environ. Res. Public Health 2020, 17, 8289 4 of 23 3.1. Demographic Variables Of the seven demographic variables, the most notable is age, which is included in all the specialties studied. Gender is only recorded in the 5000 m specialty [12]. 3.2. Aerobic Metabolism Assessment Variables In this section, the variables were classified into two groups: 1. Maximum range (VO2max, velocity at maximal oxygen uptake [vVO2max], maximum heart rate, maximum lactate, vVO2 with the University of Montreal Track Test, anaerobic capacity and oxygen deficit, etc.). 2. Submaximal range (VO2 at lactate threshold, lactate threshold, velocity at lactate levels of 2.5–3 and 4 mmol/L, RE, heart rate at individual anaerobic threshold (IAT), velocity at heart rate deflection point, VO2 and % VO2 at AT, velocity at AT, lactate level at AT and % of peak velocity at AT). Of particular note are vVO2max and VO2max, RE, understood as oxygen uptake at specific velocity, VO2 at AT and velocity at the level of 4 mmol/L lactate. Thirty-one of these studies include mL/kg/min among the variables that are associated with or are predictive factors of running performance from middle to long distance. Additionally, 24 studies include variables such as km/h, m/min, m/s associated with conditions obtained at VT2 (anaerobic threshold), velocity at heart rate deflection, IAT, ATLab (AT in laboratory test), etc. 3.3. Training Variables The training variables were grouped into two categories: quantitative (mean race duration, number of training sessions per week, miles per week, km per week, training volume, miles in 8 weeks, training in 9 weeks, years of training) and qualitative (training pace, record for 1 mile, 5 miles, 10 miles, half-marathon time and having finished a marathon). 3.4. Field Test Variables Only two studies measuring AT using the University of Montreal Track test [13], and covered distance in the Cooper test [14,15] 3.5. Anthropometric Variables These variables are classified into three categories: (i) basic measurements (height, weight, body mass index, skinfolds and muscle circumferences), (ii) body composition fractions (fat mass, fat-free mass and skeletal muscle mass) and (iii) somatotype components (endomorphy, mesomorphy and ectomorphy). Other important performance-related variables are body mass index, fat mass percentage, and skinfolds as regional indicators of adiposity associated with performance. Fifteen of the 26 studies were conducted in the half-marathon specialty by Knechtle’s research group [8,16,17]. 3.6. Other Variables Noteworthy are also the use of a biochemical variable such as transferrin levels, as well as a model based on data collection through a post-competition survey [14] and leg volume and heart rate changes during the Ruffier test recovery period [15]. 3.7. Data Management and Presentation Tables 2–5 are individual tables for each distance (5000 m, 10,000 m) and long-distance specialty (half-marathon and marathon) respectively and structured to display: Author, year of publication, sex, number of participants, athletic level, dependent variable, independent variable(s) associated with performance (correlation coefficient, p-value) or if the independent variables comprise a significant model (equation): the coefficient of determination (R2) and the standard error of the estimate (SEE), the limits of agreement of the Bland–Altman plot (only in half-marathon) and the predictive equation. Int. J. Environ. Res. Public Health 2020, 17, 8289 5 of 23 Table 2. Multiple regression models associated with performance in 5000 m races. Author Year Sex n Level Dependent Variable Independent Variable r p R2 SEE Foster 1983 1 28 Well-trained 3 Miles VO2max −0.92 Training volume Intensity Tanaka 1984 1 21 Trained 5000 m vVO2max 0.78 <0.001 0.62 nr Ramsbottom 1987 1 55 University VO2max 5000 m −0.85 <0.01 0 43 5000 m −0.80 <0.01 1987 1 55 University 5000 m RE 0.39 <0.01 0 43 RE 0.34 <0.05 Fay 1989 0 13 Mod-Highly 5000 m (m/min) Vlact 4 mMol/L (m/min) 0.94 0.940–0.97 nr VO2max (ml/kg/min) Oxygen cost of running −0.4–(−0.63) Velocity (m/min) = 0.346 (vLac 4 mMol/L) + 1.899 (VO2max) Kenney 1985 1 8 Elite 5000 (time in sec) Age + VT2 (mL/kg/min) <0.02 0.98 nr Time (sec) = 11555 − 5.1 (age) − 2.9 (VT2) Weyand 1994 1–0 22–19 Competitive 5000 m Peak O2 Def (POD) −0.4 VO2max High %VO2 AT RE at 3.6 m/s Gender (1 = male; 2 = female) Specialty Time (sec) = 0.38 (POD) − 1.29 (VO2max) + 1.25 AT(%VO2) + 4.42 (RE) + 55.9 (Gender) − 47.4 (specialty) (1 sprinter, 2 long-distance runner) + 1664.9 nr nr Takeshima 1995 1 51 Popular 5000 m (m/s) VO2 LT (ml/kg/min) 0.87 Age ARD 0.89 VO2 LT (ml/kg/min) 0.79 Age VO2 LT (ml/kg/min) 0.82 Age ARD Velocity (m/s) = 4.436 + 0.045 (VO2 LT) − 0.033 (Age) + 0.005 (ARD) 0.89 0.27 Int. J. Environ. Res. Public Health 2020, 17, 8289 6 of 23 Table 2. Cont. Author Year Sex n Level Dependent Variable Independent Variable r p R2 SEE Roecker 1998 1–0 339–88 Competitive 5000 m (m/s) vPeak (km/h) 0.91 <0.001 0.940–0.97 IAT (m/s) 0.91 % Fat Mass nr MHR (bpm) Max Lact (mMol/L) Velocity (m/s) = 3.404 + 0.683 (vPeak) + 0.274 (IAT) − 0.05 (%FM) (MHR) − 0.079 (Max Lact) Nummela 2006 1 18 Well-trained Velocity (m/s) VO2max 0.55 <0.05 MART Vel (m/s) = 0.066 (VO2max) + 0.048 (MART) − 0.549 0.728 nr Stratton 2009 1–0 17–22 Untrained 5000 m (km/h) VO2 max (ml/kg/min) 0.55 <0.01 V LT (km/h) 0.73 <0.01 V Max (km/h) 0.89 <0.01 Run velocity (km/h) = −1.124 + 0.514 (Vmax) + 0.267 (V LT) 0.812 2009 1–0 17–22 Trained 5000 m (km/h) VO2 max (ml/kg/min) 0.51 <0.01 V LT (km/h) 0.76 <0.01 V Max (km/h) 0.83 <0.01 Run velocity (km/h) = −2.629 + 0.546 (Vmax) + 0.345 (V LT) 0.738 Mendes de Souza 2014 1 10 5000 m vVO2 max Lab 0.05 0.35 nr 1 10 5000 m vVO2 max Montreal 0.002 0.66 nr Dellagrana 2015 1 23 Moderately trained 5000 (time) vVT (km/h) −0.64 0.001 RE at 11.2 km/h (L/min) 0.44 0.035 Fat-free mass (kg) 0.57 <0.005 5 km T (min) = 25.64 − 0.71 (vVT) − 3.38 (RE 11.2) + 0.21 (FFM) 0.71 0.67 r: correlation coefficient; p: significance level; R2: coefficient of determination; SEE: standard error of estimation; vVO2max: max velocity in VO2max; RE: running economy; VLact4: velocity at 4mMol/L; AT: anaerobic threshold; POD: peak oxygen deficit; LT: lactate threshold; ARD: average running duration; IAT: individual anaerobic; threshold; MHR: maximal heart rate; Max Lact: maximal lactate; MART: maximal anaerobic running test; vVO2maxLab: maximal velocity at exercise laboratory test: vVO2max Montreal: maximal velocity on Montreal field test. vVT: velocity at ventilatory threshold. Int. J. Environ. Res. Public Health 2020, 17, 8289 7 of 23 Table 3. Multiple regression models associated with performance in 10,000 m races. Author Year Sex n Level Dependent Variable Independent Variable r p R2 SEE Foster 1983 1 17 Well-trained 3 Miles VO2 max −0.94 Training volume Intensity Tanaka 1984 1 21 Trained 10,000 m vVO2 max 0.96 nr 1 21 Trained 10,000 m vAT (ml/kg/min) 0.80 <0.001 Bale 1986 1 60 Elite & Good Time 10,000 m Workouts (WO)per week −0.87 0.75 2.28 Time (min) = 44.27 − 1.44 (WO) WO + Miles (MW) per week −0.84 Time (min) = 46.32 − 0.91 (WO) − 0.11 (MW) 0.8 2.08 WO + MW + Running years (RY) −0.80 Time (min) = 46.45 −0.68 (WO) − 0.11 (MW) − 0.38 (RY) 0.83 1.92 WO + MW + RY + Ectomorphy −0.40 Time (min) = 47.93 − 0.68 (WO) − 0.10 (MW) – 0.38 (RY) − 0.68 (Ectomorphy) 0.86 1.78 Brandon 1987 Middle 10,000 (m/s) VO2max (ml/kg/min) Anaerobic Capacity (AC) Height (cm) 10,000 (m/s) = 127.39 + 0.64 (VO2) + 0.21 (AC) + 0.4 (Height) Fay 1989 0 13 Moderate 10,000 m (m/min) Vlact 4 mmol/L(m/min) 0.840–0.94 High VO2max (ml/kg/min) Vlact 2 mmol/L(m/min) 10,000 (m/min) = 0.437 (vLA 4 mmol/L) + 2.082 (VO2max) + 8.698 10000 (m/min) = 0.728 (vLac 4 mmol/L) + 57.926 10,000 (m/min) = 0.407 (vLac 2 mmol/L) + 2.276 (VO2max) + 12.706 Morgan 1989 1 10 Well-trained Time (min) VO2max −0.45 >0.05 vVO2max −0.87 <0.01 Vel at 4 mmol/L −0.82 <0.01 RE 0.64 <0.05 Int. J. Environ. Res. Public Health 2020, 17, 8289 8 of 23 Table 3. Cont. Author Year Sex n Level Dependent Variable Independent Variable r p R2 SEE Petit 1997 1 15 Trained Vel Ventilatory threshold 0.95 0.96 Vel HR def (km/h) 10,000 (km/h) = 1.03 (Vel Deflection HR) Berg 1998 1 34 Mod trained Time 10,000 m BMI and Mesomorphy 0.61 0.38 7.3 10,000 (min) = 4.12 (BMI) − 4.5 (Mesomorphy) − 29.1 0 19 Mod trained Time 10,000 m Endomorphy 0.64 0.41 6.5 10,000 (min) = 37 + 3.3 (Endomorphy) Evans 1995 0 31 Highly trained 10,000 Pace (m/min) VO2max 0.89 0.05 0.8 Lac Threshold 0.89 0.05 0.8 VO2 (ml/kg FFM/min) 0.81 0.05 0.66 VO2 in LT 0.84 0.05 0.71 Takeshima 1995 1 51 Trained 10,000 vel (m/s) VO2 in LT (ml/kg/min) 0.78 0.62 nr Age VO2 in LT 0.81 0.67 Age nr Workout (min) 10,000 (m/s) = 4.371 + 0.037 (VO2 in LT) − 0.031 (Age) + 0.005 (Workout) 0.82 0.335 r: correlation coefficient; p: significance level; R2: coefficient of determination; SEE: standard error of the estimate; VO2max: maximal oxygen uptake; vVO2max: velocity at VO2max; WO: workouts; vAT: velocity at anaerobic threshold; Lact 4: velocity at 4 mmol/L; vLact 2: velocity at 2 mmol/L; RE: running economy; vHR def: velocity at heart rate deflection; BMI: body mass index; FFM: fat-free mass; LT: lactate threshold; AT: anaerobic threshold; IAT: individual anaerobic threshold; HR: heart rate; Max Lact: maximal lactate; SK: skinfold. Int. J. Environ. Res. Public Health 2020, 17, 8289 9 of 23 Table 4. Multiple regression models associated with performance in half-marathon races. Author Year Sex n Level Dependent Variable Independent Variable r p R2 SEE L LOA to U LOA Campbell 1985 1–0 88–10 Finishers Running Speed (km/h) Age Height 0.18 ns Pulse rate 1 (PR1) −0.53 Pulse rate 2 (PR2) −0.35 km/week (K) 0.53 <0.01 Training weeks (NW) 0.4 <0.01 BMI −0.41 <0.01 Running Speed (km/h) = 21.3 +0.028 (K) − 0.31 (BMI) − 0.037 (PR2) + 0.012 (NW) 0.47 nr Roecker 1998 1–0 339–88 Competitive IAT 0.93 <0.001 Running vel at 4 mmol/L 0.91 <0.001 vVO2max 0.89 <0.001 Rüst 2011 1 84 Recreational Race time BMI 0.56 0.01 Skinfolds 0.360–0.53 0.005 Percent fat mass 0.49 0.01 Race time = 72.91 + 3.045 (BMI) − 3.884 (SRT) 0.44 nr −25.1 to 25.1 Knechtle 2011 0 42 Recreational Race time Skinfolds 0.490–0.61 <0.001 Race time = 166.7 + 1.7 (mid axilla SK) − 6.4 (SRT) 0.71 nr nr nr Muñoz 2013 1 24 Vel (km/h) Velocity 2 at 14.6 ± 2.6 km/h Blood Lactate at velocity 2 0.97 0.414 Vel Half-marathon (km/h) = V2 * 1.085 + (BLa2 * −0.282) − 0.131 nr Friedrich 2014 0 83 Recreational Race time Weight 0.63 <0.0001 Height 0.27 0.01 BMI 0.57 <0.0001 Circumferences 0.510–0.55 <0.0001 Skinfolds 0.390–0.59 <0.0001 Skeletal Muscle mass 0.24 0.03 Fat mass 0.6 <0.0001 Friedrich 2014 1 147 Popular Race time Weight 0.27 0.0009 Height −0.17 0.04 BMI 0.46 <0.0001 Arm circumference 0.37 <0.0001 Skinfolds 0.290–0.43 <0.0001 Skeletal Muscle mass −0.07 >0.05 Fat mass 0.49 <0.0001 Int. J. Environ. Res. Public Health 2020, 17, 8289 10 of 23 Table 4. Cont. Author Year Sex n Level Dependent Variable Independent Variable r p R2 SEE L LOA to U LOA Knechtle 2014 1 147 Recreational Race time (min) Percent fat mass SRT (km/h) Race time (min) = 142.7 + 1.158 (%FM) − 5.223 (SRT) 0.42 13.3 −26 to 25.8 Knechtle 2014 0 83 Recreational Race time (min) Percent fat mass SRT (km/h) Race time (min) = 168.7 + 1.077 (%FM) − 7.556 (SRT) 0.68 9.8 −19 to 19.1 Gómez 2017 1 48 Recreational Race time (min) Week training (km) WT −0.75 < 0.05 Running experience (years) RE −0.80 < 0.05 BMI 0.64 < 0.05 Sum 6 Skinfolds (mm) 0.78 < 0.05 Race time (min) = 56.83 − 0.11 WT − 0.46 RE + 1.19 BMI + 0.16 Sum6SKF 0.82 nr −9.2 to 12.2 2017 1 48 Recreational Race time (min) Peak speed (km/h) −0.92 < 0.05 RCT (km/h) −0.92 < 0.05 Race time (min) = 180.86 − 2.81 Peak speed − 2.77 RCT 0.90 nr −6.7 to 6.4 2017 1 48 Recreational Race time (min) RCT Step rate (Hz) −0.38 < 0.05 RCT Step length (m) −0.87 < 0.05 Maximal step length (m) −0.73 < 0.05 Race time (min) = 271.9 − 33.38 RCTsr − 28.38 RCTsl − 29.8 Msl 0.88 nr −9.7 to 5.7 2017 1 48 Recreational Race time (min) Peak speed (km/h) −0.92 < 0.05 RCT (km/h) −0.92 < 0.05 Running Experience (years) −0.75 < 0.05 Race time (min) = 169.54 − 2.51 Peak speed − 2.25 RCT − 0.37 RE 0.93 nr −6.7 to 6.0 Alvero-Cruz 2019 1 23 Recreational Race time (min) Cooper test (m) −0.92 <0.0001 Race time (min) = 201.26 − 0.03433 Cooper (m) 0.873 3.78 −7.5 to 7.4 2019 1 23 Recreational Race time (min) vVO2max (km/h) −0.85 < 0.0001 Weight (kg) 0.4 0.04 Race time (min) = 156.7117 − 4.7194 vVO2max − 0.3435 Weight 0.769 5.28 9.5 to 9.7 Alvero-Cruz 2020 1 177 Recreational Race time (min) Cooper test (m) −0.906 <0.0001 0 21 Recreational Race time (min) = 205.6272 − 0.0356 Cooper (m) 0.82 5.19 −10.7 to 9.7 r: correlation coefficient; p: significance level; R2: coefficient of determination; SEE: standard error of the estimate; L: Low; U: Upper, LOA: limits of agreement; nr: no reported; BMI: body mass index; IAT: individual anaerobic threshold; vVO2max: velocity at VO2max; SRT: speed running time. Int. J. Environ. Res. Public Health 2020, 17, 8289 11 of 23 Table 5. Multiple regression models associated with performance in marathon races. Author Year Sex (M/F) n Level Dependent Variable Independent Variable r p R2 SEE Foster 1975 Race Time (min) VO2max(ml/kg/min) Time (min) = 3.45 (VO2max) + 387.3 nr nr Foster 1975 Race Time (min) VO2max Training longer in last 8 w Pace (seconds/mile) Time (min) = 2.75 (VO2max) − 0.022 (miles 8w) − 1 (TL8w) + 0.146 (pace) + 319.4 nr nr Slovic 1977 Race Time (min) Best record in mile (min) (BR1) Best record in 5 miles (min) (BR5) Best record in 10 miles (min)(BR10) Miles in last 8 weeks Finisher of one marathon Training longer in last 8 w Time (min) = 0.45 (BR1min) − 7.9 (Finisher) − 0.08(Miles 8w) − 1.45 (TL8w(min) + 116.5 nr nr Slovic 1977 Race Time (min) Best record in 5 miles (min) (BR5) Miles in last 8 weeks Training longer in last 8 w Time (min) = 6.62 (BR 5min) − 0.05(Miles 8w) − 1.45 (TL8w(min)) + 42.8 nr nr Slovic 1977 Race Time (min) Best record in 10 miles (min)(BR10) Miles in last 8 weeks Training longer in last 8 w Time (min) = 2.98 (BR 10 (min) − 0.04(Miles 8w) − 1.3 (TL8w(min) + 46.6 nr nr Davis 1979 Race Time (min) VO2max(ml/kg/min) %VO2 in AT Time (h) = 7.445 − 0.0338 (VO2max) − 0.0303 (%VO2) 0.99 Int. J. Environ. Res. Public Health 2020, 17, 8289 12 of 23 Table 5. Cont. Author Year Sex (M/F) n Level Dependent Variable Independent Variable r p R2 SEE Hagan 1981 1 50 Trained Race Time (min) VO2max −0.63 Avg km WO in last 9 weeks −0.64 total km −0.67 overall WO in last 9 weeks −0.62 Mean pace (m/min) Time (min) = 525.9 + 7.09 km (kmWO) − 0.45 (WO speed m/min) − 0.17 (km 9 weeks) 0.71 −2.01 (VO2max, ml x kg−1 x min−1) − 1.24 (age, year) Foster 1983 1 25 Well-trained 26.2 miles VO2max −0.95 Training volume Intensity Bale 1985 0 36 Trained Race Time (min) workouts/week Time (min) = −4.42 (WO per week) + 218.5 nr nr 1985 0 36 Trained Race Time (min) workouts/week Ectomorphy Time (min) = −3.72 (WO per week) − 7.02 (Ectomorphy) + 242.6 nr nr 1985 0 36 Trained Race Time (min) workouts/week Ectomorphy training years (TY) Time (min) = −3.32 (WO per week) − 6.05 (Ectomorphy) − 0.85 (TY) + 240.6 nr nr Hagan 1987 0 35 Combined Race Time (min) Mean km/day 0.77 <0.001 0.59 Training pace (m/min) 0.66 <0.001 0.44 Race Time = 449.88 − 7.61 (Mean km/day) − 0.63 (Training pace m/min) 0.82 nr 0.68 18.4 0 16 Experienced Race Time (min) BMI 0.7 nr 0.49 Training pace (m/min) 0.78 <0.001 0.61 Race Time = 214.24 + 393.07 (BMI) − 0.68 (training pace m/min) 0.87 nr 0.76 12.4 0 19 Novice Race Time (min) BMI 0.31 ns 0.1 Race Time = 369.58 − 10.1 (Mean km/day) 0..69 nr 0.48 22.2 Föhrenbach 1987 1–0 34 Race Time (min) Mean km last 9 weeks vLact 2,5 (m/s) 0.880–0.99 <0.001 vLact 3 (m/s) vLact 4 (m/s) Int. J. Environ. Res. Public Health 2020, 17, 8289 13 of 23 Table 5. Cont. Author Year Sex (M/F) n Level Dependent Variable Independent Variable r p R2 SEE Noakes 1990 1 20 Race Time (min) Time in Half-M (THM) Lact AT (mmol/L) % peak Vel in AT (lact) −0.88 Time (min = 1.98 (THM) + 6.23 AT (mmol/L) − 0.46 AT % vPeak mmol/L + 33.84 Time (min) = 1.94 (THM) + 5.8 AT (mmol/L) − 0.44 AT % vPeak mmol/L + 0.39 RE at 16 km/h + 16.79 Time (min) = 1.29 % vPeak mmol/L − 10.86 vLT (km/h) + 241.3 Time (min) = −4.92 vLT (km/h) − 4.46 vPeak (km/h) + 337.8 Noakes 1990 1 20 Race Time (min) Time in Half-M Lact AnT (mmol/L) % peak Vel i nAT (lact) VO2 at 16 km/h 0.760–0.9 Race Time (min) Lact AnT (mmol/L) % peak Vel in AT (lact) Race Time (min) Vel in AnT by lact in km/h vVO2max (km/h) Takeshima 1995 1 51 Popular Mean Velocity (m/s) VO2 LT (ml/kg/min) Age Mean Duration Workouts (min) Mean Vel (m/s) = 0.038 (VO2 LT) − 0.031 (Age) + 0.005 (MDWO) + 3.707 0.93 0.199 Roecker 1998 1–0 339–88 Competitive Mean Velocity (m/s) vIAT (m/s) 0.93 <0.001 0.950–0.97 vVO2max (km/h) 0.87 <0.001 MHR Weight Mean Vel (m/s) = 0.546 (vIAT) + 0.293 (vVO2max) + 0.013 (km/week) − 0.0155 (MHR) − 0.0253 (Weight) + 3.4 Arrese 2006 0 8 Highly trained Race Time Iliac crest SK 0.76 <0.05 Abdominal SK 0.76 <0.05 Subscapular SK 0.78 <0.05 Serum ferritin (µg/L) −0.76 <0.05 Race Time = 7658.331 + 55.519 (Subscapular SK) − 4.834 (ferritin) + 34.895 (Sum 6 SK) 0.992 <0.001 2006 1 10 Highly trained Race Time Left ventricular diameter (LVD) −0.68 <0.05 Lactate at 10 km/h 0.91 <0.001 Lactate at 22 km/h Race Time = 8408.623 (lact 10 km/h) − 18.255 (LVD) + 22.522 (lact 22 km/h) 0.991 <0.001 Int. J. Environ. Res. Public Health 2020, 17, 8289 14 of 23 Table 5. Cont. Author Year Sex (M/F) n Level Dependent Variable Independent Variable r p R2 SEE Tanda 2011 1–0 21–ene Trained Pace (sec/km) K (km/week) 0.94 0.81 Pace (P) (sec/km) 0.85 Pace (sec/km) = 17.1 + 140 exp [–0.0053 K] + 0.55 (Pace) 0.81 5.77 Muñoz 2013 1 24 Vel (km/h) Velocity 1 at 13,5 ± 0,9 km/h (V1) Blood Lactate at velocity 1 0.81 0.626 Vel Marathon (km/h) = V1 1.085 + (BLa2 − 0.429) − 0.170 Tanda 2013 1 126 Recreational Pace (sec/km) Km week Pace training (sec/km) Percent body fat Pace (sec/km) = 11.03 + 98.46 exp [−0.0053 Km week] + 0.387 (Pace) + 0.1 exp [0.23 %BF] 0.81 0.64 14.3 Mooses 2013 1 20 International IAAF scoring Total time on treadmill (TtT)(sec) 0.40 66.2 IAFF score = 162.30 + 0.41 (TtT) Till 2016 1–0 40 Recreational Race Time (min) treadmill time (min) Time (min) = −3.85 (treadmill time) +351.57 0.447 Salinero 2017 1 84 Amateur Time (min) % Body fat (%BF) 0.42 <0.001 ∆ Recovery Ruffier test (RT) 0.37 <0.000 Half-marathon performance (HMP) 0.81 <0.001 Time (min) = 96.1 + 2.3 (%BF) + 62.9 (RT) + 0.023 (HMP) 0.59 nr Time (min) % Body fat (%BF) 0.42 <0.001 ∆ Recovery Ruffier test (RT) 0.37 <0.000 10 km performance (10 km P) 0.73 <0.001 Time (min) = 104.3 + 3.1 (%BF) + 67.3 (RT) + 0.045 (10 km P) 0.53 nr Esteve-Lanao 2019 1–0 8–8 Recreational Avg speed 42k (km/h) 116 days before = AnT 0.810–0.94 <0.05 Speed 42k (km/h) = SpeedAnT (km/h) 0.771 + 0.959 0.659 nr 88 days before = AnT Speed 42k (km/h) = SpeedAnT (km/h) 0.863 − 1.463 0.714 nr 60 days before = AnT Speed 42k (km/h) = SpeedAnT (km/h) 1.013 − 0.944 0.76 nr 32 days before = AeT Speed 42k (km/h) = SpeedAeT (km/h) 1.012 − 1.147 0.804 nr 11 days before = AeT Speed 42k (km/h) = SpeedAeT (km/h) 1.004 − 1.145 0.85 nr Int. J. Environ. Res. Public Health 2020, 17, 8289 15 of 23 Table 5. Cont. Author Year Sex (M/F) n Level Dependent Variable Independent Variable r p R2 SEE Keogh 2020 1–0 157–103 Recreational Time (min) Age BMI Marathon experience (ME) Predicted finish time (PFT) Diff pred vs. finish time (DPvF) Pace St deviation Sex Time (min) = −5.252 + 0.162 Age + 0.319 BMI + 0.451 ME + 0.947 PFT − 0.636 (DPvF) + 2.925 Pace − 3.232 Sex 0.858 nr r: correlation coefficient; p: significance level; R2: coefficient of determination; SEE: standard error of estimation; VO2max: Maximal oxygen uptake; %VO2AT: percentage of VO2max at anaer. threshold; Avg km WO: average km of workouts; BMI: body mass index; vLact 2.5: velocity in m/s at 2.5 mmol/L; vLact 3: velocity in m/s at 3 mmol/L; vLact 4: velocity in m/s at 4 mmol/L; AnT: anaerobic threshold; MHR: maximal heart rate; vVO2max: velocity at VO2max; LVD: left ventricular diameter. Int. J. Environ. Res. Public Health 2020, 17, 8289 16 of 23 The tables present two types of study: those without a prediction equation in which they provide the correlations between the independent variables and the dependent variable (correlation coefficient and p-value. The studies including a prediction equation are shown in the tables with the R2 value and the SEE. In Table 4 only, corresponding to the studies on the half-marathon, a further section is included, pertaining to the information on bias between the predicted and the actual time, with the limits of agreement derived from the studies by Knechtle’s [8,18,19] and other authors [14,15,20,21]. Finally, the studies with a prediction equation are presented in a highlighted text box 3.8. Variables and Models Associated with the 5000 m Event Search: The different keywords were combined as follows: “performance, performance prediction”, “performance determinants”, “anthropometric and physiological determinants”, “5000 m”, “5 km”. Appraisal: The subjects of the different studies were generally moderately trained or highly trained athletes of different athletic levels (amateur, collegiate, competitive, elite), except for the study by Stratton which includes untrained individuals [22]. Of all the studies, only a few provide coefficients for determining the independent variable [13,23–27]. The coefficients of determination ranged from 0.62 to 0.98, but none of the studies reported the standard error. Additionally, the study by Stratton has an external validation study in a subsample of subjects [22]. Synthesis: It should be noted that in all the studies, the variables most used for performance prediction are derived from determinations of aerobic metabolism. In one study the variable is the percentage of fat mass measured by anthropometry [28] and in another the fat-free mass [29]. Only one study was conducted in which the velocity at VO2max in the University of Montreal Track Test, as a field variable, is presented as a predictor variable [13]. Analysis: Table 2 presents 12 studies from 1983 to 2015 [12,13,22–26,28–32]. The most notable are the physiological variables such as VO2max [12,23,25,32] and vVO2max, [13,22,28,31] and RE measurements [12,29,30,33]. Only one study examines training variables [26]. The most important anthropometric variables are the inclusion of body composition fractions (fat mass and fat-free mass). Of the 12 studies, eight include a prediction equation [12,22–26,28,29] (Table 2). 3.9. Variables and Models Associated with the 10,000 m Event Search: The different keywords were combined as follows: “performance, performance prediction,” “anthropometric and physiological determinants,” “performance determinants,” “10,000 m,” “10 km”. Appraisal: The subjects of the different studies were generally trained athletes of different levels (amateur, competitive, elite) with the exception of the studies by Brandon [34] and Berg [35], which included only moderately trained individuals. Synthesis: In all the studies, the variables most used for prediction continue to be those derived from laboratory tests. Furthermore, these variables increase compared to the 5000 m specialty. New variables include those from training data, such as number of training sessions, miles per week and years of training [7]. In addition, anthropometric variables such as skinfolds [36] and two somatotype components are beginning to be included [35] although these equations have a low-moderate R2 (0.380–0.41). Analysis:Table 3 presents 13 studies from 1983 to 2014 [13,23,26–28,33–46]. Physiological variables such as VO2max [23,32–34,38] and vVO2max continue to be prominent [27,28,33]. Of the 13 studies, seven have a prediction equation [7,23,26,28,34,37,44]. The coefficients of determination (R2) of the equations by Bale et al. (1986) are moderately high (from 0.75 to 0.86) and are based on training variables including the number of training sessions, miles run, years of training and a somatotype component such as ectomorphy [7,38] and the studies by Fay et al. (1989) with R2 > 0.84, based on the velocity associated with metabolic variables such as lactate at 2 and 4 mmol/L and at VO2max (Table 3). Int. J. Environ. Res. Public Health 2020, 17, 8289 17 of 23 3.10. Variables and Models Associated with the Half-Marathon Event Search: The different keywords were combined as follows: “long distance runners,” “performance, performance prediction,” “anthropometric and physiological determinants,” “performance determinants,” “half-marathon”. Appraisal: The subjects of the different studies were generally at an amateur level and infrequently at a competitive level (Roecker et al., 1998) [28]. Synthesis: It should be noted that the half-marathon is not an official specialty of the Olympic Games or the World Championships, although there are national and international competitions in this event. Consequently, the largest number of individuals who practice this modality are amateur runners, with different training loads, ages and levels of experience. Multiple associations have been found between performance and anthropometric variables, but with models of moderate predictive power (R2 = 0.440–0.71) and with wide limits of agreement between the predicted time and the actual race time. Finally, two studies should be mentioned due to the high coefficient of determination (R2 = 0.84) and relatively low limits of agreement obtained through the distance covered in the Cooper test as a predictor variable [14,15]. This is a simple field test that can be introduced into training routines and can provide very useful information and Cooper’s test has a good accuracy and reliability in amateur long-distance runners [20]. Analysis: Table 4 presents 11 studies from 1985 to 2020 [8,14–16,28,47–50]. Of these 11 studies, nine were undertaken from 2011. In this section we should note the many contributions by Knechtle’s group. Multiple publications by these authors base their results on the relationships between performance in half-marathon races with anthropometric variables such as skinfolds, estimated body composition variables such as fat mass and skeletal muscle mass, and training load variables such as average training velocity [8,48,50,51] (Table 4). 3.11. Variables and Models Associated with the Marathon Event Search: The different keywords were combined as follows: “long distance runners,” “performance, performance prediction,” “anthropometric and physiological determinants,” “performance determinants” and “marathon”. Appraisal: The subjects in the different studies are generally trained and/or highly trained and at different levels (amateur, competitive, elite), with the exception of the study by Hagan which includes novice runners [41]. Synthesis: The first studies in this field, by Foster (1983) [32], Slovic (1977) [52], Davies and Thompson (1979) [53], Föhrenbach et al. (1987) [39] and Noakes et al. (1990) [43], primarily relate training variables to athletic performance. A powerful prediction model should be mentioned (Tanda, 2011) [54], which estimates race pace with a high coefficient of determination of 0.81. Analysis: Table 5 presents 21 studies from 1975 to 2020. Of note are the variables associated with exercise physiology and aerobic metabolism [28,40,41,53] as well as, to a large extent, those related to training load [26,41,52,54–56] (Table 5). 4. Discussion The main strength of this literature review is the considerable number of publications and the subsequent analysis of the variables that make up the prediction equations of each of the specialties. This analytical text invites the reader and the scholar to use the assessment methods available to evaluate athletic performance. One of the difficulties we encountered in comparing the different equations is that there is no consensus on the definition of the type of athletes, with each author having named the type of subjects involved. Therefore, we recommend unifying and clearly defining each of the athletes and their level. We also found great differences in the number of athletes participating in the studies, ranging from eight subjects [24,36] to 427 including both men and women [28]. Int. J. Environ. Res. Public Health 2020, 17, 8289 18 of 23 The dependent variables of the models found are diverse, as they are expressed as time in minutes, seconds, hours; speed in m/s, m/min, km/h and, finally, the race pace in s/km. On this issue these have been the independent variables that have defined training loads, without finding work that has influenced in a quantification of both, strength trainings [57] and high-intensity intervals [6,58] from which predictor variables can be extracted. The number of independent variables is two or three, with some equations having as many as six independent variables. A piece of data missing in almost all the studies is the variance inflation factor (VIF), which informs us of multicollinearity. Some of the possible solutions to the problem of multicollinearity are the following: improvement in the sample design by extracting the maximum information from the observed variables, elimination of the variables suspected of causing multi-collinearity and, finally, in the case of having few observations, increasing the sample size [59]. The identification of physiological variables for performance prediction has at least two important applications around sports training. The first is the evaluation of certain defining physiological characteristics related to the sports specialty and the second is associated with training (volume and intensity) in relation to the sports modality and especially with regard to metabolic and functional characteristics (capacity and power, aerobic and anaerobic). The most widely studied variables for predicting aerobic performance in running are VO2max and vVO2max, both of which are fundamentally associated with short distances such as the 5000 m and 10,000 m events [10,22,23,25,28,43]. This is likely because the intensities at which these races are executed are very close to maximal intensities and thus their close correlation. VO2max is the physiological variable that represents aerobic capacity, or in other words, the measurement of the maximum energy produced by aerobic metabolism per unit of time. Both vVO2max and VO2max would effectively be the same as they occur essentially at the same time [28,31,43,60,61]. The variables related to the submaximum level and the variable intensities that occur in these areas have been studied extensively in all specialties, except for the half-marathon [26,28,39,43,62]. This is related to the fact that the half-marathon has not been recognized in the international federative sphere and, therefore there has been no interest in its study. In the half-marathon specialty, very few studies are available: one by Campbell in 1985 [47] and another by Roecker et al. [28] Campbell finds moderate-low correlations between some basic anthropometric parameters and running pulse rate and weeks of training. Roecker et al. [28] observed high correlations (r > 0.89) between individual anaerobic threshold and running velocity at an intensity of 4 mmol/L, both physiologically very similar concepts, and vVO2max. From 2011 onwards, the following references are provided by Knechtle’s group, which published many articles linking half-marathon times with numerous anthropometric variables and with low-moderate correlation coefficients [48] and with prediction models also with moderate coefficients of determination [19]. Many studies in the literature analyse performance prediction in aerobic specialties based on the physiological parameters mentioned above. However, these studies, using simple or multiple regression models, analyse the associations between physiological parameters and aerobic performance capacity in athletes for a single distance (frequently between 1500 m and 10,000 m) [27,61,63] Based on the studies mentioned above, it has been proposed that race distance and, therefore, exercise intensity may influence the associations between physiological indicators and aerobic performance. Nonetheless, no studies have addressed aerobic performance capacity in the same athletes at different distances with two or more physiological indicators, particularly in studies with vVO2max and its respective time to exhaustion. As a result, it is not possible to draw the same conclusions for all sports specialties and at different athletic levels (amateur, highly trained, trained) [60]. The variables related to the quantity and quality of training are almost exclusive to studies undertaken in the marathon specialty and for different levels of training. A contribution of this review is the general idea that the parameters recorded at the end of the graded exercise stress test are well understood, as are the parameters associated with aerobic and anaerobic thresholds, in terms of both metabolism and gas exchange, since in the different prediction Int. J. Environ. Res. Public Health 2020, 17, 8289 19 of 23 models, variables range between 85% and 99% of the stress intensities. From our point of view, it is here, in this range of intensities where stronger associations should be sought, that would allow us to obtain more powerful models for predicting performance. Similarly, in the field of ultramarathon races, which are becoming increasingly popular, variables related to RE, associated low lactate concentrations, percentage of VO2max and the search for models that integrate genetic aspects related to muscle damage and protein synthesis capacity should be explored, as well as how to more accurately determine and calculate training load both in terms of quantity and quality. In relation to genetic studies, it has been shown that polymorphisms (about 160) in 27 genes were identified in 10,442 participants, of whom 2984 were marathon runners, leaving the variance in the result on sports performance to be studied [64]. 4.1. Practical Applications The prediction of race time in the long-distance modalities has, above all, an initial application for novice runners, who have little knowledge of their race paces, allowing them to adjust to constant paces. Running paces can be modified depending on the phase of training. The knowledge of the variables associated with performance in long-distance runners should help coaches and exercise physiologists understand and promote the search for new variables that improve the prediction of sports performance. 4.2. Future Research Directions As future lines of research, we must consider aspects that are currently known as physiological events that occur at the aerobic threshold (VT1), at the anaerobic threshold (VT2) and at maximum intensities (VO2max). At the lactate threshold, normally below 50–60% of VO2max, we know the lactate values, the energy expenditure for the race and the RE. These same parameters are also well known at the anaerobic threshold, which could be estimated to be around 85% of VO2max. We have many parameters that associate sports performance with VO2max, such as running speed, individual anaerobic threshold, and lactate levels. In addition, we know the physiological responses when reaching 100% of VO2max. Up to this point we can see what the exercise physiology studies have been based on for performance. However, we believe that there is a gap in what occurs between the aforementioned points, with regard to studying these values (percentage VO2max, RE, lactate levels, etc.). Anaerobic capacities should also be further explored, particularly as related to the 5000 and 10,000 m events. Finally, we must not forget the quantification of training load and of the molecular and genetic aspects related to human performance (see Figure 2). Int. J. Environ. Res. Public Health 2020, 17, x FOR PEER REVIEW 18 of 21 been based on for performance. However, we believe that there is a gap in what occurs between the aforementioned points, with regard to studying these values (percentage VO2max, RE, lactate levels, etc.). Anaerobic capacities should also be further explored, particularly as related to the 5000 and 10,000 m events. Finally, we must not forget the quantification of training load and of the molecular and genetic aspects related to human performance (see Figure 2). Figure 2. Proposal for the study of long-distance runners. 5. Conclusions Physiological stress assessments are almost exclusive to the short long-distance specialties (5000 m and 10,000 m). Half-marathon predictor variables are mainly anthropometric, with moderate coefficients of determination and physiological and field test variables with high coefficients R2. The most relevant variables in the marathon modality are training variables derived from the evaluation of aerobic metabolism and anthropometric parameters. Author Contributions: Conceptualization J R A C M A G G and E A C ; methodology J R A C M A G G Figure 2. Proposal for the study of long-distance runners. Int. J. Environ. Res. Public Health 2020, 17, 8289 20 of 23 5. Conclusions Physiological stress assessments are almost exclusive to the short long-distance specialties (5000 m and 10,000 m). Half-marathon predictor variables are mainly anthropometric, with moderate coefficients of determination and physiological and field test variables with high coefficients R2. The most relevant variables in the marathon modality are training variables derived from the evaluation of aerobic metabolism and anthropometric parameters. Author Contributions: Conceptualization, J.R.A.-C., M.A.G.G. and E.A.C.; methodology, J.R.A.-C., M.A.G.G. and E.A.C.; investigation, J.R.A.-C., M.A.G.G. and E.A.C.; data curation, F.A. and L.C.-G.; writing—original draft, J.R.A.-C., M.A.G.G., E.A.C., F.A. and L.C.-G.; writing—review and editing, J.R.A.-C., M.A.G.G., E.A.C., F.A., T.R., L.C.-G., P.T.N. and B.K.; supervision, J.R.A.-C., M.A.G.G., E.A.C., F.A., T.R., L.C.-G., P.T.N. and B.K.; funding acquisition, J.R.A.-C. and B.K. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Exercise Physiology Research Group CTS-132, Junta de Andalucía, Spain and Institute of Primary Care, University of Zurich, Zurich, Switzerland. Conflicts of Interest: The authors declare no conflict of interest. References 1. Pate, R.R.; Macera, C.A.; Bailey, S.P.; Bartoli, W.P.; Powell, K.E. Physiological, anthropometric, and training correlates of running economy. Med. Sci. Sports Exerc. 1992, 24, 1128–1133. [CrossRef] [PubMed] 2. Doherty, C.; Keogh, A.; Davenport, J.; Lawlor, A.; Smyth, B.; Caulfield, B. An evaluation of the training determinants of marathon performance: A meta-analysis with meta-regression. J. Sci. Med. Sport 2020, 23, 182–188. [CrossRef] [PubMed] 3. Kenneally, M.; Casado, A.; Santos-Concejero, J. The effect of periodization and training intensity distribution on middle-and long-distance running performance: A systematic review. Int. J. Sports Physiol. Perform. 2018, 13, 1114–1121. [CrossRef] [PubMed] 4. González-Mohíno, F.; Santos-Concejero, J.; Yustres, I.; González-Ravé, J.M. The Effects of Interval and Continuous Training on the Oxygen Cost of Running in Recreational Runners: A Systematic Review and Meta-analysis. Sports Med. 2020, 50, 283–294. 5. Casado, A.; Hanley, B.; Santos-Concejero, J.; Ruiz-Pérez, L.M. World-Class Long-Distance Running Performances Are Best Predicted by Volume of Easy Runs and Deliberate Practice of Short-Interval and Tempo Runs. J. Strength Cond. Res. 2019, 30. [CrossRef] 6. Laursen, P.B.; Jenkins, D.G. The scientific basis for high-intensity interval training: Optimising training programmes and maximising performance in highly trained endurance athletes. Sports Med. 2002, 32, 53–73. [CrossRef] [PubMed] 7. Bale, P.; Bradbury, D.; Colley, E. Anthropometric and training variables related to 10km running performance. Br. J. Sports Med. 1986, 20, 170–173. [CrossRef] [PubMed] 8. Knechtle, B.; Barandun, U.; Knechtle, P.; Zingg, M.A.; Rosemann, T.R.C. Prediction of half-marathon race time in recreational female and male runners. Springerplus 2014, 3, 248. [CrossRef] 9. Joyner, M.J. Modeling: Optimal marathon performance on the basis of physiological factors. J. Appl. Physiol. 1991, 70, 683–687. [CrossRef] 10. Midgley, A.W.; McNaughton, L.R.; Jones, A.M. Training to enhance the physiological determinants of long-distance running performance: Can valid recommendations be given to runners and coaches based on current scientific knowledge? Sports Med. 2007, 37, 857–880. [CrossRef] [PubMed] 11. Grant, M.J.; Booth, A. A typology of reviews: An analysis of 14 review types and associated methodologies. Health Inf. Libr. J. 2009, 26, 91–108. [CrossRef] [PubMed] 12. Weyand, P.G.; Cureton, K.J.; Conley, D.S.; Sloniger, M.A.; Liu, Y.L. Peak oxygen deficit predicts sprint and middle-distance track performance. Med. Sci. Sports Exerc. 1994, 26, 1174–1180. [CrossRef] 13. de Souza, K.M.; de Lucas, R.D.; Grossl, T.; Costa, V.P.; Guglielmo, L.G.A.; Denadai, B.S. Performance prediction of endurancerunners through laboratory and track tests. Rev. Bras. Cineantropometria Desemp Desempeho Hum. 2014, 16, 465–474. Int. J. Environ. Res. Public Health 2020, 17, 8289 21 of 23 14. Alvero-Cruz, J.R.; Carnero, E.A.; Giráldez García, M.A.; Alacid, F.; Rosemann, T.; Nikolaidis, P.T.; Knechtle, B. Cooper Test Provides Better Half-Marathon Performance Prediction in Recreational Runners Than Laboratory Tests. Front. Physiol. 2019, 5, 1349. [CrossRef] 15. Alvero-Cruz, J.R.; Standley, R.A.; Giráldez-García, M.A.; Carnero, E.A. A simple equation to estimate half-marathon race time from the cooper test. Int. J. Sports Physiol. Perform. 2020, 15, 690–695. [CrossRef] 16. Knechtle, B.; Knechtle, P.; Barandun, U.; Rosemann, T.L. Predictor variables for half marathon race time in recreational female runners. Clinics (Sao Paulo) 2011, 66, 287–291. [CrossRef] 17. Knechtle, B.; Wirth, A.; Knechtle, P.; Zimmermann, K.; Kohler, G. Personal best marathon performance is associated with performance in a 24-h run and not anthropometry or training volume. Br. J. Sports Med. 2009, 43, 836–839. [CrossRef] 18. Rüst, C.A.; Knechtle, B.; Knechtle, P.; Barandun, U.; Lepers, R.; Rosemann, T. Predictor variables for a half marathon race time in recreational male runners. Open Access J. Sports Med. 2011, 113–119. [CrossRef] 19. Knechtle, B. Relationship of anthropometric and training characteristics with race performance in endurance and ultra-endurance athletes. Asian J. Sports Med. 2014, 5, 73–90. 20. Alvero-Cruz, J.R.; Giráldez García, M.A.; Carnero, E.A. Reliability and accuracy of Cooper’s test in male long distance runners. Rev. Andal. Med. Deport. 2017, 10, 60–63. [CrossRef] 21. Gómez-Molina, J.; Ogueta-Alday, A.; Camara, J.; Stickley, C.; Rodríguez-Marroyo, J.A.; García-López, J. Predictive variables of half-marathon performance for male runners. J. Sports Sci. Med. 2017, 16, 187–194. 22. Stratton, E.; O’Brien, B.J.; Harvey, J.; Blitvich, J.; McNicol, A.J.; Janissen, D.; Paton, C.; Knez, W. Treadmill velocity best predicts 5000-m run performance. Int. J. Sports Med. 2009, 30, 40–45. [CrossRef] 23. Fay, L.; Londeree, B.; LaFontaine, T.; Volek, M. Physiological parameters related to distance running performance in female athletes. Med. Sci. Sports Exerc. 1989, 21, 319–324. [CrossRef] [PubMed] 24. Kenney, W.L.; Hodgson, J.L. Variables predictive of performance in elite middle-distance runners. Br. J. Sports Med. 1985, 19, 207–209. [CrossRef] 25. Nummela, A.T.; Paavolainen, L.M.; Sharwood, K.A.; Lambert, M.I.; Noakes, T.D.; Rusko, H.K. Neuromuscular factors determining 5 km running performance and running economy in well-trained athletes. Eur. J. Appl. Physiol. 2006, 97, 1–8. [CrossRef] [PubMed] 26. Takeshima, N.; Tanaka, K. Prediction of endurance running performance for middle-aged and older runners. Br. J. Sports Med. 1995, 29, 20–23. [CrossRef] 27. Tanaka, K.; Matsuura, Y. Marathon performance, anaerobic threshold, and onset of blood lactate accumulation. J. Appl. Physiol. 1984, 57, 640–643. [CrossRef] 28. Roecker, K.; Schotte, O.; Niess, A.M.; Horstmann, T.; Dickhuth, H.H. Predicting competition performance in long-distance running by means of a treadmill test. Med. Sci. Sports Exerc. 1998, 30, 1552–1557. [CrossRef] 29. Dellagrana, R.A.; Guglielmo, L.G.; Santos, B.V.; Hernandez, S.G.; da Silva, S.G.; de Campos, W. Physiological, anthropometric, strength, and muscle power characteristics correlates with running performance in young runners. J. Strength Cond. Res. 2015, 29, 1584–1591. [CrossRef] 30. Ramsbottom, R.; Nute, M.G.; Williams, C. Determinants of five kilometre running performance in active men and women. Br. J. Sports Med. 1987, 21, 9–13. [CrossRef] 31. Tanaka, K.; Matsuura, Y.; Matsuzaka, A.; Hirakoba, K.; Kumagai, S.; Sun, S.O.; Asano, K. A longitudinal assessment of anaerobic threshold and distance-running performance. Med. Sci. Sports Exerc. 1984, 16, 278–282. [CrossRef] 32. Foster, C. VO2max and training indices as determinants of competitive running performance. J. Sports Sci. 1983, 1, 13–22. [CrossRef] 33. Morgan, D.W.; Baldini, F.D.; Martin, P.E.; Kohrt, W.M. Ten kilometer performance and predicted velocity at VO2max among well-trained male runners. Med. Sci. Sports Exerc. 1989, 21, 78–83. [CrossRef] 34. Brandon, L.J. Physiological factors associated with middle distance running performance. Sports Med. 1995, 19, 268–277. [CrossRef] 35. Berg, K.; Latin, R.W.; Coffey, C. Relationship of somatotype and physical characteristics to distance running performance in middle age runners. J. Sports Med. Phys. Fit. 1998, 38, 253–257. 36. Arrese, A.L.; Ostáriz, E.S. Skinfold thicknesses associated with distance running performance in highly trained runners. J. Sports Sci. 2006, 24, 69–76. [CrossRef] 37. Berg, K. Endurance training and performance in runners: Research limitations and unanswered questions. Sports Med. 2003, 33, 59–73. [CrossRef] Int. J. Environ. Res. Public Health 2020, 17, 8289 22 of 23 38. Evans, S.L.; Davy, K.P.; Stevenson, E.T.; Seals, D.R. Physiological determinants of 10-km performance in highly trained female runners of different ages. J. Appl. Physiol. 1995, 78, 1931–1941. [CrossRef] 39. Föhrenbach, R.; Mader, A.; Hollmann, W. Determination of endurance capacity and prediction of exercise intensities for training and competition in marathon runners. Int. J. Sports Med. 1987, 8, 11–18. [CrossRef] [PubMed] 40. Foster, C.; Daines, E.; Hector, L.; Snyder, A.C.; Welsh, R. Athletic performance in relation to training load. Wis. Med. J. 1996, 95, 370–374. 41. Hagan, R.D.; Smith, M.G.; Gettman, L.R. Marathon performance in relation to maximal aerobic power and training indices. Med. Sci. Sports Exerc. 1981, 13, 185–189. [CrossRef] [PubMed] 42. Legaz Arrese, A.; Munguía Izquierdo, D.; Serveto Galindo, J.R. Physiological measures associated with marathon running performance in high-level male and female homogeneous groups. Int. J. Sports Med. 2006, 27, 289–295. [CrossRef] 43. Noakes, T.D.; Myburgh, K.H.; Schall, R. Peak treadmill running velocity during the VO2 max test predicts running performance. J. Sports Sci. 1990, 8, 35–45. [CrossRef] 44. Petit, M.A.; Nelson, C.M.; Rhodes, E.C. Comparison of a mathematical model to predict 10-km performance from the Conconi test and ventilatory threshold measurements. Can. J. Appl. Physiol. 1997, 22, 562–572. [CrossRef] 45. Tanda, G.; Knechtle, B. Marathon performance in relation to body fat percentage and training indices in recreational male runners. Open Access J. Sports Med. 2013, 4, 141–149. 46. Bale, P.; Rowell, S.; Colley, E. Anthropometric and training characteristics of female marathon runners as determinants of distance running performance. J. Sports Sci. 1985, 3, 115–126. [CrossRef] 47. Campbell, M.J. Predicting running speed from a simple questionnaire. Br. J. Sports Med. 1985, 19, 142–144. [CrossRef] 48. Friedrich, M.; Rüst, C.A.; Rosemann, T.; Knechtle, P.; Barandun, U.; Lepers, R.; Knechtle, B. A Comparison of Anthropometric and Training Characteristics between Female and Male Half-Marathoners and the Relationship to Race Time. Asian J. Sports Med. 2014, 5, 10–20. [CrossRef] 49. Perez, I.M.; Perez, D.M.; Gonzalez, C.C.; Esteve-Lanao, J. Prediction of race pace in long distance running from blood lactate concentration around race pace. J. Hum. Sport Exerc. 2012, 7, 763–769. [CrossRef] 50. Knechtle, B.; Knechtle, P.; Barandun, U.; Rosemann, T. Anthropometric and training variables related to half-marathon running performance in recreational female runners. Phys. Sportsmed. 2011, 39, 158–166. [CrossRef] 51. Knechtle, B.; Wirth, A.; Knechtle, P.; Rosemann, T. Training volume and personal best time in marathon, not anthropometric parameters, are associated with performance in male 100-km ultrarunners. J. Strength Cond. Res. 2010, 24, 604–609. [CrossRef] 52. Slovic, P. Empirical study of training and performance in the marathon. Res. Q. 1977, 48, 769–777. [CrossRef] 53. Davies, C.T.M.; Thompson, M.W. Aerobic performance of female marathon and male ultramarathon athletes. Eur. J. Appl. Physiol. Occup. Physiol. 1979, 41, 233–245. [CrossRef] 54. Tanda, G. Predicition of marathon performance time on the basis of training indices. J. Hum. Sport Exerc. 2011, 6, 511–520. [CrossRef] 55. Esteve-Lanao, J.; Del Rosso, S.; Larumbe-Zabala, E.; Cardona, C.; Alcocer-Gamboa, A.; Boullosa, D.A. Predicting Recreational Runners’ Marathon Performance Time During Their Training Preparation. J. Strength Cond. Res. 2019, 1. [CrossRef] 56. Lucia, A.; Esteve-lanao, J.; Oliván, J.; Gómez-Gallego, F.; San Juan, A.F.; Santiago, C.; Pérez, M.; Chamorro, C.; Foster, C. Physiological characteristics of the best Eritrean runners—Exceptional running economy. Appl. Physiol. Nutr. Metab. 2006, 31, 530–540. [CrossRef] [PubMed] 57. Blagrove, R.C.; Howatson, G.; Hayes, P.R. Effects of Strength Training on the Physiological Determinants of Middle- and Long-Distance Running Performance: A Systematic Review. Sports Med. 2018, 48, 1117–1149. [CrossRef] 58. Cabral-Santos, C.; Gerosa-Neto, J.; Inoue, D.S.; Rossi, F.E.; Cholewa, J.M.; Campos, E.Z.; Panissa, V.L.G.; Lira, F.S. Physiological Acute Response to High-Intensity Intermittent and Moderate-Intensity Continuous 5 km Running Performance: Implications for Training Prescription. J. Hum. Kinet. 2017, 11, 127–137. [CrossRef] Int. J. Environ. Res. Public Health 2020, 17, 8289 23 of 23 59. Kuhn, M.; Johnson, K. Applied Predictive Modeling; Springer: New York City, NY, USA, 2013; ISBN 9781461468493. 60. Billat, V.; Beillot, J.; Jan, J.; Rochcongar, P.; Carre, F. Gender effect on the relationship of time limit at 100% VO(2max) with other bioenergetic characteristics. Med. Sci. Sports Exerc. 1996, 28, 1049–1055. [CrossRef] 61. Billat, V.; Demarle, A.; Paiva, M.; Koralsztein, J.P. Effect of training on the physiological factors of performance in elite marathon runners (males and females). Int. J. Sports Med. 2002, 23, 336–341. [CrossRef] 62. Florence, S.; Weir, J.P. Relationship of critical velocity to marathon running performance. Eur. J. Appl. Physiol. Occup. Physiol. 1997, 75, 274–278. [CrossRef] 63. Conley, D.L.; Krahenbuhl, G.S. Running economy and distance running performance of highly trained athletes. Med. Sci. Sports Exerc. 1980, 12, 357–360. [CrossRef] 64. Moir, H.J.; Kemp, R.; Folkerts, D.; Spendiff, O.; Pavlidis, C.; Opara, E. Genes and elite marathon running performance: A systematic review. J. Sports Sci. Med. 2019, 18, 559–568. Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Predictive Performance Models in Long-Distance Runners: A Narrative Review.

11-09-2020

Alvero-Cruz, José Ramón,Carnero, Elvis A,García, Manuel Avelino Giráldez,Alacid, Fernando,Correas-Gómez, Lorena,Rosemann, Thomas,Nikolaidis, Pantelis T,Knechtle, Beat

eng

PMC9602481

Citation: Motevalli, M.; Tanous, D.; Wirnitzer, G.; Leitzmann, C.; Rosemann, T.; Knechtle, B.; Wirnitzer, K. Sex Differences in Racing History of Recreational 10 km to Ultra Runners (Part B)—Results from the NURMI Study (Step 2). Int. J. Environ. Res. Public Health 2022, 19, 13291. https://doi.org/10.3390/ ijerph192013291 Academic Editors: Stacy T. Sims and Christopher T. Minson Received: 18 August 2022 Accepted: 11 October 2022 Published: 14 October 2022 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affil- iations. Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). International Journal of Environmental Research and Public Health Article Sex Differences in Racing History of Recreational 10 km to Ultra Runners (Part B)—Results from the NURMI Study (Step 2) Mohamad Motevalli 1,2 , Derrick Tanous 1,2 , Gerold Wirnitzer 3, Claus Leitzmann 4, Thomas Rosemann 5 , Beat Knechtle 5,6,* and Katharina Wirnitzer 1,2,7 1 Department of Sport Science, Leopold-Franzens University of Innsbruck, 6020 Innsbruck, Austria 2 Department of Research and Development in Teacher Education, University College of Teacher Education, Tyrol, 6010 Innsbruck, Austria 3 adventureV & change2V, 6135 Stans, Austria 4 Institute of Nutrition, University of Gießen, 35390 Gießen, Germany 5 Institute of Primary Care, University of Zurich, 8000 Zurich, Switzerland 6 Medbase St. Gallen Am Vadianplatz, 9000 St. Gallen, Switzerland 7 Research Center Medical Humanities, Leopold-Franzens University of Innsbruck, 6020 Innsbruck, Austria * Correspondence: [email protected] Abstract: Sex differences in anatomy and physiology are the primary underlying factor for dis- tinctions in running performance. Overall participation in recreational running events has been dominated by males, although increasing female participation has been reported in recent years. The NURMI study participants filled in a survey following the cross-sectional study design with questions on sociodemographic data, running and racing motivations, training behaviors, and racing history and experience. Data analysis included 141 female and 104 male participants aged 39 (IQR 17) with a healthy median BMI (21.7 kg/m2; IQR 3.5). Statistical analyses revealed sex differences with the males performing faster at half-marathon (p < 0.001) and marathon (p < 0.001) events but no difference at ultra-marathons (p = 0.760). Mediation analyses revealed no significant sex differences in the performance of half-marathon and marathon when considering training behaviors (p > 0.05), racing history (p > 0.05), or racing experience (p > 0.05). Differences in recreational performance may be more closely related to social constraints and expectations of females rather than the physiological advantages of the male athlete. Health professionals who guide and support recreational runners as well as the runners themselves and their coaches may benefit from this study’s results in order to improve the best time performance through a deeper understanding of the areas that mediate sex differences. Keywords: running; marathon; gender; female; competition; performance; behavior; habit; endurance 1. Introduction Previous research has identified a clear sex-based difference in running performance [1–3], which underlies the separation of sexes at running competitions [4], although this dif- ference has been reported to diminish with distances beyond that of a marathon [5,6]. Runner training and preparation, however, appear to be crucial for racing performance regardless of sex [7]. To date, several studies have investigated the sex-based differences in the performance of recreational runners of specific distances [8–10], but to the best of the authors knowledge, no study has assessed sex-specific differences in recreational 10-kilometer (10 km), half-marathon (HM), marathon (M), and ultra-marathon (UM) run- ners in a single analysis. Anatomical (e.g., overall body size and weight, anthropometrics, body composition) [2,8] and physiological sex differences (e.g., aerobic capacity, muscular strength) [11] are reported to contribute to an advantage for the male runner. Thus, a biological advantage in favor of the male runner in the best time marathon performance of approximately 9-11% has been Int. J. Environ. Res. Public Health 2022, 19, 13291. https://doi.org/10.3390/ijerph192013291 https://www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2022, 19, 13291 2 of 10 reported [12–14] and appears to be due to various factors related to sex (e.g., greater body fat percentage and proportion of extremity vs. trunk fat in females, smaller hearts among females, and lower serum iron among females), potentially leading to a greater submaximal volume of oxygen consumption, increased lactate levels, and a higher heart rate when gradually increasing the running workload [1,11,12,15]. However, social circumstances may unnecessarily hinder female performance [11]. For example, overall running event participation has been dominantly male [16]. In addition, the antiquated biases that were once widely accepted in most developed countries in the 1960s may still linger across societies today regarding women’s athletic abilities, such as that long-distance endurance running is unhealthy and even harmful to the female anatomy [16]. Nowadays, running participation is reported as healthy for both sexes based on conclusive, detailed analyses [17–20]. Despite the previous misconceptions, women’s participation in running races has been steadily increasing to nearly match that of men today [21]. In addition to health [19,20,22], running event participants are often oriented towards racing through leisure or performance motivations [23,24], and participating in distance running events (10 km, HM, M, or UM) provides an opportunity for focusing on personal goals. The racing environment, the season of event, as well as individual characteris- tics such as anthropometrics, personality and motivations, nutritional status, and racing history/experience all contribute to the best time performances [7,25,26]. Therefore, the objective of this study is to analyze sex-related differences in racing history, experience, and performance of recreational 10 km to ultra-distance runners for the first time and whether training behaviors, racing history, or racing experience mediate a potential performance relationship. As there seems to be contradictions and inconsistency of results regarding the differences and similarities in endurance runners between the sexes in the available literature, especially based on the previous literature [8–10], this investigation hypothe- sizes there is a difference in best time race performance between the sexes of recreational distance runners. 2. Materials and Methods The profile of the present investigation methodology has been described previously (see Part A of the sequenced paper) [27] and detailed information for interested readers are available elsewhere [22,28]. In summary, the Nutrition and Running High Mileage (NURMI) study followed a protocol [29] approved by the ethics board of St. Gallen, Switzerland on 6 May 2015 (EKSG 14/145) and has the trial registration number of IS- RCTN73074080, which was registered retrospectively. The subjects, who were distance runners competing over 10 km to UM, were required to provide informed consent prior to participating in the NURMI study. Then, participants completed an online survey contain- ing questions on sociodemographic information and a complete profile of running- and performance-related data. The participants characteristics are presented in Table 1. The interested reader is kindly referred to the Part A publication for the participants’ recruitment and study proce- dures [27]. Figure 1 shows the enrollment and categorization of participants. The race performances and training behaviors of active female and male distance runners were described using the subsequential details: total and respective number of completed running events (HM, M/UM, 10 km); number of years of active running completed without break; racing history (age at first race; first race distance: 10 km, HM, M, UM; successful completion of HM, M, UM in last two years; ratio of completed HM/M to other events; best HM time, best M time, best UM time), running training (frequency of training, and daily and weekly span of training (hours, km) related to periodized phases and stages of training, respectively) and preparation for the main event (professional training resource; total duration). Running performance was linked to a normalized aggregate mean considering the best finishing time of HM and M and transformed to an index range (0–100). Int. J. Environ. Res. Public Health 2022, 19, 13291 3 of 10 Int. J. Environ. Res. Public Health 2022, 19, 13291 3 of 11 Figure 1. Flow Chart of Participants’ Enrollment by Race Distance and Sex. BMI—body mass index. HM—half-marathon. M/UM—marathon/ultra-marathon. 10 km—10 kilometers. The race performances and training behaviors of active female and male distance runners were described using the subsequential details: total and respective number of completed running events (HM, M/UM, 10 km); number of years of active running com- pleted without break; racing history (age at first race; first race distance: 10 km, HM, M, UM; successful completion of HM, M, UM in last two years; ratio of completed HM/M to other events; best HM time, best M time, best UM time), running training (frequency of training, and daily and weekly span of training (hours, km) related to periodized phases and stages of training, respectively) and preparation for the main event (professional training resource; total duration). Running performance was linked to a normalized ag- gregate mean considering the best finishing time of HM and M and transformed to an index range (0–100). R software (version 3.6.2 Core Team 2019: R Foundation for Statistical Computing, Vienna, Austria) was used to perform all the statistical analyses. Descriptive statistics (me- dian, interquartile range (IQR); mean, standard deviation (SD)) were used for the explor- atory analysis. PCA was used in identifying the two latent factors. Significant differences in racing activity (history, training, racing, etc.) between race distance and sex were calculated with a non-parametric test. Associations between varia- bles were performed by Chi-square test (χ²; nominal scale) and Wilcoxon test (ordinal and metric scale) and were approximated by using the F distributions with ordinary least squares. Multiple linear regression analysis and multivariate linear regression were used to test differences in the performance, health, and leisure motivations of female and male runners. The results of the regression are displayed as effect plots (95% confidence interval (95%-CI)). The level of statistical significance was set at p ≤ 0.05. Figure 1. Flow Chart of Participants’ Enrollment by Race Distance and Sex. BMI—body mass index. HM—half-marathon. M/UM—marathon/ultra-marathon. 10 km—10 kilometers. R software (version 3.6.2 Core Team 2019: R Foundation for Statistical Computing, Vienna, Austria) was used to perform all the statistical analyses. Descriptive statistics (median, interquartile range (IQR); mean, standard deviation (SD)) were used for the exploratory analysis. PCA was used in identifying the two latent factors. Significant differences in racing activity (history, training, racing, etc.) between race distance and sex were calculated with a non-parametric test. Associations between variables were performed by Chi-square test (χ2; nominal scale) and Wilcoxon test (ordinal and metric scale) and were approximated by using the F distributions with ordinary least squares. Multiple linear regression analysis and multivariate linear regression were used to test differences in the performance, health, and leisure motivations of female and male runners. The results of the regression are displayed as effect plots (95% confidence interval (95%-CI)). The level of statistical significance was set at p ≤ 0.05. 3. Results A total of 317 distance runners completed the questionnaire. Following data clearance, 72 participants did not meet the inclusion criteria; therefore, 245 runners (141 female, and 104 male) made up the final sample with a median body weight of 65 kg (IQR 14.2), height of 1.7 m (IQR 0.01), BMI of 21.7 kg/m2 (IQR 3.5), and age of 39 (IQR 17) years. The runners resided in various countries such as Germany (72%), Austria (18%), Switzerland (5%), or other (4%). There were significant differences between the sexes regarding body weight (p < 0.001), height (p < 0.001), and BMI (p < 0.001), with the males being heavier (73 kg, IQR 11.9), taller (1.8 m, IQR 0.1), and having a higher BMI (22.8, IQR 3.16). Moreover, 15 (67% female) participants were divorced (or separated), 164 (52% female) were married (or living with partner) and 66 (68% female) were single. Concerning the participants’ educational background, there were 44 females (vs. 39 males) with a high school degree or equivalent, 49 females (vs. 34 males) with a graduate degree (university level), and 31 females (vs. 22 males) with A-Level or equivalent; one female had no degree, while a total of 25 runners Int. J. Environ. Res. Public Health 2022, 19, 13291 4 of 10 provided no answer. A significant sex difference was found regarding exercise focus (p = 0.044), where 81 females (vs. 52 males) were leisure focused, with more males (44%; n = 46) focused on sports performance (30%) compared to females (30%; n = 43). A total of 154 participants were included as NURMI-runners racing at the HM or M/UM distance, including a greater proportion of males (52% male), and an additional 91 participants (26% male) were 10 km runners (p < 0.001). The males had completed more 10 km (8, IQR 15), HM (10, IQR 15), and M/UM (10, IQR 11) races than females (6, IQR 7; 6, IQR 6; 10, IQR 10, respectively). Characteristics of participants with racing motivation, history, and preferences of females and males, including racing experience, are shown in Table 1. Further details on the profile of the total sample and the sex-specific subsamples are provided in Part A [27]. Table 1. Distance Runner Characteristics, including Racing Motivation, Preferences, and Experience Displayed by Sex. Total Female Male Statistics 100% (245) 58% (141) 42% (104) Age (years) 39 (IQR 17) 37 (IQR 16) 43 (IQR 18) F(1, 243) = 7.03 p = 0.009 BMI (kg/m2) 21.7 (IQR 3.5) 20.9 (IQR 3.01) 22.8 (IQR 3.16) F(1, 243) = 28.72 p < 0.001 Marital Status Divorced/Separated 6% (15) 7% (10) 5% (5) χ2(2) = 5.32 p = 0.70 Married/With Partner 67% 164 61% (86) 75% (78) Single 27% (66) 32% (45) 20% (21) Racing Distance 10 km 37% (91) 48% (67) 23% (24) χ2(2) = 19.55 p < 0.001 HM 36% (89) 35% (49) 38% (40) M/UM 27% (65) 18% (25) 38% (40) Racing Motivation Leisure 46% (106) 50% (65) 41% (41) χ2(1) = 1.70 p = 0.193 Sport Performance 54% (125) 50% (66) 59% (59) Favorite Race Season Winter <1% (2) 2% (2) / χ2(3) = 4.45 p = 0.216 Spring 46% (106) 47% (62) 44% (44) Summer 23% (52) 18% (24) 28% (28) Autumn 31% (71) 33% (43) 28% (28) Years Active in Running 7 (IQR 7) 5 (IQR 7) 8 (IQR 11) F(1, 242) = 10.75 p = 0.001 First Race Age (years) 10 km 30 (IQ 16) 31 (IQR 16) 29 (IQR 15) F(1, 152) = 1.62 p = 0.205 HM 32 (IQR 16) 33 (IQR 15) 32 (IQR 16) F(1, 217) = 0.13 p = 0.720 M 35 (IQR 13) 35 (IQR 14) 34 (IQR 12) F(1, 136) = 0.38 p = 0.539 UM 40 (IQR 11) 41 (IQR 8) 38 (IQR 12) F(1, 46) = 0.75 p = 0.391 Total 30 (IQR 16) 31 (IQR 16) 30 (IQR 18) F(1, 240) = 0.05 p = 0.831 First Race Distance 10 km 65% (157) 70% (97) 58% (60) χ2(2) = 4.21 p = 0.122 HM 27% (65) 24% (33) 31% (32) M 9% (21) 6% (9) 12% (12) Races Completed in Total 8 (IQR 11) 7 (IQR 9) 10 (IQR 13) F(1, 243) = 6.75 p = 0.010 Ratio of HM/M to Other Completed Events 40 (IQR 50) 40 (IQR 51) 40 (IQR 48) F(1, 243) = 0.07 p = 0.791 Best Finishing Time (minutes) HM 111 ± 33 122 ± 39 98 ± 14 F(1, 215) = 72.41 p < 0.001 M 230 ± 45 252 ± 49 213 ± 32 F(1, 130) = 28.57 p < 0.001 UM 628 ± 489 662 ± 579 614 ± 454 F(1, 44) = 0.09 p = 0.760 Note. Results are presented as percentage (%), total numbers, median (IQR), and mean ± SD. χ2 statistic calculated by Pearson’s Chi-squared test and F statistic calculated by Wilxocon test. 10 km—10 kilometers. HM—half-marathon. M/UM—marathon/ultra-marathon. Int. J. Environ. Res. Public Health 2022, 19, 13291 5 of 10 No significant differences were found between the sexes in the motivation for racing (p = 0.193) or the current running motivation (p = 0.356). The favored season for running events did not differ between females and males (p = 0.216), participants most frequently preferred racing (n = 106; 42% male) in the springtime. Significant differences were found between the sexes racing history based on: (i) years fully active in running (p = 0.001), where males finished more complete years of running (8, IQR 11) than females (5, IQR 7); (ii) total races completed (p = 0.010), where males finished more events (10, IQR 13) than females (7, IQR 9); (iii) best finishing time over the HM distance (p < 0.001), in which males were faster on average (98 ± 14 min) compared to females (122 ± 39 min); (iv) best finishing time over the M distance (p < 0.001), in which males were faster on average (213 ± 32 min) compared to females (252 ± 49 min); (v) successful completion of planned M events over the previous two years (p < 0.001), where males finished more events (1, IQR 3) than females (0, IQR 1); and (vi) successful completion of planned UM events over the last two years (p < 0.001), where males finished more events (0, IQR 1) than females (0, IQR 0). No significant differences were observed for racing experience/history considering the age at the first time participating in a running event (p = 0.831), also for each race distance separately (10 km, HM, M, UM; p > 0.05), the first event race distance–whether 10 km, HM, M/UM—(p = 0.122), the best finishing time over the UM distance (females 662 ± 579 min vs. males 614 ± 454 min; p = 0.760), or the successful completion of planned HM events (p = 0.728). A sex difference was identified by a multivariate linear regression, with the male runners being significantly more sport performance motivated than the females (b = 9.6; 95% CI [0.02–19.2]; p < 0.05). No sex differences in health (b = −5.46; 95% CI [−12.9–1.93]; p > 0.05) or leisure motives (b = −2.5; 95% CI [−9.87–4.86]; p > 0.05) were identified by multivariate linear regression, as displayed by Figure 2. Int. J. Environ. Res. Public Health 2022, 19, 13291 6 of 11 Figure 2. Effect plots with 95%-CI displaying the differences between female and male running ex- ercise/racing motives (n = 231): performance, health, and leisure. Note. 95%-CIs were computed us- ing the multivariate regression analyses (Wald approximation). Multivariate linear regression analyses were used to predict the best HM and M run- ning time performance based on the following confounders: (i) the participants’ sex and training behavior–preparatory stage 3 (weekly training frequency, mileage and hours, and daily training mileage and hours), preparatory stage 4 (weekly training frequency, mile- age and hours, and daily training mileage and hours), weekly mileage of preparatory stage 1, professional advice sought, and pre-event training duration, which explained 22% of the variance (adjusted R2 = 0.22) with no significant difference found for sex (b = −0.295; 95% CI [−7.7–7.11]; p > 0.05); (ii) the participants’ sex, and racing history (years fully active in running and age at first running event), in which the model explains 16% of the variance (adjusted R2 = 0.16), and no significant difference was found for sex (b = −1.46; 95% CI [−8.93–6]; p > 0.05); and (iii) the participants’ sex and racing experience (HM races com- pleted, M races completed, the ratio of HM/M to total events completed, and total events completed), which explains 14% of the variance (adjusted R2 = 0.14) with no significant difference found for sex (b = −1.72; 95% CI [−9.14–5.7]; p > 0.05). 4. Discussion Figure 2. Effect plots with 95%-CI displaying the differences between female and male running exercise/racing motives (n = 231): performance, health, and leisure. Note. 95%-CIs were computed using the multivariate regression analyses (Wald approximation). Multivariate linear regression analyses were used to predict the best HM and M running time performance based on the following confounders: (i) the participants’ sex and training behavior–preparatory stage 3 (weekly training frequency, mileage and hours, and daily training mileage and hours), preparatory stage 4 (weekly training frequency, mileage and hours, and daily training mileage and hours), weekly mileage of preparatory stage 1, professional advice sought, and pre-event training duration, which explained 22% of the variance (adjusted R2 = 0.22) with no significant difference found for sex (b = −0.295; 95% CI [−7.7–7.11]; p > 0.05); (ii) the participants’ sex, and racing history (years fully active in running and age at first running event), in which the model explains 16% of the variance (adjusted R2 = 0.16), and no significant difference was found for sex (b = −1.46; 95% CI [−8.93–6]; p > 0.05); and (iii) the participants’ sex and racing experience (HM races completed, M races completed, the ratio of HM/M to total events completed, and Int. J. Environ. Res. Public Health 2022, 19, 13291 6 of 10 total events completed), which explains 14% of the variance (adjusted R2 = 0.14) with no significant difference found for sex (b = −1.72; 95% CI [−9.14–5.7]; p > 0.05). 4. Discussion This study aimed to identify sex differences in racing history, experience, and race performance among recreational 10 KM to UM distance runners. The most important findings were (1) males weighed more, were taller, and had a higher BMI than female runners; (2) there was no sex difference in racing motivation; however, when the running motivations (initial, current, racing) were combined with exercise focus, a significant sex difference was identified, with males being significantly more sport performance motivated; (3) males accumulated significantly more years of active running and completed significantly more races, especially for marathons and ultra-marathons within the last 2 years; (4) a significant sex difference in best time performance with males being faster on average at HM and M distances. Therefore, this investigation verifies the initial hypothesis that there is a sex difference in racing best time performance of recreational distance runners with males being significantly faster on average at HM and M events. (5) However, when considering best time performance as an index and including various mediators (training behavior, racing history, or racing experience) within multivariate linear regression models, no significant sex differences in performance were found. The findings of the anthropometric proportions of this study’s participants (weight, height, BMI) are consistent with previous reports [8], indicating that males have generally larger bodies and thus greater body weight than females. Females normally have a greater body fat percentage than males [30], and this has been reported to be a factor limiting long-distance running performance [1]. Excess weight, in particular, likely contributes to a disadvantage for the long-distance runner’s best time performance independent of sex, con- sidering that each additional gram must be actively transported, resulting in greater energy expenditure [11,20,31]. Although possible participants with obesity were excluded from this investigation based on the WHO criteria [32,33], the male runners (73 kg) remained significantly heavier than the females (59.5 kg). Likewise, the male BMI was significantly larger (+1.9 kg/m2) within the present sample. However, the BMI calculation does not take body composition into account [32,33], excluding potential confounders such as skeletal and lean body mass regarding the classification of individuals [34]. Moreover, the present sample is based on runners of predominantly long (HM) or very long (M/UM) distances who are well known to measure around the lower boundaries of body fat percentage compared to less active populations [35,36], which could indicate a greater skeletal and lean body mass among the males that is known to positively affect performance [11,20,31,34]. Indeed, reports of female runners developing eating disorders due to a strategy to min- imize body fat are not uncommon [20,37], which is linked to malnourishment and the co-occurrence of menstrual dysfunction and osteopenia [20,37]. However, this outcome may be limited only to sport performance motivated females [22]. Thus, there may be less of a psychological burden for male runners concerning body composition, which likely transfers to a higher level of psychological well-being, especially for performing in racing events [38]. Regarding the finding that motivations for racing were similar between the sexes, previous research has found males to be more sport performance oriented [39]. However, within the regression model, including all of the motives as covariates (initial running, exercise, current running, racing), the male dominance in sports performance motivation was in line with previous research [39]. This finding may be related to the fact that males generally tend to be more competitive than females, especially in athletics [40]. In addition, the present study found that the males had completed significantly more years fully active in running than their female counterparts, which is also consistent with previous findings [41]; however, this may be due to the fact that the males were significantly older in this sample. Likewise, the males had completed significantly more events, which shows that they are more dedicated and established in running and finishing races, possibly due to Int. J. Environ. Res. Public Health 2022, 19, 13291 7 of 10 their greater level of focus on competition and sport performance and age experience [39,40]. In turn, this finding is likely also related to the successful completion of marathon and ultra-marathon events within the previous two years, where the males finished significantly more long-distance events than the females. However, there were significantly more males within the present sample racing at the M/UM distances, which appears to contradict previous reports of increasing female participation in distance races [21]. Although, it could be possible that more females were participating in distance events (HM, M/UM) within this sample but had not successfully completed an event to date, which would have led to their exclusion from the present analyses. Interestingly, runners of extreme distances (UM) have been reported to separate from their partners or divorce due to a lack of support in running, considering the high amount of time spent training and performing in events [42]. Regarding the marital status of the present runner sample, no sex differences were identified even though more males were racing at the UM distance (40 males vs. 25 females). It was found that male runners performed significantly better than the females in HM and M events, considering their best times in which the males were 24 min and 39 min faster on average, respectively. This finding is consistent with the previous literature comparing sex differences in performance of recreational runners that show a 10% differ- ence at the minimum [43]. However, it is interesting to find that there was no sex-based performance difference in the best time to complete an ultra-marathon, but this analysis did not take into account the exact UM distance, which could have been any distance ≥50 km. Therefore, this result most likely suggests that the males were racing at longer UM distances than the females [2,3,5,6]. Furthermore, when controlling the best HM and M runtimes with the basic underlying factor of training behavior (including training duration, professional support, and specific variables within the main training preparatory period), no performance difference remained between the sexes. Previously, when sex differences in performances of elite runners were analyzed, males have been found to be faster regardless of matched training [12], and the present data on recreational runners appear to contradict that, possibly due to a more heterogeneous study sample in terms of performance ability (e.g., anthropometrics, skeletal muscle fiber composition) [13]. In addition, another multivariate linear regression model, including the covariates of racing history (age at first running event), indicated no sex difference in HM and M best time performance. This finding suggests that the males had a more robust background in racing and therefore running, which is inconsistent among other reports [8,44], but is reflected by the males having a significantly greater age than the females in this study. The connection to best time performance may be a result of long-term running adaptations due to a comparable involvement in completing the overarching periodization scheme for each sex, including main competitions over the years [8,45]. Lastly, a third multivariate linear regression model based on racing experience (number of completed HM and M events and the proportion of total events completed) also showed no sex difference in HM and M best time performance. Thus, being successful in completing running events is an important contributor for best time performance, regardless of sex, as seen previously [46]. Therefore, considering the best time performances of distance runners, males are generally faster; however, this sex difference was superficial in recreational distance runners when considering several modifiable areas that mediate performance, especially the training behaviors within Phase 2: the main preparatory period, considering that this model explained 22% of the variance (vs. racing history 16% vs. racing experience 14%). These findings may indicate that physiological sex differences are negligent regarding HM and M performance in recreational distance runners, and that social circumstances may play a considerable role in best time running performance. Based on the cross-sectional design of the study, the findings of the present investi- gation include limitations that must be considered, such as that no causative conclusions can be drawn from these results. The primary limitation to be addressed is that self-report is a common feature of the questionnaire approach that is well-known to result in a mis- Int. J. Environ. Res. Public Health 2022, 19, 13291 8 of 10 representation of answers, likely due to social views. As a means to limit the effect of misreporting, control questions were included throughout the survey, and, likewise, the participants were highly motivated distance runners, which may have enhanced the re- liability of their answers and the dataset. In addition, while 245 distance runners were included within the final analysis, the sample size was relatively modest, and the race distances were unevenly distributed, including more males in the longer distances (M/UM). Moreover, several factors of race events were not considered in the analyses, such as the race environment and conditions (inclement or mild weather along with temperature and relative humidity), time of day, season, and the event location; however, the best time reports were verified retrospectively. Another limitation of the present investigation is that nutritional status was not controlled for, or the type of nutrition followed by the runners, which is a well-known factor affecting performance [25,26,29]. However, the NURMI study has included participant nutritional data that have or will be reported in subsequent papers due to publication requirements. Considering the limitations of this investigation, valuable insights may be taken by female and male recreational distance running athletes, as well as their coaches, athletic trainers, exercise physiologists, physical therapists, and physicians, regarding essential factors contributing to a healthy best performance at running events. Future studies are suggested to investigate sex differences of recreational runners of long distances (HM, M, UM) in best time performance considering the main underlying focus of sport performance and controlling for mediators of running and racing history, professional support, and racing experience. 5. Conclusions This study was the first with the aim of investigating sex-associated differences in racing history, experience, and performance of recreational 10 km, HM, M, and UM dis- tance runners. The findings of the present study reveal males are faster in their best time performances at HM and M events and report a more robust background in running and racing history and experience. However, recreational female distance runners compete comparably with males considering the best time performance of HM and M events when statistically controlling for training behaviors or similar backgrounds in racing experi- ence/history. Thus, partaking in and, most importantly, finishing running events is key in improving best time performances for females and males alike. The results of this investi- gation provide a critical insight into the crucial differences in female and male recreational distance runner performance over several long distances (HM, M/UM, 10 km), which may be necessary for lifting the current societal circumstances (training/racing opportunities, performance expectations) and understandings of health and exercise professionals (athletic trainers, exercise physiologists, sports medicine doctors, physical therapists) in supporting or limiting female participation. Moreover, competitive runners, health professionals and coaches who supervise and counsel recreational runners may benefit from these findings in order to improve best time performance through a deeper understanding of the areas that mediate sex differences. Author Contributions: K.W. conceptualized and designed the study and the questionnaires together with B.K., C.L. and K.W. conducted data analysis and M.M. and D.T. provided statistical expertise. M.M., T.R., K.W. and D.T. drafted the manuscript. T.R., C.L., B.K. and K.W. critically reviewed it. G.W. provided technical support through data acquisition and data management. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: The study protocol is available online via https://springerplus. springeropen.com/articles/10.1186/s40064-016-2126-4 (accessed on 17 August 2022) and was ap- proved by the ethics board of St. Gallen, Switzerland on 6 May 2015 (EKSG 14/145). The study was conducted in accordance with the ethical standards of the institutional review board, medical professional codex, and with the 1964 Helsinki declaration and its later amendments as of 1996, the Int. J. Environ. Res. Public Health 2022, 19, 13291 9 of 10 Data Security Laws, and good clinical practice guidelines. Study participation was voluntary and could be canceled at any time without the provision of reasons or negative consequences. Informed Consent Statement: Informed consent was obtained from all individual participants included in the study considering the data collected, used, and analyzed exclusively and only in the context of the NURMI Study for scientific publication. Data Availability Statement: The data sets generated during and/or analyzed during the current study are not publicly available but may be made available upon reasonable request. Subjects will receive a brief summary of the results of the NURMI Study if desired. Acknowledgments: There are no professional relationships with companies or manufacturers who will benefit from the results of the present study. Moreover, this research did not receive any specific grant or funding from funding agencies in the public, commercial, or non-profit sectors. Conflicts of Interest: The authors declare no conflict of interest. References 1. Arrese, A.L.; Izquierdo, D.M.; Galindo, J.R.S. Physiological Measures Associated with Marathon Running Performance in High-Level Male and Female Homogeneous Groups. Int. J. Sport. Med. 2006, 27, 289–295. [CrossRef] [PubMed] 2. Peter, L.; Rust, C.A.; Knechtle, B.; Rosemann, T.; Lepers, R. Sex differences in 24-hour ultra-marathon performance—A retrospec- tive data analysis from 1977 to 2012. Clinics 2014, 69, 38. [CrossRef] 3. Knechtle, B.; Nikolaidis, P.T. Sex differences between women and men in running. In The Running Athlete; Canata, G.L., Jones, H., Krutsch, W., Thoreux, P., Vascellari, A., Eds.; Springer: Berlin/Heidelberg, Germany, 2022. [CrossRef] 4. IOC Framework on Fairness, Inclusion and Non-Discrimination on the Basis of Gender Identity and Sex Variations. Avail- able online: https://stillmed.olympics.com/media/Documents/News/2021/11/IOC-Framework-Fairness-Inclusion-Non- discrimination-2021.pdf (accessed on 20 June 2022). 5. Coast, J.R.; Blevins, J.S.; Wilson, B.A. Do gender differences in running performance disappear with distance. Can J. Appl. Physiol. 2004, 29, 139–145. [CrossRef] [PubMed] 6. Bam, J.; Noakes, T.D.; Juritz, J.; Dennis, S.C. Could women outrun men in ultramarathon races? Med. Sci. Sports Exerc. 1997, 29, 244–247. [CrossRef] [PubMed] 7. Knechtle, B.; Tanous, D.R.; Wirnitzer, G.; Leitzmann, C.; Rosemann, T.; Scheer, V.; Wirnitzer, K. Training and Racing Behavior of Recreational Runners by Race Distance—Results From the NURMI Study (Step 1). Front. Physiol. 2021, 12, 620404. [CrossRef] [PubMed] 8. Friedrich, M.; Rüst, C.A.; Rosemann, T.; Knechtle, P.; Barandun, U.; Lepers, R.; Knechtle, B. A Comparison of Anthropometric and Training Characteristics between Female and Male Half-Marathoners and the Relationship to Race Time. Asian J. Sports Med. 2013, 5, 10–20. [CrossRef] 9. Knechtle, B.; Knechtle, P.; Roseman, T.; Senn, O. Sex Differences in Association of Race Performance, Skin-Fold Thicknesses, and Training Variables for Recreational Half-Marathon Runners. Percept. Mot. Ski. 2010, 111, 653–668. [CrossRef] [PubMed] 10. Fokkema, R.; van Damme, A.A.D.N.; Fornerod, M.W.J.; de Vos, R.J.; Bierma-Zeinstra, S.M.A.; van Middelkopp, M. Training for a (half-)marathon: Training volume and longest endurance run related to performance and running injuries. Scand. J. Med. Sci. Sports 2020, 30, 1692–1704. [CrossRef] [PubMed] 11. Cheuvront, S.N.; Carter, R.; Deruisseau, K.C.; Moffatt, R.J. Running performance differences between men and women: An update. Sports Med. 2005, 35, 1017–1024. [CrossRef] 12. Hunter, S.; Stevens, A.A. Sex differences in marathon running with advanced age. Physiology or participation? Med. Sci. Sports Exerc. 2013, 45, 148–156. [CrossRef] [PubMed] 13. Hunter, S.K.; Stevens, A.A.; Magennis, K.; Skelton, K.W.; Fauth, M. Is there a sex difference in the age of elite marathon runners? Med. Sci. Sports Exerc. 2011, 43, 656–664. [CrossRef] [PubMed] 14. Sparling, P.B.; Nieman, D.C.; O’Connor, P.J. Selected Scientific Aspects of Marathon Racing. Sports Med. 1993, 15, 116–132. [CrossRef] 15. Joyner, M.J. Physiological limiting factors and distance running: Influence of gender and age on record performances. Exerc. Sport Sci. Rev. 1993, 21, 103–133. [CrossRef] 16. Scheerder, J.; Breedveld, K.; Borgers, J. Running across Europe: The Rise and Size of one of the Largest Sport Markets; Palgrave Macmillan: Hampshire, UK, 2015. 17. Hespanhol, L.C., Jr.; Pillay, J.D.; van Mechelen, W.; Verhagen, E. Meta-analyses of the effects of habitual running on indices of health in physically inactive adults. Sports Med. 2015, 45, 1455–1468. [CrossRef] [PubMed] 18. Lee, D.; Pate, R.R.; Lavie, C.J.; Sui, X.; Church, T.S.; Blair, S.N. Leisure-time running reduces all-cause cardiovascular mortality risk. J. Am. Coll. Cardiol. 2014, 64, 472–481. [CrossRef] [PubMed] 19. Wirnitzer, K.; Boldt, P.; Wirnitzer, G.; Leitzmann, C.; Tanous, D.; Motevalli, M.; Rosemann, T.; Knechtle, B. Health status of recreational runners over 10-km up to ultra-marathon distance based on data of the NURMI Study Step 2. Sci. Rep. 2022, 12, 10295. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2022, 19, 13291 10 of 10 20. Boldt, P.; Knechtle, B.; Nikolaidis, P.; Lechleitner, C.; Wirnitzer, G.; Leitzmann, C.; Wirnitzer, K. Sex Differences in the Health Status of Endurance Runners: Results from the NURMI Study (Step 2). J. Strength Cond. Res. 2019, 33, 1929–1940. [CrossRef] 21. International Association of Athletics Federations. The State of Running. Available online: https://runrepeat.com/state-of- running (accessed on 18 June 2021). 22. Wirnitzer, K.; Boldt, P.; Lechleitner, C.; Wirnitzer, G.; Leitzmann, C.; Rosemann, T.; Knechtle, B. Health Status of Female and Male Vegetarian and Vegan Endurance Runners Compared to Omnivores—Results from the NURMI Study (Step 2). Nutrients 2018, 11, 29. [CrossRef] [PubMed] 23. Masters, K.S.; Ogles, B.M.; Jolton, J.A. The development of an instrument to measure motivation for marathon running: The motivations of marathoners scales (MOMS). Res. Q. Exerc. Sport 1992, 64, 134–143. [CrossRef] 24. Wirnitzer, K.; Motevalli, M.; Tanous, D.; Wirnitzer, G.; Leitzmann, C.; Pichler, R.; Rosemann, T.; Knechtle, B. Who Is Running in the D-A-CH Countries? An Epidemiological Approach of 2455 Omnivorous, Vegetarian, and Vegan Recreational Runners—Results from the NURMI Study (Step 1). Nutrients 2022, 14, 677. [CrossRef] 25. Boullosa, D.; Esteve-Lanano, J.; Casado, A.; Peyre-Tartaruga, L.A.; da Rosa, R.G.; Coso, J.D. Factors affecting training and physical performance in recreation endurance runners. Sports 2020, 8, 35. [CrossRef] 26. Wirnitzer, K.; Motevalli, M.; Tanous, D.; Wirnitzer, G.; Leitzmann, C.; Wagner, K.-H.; Rosemann, T.; Knechtle, B. Training and Racing Behaviors of Omnivorous, Vegetarian, and Vegan Endurance Runners—Results from the NURMI Study (Step 1). Nutrients 2021, 13, 3521. [CrossRef] [PubMed] 27. Tanous, D.; Motevalli, M.; Wirnitzer, G.; Leitzmann, C.; Rosemann, T.; Knechtle, B.; Wirnitzer, K. Sex differences in training behaviors of 10 km to ultra endurance runners (Part A)—Results from the NURMI Study (Step 2). Int. J. Environ. Res. Public Health 2022, 19, 13238. [CrossRef] 28. Wirnitzer, K.; Motevalli, M.; Tanous, D.; Gregori, M.; Wirnitzer, G.; Leitzmann, C.; Hill, L.; Rosemann, T.; Knechtle, B. Supplement intake in half-marathon, (ultra-)marathon and 10-km runners—Results from the NURMI study (Step 2). J. Int. Soc. Sports Nutr. 2021, 18, 64. [CrossRef] 29. Wirnitzer, K.; Seyfart, T.; Leitzmann, C.; Keller, M.; Wirnitzer, G.; Lechleitner, C.; Rüst, C.A.; Rosemann, T.; Knechtle, B. Prevalence in running events and running performance of endurance runners following a vegetarian or vegan diet compared to non-vegetarian endurance runners: The NURMI Study. SpringerPlus 2016, 5, 458. [CrossRef] [PubMed] 30. Durnin, J.V.G.A.; Womersley, J. Body fat assessed from total body density and its estimation from skinfold thickness: Measure- ments on 481 men and women aged from 16 to 72 Years. Br. J. Nutr. 1974, 32, 77–97. [CrossRef] 31. Maldonado, S.; Mujika, I.; Padilla, S. Influence of Body Mass and Height on the Energy Cost of Running in Highly Trained Middle- and Long-Distance Runners. Endoscopy 2002, 23, 268–272. [CrossRef] 32. Word Health Organization (WHO). A Healthy Lifestyle—WHO Recommendations. 2010. Available online: http://www.euro. who.int/en/health-topics/disease-prevention/nutrition/a-healthy-lifestyle/body-mass-index-bmi (accessed on 11 May 2018). 33. Word Health Organization (WHO). Noncommunicable Diseases: Risk Factors. 2010. Available online: http://www.who.int/ gho/ncd/risk_factors/bmi_text/en/ (accessed on 11 May 2018). 34. Aloia, J.F.; Cohn, S.H.; Babu, T.; Abesamis, C.; Kalici, N.; Ellis, K. Skeletal mass and body composition in marathon runners. Metabolism 1978, 27, 1793–1796. [CrossRef] 35. Nikolaidis, P.T.; Rosemann, T.; Knechtle, B. Skinfold thickness distribution in recreational marathon runners. Int. J. Environ. Res. Public Health 2020, 17, 2978. [CrossRef] [PubMed] 36. Bale, P.; Rowell, S.; Colley, E. Anthropometric and training characteristics of female marathon runners as determinants of distance running performance. J. Sport. Sci. 1985, 3, 115–126. [CrossRef] [PubMed] 37. Deldicque, L.; Francaux, M. Recommendations for healthy nutrition in female endurance runners: An update. Front. Nutr. 2015, 2, 17. [CrossRef] [PubMed] 38. Zeiger, J.S.; Zeiger, R.S. Mental toughness latent profiles in endurance athletes. PLoS ONE 2018, 13, e0193071. [CrossRef] 39. Craft, B.B.; Carroll, H.A.; Lustyk, M.K.B. Gender differences in exercise habits and quality of life reports: Assessing the moderating effects of reasons for exercise. Int. J. Arts Soc. Sci. 2014, 2, 65–76. 40. Cashdan, E. Are men more competitive than women? Br. J. Soc. Psychol. 1998, 37 (Pt 2), 213–229. [CrossRef] [PubMed] 41. Thuany, M.; Souza, R.; Hill, L.; Mesquita, J.; Rosemann, T.; Knechtle, B.; Pereira, S.; Gomes, T. Discriminant Analysis of Anthropometric and Training Variables among Runners of Different Competitive Levels. Int. J. Environ. Res. Public Health 2021, 18, 4248. [CrossRef] [PubMed] 42. Mosko, E.M.; Waskiewicz, Z. The impact of family life on the motivations of ultramarathoners: The Karkonosze Winter Ultramarathon Case Study. Int. J. Environ. Res. Public Health 2020, 17, 6596. [CrossRef] 43. Nikolaidis, P.T.; Cuk, I.; Knechtle, B. Pacing of women and men in half-marathon and marathon races. Medicina 2019, 55, 14. [CrossRef] [PubMed] 44. Costill, D.L.; Fink, W.J.; Pollock, M.L. Muscle fiber composition and enzyme activities of elite distance runners. Med. Sci. Sports 1976, 8, 96–100. [CrossRef] 45. Kenneally, M.; Casado, A.; Santos-Concejero, J. The effect of periodization and training intensity distribution on middle- and long-distance running performance: A systematic review. Int. J. Sport. Physiol. Perform. 2018, 13, 1114–1121. [CrossRef] [PubMed] 46. Deaner, R.O.; Carter, R.E.; Joyner, M.J.; Hunter, S.K. Men are more likely than women to slow in the marathon. Med. Sci. Sports Exerc. 2015, 47, 607–616. [CrossRef]

Sex Differences in Racing History of Recreational 10 km to Ultra Runners (Part B)-Results from the NURMI Study (Step 2).

10-14-2022

Motevalli, Mohamad,Tanous, Derrick,Wirnitzer, Gerold,Leitzmann, Claus,Rosemann, Thomas,Knechtle, Beat,Wirnitzer, Katharina

eng

PMC6235347

Protocol Version 6 / 17-May-2016 Page 1 of 68 CLINICAL STUDY PROTOCOL A randomized, double-blind, placebo-controlled study to investigate the effects of recombinant human erythropoietin (NeoRecormon) on cycling performance and the occurrence of adverse events in well-trained cyclists. Short Title: Recombinant human erythropoietin effects on cycling performance. Version: 6 Date: 17-May-2016 CHDR number: CHDR1514 Toetsing Online number: NL54516.056.15 EudraCT number: 2015-003269-27 CHDR1514 Protocol Version 6 / 17-May-2016 Page 2 of 68 CONTACT DETAILS Trial Site (visit & delivery address) Centre for Human Drug Research Zernikedreef 8 2333 CL Leiden The Netherlands Telephone: + 31 71 5246 400 Fax: + 31 71 5246 499 Emergency: + 31 71 5246 444 Principal Investigator A.F. (Adam) Cohen, MD, PhD Telephone: + 31 (0)71 5246404 e-mail: [email protected] Co-Investigator J. (Koos) Burggraaf, MD, PhD Telephone: + 31 (0)71 5246418 e-mail: [email protected] Co-Investigator Dr. J.I. (Joris) Rotmans Telephone: + 31 71 5262148 e-mail: [email protected] Co-Investigator Dr. J.M.A. (Hans) Daniels Telephone: +31 (0)20 4444782 e-mail: [email protected] Co-Investigator J. A. A. C. (Jules) Heuberger, MSc Telephone: +31 (0)71 5246471 e-mail: [email protected] Co-Investigator Mr. H. (Herman) Ram Telephone: 010 - 2010154 e-mail: [email protected] Co-Investigator Dr. O. (Olivier) de Hon Telephone: 010-2010155 e-mail: [email protected] Co-Investigator T. (Thijs) Zonneveld e-mail: [email protected] Director of Clinical Operations J. M. (Ria) Kroon, BSc Telephone: + 31 (0)71 5246498 e-mail: [email protected] Statistician M. L. (Marieke) de Kam, MSc Telephone: + 31 (0)71 5246458 e-mail: [email protected] CHDR1514 Protocol Version 6 / 17-May-2016 Page 3 of 68 INDEPENDENT PHYSICIAN Prof. Dr G.J. Blauw, MD, PhD Department of Gerontology and Geriatrics LUMC Postbus 9600 2300 RC Leiden Telephone: + 31 71 5266 640 LABORATORY - CHEMISTRY C.M. Cobbaert, PhD CKCL LUMC, E2-P Albinusdreef 2 2333 ZA Leiden The Netherlands Contact person J. Verhagen LABORATORY - HAEMATOLOGY W.A.F. Marijt, MD, PhD CKHL LUMC, E1-Q Albinusdreef 2 2333 ZA Leiden The Netherlands Contact person F. Reymer LABORATORY – MICROBIOLOGY A.C.M. Kroes, MD, PhD KML LUMC, E4-P Albinusdreef 2 2333 ZA Leiden The Netherlands Contact person A.C.M. Kroes LABORATORY – DOPING DETECTION[ DoCoLab – Ugent Technologiepark 30 B-9052, Zwijnaarde Belgium Contact person Prof. Dr. ir. Peter Van Eenoo Telephone: + 32-9-3313291 e-mail: [email protected] LABORATORY – RNA EXPRESSION Leiden Genome Technology Center (LGTC) Human and Clinical Genetics Leiden University Medical Center, Postzone S4-P Postbus 9600 2300 RC Leiden Nederland CHDR1514 Protocol Version 6 / 17-May-2016 Page 4 of 68 Contact person Henk Buermans Tel: +31-71-526 9500/9522 Fax: +31-71-526 8285 E-mail: [email protected] PHARMACY (visit & delivery address) Apotheek LUMC, L0-P30 Albinusdreef 2 2333 ZA Leiden The Netherlands Telephone: + 31 71 5269111 Fax + 31 71 5262611 Pharmacist / Trial Manager Irene Teepe-Twiss, PhD / Linda van der Hulst cHDR1514 SIGNATURE PAGE . PRINCIPAL INVESTIGATOR Study Title A randomized, double-blind, placebo-controlled study to investigate the effects of recombinant human erythropoietin (NeoRecormon) on cycling performance and the occurrence of adverse events in well-trained cyclists. I acknowledge accountability for this protocol in accordance with CHDR's current procedures. A.F. (Adam) Cohen, MD, PhD Principal I nvestigator z,- 2õ l{ t Signature Date (dd Mmm yyyy) Protocol Version 6 I 17-May-2016 Page 5 of 68 cHDR1514 SIGNATURE PAGE. TRIAL SITE STAFF Centre for Human Drug Research Study Title A randomized, double-blind, placebo-controlled study to investigate the effects of recombinant human erythropoietin (NeoRecormon) on cycling performance and the occurrence of adverse events in well-trained cyclists. I acknowledge responsibility for this protocol in J.A.A.C (Jules) Heuberger, MSc Co-lnvestigator J. M. (Ria) Kroon, BSc Director Clinical Operations M. L. (Marieke) de Kam, MSc Statistieian ce with CHDR's current procedures. il- 2') '¡cr l I Qianafr ¡¡a Date (dd Mmm yyyy) ç I.rø ñala /¡{¡l l\/lmm rnn¡r¡\ 6, f¿l 2a l\4av 20)6 n /Siénature abscence Date (dd Mmm yyyy) Of Protocol Version 6 I 17 -May-2016 Page 6 of 68 CHDR1514 Protocol Version 6 / 17-May-2016 Page 7 of 68 TABLE OF CONTENTS CONTACT DETAILS ........................................................................................................................................... 2 SIGNATURE PAGE - PRINCIPAL INVESTIGATOR ......................................................................................... 5 SIGNATURE PAGE - TRIAL SITE STAFF ......................................................................................................... 6 PROTOCOL SYNOPSIS – CHDR1514 ............................................................................................................ 13 1 BACKGROUND AND RATIONALE ......................................................................................................... 23 1.1 Context............................................................................................................................................... 23 1.1.1 Need for research on doping ..................................................................................................... 23 1.1.2 Erythropoietin substances in patients and athletes ................................................................... 23 1.1.3 Examination of the evidence for the ergogenic properties of rHuEPO in cyclists ..................... 23 1.1.4 Study objectives ......................................................................................................................... 24 1.2 Non-clinical information ..................................................................................................................... 24 1.2.1 Non-clinical pharmacology ......................................................................................................... 24 1.2.2 Non-clinical pharmacokinetics and metabolism......................................................................... 24 1.2.3 Non-clinical toxicology and safety pharmacology ...................................................................... 24 1.3 Clinical information ............................................................................................................................ 24 1.3.1 Clinical pharmacology ................................................................................................................ 24 1.3.2 Clinical pharmacokinetics and metabolism ................................................................................ 24 1.3.3 Clinical toxicology and safety pharmacology ............................................................................. 24 1.4 Study rationale ................................................................................................................................... 24 1.4.1 Benefit and risk assessment ...................................................................................................... 24 1.4.2 Medical and regulatory background .......................................................................................... 25 1.4.3 Study population ........................................................................................................................ 25 1.4.4 Study design .............................................................................................................................. 25 1.4.5 Investigational drug and placebo ............................................................................................... 26 1.4.6 Dosing, safety margin calculations, stopping criteria ................................................................ 26 1.4.7 Treatment duration .................................................................................................................... 27 1.4.8 Endpoints ................................................................................................................................... 27 1.4.9 Statistical hypotheses and sample size ..................................................................................... 28 2 STUDY OBJECTIVES ............................................................................................................................... 30 2.1 Primary objective ............................................................................................................................... 30 2.2 Secondary objectives ......................................................................................................................... 30 2.3 Exploratory objectives ........................................................................................................................ 30 3 STUDY DESIGN ........................................................................................................................................ 31 3.1 Overall study design and plan ........................................................................................................... 31 3.1.1 Screening ................................................................................................................................... 31 3.1.2 Treatment and exercise tests .................................................................................................... 32 3.1.3 Training period ........................................................................................................................... 32 CHDR1514 Protocol Version 6 / 17-May-2016 Page 8 of 68 3.1.4 Competition ................................................................................................................................ 32 3.1.5 Follow-up ................................................................................................................................... 32 3.1.6 Urine doping control ................................................................................................................... 32 4 STUDY POPULATION .............................................................................................................................. 33 4.1 Subject population ............................................................................................................................. 33 4.2 Inclusion criteria ................................................................................................................................. 33 4.3 Exclusion criteria ................................................................................................................................ 33 4.4 Concomitant medications .................................................................................................................. 34 4.4.1 Mandatory concomitant supplementation .................................................................................. 34 4.4.2 Allowed concomitant medications ............................................................................................. 34 4.4.3 Prohibited concomitant medications .......................................................................................... 34 4.5 Lifestyle restrictions ........................................................................................................................... 34 4.6 Study drug discontinuation and withdrawal ....................................................................................... 35 4.6.1 Study drug interruption or discontinuation ................................................................................. 35 4.6.2 Subject withdrawal ..................................................................................................................... 35 4.6.3 Replacement policy ................................................................................................................... 35 5 INVESTIGATIONAL MEDICINAL PRODUCT .......................................................................................... 36 5.1 Investigational drug and matching placebo ....................................................................................... 36 5.2 Comparative drug .............................................................................................................................. 36 5.3 Study drug dosing scheme ................................................................................................................ 36 5.4 Study drug packaging and labelling ................................................................................................... 36 5.5 Drug accountability ............................................................................................................................ 36 5.6 Treatment assignment and blinding .................................................................................................. 36 5.6.1 Treatment assignment ............................................................................................................... 36 5.6.2 Blinding ...................................................................................................................................... 37 6 STUDY ENDPOINTS ................................................................................................................................. 38 6.1 Efficacy endpoints .............................................................................................................................. 38 6.2 Safety endpoints ................................................................................................................................ 39 6.3 Exploratory endpoints ........................................................................................................................ 40 7 STUDY ASSESSMENTS .......................................................................................................................... 41 7.1 Exercise-specific screening assessments ......................................................................................... 41 7.1.1 Exercise test .............................................................................................................................. 41 7.1.2 Questionnaire ............................................................................................................................ 41 7.2 Safety and tolerability assessments .................................................................................................. 41 7.2.1 Specific safety assessments...................................................................................................... 41 7.2.2 Vital signs................................................................................................................................... 41 7.2.3 Weight and height ...................................................................................................................... 41 7.2.4 Physical examination ................................................................................................................. 41 CHDR1514 Protocol Version 6 / 17-May-2016 Page 9 of 68 7.2.5 Electrocardiography ................................................................................................................... 41 7.2.6 Laboratory assessments ............................................................................................................ 41 7.3 Efficacy assessments ........................................................................................................................ 43 7.3.1 Exercise tests ............................................................................................................................ 43 7.3.2 Competition ................................................................................................................................ 43 7.3.3 Athlete biological passport ......................................................................................................... 43 7.3.4 Blood flow .................................................................................................................................. 43 7.4 NeoRecormon detection assessments .............................................................................................. 43 7.4.1 Urine doping screen ................................................................................................................... 43 7.5 RNA expression levels ....................................................................................................................... 43 7.6 Sequence of assessments and time windows ................................................................................... 44 7.7 Total blood volume ............................................................................................................................ 45 8 SAFETY REPORTING .............................................................................................................................. 46 8.1 Definitions of adverse events ............................................................................................................ 46 8.1.1 Intensity of adverse events ........................................................................................................ 46 8.1.2 Relationship to study drug ......................................................................................................... 46 8.1.3 Chronicity of adverse events ..................................................................................................... 46 8.1.4 Action ......................................................................................................................................... 46 8.1.5 Serious adverse events ............................................................................................................. 46 8.1.6 Suspected unexpected serious adverse reactions .................................................................... 47 8.1.7 Reporting of serious adverse events ......................................................................................... 47 8.1.8 Follow-up of adverse events ...................................................................................................... 47 8.2 Section 10 WMO event ...................................................................................................................... 47 8.3 Annual safety report or development safety update report ............................................................... 47 9 STATISTICAL METHODOLOGY AND ANALYSES ................................................................................ 49 9.1 Statistical analysis plan...................................................................................................................... 49 9.2 Protocol violations/deviations ............................................................................................................ 49 9.3 Power calculation ............................................................................................................................... 49 9.4 Missing, unused and spurious data ................................................................................................... 49 9.5 Analysis sets ...................................................................................................................................... 49 9.5.1 Safety set ................................................................................................................................... 49 9.5.2 Efficacy analysis set .................................................................................................................. 49 9.6 Subject disposition ............................................................................................................................. 49 9.7 Baseline parameters and concomitant medications .......................................................................... 50 9.7.1 Demographics and baseline variables ....................................................................................... 50 9.7.2 Medical history ........................................................................................................................... 50 9.7.3 Concomitant Medications .......................................................................................................... 50 9.7.4 Treatment compliance/exposure ............................................................................................... 50 CHDR1514 Protocol Version 6 / 17-May-2016 Page 10 of 68 9.8 Safety and tolerability endpoints ........................................................................................................ 50 9.8.1 Adverse events .......................................................................................................................... 50 9.8.2 Vital signs................................................................................................................................... 50 9.8.3 ECG ........................................................................................................................................... 51 9.8.4 Clinical laboratory tests .............................................................................................................. 51 9.9 Efficacy endpoints .............................................................................................................................. 51 9.9.1 Efficacy ...................................................................................................................................... 51 9.9.2 Inferential methods .................................................................................................................... 51 9.10 Exploratory analyses and deviations ................................................................................................. 52 9.11 Interim analyses ................................................................................................................................. 52 10 GOOD CLINICAL PRACTICE, ETHICS AND ADMINISTRATIVE PROCEDURES ............................ 53 10.1 Good clinical practice ......................................................................................................................... 53 10.1.1 Ethics and good clinical practice ............................................................................................... 53 10.1.2 Ethics committee / institutional review board ............................................................................. 53 10.1.3 Informed consent ....................................................................................................................... 53 10.1.4 Insurance ................................................................................................................................... 53 10.2 Study funding ..................................................................................................................................... 53 10.3 Data handling and record keeping ..................................................................................................... 53 10.3.1 Data collection ........................................................................................................................... 53 10.3.2 Database management and quality control ............................................................................... 54 10.4 Access to source data and documents .............................................................................................. 54 10.5 Quality control and quality assurance ................................................................................................ 54 10.5.1 Monitoring .................................................................................................................................. 54 10.6 Protocol amendments ........................................................................................................................ 54 10.6.1 Non-substantial amendment ...................................................................................................... 54 10.6.2 Substantial amendment ............................................................................................................. 54 10.7 End of study report ............................................................................................................................ 55 10.8 Public disclosure and publication policy ............................................................................................ 55 11 STRUCTURED RISK ANALYSIS ......................................................................................................... 56 12 REFERENCES ...................................................................................................................................... 57 13 APPENDIX 1 – SUMMARY OF PRODUCT CHARACTERISTICS – ENGLISH VERSION ................. 59 14 APPENDIX 2 – SUMMARY OF PRODUCT CHARACTERISTICS – DUTCH VERSION .................... 60 15 APPENDIX 3 – EUROPEAN PUBLIC ASSESSMENT REPORT ........................................................ 61 16 APPENDIX 4 – SCREENING: QUESTIONNAIRE SPORT ACTIVITIES ............................................. 62 17 APPENDIX 5 – EXERCISE TEST ......................................................................................................... 64 18 APPENDIX 6 – COMPETITION ............................................................................................................ 65 19 APPENDIX 7 – URINE DOPING CONTROL PROCEDURE ................................................................ 67 CHDR1514 Protocol Version 6 / 17-May-2016 Page 11 of 68 CHDR1514 Protocol Version 6 / 17-May-2016 Page 12 of 68 LIST OF ABBREVIATIONS AE Adverse Event ALT Alanine aminotransferase/Serum glutamic pyruvic transaminase (SGPT) ANCOVA Analysis of Covariance ANOVA Analysis of Variance ATC Anatomic Therapeutic Chemical ABP Athletes Biological Passport BMI Body Mass Index BP Blood Pressure bpm Beats per minute CCMO Central Committee on Research Involving Human Subjects; in Dutch: Centrale Commissie Mensgebonden Onderzoek CHDR Centre for Human Drug Research CIRC Cycling Independent Reform Commission CK Creatine Kinase CRF Case Report Form EC Ethics Committee (also Medical Research Ethics Committee (MREC); in Dutch: Medisch Ethische Toetsing Commissie (METC). ECG Electrocardiogram EDTA Ethylene diamine tetra-acetic acid FDA Food and Drug Administration GCP Good Clinical Practice Hb Haemoglobin Ht Haematocrit ICH International Conference on Harmonization IRB Institutional Review Board LDH Lactate dehydrogenase MedDRA Medical Dictionary for Regulatory Activities rHuEPO Recombinant Human Erythropoietin SAE Serious Adverse Event SAP Statistical Analysis Plan SD Standard Deviation SEM Standard Error of the Mean SOC System Organ Class SOP Standard Operating Procedure SST Serum Separator Tube SmPC Summary of Product Characteristics SUSAR Suspected Unexpected Serious Adverse Reaction TTE Time To Exhaustion UCI Union Cycliste Internationale WADA World Anti-Doping Agency WHO World Health Organization WMO Medical Research Involving Human Subjects Act; in Dutch: Wet Medisch- wetenschappelijk Onderzoek met Mensen. CHDR1514 Protocol Version 6 / 17-May-2016 Page 13 of 68 PROTOCOL SYNOPSIS – CHDR1514 1. Title A randomized, double-blind, placebo-controlled study to investigate the effects of recombinant human erythropoietin (NeoRecormon) on cycling performance and the occurrence of adverse events in well-trained cyclists. 2. Short Title Recombinant human erythropoietin effects on cycling performance. 3. Background & Rationale 3.1 Need for research on doping A recent report of the Union Cycliste Internationale gives an in-depth analysis of doping throughout cycling’s history, from 1890 to the present day. The report’s final conclusion is that cycling has had, and continues to have, a serious doping problem.[1] Although it could be argued that administering substances that improve performance is forbidden and nothing more needs to be known about it, these substances are apparently being used and therefore research to investigate the effects and safety of doping substances in this population is necessary. There are a number of reasons for this. First, it is often unknown if a forbidden substance really improves performance. If this is not the case the need for administration is strongly diminished. Additionally the adverse effects of such substances are often insufficiently known and athletes may be exposed to risks without being adequately informed about them. 3.2 Erythropoietin substances in patients and athletes Recombinant Human Erythropoietin (rHuEPO) is used to treat patients with anemia resulting from chronic kidney disease or chemotherapy.[2] The correction of the anemia results in an increase in exercise capacity in these patients.[2] The treatment immediately attracted the attention of athletes because they assumed that rHuEPO would also improve their exercise performance. Due to this presumption, the use of rHuEPO in athletes became very common. In 1990, the use of rHuEPO was placed on the list of prohibited substances published by the World Anti-Doping Agency (WADA).[3] At the time of the first ban there was no published evidence that rHuEPO would actually improve sports performance. 3.3 Examination of the evidence for the ergogenic properties of rHuEPO in cyclists The evidence for the effect of rHuEPO in well trained athletes is in fact sparse until today. A qualitative systematic review of the available literature was performed in 2012 to examine the evidence for the performance enhancing properties of rHuEPO in cyclists.[4] The review demonstrated that the characteristics of the study populations differed from the population suspected of rHuEPO abuse. Studies did not use well-trained cyclists, still less elite or world-class cyclists.[5] Most studies used a small number of untrained subjects and the quality of the research was often questionable. In these studies, the main studied effect was the maximal oxygen carrying capacity of blood (VO2, max) which only has a remote connection to performance in endurance sports, especially in well-trained athletes. This is in line with the knowledge that multiple factors affect performance, in which oxygen carrying capacity of the blood becomes less relevant when other factors become rate-limiting. Endurance performance may be better correlated with submaximal exercise factors. In addition, there is virtually no research on the potential adverse effects of this form of doping. To conclude, the results of this literature search showed that there is no scientific basis from which to conclude that rHuEPO has performance-enhancing properties in well-trained cyclists and that knowledge about potential side effects is lacking in full. 3.4 Aim In the current study the effect of NeoRecormon, a rHuEPO, will be investigated in a population with CHDR1514 Protocol Version 6 / 17-May-2016 Page 14 of 68 cycling performance abilities as close as possible to those of professional cyclists and in conditions closely resembling racing conditions and the required performance duration. 4. Objective(s) 4.1 Primary Objective - To explore the effects of NeoRecormon on well-trained cyclists and their cycling performance by maximal and sub-maximal exercise parameters measured during exercise tests. 4.2 Secondary Objectives - To explore the effects of NeoRecormon on well-trained cyclists and their cycling performance by the performance and outcome of a race. - To evaluate the safety of NeoRecormon in well-trained cyclists. - To evaluate the performance of doping detection methods for NeoRecormon use in well-trained cyclists. 4.3 Exploratory objectives - To explore how a standardized submaximal exercise affects gene expression patterns in well- trained individuals. - To explore the difference in RNA-profiles between individuals treated with rHuEPO and placebo - To identify potential transcripts that can be used as biomarkers for rHuEPO use - To explore correlations between changes in whole blood gene expression patterns observed before and after a submaximal exercise test in individuals and their performance 5. Design Randomized, double-blind, placebo-controlled study to investigate the effects and safety of NeoRecormon in well-trained cyclists. 6. Principal Investigator & Trial Site Principal Investigator: Prof. dr. A.F. (Adam) Cohen, MD Co-Investigators: Prof. dr. J. (Koos) Burggraaf, MD Dr. J.I. (Joris) Rotmans Dr. J.M.A. (Hans) Daniels J.A.A.C. (Jules) Heuberger, MSc Mr. H. (Herman) Ram Dr. O. (Olivier) de Hon T. (Thijs) Zonneveld Trial Site: Centre for Human Drug Research, Zernikedreef 8, 2333 CL Leiden, The Netherlands 7. Subjects / Groups A total of 48 subjects are planned to be enrolled in this study. Eligible subjects will be randomized to the NeoRecormon or Placebo treatment groups on a 1:1 basis. 8. Sample Size Justification 8.1 Power analysis based on VO2,max Based on the increase in VO2,max after administration of NeoRecormon in a previous study with moderately trained subjects an exploratory power analysis has been performed.[6] A sample size of 6 in each group will have a power of 80% to detect a difference in means of 3.8 ml/min/kg, assuming that the common standard deviation is 1.95, using a two-tailed t-test with a 0.05 two sided significance level. A review of the available literature showed that, after initial years of training, well-trained athletes maintain a plateau in their VO2,max, but continue to improve their performance.[4] This indicates that the difference between effects on VO2,max between the NeoRecormon and placebo group in well- CHDR1514 Protocol Version 6 / 17-May-2016 Page 15 of 68 trained cyclists might be smaller. The smaller the size of the difference, the larger the sample size must be to detect a significant difference. To detect a difference of 1.5ml/min/kg with a power of 80% a sample size of 22 is needed, assuming that the common standard deviation is 1.95, using a two-tailed t-test with a 0.05 two sided significance level. When taking into account a ±10% attrition rate, 24 subjects are needed in both the NeoRecormon and placebo group. 8.2 Power analysis based on Pmax/kg The power calculation in section 8.1 is based on an endpoint that only has a remote connection to performance in well-trained cyclists. A better endpoint would be power output per kilogram (P/kg) at a submaximal level, such as 80% of VO2,max. Unfortunately no studies have been performed using this endpoint, so the effect of rHuEPO on P/kg at 80% VO2,max is still unknown. The mean P/kg at 80% VO2,max of 11 male professional cyclists however, is 5.2 W/kg with a standard deviation of 0.199.[7] Using a sample size of 22 (including 10% attrition rate) and a two-tailed t-test with a 0.05 two sided significance level a difference of 0.172 W/kg can be detected with a power of 80%. This difference would mean that a professional cyclist weighing 75 kg would go from an average of 390 W at 80% VO2,max to 402.9 W. On a racing bike weighing 9 kg sitting in racing position at 25 degrees Celsius, this would produce a speed of 43.80 km/h and 44.32 km/h respectively (calculated from http://bikecalculator.com). In a flat terrain of 40 km this would result in a finish time of 54 min 48 sec and 54 min 09 sec, a difference of 39 seconds, which is very relevant in a race like to Tour the France. 9. Inclusion criteria  Well-trained (as determined by cycling history and maximal power output >4 W/kg) male subjects, 18 to 50 years old (inclusive);  Subjects must be healthy / medically stable on the basis of clinical laboratory tests, medical history, vital signs, and 12-lead ECG performed at screening, including exercise ECG.  Each subject must sign an informed consent form prior to the study. This means the subject understands the purpose of and procedures required for the study. 10. Exclusion criteria  Any clinically significant abnormality, as determined by medical history taking and physical examinations, obtained during the screening visit that in the opinion of the investigator would interfere with the study objectives or compromise subject safety.  Unacceptable known concomitant diagnoses or diseases at baseline, e.g., known cardiovascular, pulmonary, muscle, metabolic or haematological disease, renal or liver dysfunction, ECG or laboratory abnormalities, etc.  Unacceptable concomitant medications at baseline, e.g., drugs known or likely to interact with the study drugs or study assessments.  Unacceptable potential cycling performance enhancing medications at baseline, e.g. Erythropoiesis-stimulating agents, Anabolic Androgenic Steroids, Growth Hormone, Insulin, IGF-I and Beta-Adrenergic Agents or methods, e.g. altitude tents.  Blood transfusion in the past three months.  Loss or donation of blood over 500 mL within three months.  Participation in a clinical trial within 90 days of screening or more than 4 times in the previous year.  Known hypersensitivity to the treatment or drugs of the same class, or any of their excipients.  Any known factor, condition, or disease that might interfere with treatment compliance, study conduct or interpretation of the results such as drug or alcohol dependence or psychiatric disease.  Positive urine drug test at screening. CHDR1514 Protocol Version 6 / 17-May-2016 Page 16 of 68  Positive alcohol breath test at screening.  Haemoglobin (Hb) concentration > 9.8 mmol/l at screening.  Hb concentration < 8 mmol/l at screening.  Haematocrit (Ht) ≥ 48% at screening.  Being subject to WADA’s anti-doping rules, meaning being a member of an official cycling union or other sports union for competition (such as the KNWU) or participating in official competition during the study.  Positive results from serology at screening (except for vaccinated subjects or subjects with past but resolved hepatitis)  Previous history of fainting, collapse, syncope, orthostatic hypotension, or vasovagal reactions.  Any circumstances or conditions, which, in the opinion of the investigator, may affect full participation in the study or compliance with the protocol. 11. Concomitant medications The clinical results obtained so far do not indicate any interaction of NeoRecormon with other medicinal products. Mandatory supplementation:  50mg vitamin C (ascorbic acid) per day  200 mg iron (ferrofumerate) per day Allowed:  Paracetamol  Other medications that are discussed, approved and clearly documented by the investigator. Prohibited:  All substances (except NeoRecormon during treatment period) that are on the doping list and enhance cycling performance are prohibited within 6 months prior to study drug administration and during the course of the study (e.g. Other Erythropoiesis-stimulating agents, Anabolic Androgenic Steroids, Growth Hormone, Insulin, IGF-I and Beta-Adrenergic Agents). 12. Study periods The total study period will be 17 weeks. Study periods Occasion Weeks Screening + training 1x Within 6 weeks prior to treatment period Ramp exercise test*: 1x Within 2 weeks before the start of the treatment period, but after the screening Time To Exhaustion exercise test*: 1x Treatment Once a week During 8 weeks Ramp exercise test Every two weeks Training Whole period TTE exercise test: 1x During the 7th treatment week Competition 1x During the 9th week Follow-up 1x 30 Days after last visit CHDR1514 Protocol Version 6 / 17-May-2016 Page 17 of 68 *Exercise tests will be performed within 2 weeks prior to the treatment period for baseline measurements. 13. Investigational drug rHuEPO: NeoRecormon (Active substance: Epoëtine beta) Dosage: See Figure 1 for a detailed description of the dosage adjustment schedule. Administration: Subcutaneously Dosage Rationale: NeoRecormon 2000, 5000 or ≥6000 but ≤10.000 IU/week, to be able to reach the target range and adjust the dosage as soon as possible if it seems necessary from Hb or Ht results 14. Placebo Placebo: subcutanous injection of saline, (0.90% w/v NaCl) The investigational drug and its matching placebo are indistinguishable and will be packaged in the same way. Blinding will be accomplished by using the same syringes or by covering the syringes with aluminum foil. 15. Efficacy endpoints Efficacy will be assessed at the time points indicated in the Visit and Assessment Schedule (see Table 1) in four ways:  Exercise tests will measure maximal and submaximal exercise parameters (e.g. VO2,max, Pmax, VE, Vt, VO2, VCO2, see CHDR1514 Protocol Version 6 / 17-May-2016 Page 18 of 68  Table 2.)  Subjects will participate in a competition designed in such a way that it closely resembles real racing conditions. Before and during the race, physiological parameters will be measured (e.g. power, heart rate, blood pressure).  Blood will be collected at predetermined stages at each clinical visit before administration of NeoRecormon/placebo and before and during the exercise tests and competition. This blood will be used for the following Clinical Laboratory Assessments: o Haematology (e.g. markers for the Athlete’s Biological Passport) o Chemistry (e.g. creatinine phosphokinase and c-reactive protein levels) o Coagulation (e.g. D-Dimer, F1+2, e-Selectin, p-Selectin)  Laser Speckle contrast imaging for blood flow measurements. 16. Safety endpoints Safety will be assessed by: 1. Physical examination 2. Monitoring vital signs o Pulse Rate (bpm) o Systolic blood pressure (mmHg) o Diastolic blood pressure (mmHg) o Temperature measurements (ºC) 3. Electrocardiogram (ECG) o Heart Rate (HR) (bpm), PR, QRS, QT, QTcB 4. Clinical Laboratory Assessments o Haematology Ht must be <52%. If Ht level is ≥52%, therapy should be interrupted until the Ht percentage begins to fall. Hb must be below a certain level (see Figure 1). If Hb exceeds that level, therapy should be interrupted until the Hb concentration falls back into the range. o Chemistry o Urinalysis o Coagulation 5. Collection of treatment-emergent (serious) adverse events ((S)AEs) 1-4 will be assessed during screening, 2-5 will be assessed at each clinical visit and before the competition. 2 and 3 will be assessed before and during the exercise tests. (Serious) Adverse Events ((S)AEs) will be collected throughout the study. 17. Blinding This study will be performed in a double-blind fashion. The investigator, subjects and all study staff will remain blinded. A non-blinded CHDR staff member (not part of study staff) will receive report of Hb and Ht before dosing and will follow the instruction as described in Figure 1. This individual will be responsible for the dosage adjustments. Procedure for dosage adjustment: When the Hb concentration and/or Ht exceeds a certain value or Hb stays below a certain value (see Figure 1) the dose adjustment officer will issue a request for a dosage change for the subject that requires the change. This request will be for the subject that requires the change in treatment but will also be issued to a random placebo subject to preserve the blinding of the study. 18. NeoRecormon detection in urine Urine will be collected at two predetermined periods. In the second treatment week samples will be taken pre-dose (day 7/8/9), two days later (day 9/10/11), at day 11/12/13/14 before and after the exercise test and pre-dose at day 14/15/16. Additionally, one sample will be taken before and after CHDR1514 Protocol Version 6 / 17-May-2016 Page 19 of 68 the competition (week 9). These samples will be sent to a lab specialized in rHuEPO (NeoRecormon) detection in urine, according to the current protocol of the Dutch Doping Authority. 19. Statistical methodology All efficacy endpoints will be summarized (mean and standard deviation of the mean, median, minimum and maximum values) by treatment and time, and will also be presented graphically as mean over time, with standard deviation as error bars. Change from baseline results will be utilized in all data summaries. All categorical efficacy endpoints will be summarized by frequencies. The PD endpoints will be analyzed separately by mixed model analyses of variance with treatment, time and treatment by time as fixed effects, with subject and subject by time as random effect, and with the (average) baseline value as covariate for recurring measurements and a one-way ANCOVA with treatment and baseline as covariates for measurements done at baseline and the end of the treatment period. CHDR1514 Protocol Version 6 / 17-May-2016 Page 20 of 68 Figure 1. This dosing schedule must be applied before every administration of NeoRecormon/placebo during the 8 week treatment period. Ht = haematocrit, Hb = haemoglobin CHDR1514 Protocol Version 6 / 17-May-2016 Page 21 of 68 Table 1. Visit and Assessment Schedule. SCR + training Baseline measurements Treatment and exercise test weeks* Competition week FU 1, 3, 5, 7 2, 4, 6, 8 9 12 Day -42 to -3 Day -14 to -1 Day 0/1/2, 14/15/16, 28/29/30, 42/43/44 Ramp exercise tests: Day 11/12/13/14, 25/26/27/28, 39/40/41/42, 53/54/55/56 Time To Exhaustion exercise test: 46/47/48/49 Day 7/8/9, 21/22/23, 35/36/37, 49/50/51 Weekend after last exercise test 80 +- 7 Time point Assessment -6 to - 1h 0h Before ET During ET After ET -6 to -1h 0h Before competition During competition After competition Informed consent X Demographics X Inclusion & exclusion criteria X Height and Weight 10 X X X Medical history X Physical examination X X Concomitant medication X Serology X Blood sample chemistry X6 X X9 X9 X X X6 Blood sample haematology X X X X X Blood sample coagulation X3 X4 X X8 X9 X9 X X X3 Blood sample biomarker X5 X11 X11 Blood sample RNA X12 X12 Urinalysis X X X X X Urine Drug Screen, Breath Alcohol Test X ECG X X X X General symptoms X X X X X Vital Signs (HR, BP, Temp) X X X X Blood sample haemoglobin and haematocrit "online" X X Drug (-placebo) administration X X CHDR1514 Protocol Version 6 / 17-May-2016 Page 22 of 68 Ramp exercise test X1 X X TTE exercise test X X Laser Speckle (LSCI) X7 X7 X7 Blood Sample Lactate X Urine doping control X2 X2 X2 X2 X X (S)AE/Con-meds <----- continuous -----> Training <----- Whole week -----> SCR = Screening, ET = Exercise Test, ECG = Electrocardiogram, BP = Blood Pressure, HR = Heart Rate, Temp = temperature, PD = Pharmacodynamics, (S)AE = (Special) Adverse Event, FU = Follow Up, * / means ‘or’ 1 Exercise test training session. 2 Urine samples for doping control will only be collected in week 2 on day 7/8/9, at 7/8/9 +2 days, 11/12/13/14 and 14/15/16 (meaning post third dose and at day 2, 4 and 7 after that dose) 3 Short coagulation panel at screening and follow-up, full panel for all other samples 4 Before and after ramp exercise test 5 Only on day 0/1/2 and 42/43/44 6 Full chemistry panel at screening and follow-up, short panel for all other samples 7 Only at day pre-first-dose and after exercise test on 39/40/41/42 and before exercise test on 46/47/48/49 8 Only at day 0/1/2 post-dose, 1-3h 9 Not at Time to Exhaustion test, except for Blood sample chemistry 10 Height measured at screening only 11 Before and after exercise test on day 39/40/41/42 12 Before and after Time to exhaustion test on day 46/47/48/49 CHDR1514 Protocol Version 6 / 17-May-2016 Page 23 of 68 Table 2. Exercise parameters in endurance performance. Abbreviation Description Calculations / Characteristics VO2, max Maximal oxygen consumption (ml kg-1 min-1) Pmax Maximal power output (W) 1W = 1J/sec VE Respiratory minute ventilation (L/min) Vt Tidal volume (L) Rf Respiratory frequency VO2 Oxygen consumption (L/min) VCO2 Carbon dioxide production (L/min) eqVO2 Ventilatory equivalent for oxygen VE/VO2 eqVCO2 Ventilatory equivalent for carbon dioxide VE/VCO2 EE Energy Expenditure (J/sec) ([3.869 xVO2]+[1.195 x VCO2]) x (4.186/60) x 1000 GE Gross Efficiency (%) Reflects the percentage of total chemical energy expended that contributes to external work, with the remaining energy lost as heat. (Work rate (J/sec) / Energy Expenditure (J/sec)) x 100% DE Delta Efficiency (%) (∆ Work Production (J/sec)/ ∆ Energy Expenditure (J/sec)) x 100% CE Cycling economy (W L-1 min-1) Can be defined as the submaximal VO2 per unit of body weight required to perform a given task. P/VO2 Fatmax The highest absolute fat oxidation LT1 = AT = OBLA Lactate Threshold 1 = Anaerobic Threshold = Onset of Blood Lactate Accumulation The threshold where blood lactate levels raise 1mmol/L above baseline values. LT2 Lactate Threshold 2 The highest exercise intensity where lactate levels remain stable. VT1 Ventilatory Threshold 1 The first increase in VE that is proportional to the increase in VCO2 with no increase in VO2. Increase of end tidal O2 pressure. VT2 = RCP Ventilatory Threshold 2 = Respiratory Compensation Point Represents itself at a high work. VE increases far more compared to VCO2. End tidal CO2 pressure decreases. CHDR1514 Protocol Version 6 / 17-May-2016 Page 24 of 68 1 BACKGROUND AND RATIONALE 1.1 Context 1.1.1 Need for research on doping A recent report of the Union Cycliste Internationale (UCI) gives an in-depth analysis of doping throughout cycling’s history, from 1890 to the present day. The report’s final conclusion is that cycling has had, and continues to have, a serious doping problem.[1] Although it could be argued that administering substances that improve performance is forbidden and nothing more needs to be known about it, research to investigate the effects and safety of doping substances in this population is necessary. There are number of reasons for this. First, it is often unknown if a forbidden substance really enhances performance. If this is not the case the need for administration is strongly diminished. Additionally the adverse effects of such substances are often insufficiently known and athletes may be exposed to risks without being adequately informed about them. 1.1.2 Erythropoietin substances in patients and athletes Recombinant Human Erythropoietin (rHuEPO) is used to treat patients with anemia resulting from chronic kidney disease.[2] The correction of the anemia results in an increase in exercise capacity in these patients.[2] The treatment immediately attracted the attention of athletes because they assumed that rHuEPO would also improve their exercise performance. Due to this presumption, the use of rHuEPO in athletes became very common. In 1990, the use of rHuEPO was placed on the list of prohibited substances published by the World Anti-Doping Agency (WADA).[3] At the time of the first ban there was no published evidence that rHuEPO would actually improve sports performance. 1.1.3 Examination of the evidence for the ergogenic properties of rHuEPO in cyclists The evidence for the effect of rHuEPO in well trained athletes is in fact sparse until today. A qualitative systematic review of the available literature was performed in 2012 to examine the evidence for the performance enhancing properties of rHuEPO in cyclists.[4] The review demonstrated that the characteristics of the study populations differed from the population suspected of rHuEPO abuse. rHuEPO studies often used untrained or moderately trained cyclists.[5] It cannot be assumed that effects found in these studies automatically apply to well-trained cyclists. Most studies used a small number of untrained subjects and the quality of the research was often questionable. In these studies, the main studied effect was the maximal oxygen carrying capacity of blood (VO2, max) which only has a remote connection to performance in endurance sports, especially in well-trained athletes. This is in line with the knowledge that multiple factors affect performance, in which oxygen carrying capacity of the blood becomes less relevant when other factors become rate-limiting. Endurance performance may be better correlated with submaximal exercise factors. There a number of findings that support this last notion. Firstly, research into training for endurance performance shows that moderately trained athletes are able to improve VO2,max by interval and/or intensive training, whereas these training regimens do not improve VO2,max in well-trained athletes. After initial years of training, these well-trained athletes maintain a plateau in their VO2,max, but continue to improve their performance. This shows that other factors have to play an important role in endurance performance. Secondly, long exercise times during consecutive days, with the finish line as a known end-point (contrary to the ‘ open end’ of time-to-exhaustion tests) makes it crucial for cyclists to distribute their power during a race. This, combined with (team) tactics, the terrain and the effects of drag force, means that cyclists work for only a small amount of time at their peak intensities (VO2, max), the rest of the time they will work at sub-maximal intensities where submaximal exercise factors are more important. CHDR1514 Protocol Version 6 / 17-May-2016 Page 25 of 68 Finally, adverse events were never studied despite the fact that there is much evidence from patient studies that rHuEPO may cause hypertension and thrombotic events. Uncontrolled use of rHuEPO therefore involves risks for the users’ health, irrespective of such a substance being used legally or illegally. 1.1.4 Study objectives The current study will - explore the effects of NeoRecormon on cycling performance in well-trained cyclists by  performance in exercise tests  performance in a competition  measuring markers from the haematological module of the Athlete Biological Passport  measuring blood flow - evaluate the safety of NeoRecormon in well-trained cyclists. - evaluate the performance of doping detection methods for NeoRecormon use in well-trained cyclists. 1.2 Non-clinical information 1.2.1 Non-clinical pharmacology Please refer to the respective Summary of Product Characteristics (SmPC) (see appendix 1) and European Public Assessment Report (EPAR) (see appendix 3). 1.2.2 Non-clinical pharmacokinetics and metabolism Please refer to the respective SmPC and EPAR (see appendix 1 and 3). 1.2.3 Non-clinical toxicology and safety pharmacology Please refer to the respective SmPC and EPAR (see appendix 1 and 3). 1.3 Clinical information 1.3.1 Clinical pharmacology Please refer to the respective SmPC and EPAR (see appendix 1 and 3). 1.3.2 Clinical pharmacokinetics and metabolism Please refer to the respective SmPC and EPAR (see appendix 1 and 3). 1.3.3 Clinical toxicology and safety pharmacology Please refer to the respective SmPC and EPAR (see appendix 1 and 3). 1.4 Study rationale 1.4.1 Benefit and risk assessment NeoRecormon is a registered drug. The safety profile of this compound is known. Because side effects might occur and anaphylactoid reactions were observed in isolated cases (≤1/10.000) (see SmPC, appendix 1), the study drug administrations will be done in the clinic under medical supervision. Subjects will be closely monitored and will only be discharged from the unit if their medical condition allows this. Subjects will receive a dose of 2000, 5000 or ≥6000IU with a maximum of 10.000IU/week of NeoRecormon once a week for 8 weeks. The dosage depends on haemoglobin (Hb) concentration CHDR1514 Protocol Version 6 / 17-May-2016 Page 26 of 68 and haematocrit (Ht) measured prior to administration. If Ht is ≥52% dosage will be interrupted. If Ht is <52% the dosage depends on the Hb concentration (see Figure 1). The effects of NeoRecormon used in patients in an autologous blood predonation programme most closely resembles the effects of NeoRecormon in healthy volunteers. For the use of NeoRecormon in an autologous blood predonation programme, the SmPC (see appendix 1) states that the maximum dose should not exceed 1200IU/kg (or 90.000 IU for a 75kg subject) per week for subcutaneous administration. Planned doses of 2000IU, 5000IU or ≥6000IU with a maximum of 10.000IU/week of NeoRecormon will be well below this maximum dose and are therefore considered safe. The risk is considered small and therefore acceptable compared to the scientific benefit. 1.4.2 Medical and regulatory background The genetically engineered erythropoietin hormone, rHuEPO (in this research NeoRecormon), works in the same way as the natural hormone erythropoietin (EPO). EPO is a (glycoprotein) hormone primarily produced by the kidneys.[8] The kidneys secrete EPO in response to hypoxia in the renal circulation.[8] The secreted EPO binds to the EPO receptor on the erythroid progenitor cell surface and in this way activates intracellular signaling pathways that lead to erythropoiesis in the bone marrow.[8] Erythropoiesis is the proliferation and differentiation of erythroid progenitor cells to erythrocytes (red blood cells).[8] Erythrocytes are responsible for oxygen transport through the blood.[4] Due to the lack of a nucleus and other cellular machinery, erythrocytes are not able to repair themselves.[4] They have a lifespan of approximately 120 days in the circulation after which they are degraded by the spleen (2-3 million every second).[4] To keep the oxygen carrying capacity of the blood at a steady level, constant erythropoiesis is necessary.[4] The concentration of EPO in blood is relatively constant at approximately 5 pmol L-1.[4] 1.4.3 Study population Forty-eight well-trained male subjects, 18 to 50 years of age who have a Hb level >8.0 mmol/L and ≤ 9.8 mmol/L, a Ht <48% and a maximal power output > 4.0 W/kg will be included. Previous studies showed that professional cyclists have Hb concentrations between 8.25 and 10.25mmol/l and Ht levels between 39.2 and 48.1%.[9], [10] The normal Hb concentration for males is 8.75 to 11.25 mmol/l.[11;12] The rHuEPO treatment in this study will increase Hb levels of the subjects with 10-15%. Subjects will be excluded during the screening when they have Hb levels >9.8mmol/l, because the normal Hb level (11.25mmol/l) should not be exceeded (9.8 x 1.15 = 11.27 mmol/l). 1.4.4 Study design This study will be conducted in accordance with guidelines from the International Conference on Harmonization (ICH) on current Good Clinical Practice (GCP) and the ethical principles that have their origins in the Declaration of Helsinki. In order to investigate the effects and safety of NeoRecormon in well-trained cyclists a randomized, double-blind, placebo-controlled study is most appropriate. The total study period will be 17 weeks. Within 6 weeks prior to the start of the treatment, subjects will undergo a medical screening. A ramp exercise test on a cycle ergometer will be performed directly after the screening to familiarize with the test and to determine the subject’s maximal power output. A second ramp and TTE exercise test will be performed for baseline measurements within 2 weeks prior to the first administration, but after the screening. During the 8 week treatment period subjects will be administered with NeoRecormon or placebo once a week. During the treatment period subjects will follow their usual training program and will perform a ramp exercise test (see appendix 5) every two weeks. In the 7th treatment week the subjects will also perform a TTE CHDR1514 Protocol Version 6 / 17-May-2016 Page 27 of 68 exercise test (see appendix 5). After the treatment period subjects will participate in a competition (see appendix 6). A follow-up visit will be scheduled, 30 days after the last dose. Please refer to the visit and assessment schedule for a more detailed description of the study design (see Table 1). 1.4.5 Investigational drug and placebo Investigational drug: NeoRecormon (Active substance: Epoëtine beta) Dosage: 2000IU, 5000IU or ≥6000IU with a maximum of 10.000IU/week, depends on Ht percentage and Hb concentration (see Figure 1) Administration: Subcutaneously Placebo: Saline, 0.90% w/v NaCl. The investigational drug and its matching placebo are indistinguishable and will be packaged in the same way. Blinding will be accomplished by using the same syringes or by covering the syringes with aluminum foil. Eligible subjects will be randomized to one of two treatment groups on a 1:1 basis. One group will be treated with placebo and one group with NeoRecormon. 1.4.6 Dosing, safety margin calculations, stopping criteria Dose selection and adjustments Please refer to Figure 1 for a detailed dosing schedule. The review of the data and the decision on the next dose level will be made by a non-blinded CHDR staff member unconnected to the study. Dosage Rationale An initial dose of 5000IU/week will be administered because previous studies showed that with this dose the above mentioned target range should be reached.[13-16] If the upper limit of the target range is exceeded the dose will be interrupted until the Hb concentration falls back into the range. At that point, therapy should be restarted at 2000IU/week. If treatment weeks ≥ 5 and Hb < (1.10*Initial Hb) the dose will be adjusted to ≥6000IU/week (maximal 10.000IU/week), so that it is still possible to reach the target range within 8 weeks. Safety margin calculations NeoRecormon is a marketed medication. The safety profile of this treatment is well-known (see SmPC (appendix 1) and EPAR (appendix 3)). See chapter 1.4.1 for the benefit and risk assessment. Subcutaneously administered NeoRecormon has a half-life of 12-28h. There will be one week between subcutaneous administrations of NeoRecormon, which is sufficient time to wash-out treatment from a previous occasion. CHDR1514 Protocol Version 6 / 17-May-2016 Page 28 of 68 Stopping criteria Dosing will be interrupted if - Ht ≥ 52% - Hb exceeds the upper limit of the target range (see Figure 1) - there’s an unacceptable tolerability profile based on the nature, frequency, and intensity of observed AEs These criteria will be maintained by the non-blinded CHDR staff member unconnected to the study. 1.4.7 Treatment duration In this study 24 subjects will be injected with NeoRecormon and 24 subjects will be injected with placebo once a week during 8 weeks. A treatment period of 8 weeks is chosen because this will provide sufficient time to reach the target range. If the target range is not reached after 5 weeks, the dose will be adjusted. In addition, during this treatment period study objectives such as assessment of safety and efficacy can be performed adequately. 1.4.8 Endpoints Efficacy endpoints Efficacy will be assessed at the time points indicated in the Visit and Assessment Schedule (Table 1). Efficacy will be assessed by: - performance in exercise tests - performance in a competition - measuring markers from the haematological module of the Athlete Biological Passport - measuring blood flow Exercise tests A review showed that most of the research investigating the effect of rHuEPO in cyclists focused on a parameter for maximal exercise oxygen consumption, VO2,max.[4] The review showed that besides this parameter, research should also focus on parameters for sub-maximal exercise. Therefore in this study maximal as well as sub-maximal exercise parameters (see table 2) will be measured during exercise tests. Subjects will perform a ramp exercise test every two weeks during their treatment/training period and a time to exhaustion exercise test in the 7th treatment week to see if the NeoRecormon/placebo treatment has an effect on their performance. Baseline measurements for these exercise tests will be performed one week prior to the treatment period. Before and during the exercise tests blood will be collected at predetermined stages to determine if certain protein concentrations are influenced by NeoRecormon. Please refer to appendix 5 for more detailed descriptions of the exercise tests. Competition Multiple factors affect cycling performance, especially in the racing conditions seen in official competition. Therefore, in this study the effects of NeoRecormon will also be determined by performance in a competition. The competition will be designed in such a way that it closely resembles real racing conditions. During the competition certain maximal and sub-maximal exercise parameters will be measured. Before and during the competition blood will be collected at predetermined stages to determine if certain protein concentrations are influenced by CHDR1514 Protocol Version 6 / 17-May-2016 Page 29 of 68 NeoRecormon. After the competition urine will be collected for a doping control and for urinalysis for a.o. proteinuria. Please refer to appendix 6 for a more detailed description of the competition. Athlete Biological Passport According to the WADA, the athlete biological passport (ABP) is introduced to establish whether an athlete is manipulating his physiological variables without detecting a particular substance or method. The objective of this testing is to identify athletes in a haematological module and a steroidal module. The haematological module tests for certain markers in the body that identify the enhancement of oxygen transport. The steroidal module collects information on markers for steroid doping and aims to identify endogenous anabolic androgenic steroids. In this study only the haematological module will be conducted because rHuEPO only influences markers of this module. In this study an ABP will be created to determine the effect of NeoRecormon on markers of the haematological module and in addition to investigate if these markers together are really able to prove that an athlete is manipulating his physiological variables. The ABP will also be administered for safety, because based on Hb concentration and Ht (two of the haematological module markers) a dose will be selected weekly (during the treatment period). Blood to measure the haematological module markers will be collected before NeoRecormon/Placebo administration and before the competition. Blood flow LSCI measures:  basal blood flow  blood flow upon occlusion-reperfusion  Blood flow after ‘exercise’ 1.4.9 Statistical hypotheses and sample size 1.4.9.1 Power calculation based on VO2,max Null hypothesis: There’s no difference between the effects of NeoRecormon and placebo on cycling performance parameters. Based on the results of a previous study a power analysis has been performed.[6] In this study sixteen endurance-trained men (cyclists, runners and triathletes) were assigned randomly to either the rHuEPO- (n=9) or placebo- (n=7) treated groups. Both the participants and the investigators engaged in exercise testing and blood sampling were blind with respect to the group assignment. All athletes had been in regular training for several years preceding the study and continued to train throughout the period of the study but were forbidden to participate in sport events. The load of training was not different between the rHuEPO and placebo groups throughout the 4-weeks treatment period. Moderate doses of rHuEPO or placebo were injected s.c. 3 times a week for 4 weeks in the morning. The two groups also received a daily oral dose of 200mg of iron sulphate during the 4 weeks. VO2,max was evaluated before and after rHuEPO treatment by using a ramp exercise test (30W/min after a 3-min baseline of warm-up at 60W) performed to the limit of tolerance on an electrically braked cycle ergometer that controlled external power output independent of pedal cadence. The mean VO2,max before treatment was similar in rHuEPO and placebo groups (63.0±1.5 vs 64.8±2.0 ml/min/kg). VO2,max increased significantly after rHuEPO treatment in the rHuEPO group (68.4±1.9 ml/min/kg) and was significantly higher than in the placebo group (64.6±2.0 ml/min/kg, P<0.05). CHDR1514 Protocol Version 6 / 17-May-2016 Page 30 of 68 Equation to determine sample size: [17] n = 1+ 2C(s/d)2 n = sample size α = significance level 1-β = the desired power of the experiment to detect the postulated effect C = constant dependent on the value of α and β selected (see table 4) s = standard deviation of the variable d = magnitude of the difference the investigator wishes to detect Table 4. Constant C dependent on the value of α and β selected.[17] α 0.05 0.01 1-β 0.8 7.85 11.68 0.9 10.51 14.88 Factors from the previous study to calculate sample size: [6] d = 68.4 (rHuEPO-group) - 64.6 (Placebo group) = 3.8 ml/min/kg s = 1.95 1 - β = 80% α = 5% A sample size of 6 in each group will have a power of 80% to detect a difference in means of 3.8ml/min/kg, assuming that the common standard deviations is 1.95, using a two-tailed t-test with a 0.05 two sided significance level. A review of the available literature of the research investigating the effect of rHuEPO in cyclists showed that, after initial years of training, well-trained athletes maintain a plateau in their VO2,max, but continue to improve their performance further.[4] This indicates that the difference between effects on VO2,max between the NeoRecormon and placebo group in well-trained cyclists will be smaller. The smaller the size of the difference, the larger the sample size must be to detect a significant difference. To detect a difference of 1.5ml/min/kg with a power of 80% a sample size of 22 is needed, assuming that the common standard deviation is 1.95, using a two-tailed t-test with a 0.05 two sided significance level. When taking into account a ±10% attrition rate, 24 subjects are needed in both the NeoRecormon and placebo group. 1.4.9.2 Power calculation based on Pmax/kg A better endpoint would be power output per kilogram (P/kg) at a submaximal level, such as 80% of VO2,max. Unfortunately no studies have been performed using this endpoint, so the effect of rHuEPO on P/kg at 80% VO2,max is still unknown. The mean P/kg at 80% VO2,max of 11 male professional cyclists however, is 5.2 W/kg with a standard deviation of 0.199.[7] Using a sample size of 22 (including 10% attrition rate) and a two-tailed t-test with a 0.05 two sided significance level a difference of 0.172 W/kg can be detected with a power of 80%. This difference would mean that a professional cyclist weighing 75 kg would go from an average of 390 W at 80% VO2,max to 402.9 W. On a racing bike weighing 9 kg sitting in racing position at 25 degrees Celsius, this would produce a speed of 43.80 km/h and 44.32 km/h respectively (calculated from http://bikecalculator.com). In a flat terrain of 40 km this would result in a finish time of 54 min 48 sec and 54 min 09 sec, a difference of 39 seconds, which is very relevant in a race like to Tour the France. CHDR1514 Protocol Version 6 / 17-May-2016 Page 31 of 68 2 STUDY OBJECTIVES 2.1 Primary objective To explore the effects of NeoRecormon on cycling performance by  performance in exercise tests  measuring markers from the haematological module of the Athlete Biological Passport  measuring blood flow in well-trained cyclists. 2.2 Secondary objectives To explore the effects of NeoRecormon on cycling performance in a competition (road race) To evaluate the safety of NeoRecormon in well-trained cyclists To evaluate the performance of doping detection methods for NeoRecormon use in well-trained cyclists 2.3 Exploratory objectives To explore how a standardized submaximal exercise affects gene expression patterns in well-trained individuals. To explore the difference in RNA-profiles between individuals treated with rHuEPO and placebo To identify potential transcripts that can be used as biomarkers for rHuEPO use To explore correlations between changes in whole blood gene expression patterns observed before and after a submaximal exercise test in individuals and their performance CHDR1514 Protocol Version 6 / 17-May-2016 Page 32 of 68 3 STUDY DESIGN 3.1 Overall study design and plan This study will explore the effects of NeoRecormon on well-trained cyclists and their cycling performance during exercise tests and in a competition. It will consist of a screening, a training of the ramp exercise tests, 8 treatment visits, an 8 week training program, 7 exercise test visits, a competition and a follow-up visit which are outlined in the study Visit and Assessment Schedule (see Table 1). The total duration of the study for each subject will be up to 129 days divided as follows: Study periods** Days* Screening + Training ramp exercise test -42 to -2 days Ramp + TTE exercise test for baseline measurements -7 to -1 days Visit clinical unit: Treatment 0/1/2, 7/8/9, 14/15/16, 21/22/23, 28/29/30, 35/36/37, 42/43/44, 49/50/51 Visit clinical unit: Ramp exercise tests 11/12/13/14, 25/26/27/28, 39/40/41/42, 53/54/55/56 Visit clinical unit: TTE exercise test 46/47/48/49 Training 0 to 55 Competition 57 to 60 Follow-up 80 +- 7 * / means ‘or’ ** TTE = Time to exhaustion Dosage is spread over three days because it is logistically very difficult to perform safety measurements in 48 subjects and inject them afterwards with the study drug in one day. The exercise tests are spread over four days for the same reason. 3.1.1 Screening The screening phase (days -42 to -2) will only be started after full written, verbal and signed informed consent has been obtained, according to CHDR standard operating procedures. The entire screening process will last approximately 2 hours. The screening will be divided into two parts: Medical screening - Medical interview - Physical examination - 12-lead ECG - Vital signs (Heart rate, blood pressure) - Weight - Height - Urine Drug Screen (THC, morphine, benzodiazepines, cocaine, amphetamines, methamphetamines, MDMA) - Alcohol Breath Test - Blood sampling (haematology, biochemistry, serology, coagulation) - Urinalysis Training - Exercise test performed on an ergometer (see Appendix 5 – Exercise test for a more detailed description) During screening urine and blood samples will be collected from each patient for analysis as CHDR1514 Protocol Version 6 / 17-May-2016 Page 33 of 68 described in section 7.2.6. 3.1.2 Treatment and exercise tests The time schedule of study days is provided in general in the Visit and Assessment Schedule (see Table 1). On treatment days subjects will arrive at the clinical unit at a predetermined time. Before dosing standard safety measurements will be performed (see Table 1). Subjects will perform a ramp exercise test every two weeks during the treatment period. They also will perform a Time To Exhaustion exercise test in the 7th treatment week. Baseline measurements for these tests will be performed 2 weeks prior to the treatment period. During the exercise test maximal and submaximal exercise parameters will be measured. Before and during the exercise tests blood will be collected at predetermined stages. 3.1.3 Training period During the treatment period (8 weeks) subjects will maintain their usual training programme. They will record their training intensity during this period in a diary. Additionally, subjects bikes will be mounted with a Pioneer power meter that will be used for each training activity to log the trip. This information will be shared with the investigators. 3.1.4 Competition After the treatment period subjects will participate in a competition. The competition is designed in a way that it closely resembles real racing conditions and complies with the required performance duration. During the competition physiological parameters will be measured. Before and after the competition blood will be collected at predetermined stages. Please refer to appendix 6 for a detailed description of the competition. 3.1.5 Follow-up A follow-up visit will be performed 30 days after the last visit, which includes a final physical examination, safety laboratory tests (haematology, chemistry, coagulation and urinalysis) and the measurement of vital signs (see Visit and Assessment Schedule, Table 1). Subjects will also be asked if they believe they were on active or placebo treatment. A description of all procedures and analyses is included in section 7.2. 3.1.6 Urine doping control Urine will be collected at two predetermined periods. In the second treatment week samples will be taken pre-dose (day 7/8/9), two days later (day 9/10/11), at day 11/12/13/14 before and after the exercise test and pre-dose at day 14/15/16. Additionally, one sample will be taken before and after the competition (week 9). These samples will be sent to a lab (DoCoLab – Ugent, Technologiepark 30, Zwijnaarde) specialized in rHuEPO (NeoRecormon) detection in urine, according to the current protocol of the Dutch Doping Authority. Please refer to appendix 7 for a detailed description of the urine doping control procedure. CHDR1514 Protocol Version 6 / 17-May-2016 Page 34 of 68 4 STUDY POPULATION 4.1 Subject population A total of 48 subjects will be enrolled into the study following satisfactory completion of a screening visit where eligibility for the study will be checked. Subjects will be recruited via media advertisements and via advertisements at cycling associations. 4.2 Inclusion criteria Eligible subjects must meet all of the following inclusion criteria:  Well-trained (as determined by cycling history and maximal power output >4 W/kg) male subjects, 18 to 50 years old (inclusive);  Subjects must be healthy / medically stable on the basis of clinical laboratory tests, medical history, vital signs, and 12-lead ECG performed at screening, including exercise ECG.  Each subject must sign an informed consent form prior to the study. This means the subject understands the purpose of and procedures required for the study. 4.3 Exclusion criteria Eligible subjects must meet none of the following exclusion criteria:  Any clinically significant abnormality, as determined by medical history taking and physical examinations, obtained during the screening visit that in the opinion of the investigator would interfere with the study objectives or compromise subject safety.  Unacceptable known concomitant diagnoses or diseases at baseline, e.g., known cardiovascular, pulmonary, muscle, metabolic or haematological disease, renal or liver dysfunction, ECG or laboratory abnormalities, etc.  Unacceptable concomitant medications at baseline, e.g., drugs known or likely to interact with the study drugs or study assessments.  Unacceptable potential cycling performance enhancing medications at baseline, e.g. Erythropoiesis-stimulating agents, Anabolic Androgenic Steroids, Growth Hormone, Insulin, IGF-I and Beta-Adrenergic Agents.  Blood transfusion in the past three months.  Loss or donation of blood over 500 mL within three months.  Participation in a clinical trial within 90 days of screening or more than 4 times in the previous year.  Known hypersensitivity to the treatment or drugs of the same class, or any of their excipients.  Any known factor, condition, or disease that might interfere with treatment compliance, study conduct or interpretation of the results such as drug or alcohol dependence or psychiatric disease.  Positive urine drug test at screening.  Positive alcohol breath test at screening.  Haemoglobin (Hb) concentration > 9.8 mmol/l at screening.  Hb concentration < 8 mmol/l at screening.  Haematocrit (Ht) ≥ 48% at screening. CHDR1514 Protocol Version 6 / 17-May-2016 Page 35 of 68  Being subject to WADA’s anti-doping rules, meaning being a member of an official cycling union or other sports union for competition (such as the KNWU) or participating in official competition during the study.  Positive results from serology at screening (except for vaccinated subjects or subjects with past but resolved hepatitis)  Previous history of fainting, collapse, syncope, orthostatic hypotension, or vasovagal reactions.  Any circumstances or conditions, which, in the opinion of the investigator, may affect full participation in the study or compliance with the protocol. 4.4 Concomitant medications The clinical results obtained so far do not indicate any interaction of NeoRecormon with other medicinal products. All medications (prescription and over-the-counter [OTC]) taken after study screening until the end of the study will be recorded. 4.4.1 Mandatory concomitant supplementation Mandatory concomitant supplementation for all 48 subjects:  50mg vitamin C (ascorbic acid) per day  200 mg iron (ferrofumerate) per day 4.4.2 Allowed concomitant medications Allowed concomitant medications:  Paracetamol  Other medications are only allowed if they are discussed, approved and clearly documented by the investigator. 4.4.3 Prohibited concomitant medications Prohibited concomitant medications:  All substances (except NeoRecormon during treatment period) that enhance cycling performance are prohibited within 6 months prior to study drug administration and during the course of the study (e.g. Other Erythropoiesis-stimulating agents, Anabolic Androgenic Steroids, Growth Hormone, Insulin, IGF-I and Beta-Adrenergic Agents etc.). 4.5 Lifestyle restrictions In the interest of the subjects’ safety and to facilitate assessment of the treatment effect, the patients participating in this study will be requested to agree to the following restrictions during the study:  Alcohol will not be allowed from at least 24 hours before screening and before each scheduled visit, and whilst in the study unit until discharge from the study unit. CHDR1514 Protocol Version 6 / 17-May-2016 Page 36 of 68  During the study subjects are not allowed to be subject to WADA’s anti-doping rules, meaning they are for example not allowed to be a member of an official cycling union or other sports union for competition (such as the KNWU). NeoRecormon treated subjects are not allowed to become a member of such a union for 3 months after the last dose.  Subjects must maintain their usual training programme until the end of the treatment period (week 1-8). They will record their training activity in a diary and using the Pioneer power meter. Directly after the treatment period (week 9) the subjects will participate in a competition (see appendix 6). 4.6 Study drug discontinuation and withdrawal 4.6.1 Study drug interruption or discontinuation The investigator must temporally interrupt or permanently discontinue the study drug if continued administration of the study drug is believed to be contrary to the best interests of the subject. The interruption or premature discontinuation of study drug might be triggered by an Adverse Event (AE), a diagnostic or therapeutic procedure, an abnormal assessment (e.g., ECG or laboratory abnormalities), or for administrative reasons in particular withdrawal of the subject’s consent. The reason for study drug interruption or premature discontinuation must be documented. 4.6.2 Subject withdrawal Subjects have the right to withdraw from the study at any time for any reason. Should a subject decide to withdraw from the study, all efforts should be made to complete and report the observations, particularly the follow-up examinations, as thoroughly as possible. 4.6.3 Replacement policy Subjects withdrawing in the first treatment week for reasons other than adverse events or any other tolerability issues with the treatment will be replaced. CHDR1514 Protocol Version 6 / 17-May-2016 Page 37 of 68 5 INVESTIGATIONAL MEDICINAL PRODUCT 5.1 Investigational drug and matching placebo Study drug or placebo will be administered to the subjects as detailed in Table 1. NeoRecormon will be ordered as available and its matching placebo will be manufactured. The investigational drug and its matching placebo are indistinguishable and will be packaged in the same way. Blinding will be accomplished by using the same syringes or by covering the syringes with aluminum foil. 5.2 Comparative drug Placebo: Saline, 0.90% w/v NaCl. 5.3 Study drug dosing scheme Please refer to Figure 1 for a detailed dosing schedule. 5.4 Study drug packaging and labelling NeoRecormon and placebo will be acquired, packaged and labelled by the LUMC pharmacy in accordance with local regulations. Upon arrival at the pharmacy, the investigational products should be checked for damage and verify proper identity, quantity, integrity of seals and temperature conditions, and report any deviations or product complaints upon discovery. The dispensing of the study drug will be performed by the pharmacy. Study drug will be dispensed for each subject according to the randomization list. Study drug packaging will be overseen by the Leiden University Medical Centre Pharmacy and bearing a label with the identification required by local law, the protocol number, drug identification, and dosage. All subcutaneously administered drugs and placebos will be indistinguishable and will be packaged in the same way. The study drug must be stored between 2ºC and 8ºC in a secure, temperature-controlled (refrigerator) area with limited access. The vial must be kept in the outer carton, in order to protect it from light. For the purpose of ambulatory use, the unreconstituted product may be removed from the refrigerator and be stored at room temperature (not above 25ºC ) for one single period of up to 3 days. Leaving the reconstituted solution outside the refrigerator should be limited to the time necessary for preparing the injections. 5.5 Drug accountability Drug accountability will be maintained by the Leiden University Medical Centre Pharmacy and assessed by maintaining adequate study drug dispensing records. The investigator is responsible for ensuring that dosing is administered in compliance with the protocol. Delegation of this task must be clearly documented and approved by the investigator. All study drug administration will occur under medical supervision. 5.6 Treatment assignment and blinding 5.6.1 Treatment assignment Subjects must be randomized in a consecutive order starting with the lowest number. Replacement subjects will be numbered +100, e.g. subject 5 will have subject 105 as replacement, and subject 105 will have subject 205 as replacement, etc. The randomization code will be generated SAS version 9.1.3 (or a more recent version, if available) by a study-independent, CHDR statistician. To reduce the variability between active and placebo, CHDR1514 Protocol Version 6 / 17-May-2016 Page 38 of 68 block randomisation will be used, with one block of subjects aged 18-34 (inclusive) and another of subjects aged 34-50 (inclusive). The sample size of the blocks will be determined at day -14, when the included subjects have been identified. An attempt will be made to choose the highest block size. The randomization code will be unblinded/broken and made available for data analysis only after study closure, i.e., when the study has been completed, the protocol deviations determined, and the clinical database declared complete, accurate and locked. The randomization code will be kept strictly confidential. Sealed individual randomization codes, per subject and per treatment, will be placed in a sealed envelope containing the and labelled 'emergency decoding envelopes' will be kept in a safe cabinet at CHDR. 5.6.2 Blinding This study will be performed in a double-blind fashion. The investigator, study staff, subjects and monitor will remain blinded to the treatment until study closure. The investigational drug and its matching placebo are indistinguishable and will be packaged in the same way. With the exceptions described in this section, the randomization list will not be available to the investigator, study staff, subjects and monitors. The randomization list will be made available to the pharmacist preparing the study drug, to an individual responsible for dosage adjustments and to statisticians or programmers involved in preparing blinded summaries, graphs and listings to support the dose decisions. The summaries, graphs and listings provided by the statisticians or programmers will be produced in an area to which other team members do not have access. In addition a single non blinded CHDR staff member unconnected to the study will receive reports on the haemoglobin and haematocrit of the subjects. This individual will be responsible for the dosage adjustments. The investigator will receive a set of sealed emergency codes to be broken in case of emergency situations. If the identity of the study drug administered needs to be known in order to manage the subject’s condition i.e., in case of a medical emergency or in the case a SUSAR occurs, the treatment emergency code for that subject may be broken and the study drug identified. All such occurrences should be documented in the study file. Treatment emergency codes should not be broken except in emergency situations and, if possible, the investigator should be contacted before the emergency code is opened. Just prior to database lock the unused emergency code labels will be checked and a statement to the effect that all are intact (or not as the case may be) will be made on the database lock form. CHDR1514 Protocol Version 6 / 17-May-2016 Page 39 of 68 6 STUDY ENDPOINTS 6.1 Efficacy endpoints Efficacy will be assessed at the time points indicated in the Visit and Assessment Schedule (Table 1). Efficacy will be assessed by: - performance in exercise tests - performance in a competition - measuring markers from the haematological module of the Athlete Biological Passport - measuring blood flow Exercise tests All subjects will breathe during the exercise test through a facemask that will be connected to an oxymeter to collect inspired and expired gasses for analyzing: - Oxygen consumption, VO2 (L/min) - Carbon dioxide production, VCO2 (L/min) - Respiratory minute ventilation, VE (L/min) - Tidal volume, Vt (L) - Respiratory frequency, Rf - Maximal oxygen consumption, VO2,max (ml kg-1 min-1) During the exercise tests blood will be collected at predetermined stages to measure: - Lactate levels - Tissue plasminogen activator - Creatinine phosphokinase - C-reactive protein levels VO2 and VCO2 will be used to calculate: - Ventilatory equivalent for oxygen (VE/VO2), eqVO2 - Ventilatory equivalent for carbon dioxide (VE/VCO2), eqVCO2 these values will be used to determine: - Ventilatory threshold 1, VT1 - Ventilatory threshold 2, VT2 Physiological parameters that will be determined at VT1 and VT2: - Oxygen consumption, VO2 (L/min) - Oxygen consumption per kg, VO2 (L/min/kg) - Percentage of maximal oxygen consumption, %VO2max (L/min) - Power output, P (J/s) - Power output per kg, P (J/s/kg) Physiological parameters that will be determined at maximal effort: - Maximal oxygen consumption, VO2max (L/min) - Maximal oxygen consumption per kg, VO2max (L/min/kg) - Maximal power output, Pmax (J/s) - Maximal power output per kg, Pmax (J/s/kg) - Lactate values Other determinations: - Lactate threshold 1, LT1 - Lactate threshold 2, LT2 - Cycling economy, CE (W L-1 min-1) - Gross efficiency, GE (%) CHDR1514 Protocol Version 6 / 17-May-2016 Page 40 of 68 - Heart rate (bpm) - Systolic blood pressure (mmHg) - Diastolic blood pressure (mmHg) Competition Exercise parameters that will be measured and calculated during the competition: - Result in the race - Total time of the race - Power (W) - Heart rate (bpm) - Systolic blood pressure (mmHg) - Diastolic blood pressure (mmHg) Before and after the competition blood will be collected to determine if certain protein concentrations are influenced by NeoRecormon such as: - Lactate - Tissue plasminogen activator - Creatinine phosphokinase - C-reactive protein - Coagulation and activation markers (D-Dimer, F1+2, p-Selectin, e-Selectin, ThromboxaneB2, bTG, PF4, thrombomodulin) In urine: - Proteinuria Athlete Biological Passport The following Markers are considered within the ABP Haematological Module: - Haematocrit (Ht) - Haemoglobin (Hb) - Red blood cell (erythrocyte) count (RBC) - Red blood cell mass (RCM) - Reticulocytes percentage (RET%) - Reticulocyte count (RET#) - Mean corpuscular volume (MCV) - Mean corpuscular haemoglobin (MCH) - Mean corpuscular haemoglobin concentration (MCHC) - Red cell distribution width (standard deviation) (RDW-SD) - Immature reticulocyte fraction (IRF) Further calculated markers specific to the Haematological Module include OFF-hr Score (OFFS), which is a combination of Hb and RET%[18], and Abnormal Blood Profile Score (ABPS), which is a combination of Ht, Hb, RBC, RET%, MCV, MCH, and MCHC[19]. Blood flow LSCI measures:  basal blood flow  blood flow upon occlusion-reperfusion  Blood flow after ‘exercise’ 6.2 Safety endpoints Safety measurements before NeoRecormon/placebo administration Safety will be assessed by: CHDR1514 Protocol Version 6 / 17-May-2016 Page 41 of 68  Monitoring vital signs o Pulse Rate (bpm) o Systolic blood pressure (mmHg) o Diastolic blood pressure (mmHg) o Temperature measurements (ºC)  Electrocardiogram (ECG) (at rest) o Heart Rate (HR) (bpm), PR, QRS, QT, QTcB  Clinical Laboratory Assessments o Haematology Ht must be <52%. If Ht level is ≥52%, therapy should be interrupted until the Ht level begins to fall. Hb must be below 1.15*Initial Hb (see Figure 1). If Hb is above that level, dosage should be interrupted until the Hb concentration falls back into the range. Then the dosage should be restarted at 2000IU/week of the previous dose. o Chemistry o Urinalysis o Coagulation In addition a single non blinded CHDR staff member unconnected to the study will receive reports on the Hb and Ht of the subjects. This individual will be responsible for the dosage adjustments. Procedure for dosage adjustment: When the Hb and/or Ht exceeds a certain value (see 5.3) the dose adjustment officer will issue a request for a dosage change for the subject that requires the change. This request will be for the subject that requires the change in treatment but will also be issued to a random placebo subject to preserve the blinding of the study. Additional safety measures A single non-blinded CHDR staff member unconnected to the study will receive reports on Hb concentration and Ht of the subjects. This individual will be responsible for the dosage adjustments. If Ht exceeds 52% in a subject, dosage will be stopped and if necessary, ±0.5L blood will be collected from that subject to rearrange a normal Ht percentage. This subject will then be excluded from the study. (Serious) Adverse Events ((S)AEs) will be collected throughout the study. 6.3 Exploratory endpoints  RNA expression levels in venous blood before and after a submaximal exercise test CHDR1514 Protocol Version 6 / 17-May-2016 Page 42 of 68 7 STUDY ASSESSMENTS See Table 1 for the time points of the assessments. 7.1 Exercise-specific screening assessments 7.1.1 Exercise test For the exercise test at screening, a ramp protocol will be followed until exhaustion. The details of the test will be described in the exercise manual. 7.1.2 Questionnaire Please refer to appendix 4. 7.2 Safety and tolerability assessments The definitions, reporting and follow-up of AEs and SAEs are described in section 8. 7.2.1 Specific safety assessments A single non-blinded CHDR staff member unconnected to the study will receive reports on the haemoglobin (Hb) concentration and haematocrit (Ht) of the subjects. This individual will be responsible for the dosage adjustments. If Ht exceeds 52% in a subject, dosage will be stopped and if necessary, ±0.5L blood will be collected from that subject to correct the Ht percentage to a normal level. This subject will then be excluded from the study. 7.2.2 Vital signs Evaluations of systolic and diastolic blood pressure, pulse rate and temperature will be performed throughout the study. Pulse and blood pressure will be taken after 5 minutes in the supine position. Automated oscillometric blood pressures will be measured using a Dash 3000, Dash 4000, Dynamap 400 or Dynamap ProCare 400. Additionally, the pulse rate data provided by the pulse oximeter attached to the monitor. 7.2.3 Weight and height Weight (kg) will be recorded at screening, before each exercise test and the follow-up visit or upon early termination. Height (cm) will be recorded and body mass index (BMI) calculated at screening. 7.2.4 Physical examination Physical examination (i.e., inspection, percussion, palpation and auscultation) is performed during the course of the study. Clinically relevant findings that are present prior to study drug initiation must be recorded with the subject’s Medical History. Clinically relevant findings found after study drug initiation and meeting the definition of an AE (new AE or worsening of previously existing condition) must be recorded. 7.2.5 Electrocardiography 12-lead electrocardiographs (ECGs) will be obtained during the course of the study using Marquette 800/5500 or Dash3000 and stored using the MUSE Cardiology Information System. The investigator will assess the ECG recording as 'normal', 'abnormal - not clinically significant', or 'abnormal - clinically significant' and include a description of the abnormality as required. The ECG parameters assessed will include heart rate, PR, QRS, QT, and QTcB (calculated using Bazett’s method). 7.2.6 Laboratory assessments Laboratory parameters Blood and other biological samples will be collected for the following clinical laboratory tests: Lab Tests Collection & Analysis Haematology Haemoglobin [including Mean Corpuscular volume (MCV), Mean corpuscular 4 mL of venous blood in a BD Vacutainer® K2EDTA tube. CHDR1514 Protocol Version 6 / 17-May-2016 Page 43 of 68 haemoglobin (MCH), Mean corpuscular haemoglobin concentration (MCHC)], haematocrit, red cell count (RBC), Red cell distribution width (standard deviation) (RDW- SD), reticulocyte count (RET#), reticulocytes percentage (RET%), immature reticulocyte fraction (IRF), total white cell count (WBC), leukocyte differential count and Platelet count. Differential blood count, including: basophils, eosinophils, neutrophils, lymphocytes, and monocytes. Samples will be analysed by the Central Clinical Hematology Laboratory (CKHL) of Leiden University Medical Center. Chemistry and electrolytes Sodium, potassium, calcium, inorganic phosphate, total protein, albumin, triglycerides, blood urea nitrogen (BUN), creatinine, ferritin, creatinine phosphokinase (CPK)*, uric acid, total bilirubin1, alkaline phosphatase, AST, ALT, gamma-GT, LDH, C-reactive protein*. 8.5 mL or 2 mL of venous blood in a BD Vacutainer® SST Gel and Clot Activator tube. Samples will be analysed by the Central Clinical Chemistry Laboratory (CKCL) of Leiden University Medical Center. Coagulation APTT*, PT*, fibrinogen, D-dimer, F1+2, bTG, PF4, P-selectin, E-selectin, thrombomodulin, TXB2, TNF-alfa 2.7 mL of venous blood in a 3.2% citrate BD Vacutainer and 4 mL in CTAD. Samples will be analysed by Good Biomarker Sciences (GBS) in Leiden and labs to be determined. RNA expression RNA profiles 2.5 mL of venous blood in PAXgene vacuum tubes using a blood collection system analysed by Leiden Genome Technology Center (LUMC) Serology HIV1 and HIV2 antibodies, Hepatitis B antigen and Hepatitis C antibodies 5 mL of venous blood in a BD Vacutainer® SST Gel and Clot Activator tube. Samples will be analysed by the Central Clinical Microbiology Laboratory (CKML) of the Leiden University Medical Center. Urinalysis Leucocytes, blood, nitrite, protein, urobilinogen, bilirubin, pH, specific gravity, ketones, glucose. If there is a clinically significant positive result, urine will be sent to the CKCL for microscopy and/or culture. A midstream, clean-catch urine specimen will be analysed by dipstick (Multistix® 10 SG, Siemens Healthcare Diagnostics, Frimley, UK). Alcohol Alcohol Breath Test The hand-held Alco-Sensor IV meter (Honac, Apeldoorn, the Netherlands) will be used to measure the breath ethanol concentrations. Urine drug screen Cocaine, amphetamines, opiates (morphine), benzodiazepines, methamphetamine, MDMA and cannabinoids. A urine specimen will be analysed at CHDR by test kit (InstAlert, Innovacon, San Diego, USA). CHDR1514 Protocol Version 6 / 17-May-2016 Page 44 of 68 1Conjugated bilirubin will be reported only when total bilirubin is outside the reference range. *Part of the short chemistry/coagulation panel The results of the haematology assessment will be used for the safety report and for the athlete’s biological passport. 7.3 Efficacy assessments 7.3.1 Exercise tests Please refer to Appendix 5 – Exercise test for a detailed description of the exercise tests. 7.3.2 Competition Please refer to Appendix 6 – Competition for a detailed description of the competition. 7.3.3 Athlete biological passport This assessment has already been performed for safety measurements. The results will be used for the safety report and for the athlete’s biological passport. Further calculated markers specific to the Haematological Module of the ABP include OFF-hr Score (OFFS), which is a combination of Hb and RET%[17], and Abnormal Blood Profile Score (ABPS), which is a combination of Ht, Hb, RBC, RET%, MCV, MCH, and MCHC[18]. 7.3.4 Blood flow LSCI measures:  basal blood flow  blood flow upon occlusion-reperfusion  Blood flow after ‘exercise’ Forearm blood flow will be measured by non-invasivce laser speckle contrast imaging (LSCI; PeriCam PSI System, Perimed). At baseline and after 7 weeks of dosing, the flow will be measured at rest, during a five minute occlusion of the brachial artery and during reperfusion after the occlusion. Blood flow will be measured after exercise at 6 weeks. Procedures for LSCI, brachial artery occlusion-reperfusion are described in SOPs. Briefly, the subject will be seated with the left arm placed on a (table) rest. A suitable area of the volar side of the forearm will be identified. This area will be 'illuminated' by the laser and the response signal will be captured. 7.4 NeoRecormon detection assessments 7.4.1 Urine doping screen Please refer to Appendix 7 – Urine doping control procedure for a detailed description of the urine doping control procedure. 7.5 RNA expression levels In general, genes react to internal or external stimuli by becoming more active, less active or showing no response. Training stimuli induce stress resulting in changes in gene expression patterns. Gene expression (RNA-) profiles made from whole venous blood taken before and after physical exercise differ between non-trained and trained individuals. This suggests that adaptation to training can be assessed using gene expression analysis. Changes in gene expression patterns can then be used to monitor training effects. It is unclear how adaptation to training with or without use of rHuEPO will change gene expression profiles in well-trained individuals. We hypothesize that the changes in these patterns observed in samples taken before and after an exercise test are correlated with the performance and health status of individual athletes. The measurements of maximal and submaximal exercise parameters and clinical laboratory assessments of the participants in this study will be combined with gene expression profile data to investigate if specific changes in the acute response to a submaximal exercise test are correlated with better endurance CHDR1514 Protocol Version 6 / 17-May-2016 Page 45 of 68 performance or other parameters. Comparisons of these profiles in time are expected to reveal the effects of specific types of training and changes in health status. RNA profiles from whole venous blood samples will be determined using next generation sequencing. Samples for RNA-profiling will be collected around a submaximal intensity test (45 minutes, starting at 80% Power from the baseline maximal exercise test allowing changes to the power at any time by the subject himself if he wishes). Blood samples (2.5 ml) for RNA analysis will be drawn in PAXgene vacuum tubes at 2 time points, before and directly after the test. RNA-profiles made from both time points of each individual will be compared to identify changes in gene expression due to the submaximal intensity test. Profiles from the same time points will also be compared to determine interindividual variation and possible EPO treatment effects. Expression levels of the different transcripts in the RNA samples will be determined as digital counts using standard bioinformatics pipelines. Cluster analysis and principal component analysis will be performed to determine the relationships between the samples. In this way it will be possible to determine if:  The expression patterns differ at the time points before and after submaximal exercise test between individuals and between groups.  rHuEPO treatment does influence gene expression levels. Multivariate statistics will be applied to analyse and interpret complete RNA-profiles. The leukocyte composition of the original blood samples is correlated with cell-specific transcripts. Therefore, this can be derived from RNA-profiles allowing their correction for this effect. Comparison with other parameters measured during the study may indicate which changes in gene expression patterns of individuals are correlated with their performance. For all subjects RNA samples will be collected for potential analyses of gene expression. Instructions for collection, processing, handling and shipment of the samples will be outlined in the laboratory manual. Samples will be archived according to local administration regulations. Analysis will be performed outside the scope of the main study and reported separately. 7.6 Sequence of assessments and time windows Assessments will be performed in the following order, where possible: physical examination, weight and height, ECG, vital signs, blood samples, dosing. The deviations of actual time points from the expected time points will be within ten percent, calculated from the zero point (time of drug administration) or the last relevant activity. Deviations of more than 10% will be explained in a note. Pre-dose assessments are given in indicative expected times. CHDR1514 Protocol Version 6 / 17-May-2016 Page 46 of 68 7.7 Total blood volume Sample Samples taken Sample Volume* Volume Haematology 8 x 4 mL = 32 mL Chemistry 2 x 8.5 mL = 17 mL Chemistry short 16 x 2.5 mL = 40 mL Serology 1 x 5 mL = 5 mL Coagulation 17 x 6.2 mL = 105.4 mL Coagulation short 2 x 4.5 mL = 9 mL RNA expression 2 x 2.5 mL = 5 mL Biomarker 4 x 4 mL = 16 mL Hb/Hct (1mL tube) 8 x 1 mL = 8 mL Lactate (Drop) 51 x 0.8 mL = 40.8 mL * inclusive discarded volume Total blood volume/subject 278.2 mL CHDR1514 Protocol Version 6 / 17-May-2016 Page 47 of 68 8 SAFETY REPORTING 8.1 Definitions of adverse events An Adverse Event (AE) is any untoward medical occurrence in a subject who is participating in a clinical study performed. The adverse event does not necessarily have to follow the administration of a study drug, or to have a causal relationship with the study drug. An adverse event can therefore be any unfavourable and unintended sign (including an abnormal laboratory or vital sign finding), symptom, or disease temporally associated with the study participation, whether or not it is related to the study drug. 8.1.1 Intensity of adverse events The intensity of clinical AEs is graded three-point scale as defined below:  Mild: discomfort noticed but no disruption of normal daily activity;  Moderate: discomfort sufficient to reduce or affect normal daily activity;  Severe: inability to work or perform daily activity. 8.1.2 Relationship to study drug For each adverse event the relationship to drug as judged by the investigator:  Probable;  Possible;  Unlikely;  Unrelated. 8.1.3 Chronicity of adverse events The chronicity of the event will be classified by the investigator on a three-item scale as defined below:  Single occasion: single event with limited duration;  Intermittent: several episodes of an event, each of limited duration;  Persistent: event which remained indefinitely. 8.1.4 Action Eventual actions taken will be recorded. 8.1.5 Serious adverse events A Serious Adverse Event (SAE) is defined by the International Conference on Harmonization (ICH) guidelines as any AE fulfilling at least one of the following criteria:  Is fatal  Is life-threatening  Is disabling  Requires or prolongs in-patient hospitalisation  Causes congenital anomaly will be described as a SAE. Important medical events that may not be immediately life threatening or result in death or hospitalisation may be considered a serious adverse event when, based on appropriate medical judgement, they may jeopardise the subject or may require medical or surgical intervention to prevent one of the outcomes listed above. CHDR1514 Protocol Version 6 / 17-May-2016 Page 48 of 68 8.1.6 Suspected unexpected serious adverse reactions A SUSAR (Suspected Unexpected Serious Adverse Reaction) is a serious adverse event that is unexpected, (nature or severity of which is not consistent with the applicable product information (e.g., investigator's brochure for an unauthorised investigational product or summary of product characteristics for an authorised product)) and suspected (a reasonable possibility of causal relationship with investigational drug). 8.1.7 Reporting of serious adverse events SAEs and SUSAR's will be reported according to the following procedure. All SUSARs and SAEs must be reported to the investigator. In case of multi-centre studies of which a CHDR employee is the responsible investigator, the sub- investigators at sites outside CHDR must notify the investigator by telephone and in writing preferably using a standard form (e.g. CIOMS for SUSARs). The investigator must report all SAEs and SUSAR’s to the EC that approved the study, in writing as soon as practical, but at least within 15 days. Fatal and life-threatening suspected SUSAR’s should be reported within 7 calendar days, with another 8 days for completion of the report. The investigator must report all SUSAR’s to the CA, in writing as soon as practical, but at least within 15 days. Fatal and life-threatening suspected SUSAR’s should be reported within 7 calendar days, with another 8 days for completion of the report. SAE's do not have to be reported to the CA. The investigator must furthermore report all SUSAR’s to EMA's EudraVigilance database within 15 days. Fatal and life-threatening suspected SUSAR’s should be reported within 7 calendar days, with another 8 days for completion of the report. The investigator can prepare additional reports for other authorities (e.g. FDA). All SAEs and SUSAR’s must be reported on the Toetsing Online website2. Fatal and life- threatening SAEs and SUSARs must be reported within 7 calendar days. All other SAEs must be reported within 15 days. By reporting on the Toetsing Online website the EC that approved the study will also be informed. 8.1.8 Follow-up of adverse events All adverse events will be followed until they have abated, returned to baseline status or until a stable situation has been reached. Depending on the event, follow up may require additional tests or medical procedures as indicated, and/or referral to the general physician or a medical specialist. 8.2 Section 10 WMO event In accordance to section 10, subsection 1, of the WMO, the investigator will inform the subjects and the EC if anything occurs, on the basis of which it appears that the disadvantages of participation may be significantly greater than was foreseen in the research proposal. The study will be suspended pending further review by the EC, except insofar as suspension would jeopardise the subjects’ health. The investigator will ensure that all subjects are kept informed. 8.3 Annual safety report or development safety update report In addition to the expedited reporting of SUSARs, the investigator will submit, once a year throughout the clinical trial, a safety report to the EC, CA, MEB and CAs of the concerned Member States. 2 https://www.toetsingonline.nl/ CHDR1514 Protocol Version 6 / 17-May-2016 Page 49 of 68 This safety report consists of:  a list of all suspected (unexpected or expected) serious adverse reactions, along with an aggregated summary table of all reported serious adverse reactions, ordered by organ system, per study;  a report concerning the safety of the subjects, consisting of a complete safety analysis and an evaluation of the balance between the efficacy and the harmfulness of the medicine under investigation. CHDR1514 Protocol Version 6 / 17-May-2016 Page 50 of 68 9 STATISTICAL METHODOLOGY AND ANALYSES 9.1 Statistical analysis plan All safety and statistical programming is conducted with SAS 9.4 for Windows (SAS Institute Inc., Cary, NC, USA). PK variable programming is conducted with R 2.12.0 for Windows (R Foundation for Statistical Computing/R Development Core Team, Vienna, Austria, 2010. A Statistical Analysis Plan (SAP) will be written and finalized before the study closure, i.e., database closure and unblinding of the randomization code. The SAP will provide full details of the analyses, the data displays and the algorithms to be used for data derivations. The SAP will include the definition of major and minor protocol deviations and the link of major protocol deviations to the analysis sets. 9.2 Protocol violations/deviations Protocol deviations will be identified based on conditions related to the categories below:  Protocol entry criteria  Forbidden concomitant medications  Missing evaluations for relevant endpoints  Other protocol deviations occurring during study conduct. Major protocol deviations will be identified before the study closure, and listed where appropriate. 9.3 Power calculation Please refer to section 1.4.9. 9.4 Missing, unused and spurious data All missing or incomplete safety and efficacy data, including dates and times, are treated as such. Missing test results or assessments will not be imputed. Missing efficacy data, indicated as ‘M’ in the data listing, will be estimated within the statistical mixed model using SAS PROC MIXED. For graphical and summary purposes efficacy and safety values below the limit of quantification will be set to half (½) of the limit of quantification. For analysis no undetermined values will be replaced. The handling of missing, unused and spurious data will be documented in the study report. 9.5 Analysis sets Data of all subjects participating in the study will be included in the analyses if the data can meaningfully contribute to the objectives of the study. 9.5.1 Safety set The safety population will be defined as all subjects who were validated (randomised) and received at least 1 dose of study treatment. 9.5.2 Efficacy analysis set The analysis population for efficacy is defined as all subjects who were validated (randomised), received at least one dose of study treatment, and have at least one post-baseline assessment of the parameter being analysed. 9.6 Subject disposition The following subject data will be summarized by treatment and overall:  Number and percentage of subjects enrolled in each analysis set for all randomized subjects;  Number and percentage of subjects who completed the study or prematurely discontinued from the investigational period by reasons for discontinuation, to be tabulated for each analysis set. CHDR1514 Protocol Version 6 / 17-May-2016 Page 51 of 68 Subject disposition will be listed. 9.7 Baseline parameters and concomitant medications 9.7.1 Demographics and baseline variables Continuous demographic variables (e.g., age, height, weight, BMI) will be summarized by descriptive statistics (n, mean, SD, median, Min, Max).Qualitative demographic characteristics (sex, race/ethnicity) will be summarized by counts and percentages. The results of the sport activities questionnaire at screening will only be listed. 9.7.2 Medical history Medical history will only be listed. 9.7.3 Concomitant Medications Previous and concomitant medications will be coded according to the World Health Organization (WHO) drug code and the anatomical therapeutic chemical (ATC) class code. All concomitant medications will be displayed in a listing. 9.7.4 Treatment compliance/exposure Exposure to study treatment is described in terms of duration of treatment and average infusion rate. The average infusion rate (mL/hr) is summarized by mean, SD, median, Q1, Q3, Min, Max. 9.8 Safety and tolerability endpoints The safety set is used to perform all safety analyses. Baseline is defined as the last value prior to dosing. Change from baseline will be calculated for all continuous safety parameters. 9.8.1 Adverse events The AE coding dictionary for this study will be Medical Dictionary for Regulatory Activities (MedDRA). It will be used to summarize AEs by primary system organ class (SOC) and preferred term (PT). All adverse events will be displayed in listings. A treatment-emergent adverse event (TEAE) is defined as an adverse event observed after starting administration of the specific treatment, and prior to the start of another treatment, if any OR up to 5 days (96 hours) after study drug administration. If a subject experiences an event both prior to and after starting administration of a treatment, the event will be considered a TEAE (of the treatment) only if it has worsened in severity (i.e., it is reported with a new start date) after starting administration of the specific treatment, and prior to the start of another treatment, if any. All TEAEs collected during the investigational period will be summarized. The number of and/or the number of subjects with treatment emergent AEs will be summarized by: 1. treatment, MedDRA SOC and PT; 2. treatment, MedDRA SOC, PT and severity; 3. treatment, MedDRA SOC, PT and drug relatedness. 9.8.2 Vital signs At each time point, absolute values and change from baseline of supine BP and HR will be summarized with n, mean, SD, SEM, median, Min, and Max values. The number of available observations and out-of-range values (absolute and in percentage) will be presented. Values outside the reference range will be flagged in the listing. ‘H’ and ‘L’, denoting values above or below the investigator reference range (when present), will flag out-of-range results. CHDR1514 Protocol Version 6 / 17-May-2016 Page 52 of 68 9.8.3 ECG At each time point, absolute values and change from baseline of ECG numeric variables will be summarized with n, mean, SD, SEM, median, Min, and Max values. The number of available observations and out-of-range values (absolute and in percentage) will be presented. Values outside the investigator’s normal range will be flagged in the listing. ‘H’ and ‘L’, denoting values above or below the investigator reference range (when present), will flag out-of-range results. 9.8.4 Clinical laboratory tests At each time point, absolute values and change from baseline of clinical laboratory variables will be summarized with n, mean, SD, SEM, median, Min, and Max values. The number of available observations and out-of-range values (absolute and in percentage) will be presented. All laboratory data (including re-check values if present) will be listed chronologically. ‘H’ and ‘L’, denoting values above or below the investigator reference range (when present), will flag out-of-range results. 9.9 Efficacy endpoints 9.9.1 Efficacy The final analysis will be preceded by a blind data review which consists of individual graphs per visit by time of all efficacy measurements by time. The graphs will be used to detect outliers and measurements unsuitable for analysis. The efficacy parameters will be listed by treatment, subject, visit and time. Individual graphs by time will be generated. All measured PD endpoints will be summarised (n, mean, SD, SEM, median, Min and Max values) by treatment and time, and will also be presented graphically as mean over time, with standard deviation as error bars. All efficacy endpoints will be summarised (mean, SD, SEM, median, Min and Max values) by treatment, and will also be presented graphically as mean in a bargraph, with standard deviation as error bars. Parameters will initially be analyzed without transformation, but if the data suggest otherwise, log- transformation may be applied. Log-transformed parameters will be back-transformed after analysis where the results may be interpreted as percentage change. To establish whether significant treatment effects can be detected on the measured efficacy parameters, each parameter that is measured repeatedly will be analyzed with a mixed model analysis of covariance (ANCOVA) with treatment, time and treatment by time as fixed factors and subject, subject by treatment and subject by time as random factors and the (average) baseline measurement as covariate. Parameters that are only measured at baseline and at the end of the treatment phase will be analysed with a one-way ANCOVA with treatment and baseline as covariate. Single measured efficacy parameters will be analyzed with an unpaired t-test. The Kenward-Roger approximation will be used to estimate denominator degrees of freedom and model parameters will be estimated using the restricted maximum likelihood method. The general treatment effect and specific contrasts will be reported with the estimated difference and the 95% confidence interval, the least square mean estimates and the p-value. Graphs of the Least Squares Means(LSM) estimates over time by treatment will be presented with 95% confidence intervals as error bars, as well as change from baseline LSM estimates. The following contrasts will be calculated within the model:  NeoRecormon - Placebo 9.9.2 Inferential methods Null hypothesis: There’s no difference between the effects of NeoRecormon and placebo on cycling performance CHDR1514 Protocol Version 6 / 17-May-2016 Page 53 of 68 parameters. 9.10 Exploratory analyses and deviations Exploratory data-driven analyses can be performed with the caveat that any statistical inference will not have any confirmatory value. Deviations from the original statistical plan will be documented in the clinical study report. 9.11 Interim analyses No interim analysis is planned. CHDR1514 Protocol Version 6 / 17-May-2016 Page 54 of 68 10 GOOD CLINICAL PRACTICE, ETHICS AND ADMINISTRATIVE PROCEDURES 10.1 Good clinical practice 10.1.1 Ethics and good clinical practice The investigator will ensure that this study is conducted in full compliance with the protocol, the principles of the Declaration of Helsinki, ICH GCP guidelines, and with the laws and regulations of the country in which the clinical research is conducted. 10.1.2 Ethics committee / institutional review board The investigator will submit this protocol and any related documents to an Ethics Committee (EC) and the Competent Authority (CA). Approval from the EC and the statement of no objection from the CA must be obtained before starting the study, and should be documented in a dated letter/email to the investigator, clearly identifying the trial, the documents reviewed and the date of approval. A list of EC members must be provided, including the functions of these members. If study staff were present, it must be clear that none of these persons voted. Modifications made to the protocol after receipt of the EC approval must also be submitted as amendments by the investigator to the EC in accordance with local procedures and regulations. 10.1.3 Informed consent It is the responsibility of the investigator to obtain written informed consent from each individual participating in this study after adequate explanation of the aims, methods, objectives and potential hazards of the study. The investigator must also explain to the subjects that they are completely free to refuse to enter the study or to withdraw from it at any time for any reason. The Informed Consent and Subject Information will be provided in dutch. 10.1.4 Insurance The investigator has a liability insurance which is in accordance with article 7, subsection 6 of the WMO. The investigator has an insurance which is in accordance with the legal requirements in the Netherlands (Article 7 WMO and the Measure regarding Compulsory Insurance for Clinical Research in Humans of 23rd June 2003). This insurance provides cover for damage to research subjects through injury or death caused by the study.  € 650,000.- (i.e., six hundred and fifty thousand Euro) for death or injury for each subject who participates in the Research;  € 5,000,000.- (i.e., five million Euro) for death or injury for all subjects who participate in the Research;  € 7,500,000.- (i.e., seven million five hundred thousand Euro) for the total damage incurred by the organisation for all damage disclosed by scientific research for the Sponsor as ‘verrichter’ in the meaning of said Act in each year of insurance coverage. The insurance applies to the damage that becomes apparent during the study or within 4 years after the end of the study. 10.2 Study funding CHDR is the sponsor of the study and is funding the study. 10.3 Data handling and record keeping 10.3.1 Data collection Data will be recorded on paper and/or electronic data collection forms and will be entered after quality control in a Promasys database for subsequent tabulation and statistical analysis. The data will be handled confidentially and if possible anonymously. CHDR1514 Protocol Version 6 / 17-May-2016 Page 55 of 68 A Subject Screening and Enrolment Log will be completed for all eligible or non-eligible subjects with the reasons for exclusion. 10.3.2 Database management and quality control A quality control check will be done by CHDR staff using data entry progress checks and database listings (blind data review). Errors with obvious corrections will be corrected before database lock. Results of computer tests and electronically captured questionnaires, clinical laboratory and efficacy analyses will be sent electronically to CHDR and loaded into the database. After the database has been declared complete and accurate, the database will be locked. Any changes to the database after that time can only be made by joint written agreement between the investigator and the statistician. 10.4 Access to source data and documents All study data will be handled confidentially. The investigator will retain the originals of all source documents generated at CHDR for a period of 2 years after the report of the study has been finalised, after which all study-related documents will be archived (at a minimum) on micro-film which will be kept according to GCP regulations. The investigator will permit trial-related monitoring, audits, EC review and regulatory inspections, providing direct access to source data and documents. 10.5 Quality control and quality assurance This study will be conducted according to applicable Standard Operating Procedures (SOPs). Quality assurance will be performed under the responsibility of CHDR’s Quality Assurance manager. 10.5.1 Monitoring An initiation visit will be performed before the first subject is included. Monitoring visits and contacts will occur at regular intervals thereafter, according to a frequency defined in the study-specific monitoring plan. A close-out visit will be performed after study closure. 10.6 Protocol amendments Any change to a protocol has to be considered as an amendment. 10.6.1 Non-substantial amendment Administrative or logistical minor changes require a non-substantial amendment. Such changes include but are not limited to changes in study staff or contact details or minor changes in the packaging or labelling of study drug. Non-substantial amendments will be approved (signed) by the investigator(s) and will be recorded and filed by the investigator but will not be notified to the EC and the CA. The implementation of a non-substantial amendment can be done without notification to the appropriate EC or CA. It does not require their approval. The following amendments will be regarded as non-substantial: - change in timing of the samples; - changes in assay-type and / or institution where an assay will be performed, provided that validated assays will be used; - editorial changes to the volunteer information sheets; - determination of additional parameters in already collected materials, which are in agreement with the study objectives and do not provide prognostic or genetic information; - other statistical analyses than described in the protocol. 10.6.2 Substantial amendment Significant changes require a substantial amendment. Significant changes include but are not limited to: new data affecting the safety of subjects, change of the objectives/endpoints of the study, CHDR1514 Protocol Version 6 / 17-May-2016 Page 56 of 68 eligibility criteria, dose regimen, study assessments/procedures, treatment or study duration, with or without the need to modify the core Subject Information and Informed Consent Form. Substantial amendments are to be approved by the appropriate EC and the CA will need to provide a ‘no grounds for non-acceptance’ notification prior to the implementation of the substantial amendment. Urgent amendment An urgent amendment might become necessary to preserve the safety of the subjects included in the study. The requirements for approval should in no way prevent any immediate action being taken by the investigators in the best interests of the subjects. Therefore, if deemed necessary, an investigator can implement an immediate change to the protocol for safety reasons. This means that, exceptionally, the implementation of urgent amendments will occur before submission to and approval by the EC(s) and CA. 10.7 End of study report The investigator will notify the EC and the CA of the end of the study within a period of 90 days. The end of the study is defined as the last subject’s last visit. In case the study is ended prematurely, the investigator will notify the EC and the CA within 15 days, including the reasons for the premature termination. Within one year after the end of the study, the investigator will submit a final study report with the results of the study, including any publications/abstracts of the study, to the EC and the CA. The principal investigator will be the signatories for the study report. 10.8 Public disclosure and publication policy In accordance with standard editorial and ethical practice, the results of the study will be published. The authorship guidelines of the Vancouver Protocol3 will be followed regarding co-authorship. The principal investigator will have the opportunity to review the analysis of the data and to discuss the interpretation of the study results prior to publication. 3 http://www.icmje.org/ CHDR1514 Protocol Version 6 / 17-May-2016 Page 57 of 68 11 STRUCTURED RISK ANALYSIS Please refer to the respective Summary of Product Characteristics (SmPC, see appendix 1) and the European Public Assessment Report (EPAR, see appendix 3). CHDR1514 Protocol Version 6 / 17-May-2016 Page 58 of 68 12 REFERENCES References 1. Dr.Dick Marty, Mr.Peter Nicholson, Prof.Dr.Ulrich Haas. Report to the president of the Union Cycliste Internationale. Cycling Independent Reform Commision 2015. 2. Nimesh SA Patel, Massimo Collino, Muhammad M Yaqoob, Christoph Thiemermann. Erythropoietin in the intensive care unit: beyond treatment of anemia. Annals of Intensive Care 2011; 1: 1-9. 3. [online] Available at http://www.wada.-ama.org/e. World Anti-Doping Agency 2015. 4. Heuberger JA, Cohen Tervaert JM, Schepers FM, Vliegenthart AD, Rotmans JI, Daniels JM, Burggraaf J, Cohen AF. Erythropoietin doping in cycling: lack of evidence for efficacy and a negative risk-benefit. Br J Clin Pharmacol 2013; 75: 1406-1421. 5. Jeukendrup AE, Craig NP, Hawley JA. The bioenergetics of World Class Cycling. J Sci Med Sport. 2000; 3: 414-433. 6. Connes P, Perrey S, Varray A, Prefaut C, Caillaud C. Faster oxygen uptake kinetics at the onset of submaximal cycling exercise following 4 weeks recombinant human erythropoietin (r-HuEPO) treatment. Pflugers.Arch. 2003; 447: 231-238. 7. Lucia A, Hoyos J, Perez M, Santalla A, Chicharro JL. Inverse relationship between VO2max and economy/efficiency in world-class cyclists. Med Sci.Sports Exerc. 2002; 34: 2079-2084. 8. Norio Suzuki. Erythropoietin Gene Expression: Developmental-Stage Specificity, Cell-Type Specificity, and Hypoxia Inducibility. Tohoku J.Exp.Med. 2015; 235: 233-240. 9. Lucia A, Hoyos J, Santalla A, Perez M, Chicharro JL. Curvilinear VO(2):power output relationship in a ramp test in professional cyclists: possible association with blood hemoglobin concentration. Jpn J Physiol. 2002; 52: 95-103. 10. Heinicke K, Wolfarth B, Winchenbach P, Biermann B, Schmid A, Huber G, Friedmann B, Schmidt W. Blood volume and hemoglobin mass in elite athletes of different disciplines. Int J Sports Med 2001; 22: 504-512. 11. Henny H.Billett. Hemoglobin and hematocrit.In: Clinical Methods: The History, Physical, and Laboratory Examinations. 3rd edition. 1990: 718-719. 12. Online available at http://www.blood.ca/en/blood/hemoglobin. Canadian Blood Services 2015. 13. Wilkerson DP, Rittweger J, Berger NJ, Naish PF, Jones AM. Influence of recombinant human erythropoietin treatment on pulmonary O2 uptake kinetics during exercise in humans. J Physiol. 2005; 568: 639-652. 14. Thomsen JJ, Rentsch RL, Robach P, Calbet JA, Boushel R, Rasmussen P, Juel C, Lundby C. Prolonged administration of recombinant human erythropoietin increases submaximal performance more than maximal aerobic capacity. Eur.J Appl.Physiol. 2007; 101: 481-486. 15. Lundby C, Achman-Andersen NJ, Thomsen JJ, Norgaard AM, Robach P. Testing for recombinant human erythropoietin in urine: problems associated with current anti-doping testing. J Appl.Physiol. 2008; 105: 417-419. CHDR1514 Protocol Version 6 / 17-May-2016 Page 59 of 68 16. Birkeland KI, Stray-Gundersen J, Hemmersbach P, Hallen J, Haug E, Bahr R. Effect of rhEPO administration on serum levels of sTfR and cycling performance. Med Sci.Sports.Exerc. 2000; 32: 1238-1243. 17. Ralph B.Dell, Steve Holleran, Rajasekhar Ramakrishnan. Sample Size Determination. ILAR journal 2002; 43: 207-213. 18. Gore CJ, Parisotto R, Ashenden MJ, Stray-Gundersen J, Sharpe K, Hopkins WG, Emslie KR, Howe C, Trout GJ, Kazlauskas R, Hahn AG. Second-generation blood tests to detect erythropoietin abuse by athletes. Haematologica 2003; 88: 333-343. 19. Sottas PE, Robinson N, Giraud S, Taroni F, Kamber M, Mangin P, Saugy M. Statistical Classification of Abnormal Blood Profiles in Athletes. The International Journal of Biostatistics 2006; 2: 3. CHDR1514 Protocol Version 6 / 17-May-2016 Page 60 of 68 13 APPENDIX 1 – SUMMARY OF PRODUCT CHARACTERISTICS – ENGLISH VERSION CHDR1514 Protocol Version 6 / 17-May-2016 Page 61 of 68 14 APPENDIX 2 – SUMMARY OF PRODUCT CHARACTERISTICS – DUTCH VERSION CHDR1514 Protocol Version 6 / 17-May-2016 Page 62 of 68 15 APPENDIX 3 – EUROPEAN PUBLIC ASSESSMENT REPORT CHDR1514 Protocol Version 6 / 17-May-2016 Page 63 of 68 16 APPENDIX 4 – SCREENING: QUESTIONNAIRE SPORT ACTIVITIES Sport activities in the past 3 years: Type of sport Times a week Distance and duration per training/competition Mean velocity Since when 1 2 3 4 5 Sport activities in the past 6 weeks: Type of sport Times a week Distance and duration per training/competition Mean velocity Since when 1 2 3 4 5 Dutch: Sport activiteiten in de afgelopen 3 jaar: Type sport Aantal keer per week Afstand en duur van training/competitie Gemiddelde snelheid (km/u) Sinds wanneer 1 2 3 4 5 Sport activiteiten in de afgelopen 6 weken: CHDR1514 Protocol Version 6 / 17-May-2016 Page 64 of 68 Type sport Aantal keer per week Afstand en duur van training/competitie Gemiddelde snelheid (km/u) Sinds wanneer 1 2 3 4 5 CHDR1514 Protocol Version 6 / 17-May-2016 Page 65 of 68 17 APPENDIX 5 – EXERCISE TEST During this study the exercise tests will be performed on a cycle ergometer. The subjects will cycle at a self-selected pedal rate (between 70 and 90 rpm) and this pedal rate along with the saddle and handlebar heights will be recorded and reproduced in subsequent tests. All subjects will breathe through a facemask that is connected to an oxymeter to collect inspired and expired gasses for analyzing oxygen consumption, carbon dioxide production, minute ventilation, tidal volume and breathing frequency. Before each test, the oxymeter will be calibrated to ensure valid measurements. Blood pressure and heart rate will also be monitored during the tests. The blood pressure and heart rate measuring devices, cycle ergometer and oxymeter will be connected to a computer to measure and analyse all data. 18.1 Ramp exercise test Protocol A ramp protocol will be followed until exhaustion, which will take approximately 1 hour. The details of the ramp exercise test will be described in the exercise test manual. 18.2 Time to exhaustion test Protocol This exercise test will be performed according to a protocol looking at performance at submaximal intensity. It will last approximately 1 hour. The details of time to exhaustion test will be described in the exercise test manual. 18.3 Statistics All data will be presented as means ± SD (or standard error). First, a two-way factor (group versus respectively VO2, VO2/kg, %VO2max, W and W/kg) ANOVA with repeated measures will be performed to determine if an interactive effect exists between both groups and mentioned parameters. These parameters will be analyzed at VT1, VT2 and maximal intensity. The ANOVA will be conducted to determine whether the factor group will significantly influence these physiological parameters. When a significant F-ratio is detected, comparisons for unpaired data will be performed (Student’s t-tests) to locate the differences (2-tailed). Maximal lactate values between both groups will also be compared using an unpaired Student’s t-test. VO2 at the ventilatory and lactate thresholds will be used to compare both thresholds by using a paired Student’s t-test. Statistical significance will be set at P < 0.05. CHDR1514 Protocol Version 6 / 17-May-2016 Page 66 of 68 18 APPENDIX 6 – COMPETITION Subjects will climb the Mont Ventoux in an open course via Bédoin after a 150km closed course. A blood sample of each subject will be collected in Bédoin (at the end of the closed course) and at the finish. A power meter will be incorporated into each bike. Figure 2. Detailed description of the Mont Ventoux climb via Bédoin.1 CHDR1514 Protocol Version 6 / 17-May-2016 Page 67 of 68 Figure 3. Characteristics of the Mont Ventoux climb via Bédoin.1 References 1. http://www.clubcinglesventoux.org/en/roads/1-bedoin.html CHDR1514 Protocol Version 6 / 17-May-2016 Page 68 of 68 19 APPENDIX 7 – URINE DOPING CONTROL PROCEDURE Urine collection for doping detection will be done at predetermined moments as specified in the Schedule of Assessments. For each sample, a minimum of 50 mL urine will be collected, which will be stored at -20 degrees Celcius or below before shipment to the Doping Lab. The doping lab will process the samples in a blind fashion as per the WADA Technical document published online at https://www.wada-ama.org/en/resources/science-medicine/td2014-epo. Samples will be analysed using two methods, being the SAR-PAGE and Isoelectrofocusing (IEF) as described in this technical document.

Repeatability and predictive value of lactate threshold concepts in endurance sports.

11-14-2018

Heuberger, Jules A A C,Gal, Pim,Stuurman, Frederik E,de Muinck Keizer, Wouter A S,Mejia Miranda, Yuri,Cohen, Adam F

eng

PMC10723660

RESEARCH ARTICLE Relationship between methods of monitoring training load and physiological indicators changes during 4 weeks cross-country skiing altitude training Yichao Yu1,3☯, Dongye Li2,3☯, Yifan LuID2,3*, Jing Mi1 1 The School of Sports Coaching, Beijing Sports University, Beijing, China, 2 The School of Sports Medicine and Rehabilitation, Beijing Sports University, Beijing, China, 3 Laboratory of Sports Stress and Adaptation of General Administration of Sport, Beijing Sport University, Beijing, China ☯ These authors contributed equally to this work. * [email protected] Abstract This study aimed to: (i) analyze the load characteristics of 4 weeks cross-country skiing alti- tude training; (ii) analyze the relationships between methods of monitoring training load and physiological indicators changes of elite male Chinese cross-country skiers during this period. Practitioners collected load data during 4 weeks of altitude training camp. Partici- pants performed maximal oxygen uptake, lactate threshold, body composition, and skierg power test before and after the training camp to investigate the changes in physiological per- formance. Edwards TRIMP, Lucia TRIMP, and session rating of perceived exertion were collected as internal load. Training distance, time recorded by the Catapult module were col- lected as external load. The result revealed a " pyramid " pattern in the load characteristics during the altitude training camp. The correlation between luTRIMP and percent change in physiological indicators was highest. Percentage changes in lactate threshold velocity (r = .78 [95% CI -.01 to .98]), percentage changes in lactate threshold HR (r = .71 [95% CI .14- .99]), percentage changes in maximum HR (r = .83 [95% CI .19–1.00]), percentage changes in skierg power-to-weight ratio (r = .75 [95% CI -.28 to .98]) had very large relationships with luTRIMP. In cross-country skiing altitude training, training loads should be reasonably con- trolled to ensure that athletes do not become overly fatigued. Methods of training load moni- toring that combine with athletes’ physiological characteristics and program characteristics have the highest dose-response relationships, it is an important aspect of cross-country ski training load monitoring. The luTRIMP could be a good monitoring tool in cross-country ski- ing altitude training. Introduction Cross-country skiing is a typical endurance sports, which requires high demands on the physi- ological performance of athletes [1–3]. The purpose of the training is to allow athletes to PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 1 / 15 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Yu Y, Li D, Lu Y, Mi J (2023) Relationship between methods of monitoring training load and physiological indicators changes during 4 weeks cross-country skiing altitude training. PLoS ONE 18(12): e0295960. https://doi.org/10.1371/journal. pone.0295960 Editor: Rafael Franco Soares Oliveira, Instituto Polite´cnico de Santare´m: Instituto Politecnico de Santarem, PORTUGAL Received: July 27, 2023 Accepted: December 2, 2023 Published: December 15, 2023 Copyright: © 2023 Yu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: Data cannot be shared publicly because in order to protect the privacy of national team athletes, all data in this study are protected by the Chinese Winter Sports Management Centre. Data are available from the Ethics Committee of the Beijing sport University (contact: [email protected]) for researchers who meet the criteria for access to confidential data. Funding: This research was funded by the National key research and development program of China develop positive physiological adaptations in response to adapted training load stimuli and improve performance levels. Physiological indicators are an objective expression of this physi- ological adaptation [4]. The required by modern training concepts is gradual, periodic, and specific [5], and its essence is to stimulate the athlete’s body to produce the best adaptation through the appropriate load (including volume and intensity) [6]. Practitioners need to moni- tor skiers’ training load and physiological stress adaptation during different training period, which allows skiers to maintain a high level of physiological performance and avoid a severe drop in performance due to overtraining [7–11]. With the popularity of various types of training monitoring equipment and the use of advanced training analysis software, it is possible to systematically analyze the training load of skiers. Training load among cross-country skiing is the cumulative stress an individual is sub- jected to in competition or training throughout training period [12]. According to the source of monitored indicators, there are two classification forms of load: external and internal [13]. Similar to most endurance sports, heart rate changes during cross-country skiing training are key to load monitoring [14, 15]. In the last century, Banister pioneered the Training Impulse (TRIMP) to help practitioners with load monitoring, which is a combination of training time, heart rate during training period [16]. Since then, Edwards [17] and Lucia [18] proposed two interval TRIMP calculation methods, where a linear weighting factor is used to weight the training time for the delineated heart rate intervals. On this basis, Manzi [19] have proposed TRIMP calculations based on heart rate and lactate changes. In addition to heart rate based load monitoring methods, session rating of perceived exertion (sRPE) [15, 20] and external load information collected by global positioning system (GPS) [21, 22] are also applied in ski training practice. These internal or external load-based quantification methods can help prac- titioners visualize the training loads to which athletes are exposed. It is easier than ever for practitioners to access training-related load data with advances in technology. For load monitoring to be maximally effective, it is imperative that the employed methods are pertinent to significant physiological outcomes. But there is a great deal of uncer- tainty about the relationship between physiological performance change and these load data. Practitioners should select an appropriate load monitoring method based on the relationship between physiological performance and methods of training load, thus allowing for a more proactive approach to the development of training programmes [23]. In team sports such as soccer, some researchers have explored this phenomenon. Oliveira et al. [24] analyzed the rela- tionship between wellness and training and match load in 13 professional male soccer players, they found that the intensity of training on the match day was correlated with indicators of sleep quality and fatigue the next day. Costa et al. [25] did not find the within-subject relation- ship between sleep indicators and HRV with training and match load in 20 elite female soccer players, the use of statistical methods may be a possible reason for the difference. In endurance sports, this relationship can be assessed by evaluating athletes’ physiological performance change before and after training period. A study of 8 distance runners found that athletes showed a significant increase in lactate threshold velocity after training. There were essentially relationships between improvements of lactate threshold velocity and weekly individualized TRIMP (iTRIMP). But weaker relationships between Banister TRIMP (bTRIMP) and velocity improvements at the lactate threshold [19]. Sanders et al. [26] found that a load monitoring approach incorporating the individual physiological characteristics of the athlete has the most substantial dose-response relationship in 15 well-trained cyclists. The difference of load moni- toring means and the difference of sports can be one of the reasons for the different dose- response relationships. Previous research has established a relationship between training load and physiological performance that follows a dose-response relationship. However, this relationship is complex PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 2 / 15 (2018YFC2000603). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. and influenced by multiple factors, and it remains unclear whether the same holds true for cross-country ski training. This topic warrants further investigation, as cross-country skiing differs significantly from summer sports. Shedding light on this dose-response relationship can offer coaches valuable reference and data support. While a great deal of research in the past has focused on the Nordic countries, research on Chinese cross-country skiers could also enrich the training experience of emerging countries in cross-country ski training. Therefore, the purpose of this study was to: (i) analyze the load characteristics of 4 weeks cross-country skiing altitude training; (ii) analyze the relationships between methods of moni- toring training load and physiological indicators changes of elite male Chinese cross-country skiers during this period with an observation approach. Based on prior research in endurance sport, it was hypothesised that the training load characteristics conform to the classical endur- ance training model. Additionally, a load monitoring method that considers the physiological characteristics of the individual athlete is expected to provide the best dose-effect relationship. Materials and methods Participants and study design The study followed a descriptive, observational design, highlighting the relationship between methods of monitoring training load and physiological indicators changes during 4 weeks cross-country skiing altitude training (Fig 1). This study included 8 male athletes (age: 20.8 ±1.1 years; body mass 69.7±5.1 kg and height 179.6±5.9 cm) from the Chinese national cross- country skiing team, all of whom were elite level [27]. Although the sample size is small, it is something that often occurs in real-world studies of elite athletes, which is supported by previ- ous studies [11, 26]. They spent 4 weeks at the Chinese national snow sports training base (sea level: 1510–1700 m, 44.5˚ N) from 3 May 2021 to 30 May 2021. Athletes follow a regular training plan assigned by the coaching of the national training team [28, 29], Table 1 shows the training plan example for reference. Fig 1. The experimental procedure of this study. https://doi.org/10.1371/journal.pone.0295960.g001 PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 3 / 15 During this period the athletes’ training and life were systematically monitored for training load. The daily training load data was collated by the scientific staff filled in and summarized according to the relevant templates, including the collection and calculation of internal and external load data such as heart rate, sRPE, TRIMP, training time and distance. Before and after the altitude training, athletes did physiological performance test at the Beijing Erqi National Ice and Snow Research and Training Base (55m above sea level). The tests include maximal oxygen up-take tests, lactate threshold tests, body composition tests and skierg power tests. Athletes were examined by national team doctors to ensure that there were no injuries or illnesses and were in good physical condition. All participants freely completed an informed consent form before to this investigation, expressing their desire to voluntarily engage in this study. Physiological performance test In this study the lactate threshold test was completed prior to the maximal oxygen uptake (VO2max) test with a 5 minutes interval [30–32]. All participants performed an adequate warm-up prior to before the test. Using the treadmill (RL2500E, Rodby, So¨dertalje, Sweden) to evaluate the lactate threshold. The treadmill started at 7 km/h, the inclination angle was 10.5% and constant throughout, the speed increased by 1km/h every 5 minutes and the athlete has a 30 second rest period before each acceleration [30–32]. The average heart rate for the last 30 seconds of each phase was the athlete’s heart rate for that level. The athlete’s lactate is collected immediately at each step of the test and measured using a benchtop lactate meter (Boisen, EKF Industrial Electronics, Magdeburg, Germany). Ask athletes at the end of each step about their subjective feelings of fatigue (a 0–10 RPE scale). The lactate threshold refers to lactate level of 4 mmol L−1, treadmill speed at 4 mmol L−1 was calculated using linear interpolation [33]. VO2max test by athletes assessed using a portable gas metabolometer (MetaMax 3B, Cortex, Leipzig, Germany) 5 minutes after the lactate threshold test [30–32]. The treadmill was inclined at the same angle as the lactate threshold test, with the starting speed 1km/h lower Table 1. Example of a weekly training plan. Day Morning session Afternoon session Day 1 Free technique skiing 100 min: warm-up 45 min; aerobic training (15 seconds sprint + 30 seconds relax) * 5 times * 3 sets, anaerobic threshold velocity training; cool down 20 min Running 30 min; maximal strength training: 4 exercises for upper body * 3 sets + 2 exercises for lower body * 4 sets Day 2 Lactate Threshold Specialized Training Free technique roller-skiing 100 min; core training 30 min; 15–20 minutes stretching Day 3 Classical technique skiing 120 min Running 45 min; free technique ski-specific exercises 30 min; specialized strength training 15 min; swimming 30 min Day 4 Free technique roller-skiing 110 min: warm-up 30 min; uphill 10 sets, relax for 3 minutes between sets; cool down 20 min Classical technique roller-skiing 80 min; core training 20 min Day 5 Classical technique roller-skiing 150 min Running 30 min; slow strength 3 exercises for upper body * 5 sets, 2 exercises for lower body * 5 sets, 45s of exercise between sets, 30s intervals Day 6 Free technique skiing interval Training: warm-up 25 min; 5km * 2 sets, 3–4 min relaxed skating between sets; 10 minutes skating * 2–3 sets, 3–4 minutes relaxed skating between sets; cool down 20 min Running 25 min; circuit strength training, 8 exercises * 2 sets, 30 seconds exercise/30 seconds rest Day 7 Relax Relax https://doi.org/10.1371/journal.pone.0295960.t001 PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 4 / 15 than the lactate threshold speed and the speed increased by 1km/h per minute until the athlete was exhausted. During the maximal oxygen uptake test, the athlete continuously wore a venti- lation mask to measure the athlete’s oxygen uptake. The athlete’s heart rate was record by a heart rate belt (H10, Polar, Finland), the lactate concentration was measured 1 minute after the test and the RPE (a 0–10 scale) was recorded. The athlete’s maximum oxygen uptake (the average of the two highest and consecutive 30 s measurements), maximum HR (the highest 5s heart rate measurement), final treadmill speed and respiratory exchange ratio (RER) were recorded. Athletes did body composition test on the morning of the test day, which used a dualenergy X-ray bone density analyzer (Luna iDXA, General Electric Company, Schenectady, NY, USA) to analyze the muscle mass of the athletes’ upper body. Athletes performed a 30s power test using a ski ergometer (SKIERG, CONCEPT2 Inc., USA). The test was preceded by a 25-minute jogging warm-up led by a fitness coach, a 60m x 5 sets of acceleration runs, and a 1 minute x 3 sets of 75% maximum intensity warm-up using a ski ergometer. The skier was set to gear seven and the athletes performed the 30s ski ergometer test to the best of their ability. A researcher recorded the athlete’s average power output over 30 seconds and collated the records, the final result is a power to weight ratio. Methods of training load monitoring All training load data of the athletes were counted and recorded by the scientific staff accompa- nying the team. External load such as training distance, time and speed was obtained and recorded by the Catapult module (Catapult Sports, Melbourne, Australia). The module has a GPS sensor, which can be connected to the GPS satellite system, with a sampling rate was 10Hz; it can be used for more than 5 hours for continuous training and competition monitor- ing [34]. Each athlete wore a special undershirt that places the module in the middle of the shoulder blades. The module was turned on 10 minutes before training to ensure good signal reception. The same module was used by every athlete during the training camp for standardi- zation. The data were collected and analyzed after each training session by using the corporate software (Catapult openfield, Melbourne, Australia). Athletes wore heart rate bands and watches to monitor heart rate load-related data during daily training. In this study the athlete’s maximal heart rate and lactate threshold were obtained through physiological testing, therefore two methods of TRIMP calculation based on different heart rate zones standard were considered. The first TRIMP calculation method in this study is the approach eTRIMP proposed by Edward (Eq 1) [17]. eTRIMP ¼ 1  T1 þ 2  T2 þ 3  T3 þ 4  T4 þ 5  T5 ð1Þ T1: duration time when 50%HRmax <HR <60%HRmax; T2: duration time when 60% HRmax <HR <70%HRmax; T3: duration time when 70%HRmax <HR <80%HRmax; T4: duration time when 80%HRmax <HR <90%HRmax; T5: duration time when 90%HRmax <HR <100%HRmax. The second TRIMP calculation method is the way luTRIMP proposed by Lucia [18], as shown in Eq 2. luTRIMP ¼ 1  T1 þ 2  T2 þ 3  T3 ð2Þ T1: duration time when <HR1; T2: duration time when HR1 <HR <HR2; T3: duration time when >HR2. The HR1 corresponding to heart rate when blood lactate is 2 mmol L−1; the HR2 corresponding to heart rate when blood lactate is 4 mmol L−1. PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 5 / 15 All participants were asked for RPE 30 minutes after the training, which were tallied using a 0–10 subscale and recorded, as shown in Eq 3 [7, 20]. sRPE ¼ RPE  duration time ðminÞ ð3Þ The rate of the physiological indicators change was calculated by Eq 4. Physiological Indicators Changes in Percent %D ð Þ ¼ PostTest Discussion The purpose of this study were to: (i) analyze the load characteristics of 4 weeks cross-country ski- ing altitude training.; (ii) analyze the relationships between methods of monitoring training load and physiological indicators changes of elite Chinese cross-country skiers. The result revealed a " pyramid " pattern in the load characteristics of 8 Chinese male cross-country skiers during the alti- tude training cycle. The highest relationships were found in luTRIMP, the results of this study show that load monitoring methods that combine program characteristics and individual physio- logical characteristics have the highest relationships with physiological performance change, as opposed to internal and external load metrics that only combine average exercise intensity. The essence of sports training is the precise control of the training load, and detailed statis- tics on the training load can help to understand the athletes’ training situation in a certain cycle and make timely adjustments accordingly [21]. The load statistics show a " pyramid " training pattern in terms of overall training load and intensity during this period. Compared to similar training cycles for world elite cross-country skiers (polarised training pattern), endurance training time is relatively low at 1.4h (approximately 15.7h for the international elite level), strength training time is 1.6h higher per week (approximately 0.7h for the interna- tional elite level) and speed training time is similar. In terms of training load intensity, the average weekly LIT is 1.2h lower than that of international elite athletes (about 14.4h), the MIT Fig 2. Endurance distance and total training time during training camp. https://doi.org/10.1371/journal.pone.0295960.g002 PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 7 / 15 Fig 3. Training distribution during training camp. https://doi.org/10.1371/journal.pone.0295960.g003 Fig 4. Training intensity during training camp. https://doi.org/10.1371/journal.pone.0295960.g004 PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 8 / 15 is about 0.3h higher per week (about 0.5h for international elite level) and the HIT is about 0.5h lower per week (about 0.8h for international elite level). The reasons for the discrepancy with elite international athletes may be closely related to the short years of ski-specific training. Due to their short training period, athletes have not yet developed good glide economy and are prone to overrange glide heart rates during the long, low-intensity aerobic glides required by the training plan [2, 3]. The long duration of strength training is due to the fact that the national cross-country skiing coach, in response to the lack of specific strength in our athletes, has designed special strength training after long periods of low intensity aerobic training to help the athletes to achieve the appropriate strength requirements as much as possible. Scientific assessment of the relationship between external load, internal load and sports perfor- mance can help coaches and sports researchers to better examine the dose-response relationships resulting from sports training and to make a meaningful impact on training [10]. The strongest dose-response relationship between luTRIMP and physiological performance was found in this study, where it was shown that internal loads in cross-country skiers were closely related to maxi- mal oxygen uptake and lactate threshold heart rate, and that changes in these in-ternal load moni- toring methods were correlated to a moderate to high degree with changes in physiological performance [35]. Consistent results were also seen in the hurling [36] and soccer players [37]. The correlation between eTRIMP and physiological indicators changes in this study was much less than luTRIMP. The biggest reason is that eTRIMP was obtained based on data analysis of a certain group of athletes and not based on individual physiological performance tests, and the division of heart rate intervals deviated from the sports characteristics of cross-country skiing [26]. Some researchers have pointed out the limitations of load monitoring tools that do not incorporate the physiological characteristic points of the sport, as the intensity of an athlete’s train- ing can change depending on the environment, weather and tactics. In an experiment with cyclists, researchers found a weak dose-response relationship between bTRIMP and exercise per- formance based solely on the generalized exercise blood lactate equation [38]. sRPE as a load monitoring indicator has been widely used in the field of sports training [20, 39]. However, the dose-response relationship between sRPE and physiological indicators changes in this study was still lower than that between heart rate load-related metrics, which is Table 2. Physiological performance test result. Pre-Test Post-Test Mean Difference[95%Cl] ES Lat velocity(m
s−1) 3.02 ± 0.18 3.27 ± 0.24* 0.24 [0.05–0.44] 1.18 Lat HR (beats
min−1) 179.5 ± 9.8 188.5 ± 9.2** 9.00 [5.08–12.92] 0.95 VO2max (mL
min−1
kg−1) 73.74 ± 3.63 71.12 ± 3.14 -2.61 [-6.12 to 0.90] 0.77 VO2max (L
min−1) 4.82 ± 0.36 4.64 ± 0.43 -0.17 [-0.33 to 0.01] 0.45 Maximum HR in VO2max Test (beats
min−1) 197.4 ± 11.6 198.9 ± 9.1 1.50 [-3.72 to 6.72] 0.14 RER 1.23 ± 0.09 1.11 ± 0.03** -0.12 [-0.23 to -0.01] 1.79 Maximum La in VO2max Test(mmol
L−1) 13.31 ± 1.42 11.33 ± 2.79 -1.99 [-4.49 to 0.52] 0.89 Muscle mass whole body(kg) 55.66 ± 4.01 56.01 ± 4.02 0.36 [-0.04 to 0.76] 0.09 Muscle mass upper body(kg) 6.54 ± 0.42 6.74 ± 0.36** 0.20 [0.09–0.31] 0.51 Muscle mass trunk(kg) 27.20 ± 2.18 27.47 ± 2.00 0.27 [-0.62 to 1.16] 0.13 Muscle mass lower body(kg) 18.60 ± 1.82 18.57 ± 1.68 -0.03 [-7.34 to 0.68] 0.02 Skierg power-to-weight ratio(watt
kg-1) 4.38 ± 0.18 4.66 ± 0.16* 0.28 [-0.08 to 0.47] 1.64 * indicates a significant difference compared to pre-test ( * p < 0.05 ** p < 0.01) Abbreviations: Lat = lactate threshold, VO2max = maximal oxygen uptake, HR = heart rate, RER = respiratory exchange ratio. https://doi.org/10.1371/journal.pone.0295960.t002 PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 9 / 15 consistent with the better relationship between heart rate based load monitoring methods and oxygen consumption than RPE based methods derived by Wallace et al. [40]. It should also be considered that the short-term training intervention in this study and the relatively high alti- tude may have had an additional effect [41, 42]. It should also be considered that cross-country skiing differs from other sport groups in that the training pattern of prolonged exposure to cold temperatures may lead to a bias in the athletes’ subjective perception of fatigue, which in turn may lead to a bias in the RPE values compared to laboratory tests, resulting in an error in the sRPE monitoring of training load [43]. It should also be noted that load monitoring in this study was limited to the training period, with evidence suggesting that sRPE monitoring values during competition better reflect an athlete’s true internal load level. Furthermore, a signifi- cant difference was found in the psychological state of athletes between training and competi- tion, with the former being more monotonous and the latter generating stronger emotional responses, leading to notable psychological differences. Differences in training and competi- tion intensity can result in varying effects of load quantification [44]. The correlation between load monitoring methods and changes in physiological perfor- mance in this study, whilst moderate, is far less than has been found in practice in other proj- ects. Temperatures, wind speeds, snow quality and snow waxes in alpine environments differ from those in plains, and the accumulation of fatigue due to low oxygen exposure may lead to biased results. Differences in the duration of physiological performance tests may also contrib- ute to this result, with some studies showing a stronger effect relationship between longer duration physiological tests and load quantifiers [9]. Data collection from sports training prac- tice is more difficult to control than laboratory research design, which inevitably leads to vari- ability in the data collected [45]. Considering the above factors, the relationships in this study should be interpreted with caution. Therefore, in future studies that wish to apply load moni- toring methods to measure or analyze the dose effects of physiological performance in cross- country skiers in a given cycle, more consideration should be given to the influence of complex confounding factors on the results [46]. Limitations As an observational study, this study still has a few flaws in it. Firstly, the sample size in this study was relatively small due to design limitations, which is usually the norm in real-world Table 3. Relationship between methods of training load quantification and physiological indicators changes in percentage. Total Distance Total Time luTRIMP eTRIMP sRPE %Δ Lat Velocity .49[-.51 to .39] .60[-.13 to .93] .78*[-.01 to .98] .42[-.52 to .86] .62[-.31 to .94] %Δ Lat HR .44[-.37 to .92] .49[-.28 to .97] .71*[0.14–0.99] .38[-.39 to .91] .52[-.12 to .92] %Δ VO2maxRl .68[.31-.95] .39[-.52 to .97] .59[-.19 to .96] .61[-.06 to .99] .50[.01-.92] %Δ VO2maxAb .69[.33-.97] .42[-.54 to .93] .63[-.26 to .91] .57[-.14 to .98] .61[.21-.94] %Δ HRmax .81*[.36-.96] .61[-.20 to .98] .83*[.19–1.00] .79*[.29-.97] .66[-.24 to .97] %Δ Skierg W/Kg .62[-.34 to .94] .60[-.10 to .92] .75*[-.28 to .98] .50[-.28 to .86] .72*[-.41 to .96] %Δ Upperbody Muscle Mass .54[-.13 to .96] .78*[.44-.99] .62[-.11 to .94] .65[-.02 to .95] .53[-.34 to .97] * Indicates a significant at the 0.05 level (2-tailed) ** indicates significant at the 0.01 level (2-tailed). Abbreviations: %Δ Lat Velocity = percentage changes in lactate threshold velocity, %Δ Lat HR = percentage changes in lactate threshold HR, %Δ VO2maxRl = percentage changes in maximal oxygen uptake relative values, %Δ VO2maxAb = percentage changes in maximal oxygen uptake absolute values, %Δ HRmax = percentage changes in maximum HR, %Δ Skierg W/Kg = percentage changes in skierg power-to-weight ratio, %Δ Upperbody Muscle Mass = percentage changes in upperbody Muscle Mass, sRPE = session rating of perceived exertion, eTRIMP = Edwards training impulse, luTRIMP = Lucia training impulse. https://doi.org/10.1371/journal.pone.0295960.t003 PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 10 / 15 sports science research of elite athletes. Additionally, the training period only lasted for 4 weeks, which could limit the study’s ability to fully capture the physiological adaptations to altitude training, as well as the acute and chronic effects of applied training. Finally, the load monitoring methods used in this study were very limited due to practical constraints and more different load monitoring methods should be considered for future analysis. Therefore, future research should aim to expand the sample size and increase the number of data points to improve statistical analysis. Linear mixed models could be considered to mitigate the impact of individual differences on the outcomes [25]. It is important to note that competitive sports prioritise results, hence more attention should be given to sports performance indicators. Conclusion The study revealed a " pyramid " pattern in the load characteristics of 8 Chinese male cross- country skiers during the 4 weeks altitude training short cycle in preparation for the Beijing Fig 5. Relationship Between Average Weekly luTRIMP and Physiological Indicators Changes in Percent-age(N = 8), (a) %Δ Lat Velocity, (b) %Δ Lat HR, (c) % Δ HRmax, (d) %Δ Skierg W/Kg. Abbreviations: %Δ Lat Velocity = percentage changes in lactate threshold velocity, %Δ Lat HR = percentage changes in lactate threshold HR, %Δ HRmax = percentage changes in maxi-mum HR, %Δ Skierg W/Kg = percentage changes in skierg power-to-weight ratio, luTRIMP = Lucia training impulse. https://doi.org/10.1371/journal.pone.0295960.g005 PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 11 / 15 Winter Olympics, with a lower duration of LIT and a higher duration of MIT than at the elite international level. In the process of training on the altitude, it is important to control LIT while improving the other basic skills of the athletes to avoid fatigue and other phenomena. Training load monitoring methods that combine with program characteristics and individual physiological characteristics have the highest dose-response relationships with physiological indicators changes. Indicating that this is an important factor in cross-country skiing training load monitoring, luTRIMP could be used as a good internal load monitoring tool in cross- country skiing. The use of non-invasive load monitoring methods incorporating individual physiological characteristics of athletes should be considered in future cross-country skiing training monitoring to reveal changes in athletes’ physiological performance after a particular training intervention, in order to improve cross-country skiing training monitoring and load evaluation systems. Practical applications Training load monitoring can better help practitioners understand the load characteristics during a cycle, especially in specific training environments. For instance, when coaches iden- tify that athletes are experiencing high loads based on the monitoring outcomes, they should timely modify the training programme to avoid overreaching or overtraining. It is also valu- able for practitioners to strengthen the dose-respond relationship between load quantification methods and physiological indicators. This can enable the practitioners to understand the physiological adaptations that will occur in the athlete through training loads alone. Although the relationship can only be established for a specific group of athletes, providing an evidence- based framework can aid in the development of training programs. Moreover, this relationship might be impacted due to cold weather in winter sports. Practitioners ought to choose load monitoring methods that are specific to the training environment and program characteristics. Future research should consider better statistical methods based on more data to evaluate the dose-response relationship between longer training and more physiological indicators in cross-country skiing. Acknowledgments Thanks to the athletes who participated in the study and the Chinese National Olympic Committee. Author Contributions Conceptualization: Yifan Lu, Jing Mi. Data curation: Yichao Yu, Dongye Li. Formal analysis: Yifan Lu, Jing Mi. Funding acquisition: Yifan Lu. Investigation: Yichao Yu, Dongye Li. Methodology: Yifan Lu, Jing Mi. Project administration: Yichao Yu, Dongye Li, Yifan Lu. Resources: Yichao Yu, Dongye Li, Yifan Lu. Software: Yichao Yu, Dongye Li. Supervision: Jing Mi. PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 12 / 15 Writing – original draft: Yichao Yu, Dongye Li. Writing – review & editing: Yichao Yu, Dongye Li, Yifan Lu, Jing Mi. References 1. Sandbakk O, Hegge AM, Losnegard T, Skattebo O, Tonnessen E, Holmberg HC. The Physiological Capacity of the World’s Highest Ranked Female Cross-country Skiers. Med Sci Sports Exerc. 2016; 48 (6):1091–100. Epub 2016-6-1. https://doi.org/10.1249/MSS.0000000000000862 PMID: 26741124. 2. Sandbakk O, Holmberg HC. Physiological Capacity and Training Routines of Elite Cross-Country Ski- ers: Approaching the Upper Limits of Human Endurance. Int J Sports Physiol Perform. 2017; 12 (8):1003–1011. Epub 2017-9-1. https://doi.org/10.1123/ijspp.2016-0749 PMID: 28095083. 3. Sandbakk O, Holmberg HC, Leirdal S, Ettema G. The physiology of world-class sprint skiers. Scand J Med Sci Sports. 2011; 21(6):e9–16. Epub 2011-12-1. https://doi.org/10.1111/j.1600-0838.2010.01117. x PMID: 20500558. 4. Lianshi F. The Functional Diagnosis Methods in El ite Athlete and Problems. Shanghai Sport Science Research. 2003;(03):49–54. PubMed. 5. Gibson M, Pettitt R. Science and Practice of Strength Training. J Athl Train. 1998;33Epub 1998-1-1. PubMed. 6. Chandler J, Brown L. Conditioning for Strength and Human Performance, 3rd Edition. 2019. 7. Foster C, Rodriguez-Marroyo JA, de Koning JJ. Monitoring Training Loads: The Past, the Present, and the Future. Int J Sports Physiol Perform. 2017; 12(Suppl 2):S22–S28. Epub 2017-4-1. https://doi.org/ 10.1123/ijspp.2016-0388 PMID: 28253038. 8. Holmberg HC. The elite cross-country skier provides unique insights into human exercise physiology. Scand J Med Sci Sports. 2015;25 Suppl 4:100–9. Epub 2015-12-1. https://doi.org/10.1111/sms.12601 PMID: 26589123. 9. Meeusen R, Duclos M, Foster C, Fry A, Gleeson M, Nieman D, et al. Prevention, diagnosis, and treat- ment of the overtraining syndrome: joint consensus statement of the European College of Sport Science and the American College of Sports Medicine. Med Sci Sports Exerc. 2013; 45(1):186–205. Epub 2013- 1-1. https://doi.org/10.1249/MSS.0b013e318279a10a PMID: 23247672. 10. Mike RM, Mike M. Monitoring Training and Performance in Athletes: Human Kinetics, Inc.;Human Kinetics. 2017. 11. Xudan C, Lijuan M, Bei Z, Yongming L, Xiaoping C. Effects of Three Weeks 1 550 m Moderate Altitude Training on Physiological Capacity and Body Composition of High Level Young Male Cross-Country Skiers. CHINA SPORT SCIENCE AND TECHNOLOGY. 2022; 58(01):30–37. https://doi.org/10.16470/ j.csst.2021151 PubMed. 12. Gabbett TJ, Whyte DG, Hartwig TB, Wescombe H, Naughton GA. The relationship between workloads, physical performance, injury and illness in adolescent male football players. Sports Med. 2014; 44 (7):989–1003. Epub 2014-7-1. https://doi.org/10.1007/s40279-014-0179-5 PMID: 24715614. 13. Mujika I. The alphabet of sport science research starts with Q. Int J Sports Physiol Perform. 2013; 8 (5):465–6. Epub 2013-9-1. https://doi.org/10.1123/ijspp.8.5.465 PMID: 24058942. 14. Mujika I. Quantification of Training and Competition Loads in Endurance Sports: Methods and Applica- tions. Int J Sports Physiol Perform. 2017; 12(Suppl 2):S29–S217. Epub 2017-4-1. https://doi.org/10. 1123/ijspp.2016-0403 PMID: 27918666. 15. Talsnes RK, Berdal T, Brattebø JM, Seeberg T, Losnegard T, Kocbach J, et al. Comprehensive analy- sis of performance, physiological, and perceptual responses during an entire sprint cross-country skiing competition. Eur J Appl Physiol. 2023; Epub 2023-10-7. https://doi.org/10.1007/s00421-023-05326-w PMID: 37804364. 16. Banister EW, Carter JB, Zarkadas PC. Training theory and taper: validation in triathlon athletes. Eur J Appl Physiol Occup Physiol. 1999; 79(2):182–91. Epub 1999-1-1. https://doi.org/10.1007/ s004210050493 PMID: 10029340. 17. Edwards S. The Heart Rate Monitor Book; 1993. 18. Lucia A, Hoyos J, Santalla A, Earnest C, Chicharro JL. Tour de France versus Vuelta a España: which is harder? Med Sci Sports Exerc. 2003; 35(5):872–8. Epub 2003-5-1. https://doi.org/10.1249/01.MSS. 0000064999.82036.B4 PMID: 12750600. 19. Manzi V, Iellamo F, Impellizzeri F, D’Ottavio S, Castagna C. Relation between individualized training impulses and performance in distance runners. Med Sci Sports Exerc. 2009; 41(11):2090–6. Epub 2009-11-1. https://doi.org/10.1249/MSS.0b013e3181a6a959 PMID: 19812506. PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 13 / 15 20. Foster C, Florhaug JA, Franklin J, Gottschall L, Hrovatin LA, Parker S, et al. A new approach to monitor- ing exercise training. J Strength Cond Res. 2001; 15(1):109–15. Epub 2001-2-1. PMID: 11708692. 21. Halson SL. Monitoring training load to understand fatigue in athletes. Sports Med. 2014; 44 Suppl 2 (Suppl 2): S139–47. Epub 2014-11-1. https://doi.org/10.1007/s40279-014-0253-z PMID: 25200666. 22. Talsnes RK, Solli GS, Kocbach J, Torvik PØ, Sandbakk Ø. Laboratory- and field-based performance- predictions in cross-country skiing and roller-skiing. PLoS One. 2021; 16(8):e0256662. Epub 2021-1- 20. https://doi.org/10.1371/journal.pone.0256662 PMID: 34428258. 23. Akubat I. TRAINING LOAD MONITORING IN SOCCER. 2014. 24. Oliveira R, Cana´rio-Lemos R, Peixoto R, Vilac¸a-Alves J, Morgans R, Brito JP. The relationship between wellness and training and match load in professional male soccer players. PLoS One. 2023; 18(7): e0289374. Epub 2023-1-20. https://doi.org/10.1371/journal.pone.0289374 PMID: 37523395. 25. Costa J, Figueiredo P, Nakamura F, Rago V, Rebelo A, Brito J. Intra-individual variability of sleep and nocturnal cardiac autonomic activity in elite female soccer players during an international tournament. PLoS One. 2019; 14(9):e0218635. Epub 2019-1-20. https://doi.org/10.1371/journal.pone.0218635 PMID: 31527865. 26. Sanders D, Abt G, Hesselink M, Myers T, Akubat I. Methods of Monitoring Training Load and Their Relationships to Changes in Fitness and Performance in Competitive Road Cyclists. Int J Sports Phy- siol Perform. 2017; 12(5):668–675. Epub 2017-5-1. https://doi.org/10.1123/ijspp.2016-0454 PMID: 28095061. 27. McKay A, Stellingwerff T, Smith ES, Martin DT, Mujika I, Goosey-Tolfrey VL, et al. Defining Training and Performance Caliber: A Participant Classification Framework. Int J Sports Physiol Perform. 2022; 17(2):317–331. Epub 2022-2-1. https://doi.org/10.1123/ijspp.2021-0451 PMID: 34965513. 28. Grushin AA, Nageykina SV. Elite female cross-Country skiers1 training for major international sport competitions in mid-altitude areas. 2016:66–69. Epub 2016-1-1. PubMed. 29. Tonnessen E, Sylta O, Haugen TA, Hem E, Svendsen IS, Seiler S. The road to gold: training and peak- ing characteristics in the year prior to a gold medal endurance performance. PLoS One. 2014; 9(7): e101796. Epub 2014-1-20. https://doi.org/10.1371/journal.pone.0101796 PMID: 25019608. 30. Talsnes RK, Hetland TA, Cai X, Sandbakk O. Development of Performance, Physiological and Techni- cal Capacities During a Six-Month Cross-Country Skiing Talent Transfer Program in Endurance Ath- letes. Front Sports Act Living. 2020; 2:103. Epub 2020-1-20. https://doi.org/10.3389/fspor.2020.00103 PMID: 33345092. 31. Xudan C, Lijuan M, Bei Z, Kjøsen TR, SANDBAKK Ø, Tor-Arne H, et al. The Development of Physical Capacity of Talent-Transferring Athletes in Long Term Cross-Country Skiing Training—Based on Sports Physiological Evaluations. CHINA SPORT SCIENCE. 2021; 41(08):3–13. Epub 2021-8-15. PubMed. 32. Yu Y, Wang R, Li D, Lu Y. Monitoring Physiological Performance over 4 Weeks Moderate Altitude Train- ing in Elite Chinese Cross-Country Skiers: An Observational Study. Int J Environ Res Public Health. 2022;20(1)Epub 2022-12-24. https://doi.org/10.3390/ijerph20010266 PMID: 36612586. 33. Sjodin B, Jacobs I, Svedenhag J. Changes in onset of blood lactate accumulation (OBLA) and muscle enzymes after training at OBLA. Eur J Appl Physiol Occup Physiol. 1982; 49(1):45–57. Epub 1982-1- 19. https://doi.org/10.1007/BF00428962 PMID: 6213407 34. Parmley J, Jones B, Whitehead S, Rennie G, Hendricks S, Johnston R, et al. The speed and accelera- tion of the ball carrier and tackler into contact during front-on tackles in rugby league. J Sports Sci. 2023:1–9. Epub 2023-11-5. https://doi.org/10.1080/02640414.2023.2273657 PMID: 37925647. 35. Mourot L, Fabre N, Andersson E, Willis S, Buchheit M, Holmberg HC. Cross-country skiing and postex- ercise heart-rate recovery. Int J Sports Physiol Perform. 2015; 10(1):11–6. Epub 2015-1-1. https://doi. org/10.1123/ijspp.2013-0445 PMID: 24806737. 36. Malone S, Hughes B, Collins K, Akubat I. Methods of Monitoring Training Load and Their Association With Changes Across Fitness Measures in Hurling Players. J Strength Cond Res. 2020; 34(1):225– 234. Epub 2020-1-1. https://doi.org/10.1519/JSC.0000000000002655 PMID: 29985218. 37. Akubat I, Patel E, Barrett S, Abt G. Methods of monitoring the training and match load and their relation- ship to changes in fitness in professional youth soccer players. J Sports Sci. 2012; 30(14):1473–80. Epub 2012-1-20. https://doi.org/10.1080/02640414.2012.712711 PMID: 22857397. 38. Borresen J, Lambert MI. The quantification of training load, the training response and the effect on per- formance. Sports Med. 2009; 39(9):779–95. Epub 2009-1-20. https://doi.org/10.2165/11317780- 000000000-00000 PMID: 19691366. 39. Haddad M, Stylianides G, Djaoui L, Dellal A, Chamari K. Session-RPE Method for Training Load Moni- toring: Validity, Ecological Usefulness, and Influencing Factors. Front Neurosci. 2017; 11:612. Epub 2017-1-20. https://doi.org/10.3389/fnins.2017.00612 PMID: 29163016. PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 14 / 15 40. Wallace LK, Slattery KM, Impellizzeri FM, Coutts AJ. Establishing the criterion validity and reliability of common methods for quantifying training load. J Strength Cond Res. 2014; 28(8):2330–7. Epub 2014- 8-1. https://doi.org/10.1519/JSC.0000000000000416 PMID: 24662229. 41. Garvican LA, Pottgiesser T, Martin DT, Schumacher YO, Barras M, Gore CJ. The contribution of hae- moglobin mass to increases in cycling performance induced by simulated LHTL. Eur J Appl Physiol. 2011; 111(6):1089–101. Epub 2011-6-1. https://doi.org/10.1007/s00421-010-1732-z PMID: 21113616. 42. Pinot J, Grappe F. A six-year monitoring case study of a top-10 cycling Grand Tour finisher. J Sports Sci. 2015; 33(9):907–14. Epub 2015-1-20. https://doi.org/10.1080/02640414.2014.969296 PMID: 25357188. 43. Morton RH, Fitz-Clarke JR, Banister EW. Modeling human performance in running. J Appl Physiol (1985). 1990; 69(3):1171–7. Epub 1990-9-1. https://doi.org/10.1152/jappl.1990.69.3.1171 PMID: 2246166 44. van Erp T, Foster C, de Koning JJ. Relationship Between Various Training-Load Measures in Elite Cyclists During Training, Road Races, and Time Trials. Int J Sports Physiol Perform. 2019; 14(4):493– 500. Epub 2019-4-1. https://doi.org/10.1123/ijspp.2017-0722 PMID: 30300025. 45. Hoogeveen AR. The effect of endurance training on the ventilatory response to exercise in elite cyclists. Eur J Appl Physiol. 2000; 82(1–2):45–51. Epub 2000-5-1. https://doi.org/10.1007/s004210050650 PMID: 10879442. 46. Vermeire KM, Van de Casteele F, Gosseries M, Bourgois JG, Ghijs M, Boone J. The Influence of Differ- ent Training Load Quantification Methods on the Fitness-Fatigue Model. Int J Sports Physiol Perform. 2021; 16(9):1261–1269. Epub 2021-9-1. https://doi.org/10.1123/ijspp.2020-0662 PMID: 33691278. PLOS ONE Relationship between methods of monitoring training load and physiological indicators changes PLOS ONE | https://doi.org/10.1371/journal.pone.0295960 December 15, 2023 15 / 15

Relationship between methods of monitoring training load and physiological indicators changes during 4 weeks cross-country skiing altitude training.

12-15-2023

Yu, Yichao,Li, Dongye,Lu, Yifan,Mi, Jing

eng

PMC3081141

Subjects no. Genotype Gender Age (years) CAG repeat number Age at onset Independance scale score Functional capacity 1 HD M 40 55 31 70 4 2 HD M 36 50 35 90 11 3 HD M 55 43 40 70 6 4 HD M 53 47 49 100 13 5 HD M 58 42 50 50 2 6 Control M 36 7 Control M 50 8 Control M 48 9 Control M 27 10 Control M 48 Supplemental Table 1. Demographic and genetic data of HD patients and control subjects. Muscle biopsies were used for myoblast culture.

Low anaerobic threshold and increased skeletal muscle lactate production in subjects with Huntington's disease.

10-07-2010

Ciammola, Andrea,Sassone, Jenny,Sciacco, Monica,Mencacci, Niccolò E,Ripolone, Michela,Bizzi, Caterina,Colciago, Clarissa,Moggio, Maurizio,Parati, Gianfranco,Silani, Vincenzo,Malfatto, Gabriella

eng

PMC10000870

Citation: Casado, A.; Foster, C.; Bakken, M.; Tjelta, L.I. Does Lactate-Guided Threshold Interval Training within a High-Volume Low-Intensity Approach Represent the “Next Step” in the Evolution of Distance Running Training? Int. J. Environ. Res. Public Health 2023, 20, 3782. https://doi.org/10.3390/ ijerph20053782 Academic Editors: António Carlos Souse and Paul B. Tchounwou Received: 27 December 2022 Revised: 16 February 2023 Accepted: 17 February 2023 Published: 21 February 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). International Journal of Environmental Research and Public Health Review Does Lactate-Guided Threshold Interval Training within a High-Volume Low-Intensity Approach Represent the “Next Step” in the Evolution of Distance Running Training? Arturo Casado 1,* , Carl Foster 2, Marius Bakken 3 and Leif Inge Tjelta 4 1 Center for Sport Studies, Rey Juan Carlos University, 28933 Madrid, Spain 2 Department of Exercise and Sport Science, University of Wisconsin-LaCrosse, La Crosse, WI 54601, USA 3 Søm Medical Center, 4637 Kristiansand, Norway 4 Departament of Education and Sports Science, University of Stavanger, 4021 Stavanger, Norway * Correspondence: [email protected]; Tel.: +34-914888401 Abstract: The aim of the present study was to describe a novel training model based on lactate- guided threshold interval training (LGTIT) within a high-volume, low-intensity approach, which characterizes the training pattern in some world-class middle- and long-distance runners and to review the potential physiological mechanisms explaining its effectiveness. This training model consists of performing three to four LGTIT sessions and one VO2max intensity session weekly. In addition, low intensity running is performed up to an overall volume of 150–180 km/week. During LGTIT sessions, the training pace is dictated by a blood lactate concentration target (i.e., internal rather than external training load), typically ranging from 2 to 4.5 mmol·L−1, measured every one to three repetitions. That intensity may allow for a more rapid recovery through a lower central and peripheral fatigue between high-intensity sessions compared with that of greater intensities and, therefore, a greater weekly volume of these specific workouts. The interval character of LGTIT allows for the achievement of high absolute training speeds and, thus, maximizing the number of motor units recruited, despite a relatively low metabolic intensity (i.e., threshold zone). This model may increase the mitochondrial proliferation through the optimization of both calcium and adenosine monophosphate activated protein kinase (AMPK) signaling pathways. Keywords: running; performance; physiological adaptations; endurance sports; lactate; training monitoring 1. Introduction On 7 August 2021, 20-year-old Norwegian middle-distance runner Jakob Ingebrigtsen won the 1500 m Olympic title in Tokyo while breaking the Olympic and European records with a time of 3:28.32 (min:s). He also has won the World 5000 m and European 1500 m, 3000 m, 5000 m, and cross-country titles and owns the current indoor 1500 m world record (3:30.60 (min:s)). Further, his brothers Henrik and Filip, also Olympians, won the European 1500 m championships in 2012 and 2016, respectively. Their training pattern was described in a recent article [1] and is considered critical for their development as athletes. While it does not differ greatly from usual training modes in world-class runners [2,3], there is one specific characteristic which makes it unique and innovative: they were typically measuring their blood lactate concentration ([BLa]) during most of their high-intensity training sessions with the intent of matching a specific physiological intensity [1]. The main physiological performance determinants which account for success in dis- tance running events are: maximal oxygen uptake (VO2max) [4–6]; running economy (RE), defined as steady-state VO2 at a given submaximal speed or as the VO2 per unit of distance [5,7–9]; the ability to sustain a high percentage of VO2max during competition (% VO2max) [10–12]; the lactate threshold (LT), defined either as the velocity at which a non-linear increase in blood lactate occurs, the maximal lactate steady-state (MLSS), Int. J. Environ. Res. Public Health 2023, 20, 3782. https://doi.org/10.3390/ijerph20053782 https://www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2023, 20, 3782 2 of 15 or the velocity corresponding to a blood lactate concentration of 4 mmol·L−1 [13]; ve- locity at LT (vLT)/MLSS [14,15]; and the minimum velocity needed to achieve VO2max (vVO2max) [6,16,17]. To improve distance running performance, the training stimulus must enhance one or more of these factors [18]. The training stimulus represents the interaction among training volume (km per week), training frequency, and training intensity designed to enhance the aforementioned performance physiological determinants and performance in distance runners (19). The ideal relationship among these three training variables has, through several decades, been a topic of discussion in both the scientific [19–25] and coaching [26–30] literature. However, it remains unclear whether selecting the absolute training intensity com- posing the training stimulus through the control of an internal training load marker (i.e., blood lactate concentration) to match specific metabolic (relative) intensities may represent a training pattern optimizing the improvement of performance and its physiological deter- minants in distance runners. Accordingly, the present article aims to describe this training model and its similarities with those considered optimal according to the current scientific literature and examine the potential physiological mechanisms which may support its effectiveness. It would encourage the conduction of further intervention studies testing its influence on performance and its physiological determinants. If this model represents a more efficient training approach than those currently accepted, it may be useful for coaches and athletes, thereby optimizing performance in the latter. 2. Historical Trends in Distance Runners’ Training Principles During the last 100 years, the training principles used by middle- and long-distance runners have been inspired by training theories that provided success for contemporary outstanding runners. To a lesser extent, principles derived from physiological research have contributed to our understanding of how to train runners. In the 1920s and 1930s, international distance running was dominated by Finnish runners. The Finnish sports professor Lauri Pikhala inspired Pavo Nurmi (nine-time Olympic champion from 1920–1928 in events ranging from 1500 m to 10,000 m and cross country) and other Finnish runners with training principles he brought home from the United States. Their training system during the spring and summer seasons was a precursor to interval training [27]. Nurmi could, for instance, incorporate 6 × 400 m in 60 s into a slow run of 10 to 20 km in the forest [31]. The term “interval training” was introduced in the 1930s by the German coach Woldmar Gerschler and physician Herbert Reindel [27]. Their interval training represented a way to quantify the training load on the basis of repetitive runs to a heart rate of 180 beats/min, with a recovery interval to a heart rate of 120 beats/min. An interval training session consisted of repetitions of shorter (100 m to 400 m) runs. Gerschler was the coach of elite German middle-distance runners, such as Rudolf Harbig, who broke the 800 m world record in 1939 with a time of 1:46.6 (min:s). Importantly, in the 1930s, many years before the advent of portable heart rate monitors, accurately measuring a heart rate of 180 was nearly impossible, and the rationale for choosing “run to 180, recover to 120” is lost to history. Gösta Holmér was the coach of the Swedish runners Gunder Hägg and Arne Anderson who set numerous world records (WR) over distances from 1500 m to 5000 m in the 1940s. Holmér developed “fartlek”, which consisted of intensive efforts of varying distance and duration, interspersed with slower running [32]. It was very similar to Gerschler’s interval training but less formally organized and often conducted “by feel” in the forests rather than on a track. Czech runner Emil Zatopek, multiple-time Olympic champion in events from 5000 m to marathon from 1948 to 1952, typically performed an interval training regime consisting of a very high number of repetitions over 400 m (i.e., 60 × 400 m or 40 × 400 m with a recovery period between repetitions typically of a 200 m jog). The pace used and effort made during these repetitions were submaximal [27]. Int. J. Environ. Res. Public Health 2023, 20, 3782 3 of 15 An interval training regime was also used by Mihály Igloi, who coached Hungarian Sandor Iharos. Iharos broke WRs in the events ranging from 1500 m to 10,000 m during the 1950s. Training intensity during their intervals was higher than that used by Gerschler [27]. In the 1960s, the New Zealand coach Arthur Lydiard criticized the hard interval training regimes, primarily on the grounds that predicting when peak performance would occur was difficult. Lydiard proposed that effective distance running training should be founded on the basis of high volume of continuous low- to moderate-intensity running. He coached his countrymen Peter Snell (three-time Olympic gold medallist in 800 m and 1500 m between 1960 and 1964) and Murray Halberg (5000 m gold in 1960). His training philosophy involved a periodized training pattern. Three main training periods were completed: a 10–12 week preparation period which consisted mainly of high mileage of easy continuous running targeted at reaching 100 miles (160 km) per week, a 6–8 week period characterized by a high volume of hill running, and a 10–12 week competitive period consisting mainly of track interval training at or near race pace leading up to the main competition of the year [26]. In particular, the net effort during the competition period was fairly low, based on Lydiard’s saying, “you can’t train hard and race hard at the same time”. In the same general timeframe, German coach and physician Ernst Van Aaken proposed the Pure Endurance Training Method, which was based on very similar principles as those proposed by Lydiard, but without fixing a specific training volume (i.e., 100 miles per week), using hill repetitions and developing a periodized pattern. Van Aaken coached German runner Harald Norpoth, who achieved a silver medal at the 1964 Olympic Games in the 5000 m event [29]. In the 1970s and 1980s, many athletes who competed at an international level in distance running used a training regime based on Lydiard’s high volume of continuous training principle, but in contrast to Lydiard, they also incorporated sessions of interval training during the preparation period [33–35]. The “hard day–easy day” approach to training system is usually attributed to University of Oregon coaches Bill Bowerman and Bill Dellinger (bronze medallist in 5000 m in the 1964 Olympic Games), in which two to three high intensity interval sessions per week were separated by easier days (some with a training volume of <5 km/day) with continuous running [30,36]. From the 1970s and 1980s to the present day, most athletes have used a training regime consisting of two to five weekly sessions of interval training and/or longer tempo runs combined with a relatively high volume of easy and moderate intensity continuous run- ning [33,34,37,38]. A variety of sources have reported that successful distance runners have typically run between 120 and 250 km per week distributed across 11 to 18 sessions [37–41]. Most of these training characteristics have been determined through a ‘trial and error’ approach rather than by the outcomes of intervention studies. Furthermore, apart from Gerschler, internal physiological intensity control during high intensity interval sessions has rarely been proposed as a training strategy to improve performance. 3. External and Internal Training Load in Distance Running The training load, which refers to the interaction between training intensity and training volume, can be understood as either external (i.e., measurable aspects of training occurring externally to the athlete such as volume or intensity (i.e., running speed)) or internal (actual psychophysiological response that the body initiates to cope with the requirements elicited by the external load) [42]. Therefore, external load refers to the actual distance covered and speed achieved during a given training session. In turn, internal load can be measured through the monitoring of heart rate or [BLa]. While external training load represents an important reference to understand the performance evolution during the training process [3], it is generally believed that internal load may be the most accurate indicator of the effort for distance runners [22] as well as for other sports [42]. Accordingly, measuring internal training load (i.e., [BLa]) during training and using that information to control the absolute training intensity (i.e., speed or duration of repetitions) in order to Int. J. Environ. Res. Public Health 2023, 20, 3782 4 of 15 achieve the most optimal stimulus represents a conceptually attractive training protocol which agrees with current recommendations [42]. 4. Training Volume and Intensity Distribution Analysis in Runners Based on Their Internal Response to Exercise The aerobic–anaerobic transition as a framework for predicting performance in en- durance events was introduced in 1979 by Kindermann et al. [43]. During the last five decades, this framework has been espoused and updated by several scientists using either gas exchange or [BLa] markers [14,40,44,45]. During the last 50–60 years, several definitions related to the LT parameter have been presented [14]. Today it is common to refer to two breakpoints from a plot of the [BLa] during an incremental exercise test in a laboratory. The first threshold (LT1) was named aerobic threshold by Skinner and McLellan [44] and refers to the upper limit of aerobic metabolism. Intensities up to this point could last for hours. The second threshold or second lactate threshold (LT2) that has also been associated with the MLSS is known as the highest constant workload during continuous dynamic work, where there is an equilibrium between lactate production and lactate elimination [14,41,46,47]. At a slightly higher intensity than MLSS, the critical power (CP) concept, which is related to the hyperbolic relationship between speed or power output and the duration for which that speed or power output can be sustained, is an alternative approach to defining the maximal metabolic steady state [48]. According to these concepts, three training intensity zones (see Table 1) for endurance athletes are commonly used [22,49]. Zone 1 represents speeds below first ventilatory threshold or 2 mmol·L−1 [BLa]. Zone 2 is represented by speeds between the two ventilatory thresholds or 2 and 4.5 mmol·L−1 [BLa] (vLT1 and vLT2, respectively). Zone 3 represents speeds above vLT2 [50]. However, this classification does not differentiate between low- and high-intensity Zone 2 training, nor does it demarcate the different intensity zones that are in Zone 3, such as lactate tolerance and sprint training, being both above the VO2max intensity. Table 1. Intensity scale for distance runners. Scale [BLa] HR VO2max RPE Training Methods 6-Zone 3-Zone mmol·L−1 % Max % 6–20 SST (6) 3 n/a n/a n/a n/a Sprint VHIT (5) 3 8–18 >97 94–140 18–20 Lactate tolerance (i.e., 800 m and 1500 m pace) HIT (4) 3 4.5–8 92–97 88–94 16–18 Intensive aerobic interval (i.e., 5000 m pace) MIT (3) 2 3.5–4.5 87–92 84–88 14–16 Threshold training: interval running (10,000 m pace) MIT (2) 2 2–3.5 82–87 80–84 12–14 Threshold training: continuous/interval running (marathon pace) LIT (1) 1 0.7–2 62–82 55–80 9–12 Easy and moderate continuous running [BLa]: Blood lactate concentration; HR: heart rate; VO2max: maximal oxygen uptake; RPE: rate of perceived exertion according to original Borg scale; SST: short sprint training; VHIT; very-high-intensity training, HIT: high-intensity training; MIT: moderate-intensity training; LIT: low-intensity training; n/a: not applicable; numbers in parentheses in the first column refer to each zone of the 6-zone scale and numbers in the second column refer to each zone of the 3-zone scale. Furthermore, the transition between the different intensity zones does not follow clearly defined limits and are not anchored on exactly defined physiological markers [22]. The relationship between HR and [BLa] will also vary among different runners and in the same athlete across different training periods or seasons [51]. Table 1 describes the type of training performed, typical [BLa], typical % of HRmax, and % VO2max in the various zones for well-trained distance runners. Table 1 uses the intensity scales (i.e., three- and six-zone models) that will be referred to in this article and is elaborated upon according to previous suggestions [1,52,53]. Further mentions of training zones in the present article are referred to by the six-zone scale as z1, z2, . . . , and z6. Int. J. Environ. Res. Public Health 2023, 20, 3782 5 of 15 In order to analyze the effect of particular combinations of training volume and intensity in each of these zones, different training intensity distribution models (TID) have been described. 1. The pyramidal model is characterized by a decreasing training volume from z1 to z2 and z3, respectively. Approximately 70–80% of volume is covered in z1, with the remaining 20–30% in z2 and z3 [50]. 2. The polarized model is characterized by the completion of approximately 80% of the volume at z1, with most of the remaining 20% covered at z3 and as little training as possible in z2 [50]. 3. The threshold model features a greater proportion of overall volume conducted in z2 (i.e., >35%) than other models. According to recent reviews [3,54], either polarized or pyramidal approaches improved performance in distance runners to a greater extent than other models, which was also the main conclusion of a recent debate regarding which of the two models was more effective [24,25]. However, a more recent review reported that a pyramidal approach was typically adopted more often in highly trained and elite distance runners, despite the fact that polarized TID also appears to be effective [55]. Most importantly, a high-volume low-intensity approach is carried out in both the pyramidal and polarized TID models. 5. Physiological and Performance Development Using Lactate-Guided Threshold Interval Training (LTIT) within a High-Volume Low-Intensity Approach 5.1. Physiological Mechanisms Underpinning the Effectiveness of the Use of High Training Volume at Low Intensity Different hypotheses have been proposed to explain the underpinning mechanisms regarding the reason why a great proportion (70–80%) of overall training volume conducted at low intensity yields optimal performance development in endurance athletes who will race at comparatively high intensities (e.g., low specificity of training). The improvement of endurance performance through high volume of low/moderate continuous training is generated by sustaining increased cardiac output over a prolonged time (therefore augmenting oxygen delivery to working skeletal muscle) and by increased capacity for the oxidative metabolism through mitochondrial biogenesis and capillarization in Type I skeletal muscle fibers [56,57]. Importantly, the mozaic architecture of human skeletal muscle dictates that increased capillarization in Type I skeletal muscle fibers also serves to augment O2 delivery in Type II muscle fibers. Two primary signaling pathways for mitochondrial proliferation (both convergent on PGC1-α expression) exist. One is based on calcium signaling, which is more likely used with high-volume training [57,58], and the other is based on signaling derived from adenosine monophosphate (AMP)-activated protein kinase (AMPK) pathway, which is more likely used with high-intensity training, as [ATP] and AMP levels are reduced and increased, respectively [59,60]. As recruiting certain motor units elicited during competitive intensity exercise is needed in order to generate adaptative responses leading to increase mitochondrial density and aerobic metabolism, it can be achieved through the completion of at least a modicum of high-intensity training. The fact that most studies conclude that most of the training volume in distance runners should be covered at easy intensity to optimize performance development implies that adaptive potential of calcium signaling pathway is much larger than that of the AMPK signaling pathway. Accordingly, only relatively small training volume of the latter is needed to reach saturation in the adaptive response using this pathway [58,61]. Alternatively, evidence suggests that some homeostatic disturbances leading to failure to adapt to training (i.e., overtraining or non-functional overreaching) may be related to either inflammatory responses [62] or that slow autonomic recovery following high intensity training [63] may be caused by monotonic loads of high intensity training. These disturbances could lead to a reduction of the capacity for aerobic ATP generation through deficiencies in the mitochondrial electron transport chain or selective delivery of blood flow and/or reductions in maximal cardiac output [24]. Despite the mechanisms involved, Int. J. Environ. Res. Public Health 2023, 20, 3782 6 of 15 quasi-experimental observations [64,65] have suggested the negative effects of an excessive amount of high intensity training. The optimal combination of low- and high-intensity training is typically achieved with a hard day–easy day pattern which avoids monotony during the training process and may act to ensure a sufficient recovery period and to prevent non-functional overreaching. This may augment adaptive responses, such as gene expression for mitochondrial proliferation [50,66]. This specific training pattern is adopted by well-trained and elite long- and middle-distance runners [52,55,64,67–70]. However, evidence of the exact balance of different types of training on mitochondrial adaptive responses is limited. Particularly in already well-trained athletes, the range of options for achieving additional adaptive responses seems likely to be relatively small. Given the large volume of low-intensity training already performed by high-level athletes, further adaptive responses may largely lie in optimizing adaptive responses in Type II muscle fibers. 5.2. Physiological Mechanisms Explaining the Effectiveness of LT2 Intensity Training It is widely accepted that lactate metabolism serves as a useful index [40,71], although not likely as a cause [72], of muscular fatigue and that a strong correlation exists between lactate accumulation and level of performance in endurance events [15,73–76]. The rela- tionship between running intensity/speed and [BLa] is widely used to predict and identify performance in distance runners [14,40,74]. A strong correlation between the speed at vLT2/vMLSS and performance in long-distance running has been consistently observed, regardless of the method used to determine these physiological variables [73,77–79]. In this sense, Tjeta et al. [51] demonstrated that VO2max, RE, and %VO2max explained 89% of the variation in vLT2 among distance runners of national to international level. According to Billat et al. [41], vLT2/vMLSS is a running speed that a well-trained distance runner can sustain for approximately one hour (half-marathon pace for elite runners). Similarly, Roecker et al. [74] found that vLT2/vMLSS was slightly faster than half-marathon pace in 427 competitive runners. This was especially the case for the best runners. As vLT2 during continuous running is close to half-marathon pace, continuous tempo runs from 8–20 km are classified as threshold training in z2 and z3. Tempo runs have been included in the train- ing regime of distance runners from the 1970s up to now [39,70,80,81]. Casado et al. [82] found that elite Kenyan distance runners performed more of their total training volume as tempo runs compared with that in the best Spanish distance runners. The combination of high training volumes in z1 with moderate volumes in z2 and z3 is a very common pattern in contemporary distance runners. It generated improvements in performance [64,83] or has been associated with very high performance in highly trained and elite middle- and long-distance runners [41,67–69]. Furthermore, the use of this ap- proach was reported to be related to either high levels [67,69,70] or an improvement in RE [64,83]. Some research also found either improvements in [64,83–85] or were related to high levels of vVO2max [41,69,70]. A few studies, using high volumes in z1 and moderate volumes in z2 and z3, were associated with high levels of VO2max [67,70,86]. Studies us- ing this training pattern also found either improvements in [83,85] or were related with high levels of vLT2 [41,67,69,70]. In any case, there are comparatively few contempo- rary elite runners who have a total training volume <100 km/week, and most perform >160 km/week [53,55]. This approach has, in most cases, one primary characteristic in common, a high proportion of z2 and z3 training was covered at intensities at or near vLT2 (i.e., high intensity within z2–z3) [41,64,67–69,83,87]. The underpinning mechanisms explaining the relationship between training near/at vLT2 and the development of performance and its physiological determinants are not clear. However, it has been hypothesized that the use of this specific exercise intensity improves muscle specific adaptations, including clearing of lactate as opposed to reducing lactate production [88]. Since only recruited motor units are likely to experience increases in mitochondrial number and capillary density, with the exception that increases in capillary density in Type I muscle fibers may benefit O2 delivery to Type II muscle fibers, it may be speculated that training near vLT2 optimizes the number of motor units recruited Int. J. Environ. Res. Public Health 2023, 20, 3782 7 of 15 without having to accept the consequences of elevated levels of catecholamines likely to be experienced with z4 training. It is also important to consider that the speed associated with a [BLa] of 4 mmol·L−1 is somewhat specific to the pace of competitions in the 10–20 km range, which represents a large percentage of available competitions. Additionally, this velocity can be thought of as «speed work» for marathon runners. Sjodin et al. [76] tried to elucidate the effects of training at the speed associated with onset of [BLa] (vOBLA or speed associated with a [BLa] of 4 mmol·L−1) and the mechanisms involved explaining those effects in eight well-trained middle-distance runners. After the addition of one weekly training session consisting of 20 min of continuous running at vOBLA to their usual training regime for 8 weeks, the rate of glycogenolysis during exercise decreased (i.e., reduction of phosphofructokinase/citrate synthase ratio), while the potential to oxidize pyruvate and/or lactate increased (i.e., increased relative activity of heart-specific lactate dehydrogenase). These enzymatic changes were accompanied by an increase in vOBLA and/or a decrease of [BLa] at a same absolute speed. This specific training effect is displayed in Figure 1, which illustrates the evolution/displacement to the right of the lactate/speed curve yielded from an incremental intensity test (see Figure 1). vLT2 and the development of performance and its physiological determinants are not clear. However, it has been hypothesized that the use of this specific exercise intensity improves muscle specific adaptations, including clearing of lactate as opposed to reducing lactate production [88]. Since only recruited motor units are likely to experience increases in mitochondrial number and capillary density, with the exception that increases in capil- lary density in Type I muscle fibers may benefit O2 delivery to Type II muscle fibers, it may be speculated that training near vLT2 optimizes the number of motor units recruited without having to accept the consequences of elevated levels of catecholamines likely to be experienced with z4 training. It is also important to consider that the speed associated with a [BLa] of 4 mmol·L−1 is somewhat specific to the pace of competitions in the 10–20 km range, which represents a large percentage of available competitions. Additionally, this velocity can be thought of as «speed work» for marathon runners. Sjodin et al. [76] tried to elucidate the effects of training at the speed associated with onset of [BLa] (vOBLA or speed associated with a [BLa] of 4 mmol·L−1) and the mechanisms involved explaining those effects in eight well-trained middle-distance runners. After the addition of one weekly training session consisting of 20 min of continuous running at vOBLA to their usual training regime for 8 weeks, the rate of glycogenolysis during exercise decreased (i.e., reduction of phosphofructokinase/citrate synthase ratio), while the potential to oxi- dize pyruvate and/or lactate increased (i.e., increased relative activity of heart-specific lac- tate dehydrogenase). These enzymatic changes were accompanied by an increase in vO- BLA and/or a decrease of [BLa] at a same absolute speed. This specific training effect is displayed in Figure 1, which illustrates the evolution/displacement to the right of the lac- tate/speed curve yielded from an incremental intensity test (see Figure 1). Figure 1. Blood lactate concentration changes between two different incremental intensity tests char- acterized by a displacement of the lactate/speed curve to the right after including a certain amount of training at the velocity associated with the second lactate threshold during a training period in a hypothetical distance runner. In addition, these authors [76] found that the runners who were able to maintain [BLa] at 4 mmol·l−1 during the 20 min runs experienced greater performance improvement after the training period than runners who allowed [BLa] to “drift”. These data are the first to suggest that relatively tight control of [BLa] during training might be advanta- geous. Figure 1. Blood lactate concentration changes between two different incremental intensity tests characterized by a displacement of the lactate/speed curve to the right after including a certain amount of training at the velocity associated with the second lactate threshold during a training period in a hypothetical distance runner. In addition, these authors [76] found that the runners who were able to maintain [BLa] at 4 mmol·L−1 during the 20 min runs experienced greater performance improvement after the training period than runners who allowed [BLa] to “drift”. These data are the first to suggest that relatively tight control of [BLa] during training might be advantageous. 5.3. Potential Benefits of Lactate-Guided Threshold Interval Training In any case, the association between this physiological intensity (i.e., vLT2 or vOBLA) and speed is usually assumed when the run is continuous. However, manipulating the variables composing an interval training session (i.e., repetition velocity, duration, and inter-repetition recovery time) to match vLT2/vMLSS through [BLa] monitoring during the session may allow for the adoption of faster speeds (i.e., faster than those derived from continuous runs) and, thus, optimize the adaptive potential of muscle-fiber-type- specific adaptations required for race pace achievement (i.e., in middle-distance runners). In this sense, Kristensen et al. [89] demonstrated that an interval training program using a higher intensity than that derived from continuous exercise yielded a greater activation of AMP-activated protein kinase in Type II muscle fibers. In this way, conducting training Int. J. Environ. Res. Public Health 2023, 20, 3782 8 of 15 in z2 and z3 while recruiting Type II muscle fibers may provide the mechanical and metabolic advantages both of running close to race pace and at LT2 intensity, respectively. Furthermore, there is an additional advantage of covering interval training at LT2 intensity rather than in z4, which is related to fatigue generation. Burnley et al. [90] found that isometric quadriceps contractions conducted at 10% above the critical torque (i.e., just above LT2 intensity in z4) generated a rate of global and peripheral fatigue four to five times greater than that yielded by the same contractions at 10% below of critical torque (i.e., just below LT2 intensity in z3). These findings agree with the existence of a threshold in fatigue development dependent on whether exercise is carried out at, just below, or just above LT2 intensity. Accordingly, distance runners may benefit from covering some of their interval training sessions at z3 but at faster absolute speeds than vLT2 (assessed through a continuous incremental test) rather than in z4. Nonetheless, this should be done through short duration repetitions so that [BLa] does not progressively rise, as by doing so runners would be able to recover faster from ‘high-intensity’ training sessions. However, the use of intensities within z4–z5 has also been found to be useful in performance development in distance runners (2, 82). A recent systematic review by Rosenblat et al. [91] determined that high-intensity interval training at or below intensities of VO2max allows the improvement in central factors influencing VO2max, such as plasma volume, left ventricular mass, maximal stroke volume, and maximal cardiac output. However, peripheral factors influencing VO2max, such as skeletal muscle capillary density, maximal citrate synthase activity, and mitochondrial respiratory capacity in Type II fibers can be developed through sprint interval training (i.e., 30 s repetitions) [91]. Therefore, given that these physiological adaptations may not all be achieved through lower intensity training (especially those derived from sprint interval training), a certain but tolerable [65] amount of high intensity training within z4–z6 is also needed to improve performance optimally in distance runners. 6. Putting This Training Model into Practice These theoretical physiological advantages derived from LGTIT within a high-volume low-intensity model are attributed as beneficial by current Norwegian middle- and long- distance runners specialized in events ranging from 1500 m to 10,000 m. In the late 1990s, Marius Bakken (co-author of the present article), a Norwegian elite 5000 m runner, started to test a new training model on himself, which consisted of accumulating a high volume of training at an easy pace, a moderate volume of interval training at threshold intensity while controlling the pace through [BLa] testing during the session and including a low volume of interval training in z5 [92]. He typically covered 180 km overall, conducted four interval training sessions (i.e., two double sessions through a hard day–easy day pattern) at threshold intensity (i.e., at [BLa] ranging from 2 to 4.5 mmol·L−1 depending on the specific goal of the session) and one session at z5 per week [92]. Bakken experienced that when following LGTIT, he could perform a much higher training volume compared with that when he carried out interval training in z4. On the assumption that a higher total volume of training is associated with larger adaptive responses, this pattern might be thought of as beneficial. This assumption also agrees with findings of Burnley et al. [90] on the reduced fatigue generation at LT2 intensity when compared with that yielded by z4 training. Bakken developed this training model through a ‘trial and error’ approach and achieved a personal best time in 5000 m of 13:06.39 (min:s), which remains as the second all-time best Nordic best. He transmitted his training knowledge and experience to Gjert Ingebrigtsen, father and former coach of the three Ingebrigtsen brothers, who developed it for the achievement of their well-known athletic performances [92]. Bakken’s approach became a model for contemporary Norwegian runners, and much of the success of Norwegian huge runners at present is based on Bakken’s training principles. For example, Tokyo 2021 triathlon Olympic champion, Norwegian Kristian Blummenfelt, also used LGTIT [92]. This model has been developed within a successful system of endurance training. Norway, with a population of only 5.5 million, has similar men’s national records in distance running events to those of the USA: 1:42.58, 3:28.32, 7:27.05, and 12:48.45 (min:s) and 2:05:48 (h:min:s) for Int. J. Environ. Res. Public Health 2023, 20, 3782 9 of 15 the 800, 1500, 3000, and 5000m and marathon, respectively. For women, Norwegians have held some of the previous 3000, 5000, and 10,000 m and marathon world records. They also achieved the top national medal count for the cross-country and biathlon skiing events at the 2022 Winter Olympic Games, and both the triathlon 2019 and 2021 World Champion (Gustav Iden) and the aforementioned 2021 Olympic champion (Blummenfelt) are Norwegians [BLa] measurement and scientific testing are/were part of their training processes in most of these athletes. Interval-training performed with lactate values in z2 and z3 is also classified as thresh- old training even though the absolute speed at which they are performed can be faster than half-marathon pace. This is especially the case for shorter intervals, and the authors of this article have observed international level distance runners showing 20–25 × 400 m in 64 s average recovering 30 s between repetitions (13:20 (min:s) pace for 5000 m and 26:40 (min:s) for 10,000 m) and 20 × 400 m in 62 s average recovering 60 s between repeti- tions (12:55 (min:s) pace for 5000 m and, therefore, much faster than half-marathon pace), with [BLa] remaining below 4 mmol·L−1. The reason why this can be achieved is that duration of the running time/distance is too short for [BLa] to rise above LT2, and the rest period between repetitions is long enough for [BLa] to return to levels near LT1 but not long enough to decrease under that threshold. It has been reported that the Ingebrigtsen brothers conducted LGTIT over distances from 2000 m to 3000 m at close to half-marathon pace as well as over distances from 400 m to 1000 m at paces between 5000 m and 10,000 m race paces. The volume of this LGTIT sessions ranges between 8 and 12 km, and the recovery time between repetitions ranges between 20 s and 1.5 min. They often covered two LGTIT sessions in the same day and a fifth specific session at a much higher intensity in z4 or z5 (i.e., 20 × 200 m uphill jogging back in 70 s) (1, 67, 92). Their training intensity has been tightly controlled via measures of heart rate and [BLa] during all interval sessions (1). While the extensive use of LGTIT (i.e., up to four sessions per week) represents a novelty in the training of elite distance runners, several studies have reported the combined use of LT2 and z4/z5 training during the training week. For example, runners may conduct two (or more) different interval training sessions per week covered at LT2 and VO2max intensities, respectively (41, 68–70, 85). On the one hand, the addition of a greater number (i.e., two or three) of ‘high-intensity’ sessions to those typically observed in highly trained and elite runners may represent an advantage in training adaptation, as assimilating this higher training load may provide greater performance improvements. On the other hand, it also may represent an increased risk of injury/overtraining syndrome. Furthermore, the characteristics of LGTIT are different from those accepted in the current literature in distance runners given that traditionally LT2 training is conducted as continuous runs at much slower absolute speeds (31). Furthermore, the use of one sprint training session as well as some strength training sessions have been suggested as part of this training model [92]. In addition, it has been reported that this model involved the completion of a high training volume (i.e., 157–185 km/week) [67,92], which also agrees with the accepted efficacy of high training volume in elite distance runners [2,55,82]. However, the longest run does not exceed 21 km [92]. Finally, while no mention of the periodization approach adopted by these runners through this training model exists, the authors’ personal observations suggest that this training pattern involves the use of a traditional periodization approach, as observed in other elite distance runners [55]. Furthermore, during the competitive period, the z5 hill interval training session should be partly substituted for track workouts targeting competition pace at high [BLa] (i.e., from 5 to 10 mmol·L−1), and two LGTIT sessions are removed from the weekly plan. In this way, the goal during the competitive period is to achieve the minimum dose of threshold work which can sustain the previously developed aerobic base allowing for the completion of high volumes of competition pace above z3. This would be consistent with the current literature regarding optimal training periodization in highly trained and elite distance runners and shows a trend from a pyramidal TID during the preparatory period towards a polarized TID during the competitive period [38,55,69,85]. Int. J. Environ. Res. Public Health 2023, 20, 3782 10 of 15 The main goal of the present approach is to improve the speed while keeping [BLa] (and heart rate) stable during LGTIT sessions across the season. An example of speed and physiological responses (i.e., [BLa] and heart rate) responses during three similar LGTIT sessions conducted by Bakken during the 2003–2004 season, leading to his former Nordic 5000 m record of 13:06.39 (min:s) is highlighted in Figure 2 and shows the dramatic fitness improvement derived from the use of the present training model. Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 11 of 17 Figure 2. Average speed (A) and heart rate (C) per repetition, and post-repetition blood lactate con- centration (B) during three lactate-guided threshold interval training sessions conducted by Marius Bakken across the 2003–2004 season, leading to his former 5000 m Nordic record of 13:06.39 (min:s). Six ×1000, seven × 1000, and five × 1000 m with a recovery time between repetitions of one min were completed in December 2003 (mid-preparation period), February 2004 (late-preparation period), and June 2004 (competitive period), respectively. Rather than a revolutionary training model, it seems much more the result of an evo- lutionary pattern, as it is based on training practice which has been developed during the last 100 years of history of training in distance runners. Gerschler trained his athletes within a specific heart rate range; Zatopek covered interval training at submaximal paces and effort; Lydiard and Van Aaken established the need for developing a big aerobic base Figure 2. Average speed (A) and heart rate (C) per repetition, and post-repetition blood lactate con- centration (B) during three lactate-guided threshold interval training sessions conducted by Marius Bakken across the 2003–2004 season, leading to his former 5000 m Nordic record of 13:06.39 (min:s). Six × 2000, seven × 2000, and five × 2000 m with a recovery time between repetitions of one min were completed in December 2003 (mid-preparation period), February 2004 (late-preparation period), and June 2004 (competitive period), respectively. Int. J. Environ. Res. Public Health 2023, 20, 3782 11 of 15 Rather than a revolutionary training model, it seems much more the result of an evolutionary pattern, as it is based on training practice which has been developed during the last 100 years of history of training in distance runners. Gerschler trained his athletes within a specific heart rate range; Zatopek covered interval training at submaximal paces and effort; Lydiard and Van Aaken established the need for developing a big aerobic base through high training volumes at an easy pace; and Bowerman demonstrated the usefulness of a hard day–easy day basis. These characteristics were implemented during the training process of the Ingebrigtsen brothers. Other coaches and researchers also assisted in the development of an evidence-based and traditional training pattern, which helped these Norwegian coaches and scientists to generate this new and effective training model for distance runners. An example of one training week in which this training model is being used is described in Table 2. Table 2. Sample training week. Adapted from Bakken [92]. Morning Evening Monday 15 km (z1) 12 km (z1). Sprints (z5) and technique. Tuesday 5 km (z1). 5 × 6 min at 2.5 mmol·L−1 recovering (r.) 1 min between repetitions (z2). 2 km (z1) 5 km (z1). 10 × 1000 m at 3.5 mmol·L−1 recovering 1 min between repetitions (z2). 2 km (z1). Wednesday 16 km (z1). Strength training. 10 km (z1). Sprints (z5) and technique. Thursday 5 km (z1). 5 × 2 km at 2.5 mmol·L−1 recovering 1 min between repetitions (z2). 2 km (z1). 5 km (z1). 25 × 400 m at 3.5 mmol·L−1 recovering 30 s between repetitions (z2). 2 km (z1). Friday 15 km (z1). Rest. Saturday 5 km (z1). 20 × 200 m uphill at 8 mmol·L−1 recovering 70 s jogging back (z4). 2 km (z1). 10 km (z1). Sunday 21 km (z1). Rest. Z1–5: Zone 1 to Zone 5 according to the 6-zone scale; mmol·L−1 is a measure of blood lactate concentration. 7. Limitations, Future Studies, and Practical Applications The present article examined the current training regime of some of the best run- ners in the world and its derived potential physiological benefits on the basis of only observational studies and reports. Therefore, the assumptions stated previously should be taken cautiously since no controlled studies have tested the efficacy of this training model. Furthermore, whereas [BLa] ranges for training zones were suggested according to current recommendations [1,52,53] allowing for interindividual variability, specific values demarcating zones should be detected for each athlete through physiological tests [14]. Additionally, the training characteristics and its effects on performance and its develop- ment have been described only in 1500 m and 5000 m runners. Its applicability in other endurance events, such as the marathon, remains uncertain. However, our article presented sufficient evidence showing that these training characteristics display agreement with those reported in the current scientific literature in highly trained and elite distance runners. In addition, their differences may, in fact, be considered advantages of this new training approach from a physiological perspective: 1. The allowance of a greater number of ‘high-intensity’ sessions compared with adopt- ing a usual z4 interval training-based approach. 2. Achieving pre-established goals of internal load during the training session. 3. The possibility of adjusting and individualizing the specific training sessions within the model framework in a periodized approach (i.e., month by month, year by year, etc.). In this way, it is possible to accurately monitor not only the training adaptations being achieved without the need of specific tests but also the response to the different sessions through [BLa] measurements and make individual adjustments to the training program on the basis of this information. 4. Adaptation to altitude training while preventing excessive internal training loads derived from low air’s O2 partial pressure, given that [BLa] monitoring ensures that internal load remains at the pre-established levels. Int. J. Environ. Res. Public Health 2023, 20, 3782 12 of 15 For these reasons, new interventions comparing the physiological and performance effects of the previously described training characteristics with those of traditional training methods in highly trained distance runners are particularly encouraged. In this way, this new training model may represent an evolution of the training characteristics of highly trained and elite distance runners, and if future studies demonstrate its efficacy and safety, it may be implemented in other runners. Training characteristics and intensity distribution characterizing this training model and its derived potential physiological benefits are illustrated in Figure 3. Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 13 of 17 For these reasons, new interventions comparing the physiological and performance effects of the previously described training characteristics with those of traditional train- ing methods in highly trained distance runners are particularly encouraged. In this way, this new training model may represent an evolution of the training characteristics of highly trained and elite distance runners, and if future studies demonstrate its efficacy and safety, it may be implemented in other runners. Training characteristics and intensity distribution characterizing this training model and its derived potential physiological benefits are illustrated in Figure 3. Figure 3. Training characteristics and intensity distribution characterizing the training methodology described in the present article and its derived potential physiological mechanisms leading to per- formance improvement. LT1: first lactate threshold; LT2: second lactate threshold; vLT2: speed as- sociated to second lactate threshold; VO2max: maximum oxygen uptake; vVO2max: minimum speed needed to achieve maximum oxygen uptake; z1–6: Zone 1 to Zone 6 according to the 6-zone scale; AMPK: Adenosine monophosphate activated protein kinase; and PGC1-α: Peroxisome proliferator- activated receptor-γ coactivator. Author Contributions: A.C., L.I.T., M.B. and C.F. contributed to the design of the paper; A.C. and L.I.T. prepared the first draft of the manuscript. All authors have read and agreed to the published version of the manuscript. Funding: No funding was provided for the completion of the present study. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Acknowledgments: The present article is dedicated to the memory of one of its coauthors, the re- cently deceased Leif Inge Tjelta, who significantly contributed to the development of distance run- ning training research in Norway and worldwide. Conflicts of Interest: Arturo Casado, Carl Foster, Marius Bakken, and Leif Inge Tjelta have no con- ficts of interest relevant to the content of this article. References Figure 3. Training characteristics and intensity distribution characterizing the training methodol- ogy described in the present article and its derived potential physiological mechanisms leading to performance improvement. LT1: first lactate threshold; LT2: second lactate threshold; vLT2: speed associated to second lactate threshold; VO2max: maximum oxygen uptake; vVO2max: mini- mum speed needed to achieve maximum oxygen uptake; z1–6: Zone 1 to Zone 6 according to the 6-zone scale; AMPK: Adenosine monophosphate activated protein kinase; and PGC1-α: Peroxisome proliferator-activated receptor-γ coactivator. Author Contributions: A.C., L.I.T., M.B. and C.F. contributed to the design of the paper; A.C. and L.I.T. prepared the first draft of the manuscript. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Acknowledgments: The present article is dedicated to the memory of one of its coauthors, the recently deceased Leif Inge Tjelta, who significantly contributed to the development of distance running training research in Norway and worldwide. Conflicts of Interest: Arturo Casado, Carl Foster, Marius Bakken, and Leif Inge Tjelta have no conficts of interest relevant to the content of this article. Int. J. Environ. Res. Public Health 2023, 20, 3782 13 of 15 References 1. Tjelta, L.I. Three Norwegian Brothers All European 1500 m Champions: What Is the Secret? Int. J. Sport. Sci. Coach. 2019, 14, 694–700. [CrossRef] 2. Casado, A.; Hanley, B.; Santos-Concejero, J.; Ruiz-Pérez, L.M. World-Class Long-Distance Running Performances Are Best Predicted by Volume of Easy Runs and Deliberate Practice of Short-Interval and Tempo Runs. J. Strength Cond. Res. 2021, 35, 2525–2531. [CrossRef] 3. Kenneally, M.; Casado, A.; Santos-Concejero, J. The Effect of Periodization and Training Intensity Distribution on Middle-and Long-Distance Running Performance: A Systematic Review. Int. J. Sport. Physiol. Perform. 2018, 13, 1114–1121. [CrossRef] [PubMed] 4. Foster, C. VO2max and Training Indices as Determinants of Competitive Running Performance. J. Sport. Sci. 1983, 1, 13–22. [CrossRef] 5. Noakes, T.D.; Myburgh, K.H.; Schall, R. Peak Treadmill Running Velocity during the Vo2 Max Test Predicts Running Performance. J. Sport. Sci. 1990, 8, 35–45. [CrossRef] 6. Ingham, S.A.; Whyte, G.P.; Pedlar, C.; Bailey, D.M.; Dunman, N.; Nevill, A.M. Determinants of 800-m and 1500-m Running Performance Using Allometric Models. Med. Sci. Sport. Exerc. 2008, 40, 345–350. [CrossRef] 7. Conley, D.L.; Krahenbuhl, G.S. Running Economy and Distance Running Performance of Highly Trained Athletes. Med. Sci. Sport. Exerc. 1980, 12, 357–360. [CrossRef] 8. Morgan, D.W.; Martin, P.E.; Krahenbuhl, G.S. Factors Affecting Running Economy. Sport. Med. 1989, 7, 310–330. [CrossRef] 9. Foster, C.; Lucia, A. Running Economy: The Forgotten Factor in Elite Performance. Sport. Med. 2007, 37, 316–319. [CrossRef] [PubMed] 10. Bassett, D.R.; Howley, E.T. Limiting Factors for Maximum Oxygen Uptake and Determinants of Endurance Performance. Med. Sci. Sport. Exerc. 2000, 32, 70–84. [CrossRef] 11. Noakes, T. Physiological Capacity of the Elite Runner. In Running&Science—In An Interdisciplinary Perspective; Bangsbo, J., Larsen, H.B., Eds.; Institute of Exercise and Sport Sciences, University of Copenhagen: Copenhagen, Denmark, 2001; pp. 19–47. 12. Jones, A.M.; Kirby, B.S.; Clark, I.E.; Rice, H.M.; Fulkerson, E.; Wylie, L.J.; Wilkerson, D.P.; Vanhatalo, A.; Wilkins, B.W. Physiological Demands of Running at 2-Hour Marathon Race Pace. J. Appl. Physiol. 2021, 130, 369–379. [CrossRef] 13. Billat, L.V. Use of Blood Lactate Measurements for Prediction of Exercise Performance and for Control of Training. Recommenda- tions for Long-Distance Running. Sport. Med. 1996, 22, 157–175. [CrossRef] 14. Faude, O.; Kindermann, W.; Meyer, T. Lactate Threshold Concepts. Sport. Med. 2009, 39, 469–490. [CrossRef] [PubMed] 15. Tolfrey, K.; Hansen, S.A.; Dutton, K.; McKee, T.; Jones, A.M. Physiological Correlates of 2-Mile Run Performance as Determined Using a Novel on-Demand Treadmill. Appl. Physiol. Nutr. Metab. 2009, 34, 763–772. [CrossRef] 16. Daniels, J.; Scardina, N. Interval Training and Performance. Sport. Med. 1984, 1, 327–334. [CrossRef] 17. Bosquet, L.; Léger, L.; Legros, P. Methods to Determine Aerobic Endurance. Sport. Med. 2002, 32, 675–700. [CrossRef] 18. Joyner, M.J.; Coyle, E.F. Endurance Exercise Performance: The Physiology of Champions. J. Physiol. 2008, 586, 35–44. [CrossRef] [PubMed] 19. Midgley, A.W.; McNaughton, L.R.; Jones, A.M. Training to Enhance the Physiological Determinants of Long-Distance Running Performance: Can Valid Recommendations Be given to Runners and Coaches Based on Current Scientific Knowledge? Sport. Med. 2007, 37, 857–880. [CrossRef] 20. Brandon, L.J. Physiological Factors Associated with Middle Distance Running Performance. Sport. Med. 1995, 19, 268–277. [CrossRef] 21. Seiler, K.S.; Kjerland, G.Ø. Quantifying Training Intensity Distribution in Elite Endurance Athletes: Is There Evidence for an “Optimal” Distribution? Scand. J. Med. Sci. Sport. 2006, 16, 49–56. [CrossRef] [PubMed] 22. Seiler, S.; Tønnessen, E. Intervals, Thresholds, and Long Slow Distance: The Role of Intensity and Duration in Endurance Training. Training 2009, 13, 32–53. 23. Grant, G. No Short-Cuts to the Top. In Athletics Weekly; Descartes. Publishing Ltd.: Peterborough, UK, 2009; p. 26. 24. Foster, C.; Casado, A.; Esteve-Lanao, J.; Haugen, T.; Seiler, S. Polarized Training Is Optimal for Endurance Athletes. Med. Sci. Sport. Exerc. 2022, 54, 1028–1031. [CrossRef] [PubMed] 25. Burnley, M.; Bearden, S.E.; Jones, A.M. Polarized Training Is Not Optimal for Endurance Athletes. Med. Sci. Sport. Exerc. 2022, 54, 1032–1034. [CrossRef] 26. Lydiard, A.; Gilmour, G. Running to the Top; Meyer & Meyer Sport: Aachen, Germany, 1997. 27. Brook, N. Endurance Running Events; British Athletics Federation: Birmingham, England, 1992. 28. Martin, D.E.; Coe, P.N. Training Distance Runners; Leisure Press: Champaign, IL, USA, 1991. 29. Van Aaken, E. Van Aaken Method: Finding the Endurance to Run Faster and Live Healthier; World Publications: Mountain View, CA, USA, 1976. 30. Dellinger, B.; Freeman, B. The Competitive Runners Book; Collier Books: Springfield, OH, USA, 1984. 31. Billat, V.L. Interval Training for Performance: A Scientific and Empirical Practice. Special Recommendations for Middle- and Long-Distance Running. Part I: Aerobic Interval Training. Sport. Med. 2001, 31, 75–90. [CrossRef] 32. Holmer, G. Veien Til Rekorden; Instruksjonsbok i Friidrett: Oslo, Norway, 1947. 33. Temple, C. Cross Country and Road Running; Stanley Paul: London, UK, 1980. 34. Burfoot, A. Training the Hard/Easy Way. Run. World 1981, 16, 57–105. Int. J. Environ. Res. Public Health 2023, 20, 3782 14 of 15 35. Ferreira, P. Experience in Oporto. Track Field Q. Rev. 1983, 83, 38–41. 36. Moritani, T.; DeVries, H.A. Neural Factors versus Hypertrophy in the Time Course of Muscle Strength Gain. Am. J. Phys. Med. 1979, 58, 115–130. 37. Casado, A.; Tjelta, L.I. Training Volume and Intensity Distribution among Elite Middle- and Long-Distance Runners. In The Science and Practice of Middle- and Long-Distance Running; Blagrove, R., Hayes, P., Eds.; Routledge: New York, NY, USA, 2021. 38. Tjelta, L.I. The Training of International Level Distance Runners. Int. J. Sport. Sci. Coach 2016, 11, 122–134. [CrossRef] 39. Kaggestad, J. So Trainiert Ingrid Kristiansen 1986. Leichtatletik 1987, 38, 831–834. 40. Jones, A.M. The Physiology of the World Record Holder for the Women’s Marathon. Int. J. Sport. Sci. Coach 2006, 1, 101–116. [CrossRef] 41. Billat, V.; Lepretre, P.M.; Heugas, A.M.; Laurence, M.H.; Salim, D.; Koralsztein, J.P. Training and Bioenergetic Characteristics in Elite Male and Female Kenyan Runners. Med. Sci. Sport. Exerc. 2003, 35, 297–304. [CrossRef] 42. Impellizzeri, F.M.; Marcora, S.M.; Coutts, A.J. Internal and External Training Load: 15 Years On. Int. J. Sport. Physiol. Perform. 2019, 14, 270–273. [CrossRef] [PubMed] 43. Kindermann, W.; Simon, G.; Keul, J. The Significance of the Aerobic-Anaerobic Transition for the Determination of Work Load Intensities during Endurance Training. Eur. J. Appl. Physiol. Occup. Physiol. 1979, 42, 25–34. [CrossRef] [PubMed] 44. Skinner, J.S.; McLellan, T.H. The Transition from Aerobic to Anaerobic Metabolism. Res. Q Exerc. Sport. 1980, 51, 234–248. [CrossRef] [PubMed] 45. Meyer, T.; Lucía, A.; Earnest, C.P.; Kindermann, W. A Conceptual Framework for Performance Diagnosis and Training Prescription from Submaximal Gas Exchange Parameters—Theory and Application. Int. J. Sport. Med. 2005, 26, S38–S48. [CrossRef] [PubMed] 46. Midgley, A.W.; McNaughton, L.R.; Wilkinson, M. Is There an Optimal Training Intensity for Enhancing the Maximal Oxygen Uptake of Distance Runners? Sport. Med. 2006, 36, 117–132. [CrossRef] [PubMed] 47. Beneke, R.; Leithäuser, R.M.; Ochentel, O. Blood Lactate Diagnostics in Exercise Testing and Training. Int. J. Sport. Physiol. Perform. 2011, 6, 8–24. [CrossRef] 48. Jones, A.M.; Burnley, M.; Black, M.I.; Poole, D.C.; Vanhatalo, A. The Maximal Metabolic Steady State: Redefining the ‘Gold Standard’. Physiol. Rep. 2019, 7, e14098. [CrossRef] 49. Stöggl, T.L.; Sperlich, B. The Training Intensity Distribution among Well-Trained and Elite Endurance Athletes. Front. Physiol. 2015, 6, 295. [CrossRef] 50. Seiler, S. What Is Best Practice for Training Intensity and Duration Distribution in Endurance Athletes? Int. J. Sport. Physiol. Perform. 2010, 5, 276–291. [CrossRef] 51. Tjelta, L.I.; Tjelta, A.R.; Dyrstad, S.M. Relationship between Velocity at Anaerobic Threshold and Factors Affecting Velocity at Anaerobic Threshold in Elite Distance Runners. IJASS Int. J. Appl. Sport. Sci. 2012, 24, 8–17. [CrossRef] 52. Haugen, T.; Sandbakk, O.; Enoksen, E.; Seiler, S.; Tonnessen, E. Crossing the Golden Divide: The Science Nd Practice of Training World-Class 800- and 1500-m Runners. Sport. Med. 2020, in press. [CrossRef] [PubMed] 53. Haugen, T.; Sandbakk, Ø.; Seiler, S.; Tønnessen, E. The Training Characteristics of World-Class Distance Runners: An Integration of Scientific Literature and Results-Proven Practice. Sport. Med. Open 2022, 8, 46. [CrossRef] [PubMed] 54. Campos, Y.; Casado, A.; Vieira, J.G.; Guimarães, M.; Sant’Ana, L.; Leitão, L.; da Silva, S.F.; Silva Marques De Azevedo, P.H.; Vianna, J.; Domínguez, R. Training-Intensity Distribution on Middle- And Long-Distance Runners: A Systematic Review. Int. J. Sport. Med. 2021, 43, 305–316. [CrossRef] 55. Casado, A.; González-Mohíno, F.; González-Ravé, J.M.; Foster, C. Training Periodization, Methods, Intensity Distribution, and Volume in Highly Trained and Elite Distance Runners: A Systematic Review. Int. J. Sport. Physiol. Perform. 2022, 17, 1–14. [CrossRef] 56. Egan, B.; Zierath, J.R. Exercise Metabolism and the Molecular Regulation of Skeletal Muscle Adaptation. Cell Metab. 2013, 17, 162–184. [CrossRef] 57. Van der Zwaard, S.; Brocherie, F.; Jaspers, R.T. Under the Hood: Skeletal Muscle Determinants of Endurance Performance. Front. Sport. Act. Living 2021, 3, 719434. [CrossRef] [PubMed] 58. Bishop, D.J.; Botella, J.; Granata, C. CrossTalk Opposing View: Exercise Training Volume Is More Important than Training Intensity to Promote Increases in Mitochondrial Content. J. Physiol. 2019, 597, 4115–4118. [CrossRef] 59. MacInnis, M.J.; Gibala, M.J. Physiological Adaptations to Interval Training and the Role of Exercise Intensity. J. Physiol. 2017, 595, 2915–2930. [CrossRef] 60. Gibala, M.J.; McGee, S.L.; Garnham, A.P.; Howlett, K.F.; Snow, R.J.; Hargreaves, M. Brief Intense Interval Exercise Activates AMPK and P38 MAPK Signaling and Increases the Expression of PGC-1α in Human Skeletal Muscle. J. Appl. Physiol. 2009, 106, 929–934. [CrossRef] 61. MacInnis, M.J.; Skelly, L.E.; Gibala, M.J. CrossTalk Proposal: Exercise Training Intensity Is More Important than Volume to Promote Increases in Human Skeletal Muscle Mitochondrial Content. J. Physiol. 2019, 597, 4111–4113. [CrossRef] [PubMed] 62. Meeusen, R.; Duclos, M.; Foster, C.; Fry, A.; Gleeson, M.; Nieman, D.; Raglin, J.; Rietjens, G.; Steinacker, J.; Urhausen, A. Prevention, Diagnosis, and Treatment of the Overtraining Syndrome: Joint Consensus Statement of the European College of Sport Science and the American College of Sports Medicine. Med. Sci. Sport. Exerc. 2013, 45, 186–205. [CrossRef] 63. Seiler, S.; Haugen, O.; Kuffel, E. Autonomic Recovery after Exercise in Trained Athletes: Intensity and Duration Effects. Med. Sci. Sport. Exerc. 2007, 39, 1366–1373. [CrossRef] [PubMed] 64. Billat, V.L.; Flechet, B.; Petit, B.; Muriaux, G.; Koralsztein, J.P. Interval Training at VO(2max): Effects on Aerobic Performance and Overtraining Markers. Med. Sci. Sport. Exerc. 1999, 31, 156–163. [CrossRef] [PubMed] Int. J. Environ. Res. Public Health 2023, 20, 3782 15 of 15 65. Esteve-Lanao, J.; Foster, C.; Seiler, S.; Lucia, A. Impact of Training Intensity Distribution on Performance in Endurance Athletes. J. Strength Cond. Res. 2007, 21, 943–949. [CrossRef] 66. Foster, C. Monitoring Training in Athletes with Reference to Overtraining Syndrome. Med. Sci. Sport. Exerc. 1998, 30, 1164–1168. [CrossRef] 67. Tjelta, L.I. A Longitudinal Case Study of the Training of the 2012 European 1500 m Track Champion. Int. J. Appl. Sport. Sci. 2013, 25, 11–18. [CrossRef] 68. Tjelta, L.I.; Enoksen, E. Training Characteristics of Male Junior Cross Country and Track Runners on European Top Level. Int. J. Sport. Sci. Coach. 2010, 5, 193–203. [CrossRef] 69. Kenneally, M.; Casado, A.; Gomez-Ezeiza, J.; Santos-Concejero, J. Training Characteristics of a World Championship 5000-m Finalist and Multiple Continental Record Holder Over the Year Leading to a World Championship Final. Int. J. Sports Physiol. Perform. 2021, 17, 1–5. [CrossRef] 70. Kenneally, M.; Casado, A.; Gómez-Ezeiza, J.; Santos-Concejero, J. Training Intensity Distribution Analysis by Race Pace vs. Physiological Approach in World-Class Middle- and Long-Distance Runners. Eur. J. Sport. Sci. 2020, 21, 819–826. [CrossRef] 71. Robergs, R.A.; Ghiasvand, F.; Parker, D. Biochemistry of Exercise-Induced Metabolic Acidosis. Am. J. Physiol. -Regul. Integr. Comp. Physiol. 2004, 287, R502–R516. [CrossRef] [PubMed] 72. Brooks, G.A. Lactate. Sport. Med. 2007, 37, 341–343. [CrossRef] [PubMed] 73. Maffulli, N.; Capasso, G.; Lancia, A. Anaerobic Threshold and Performance in Middle and Long Distance Running. J. Sport. Med. Phys. Fit. 1991, 31, 332–338. 74. Roecker, K.; Schotte, O.; Niess, A.M.; Horstmann, T.; Dickhuth, H.H. Predicting Competition Performance in Long-Distance Running by Means of a Treadmill Test. Med. Sci. Sport. Exerc. 1998, 30, 1552–1557. [CrossRef] 75. Jones, A.M.; Carter, H. The Effect of Endurance Training on Parameters of Aerobic Fitness. Sport. Med. 2000, 29, 373–386. [CrossRef] [PubMed] 76. Sjodin, B.; Jacobs, I.; Svedenhag, J. Changes in Onset of Blood Lactate Accumulation (OBLA) and Muscle Enzymes after Training at OBLA. Eur. J. Appl. Physiol. Occup. Physiol. 1982, 49, 45–57. [CrossRef] 77. Yoshida, T.; Udo, M.; Iwai, K.; Chida, M.; Ichioka, M.; Nakadomo, F.; Yamaguchi, T. Significance of the Contribution of Aerobic and Anaerobic Components to Several Distance Running Performances in Female Athletes. Eur. J. Appl. Physiol. Occup. Physiol. 1990, 60, 249–253. [CrossRef] [PubMed] 78. Nicholson, R.M.; Sleivert, G.G. Indices of Lactate Threshold and Their Relationship with 10-Km Running Velocity. Med. Sci. Sport. Exerc. 2001, 33, 339–342. [CrossRef] 79. Billat, V.L.; Sirvent, P.; Py, G.; Koralsztein, J.-P.; Mercier, J. The Concept of Maximal Lactate Steady State. Sport. Med. 2003, 33, 407–426. [CrossRef] 80. Tjelta, L.I.; Enoksen, E. Training Volume and Intensity. In Running and Science—In An Interdisciplinary Perspective; Bangsbo, J., Larsen, H.B., Eds.; Institute of Exercise and Sport Sciences, University of Copenhagen: Copenhagen, Denmark, 2001; pp. 149–177. 81. Tjelta, L.; Tønnessen, E.; Enoksen, E. A Case Study of the Training of Nine Times New York Marathon Winner Grete Waitz. Int. J. Sport. Sci. Coach. 2014, 9, 139–157. [CrossRef] 82. Casado, A.; Hanley, B.; Ruiz-Pérez, L.M. Deliberate Practice in Training Differentiates the Best Kenyan and Spanish Long-Distance Runners. Eur. J. Sport. Sci. 2020, 20, 887–895. [CrossRef] [PubMed] 83. Enoksen, E.; Shalfawi, S.A.I.; Tønnessen, E. The Effect of High-vs. Low-Intensity Training on Aerobic Capacity in Well-Trained Male Middle-Distance Runners. J. Strength Cond. Res. 2011, 25, 812–818. [CrossRef] 84. Galbraith, A.; Hopker, J.; Cardinale, M.; Cunniffe, B.; Passfield, L. A 1-Year Study of Endurance Runners: Training, Laboratory Tests, and Field Tests. Int. J. Sport. Physiol. Perform. 2014, 9, 1019–1025. [CrossRef] [PubMed] 85. Filipas, L.; Bonato, M.; Gallo, G.; Codella, R. Effects of 16 Weeks of Pyramidal and Polarized Training Intensity Distributions in Well-trained Endurance Runners. Scand. J. Med. Sci. Sport. 2021, 32, 498–511. [CrossRef] [PubMed] 86. Robinson, D.M.; Robinson, S.M.; Hume, P.A.; Hopkins, W.G. Training Intensity of Elite Male Distance Runners. Med. Sci. Sport. Exerc. 1991, 23, 1078–1082. [CrossRef] 87. Ingham, S.A.; Fudge, B.W.; Pringle, J.S. Training Distribution, Physiological Profile, and Performance for a Male International 1500-m Runner. Int. J. Sport. Physiol. Perform. 2012, 7, 193–195. [CrossRef] [PubMed] 88. Philp, A.; Macdonald, A.L.; Carter, H.; Watt, P.W.; Pringle, J.S. Maximal Lactate Steady State as a Training Stimulus. Int. J. Sport. Med. 2008, 29, 475–479. [CrossRef] 89. Kristensen, D.E.; Albers, P.H.; Prats, C.; Baba, O.; Birk, J.B.; Wojtaszewski, J.F.P. Human Muscle Fibre Type-Specific Regulation of AMPK and Downstream Targets by Exercise. J. Physiol. 2015, 593, 2053–2069. [CrossRef] 90. Burnley, M.; Vanhatalo, A.; Jones, A.M. Distinct Profiles of Neuromuscular Fatigue during Muscle Contractions below and above the Critical Torque in Humans. J. Appl. Physiol. 2012, 113, 215–223. [CrossRef] 91. Rosenblat, M.A.; Granata, C.; Thomas, S.G. Effect of Interval Training on the Factors Influencing Maximal Oxygen Consumption: A Systematic Review and Meta-Analysis. Sport. Med. 2022, 52, 1329–1352. [CrossRef] 92. Bakken, M. The Norwegian Model of Lactate Threshold Training and Lactate Controlled Approach to Training. Available online: www.mariusbakken.com/the-norwegian-model.html (accessed on 4 April 2022). Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Does Lactate-Guided Threshold Interval Training within a High-Volume Low-Intensity Approach Represent the "Next Step" in the Evolution of Distance Running Training?

02-21-2023

Casado, Arturo,Foster, Carl,Bakken, Marius,Tjelta, Leif Inge

eng

PMC5451085

RESEARCH ARTICLE Wearing lower-body compression garment with medium pressure impaired exercise- induced performance decrement during prolonged running Sahiro Mizuno1, Mari Arai2☯, Fumihiko Todoko2☯, Eri Yamada2☯, Kazushige Goto3* 1 Graduate School of Sports and Health Science, Ritsumeikan University, Kusatsu, Shiga, Japan, 2 DESCENTE Ltd., Osaka, Japan, 3 Faculty of Sports and Health Science, Ritsumeikan University, Kusatsu, Shiga, Japan ☯ These authors contributed equally to this work. * [email protected] Abstract Objective To investigate the effect of wearing a lower body compression garment (CG) exerting differ- ent pressure levels during prolonged running on exercise-induced muscle damage and the inflammatory response. Methods Eight male participants completed three exercise trials in a random order. The exercise con- sisted of 120 min of uphill running at 60% of VO2max. The exercise trials included 1) wearing a lower-body CG with 30 mmHg pressure [HIGH]; 2) wearing a lower-body CG with 15 mmHg pressure [MED]; and 3) wearing a lower-body garment with < 5 mmHg pressure [CON]. Heart rate (HR), and rate of perceived exertion for respiration and legs were moni- tored continuously during exercise. Time-course change in jump height was evaluated before and immediately after exercise. Blood samples were collected to determine blood glucose, lactate, serum creatine kinase, myoglobin, free fatty acids, glycerol, cortisol, and plasma interleukin-6 (IL-6) concentrations before exercise, 60 min of the 120 min exercise period, immediately after exercise, and 60 min after exercise. Results Jump height was significantly higher immediately after the exercise in the MED trial com- pared with that in the HIGH trial (P = 0.04). Mean HR during the 120 min exercise was signif- icantly lower in the MED trial (162 ± 4 bpm) than that in the CON trial (170 ± 4 bpm, P = 0.01). Plasma IL-6 concentrations increased significantly with exercise in all trials, but the area under the curve during exercise was significantly lower in the MED trial (397 ± 58 pg/ ml
120 min) compared with that in the CON trial (670 ± 86 pg/ml
120 min, P = 0.04). PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 1 / 12 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Mizuno S, Arai M, Todoko F, Yamada E, Goto K (2017) Wearing lower-body compression garment with medium pressure impaired exercise- induced performance decrement during prolonged running. PLoS ONE 12(5): e0178620. https://doi. org/10.1371/journal.pone.0178620 Editor: Massimo Sacchetti, Universita degli Studi di Roma ’Foro Italico’, ITALY Received: December 27, 2016 Accepted: May 16, 2017 Published: May 31, 2017 Copyright: © 2017 Mizuno et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper (not within Supporting Information files). Funding: This work was supported by the Ishimoto Memorialfoundation for the Promotion of Sports Science (http://www.descente.co.jp/ishimoto/) and the Japan Society for the Promotion of Science, Grant number; 15K12660 (http://www.jsps.go.jp/ english/e-grants/index.html). The funders had no role in study design, data collection and analysis, Conclusion Wearing a lower body CG exerting medium pressure (approximately 15 mmHg) significantly attenuated decrease in jump performance than that with wearing a lower body CG exerting high pressure (approximately 30 mmHg). Furthermore, exercise-induced increases in HR and the inflammatory response were significantly smaller with CG exerted 15mmHg than that with garment exerted < 5 mmHg. Introduction During the past two decades, use of compression garment (CG) during exercise has been prev- alent among a variety of team sport athletes as a strategy to enhance performance. Several stud- ies have determined the beneficial effect of wearing CG on performance during exercise [1–3]. However, inconsistent results have been reported in the literature because of differences in experimental conditions (e.g., exercise duration, exercise intensity and modality, CG pressure level exerted and covered area) among studies. These differences in methodological criteria may have masked the effectiveness of CG [4,5]. To date, a number of studies focused on the effect of CG on exercise performance during running [6]. For instance, reduced energy cost of running, i.e. decrease in oxygen uptake (VO2) slow component [1], and increased time to exhaustion [2] were observed when wearing CG during submaximal running, whereas previous studies still contain conflicting results for benefit of wearing CG during running [7–9]. Possible explanations for these discrepancies may be due to differences in the levels of compressive pressure applied among studies [5]. Considering that higher compression (> 35 mmHg) may reduce local blood circulation around muscle [10], moderate compressive pressures (15–20 mmHg) may be appropriate for improving exercise performance. In this regard, Ali et al. [3] determined the effects of different levels of pressure exerted by CG (covered knee to ankle) on 10-km time trial performance. As a result, two different types of CG, exerting 12 or 18 mmHg pressure at the knee, significantly attenuated the exercise-induced decrease in jump performance compared with the control trial (0 mmHg), while CG with high pressure (23 mmHg at the knee) did not improve perfor- mance. Furthermore, Miyamoto et al. [11] showed that wearing CG applying 15 or 20 mmHg to thigh significantly attenuated exercise-induced increase in skeletal muscle proton transverse relaxation times, which reflect the intramuscular pH and Pi levels, compared with wearing CG applying 8 or 25 mmHg. The finding suggested that optimal pressure intensity would exist to obtain beneficial effects of wearing CG. However, similar studies using different levels of CG pressures did not show changes in running performance [12] or oxygen uptake [13]. Taken together, appropriate levels of compressive pressure to improve running performance remain unclear. Prolonged running, such as marathon, elicits marked muscle damage and inflammation, leading to impaired muscular function [14–16]. Del Coso et al. [16] reported that a decrease in running velocity over a marathon was significantly correlated with an increase in blood myo- globin (Mb) concentration immediately after exercise. Considering that sustained external pressure while wearing CG attenuates mechanical stress by reducing muscle oscillation [17], the use of CG during exercise might attenuate increases in muscle damage markers and inflammatory cytokines. Another mechanism underlying effectiveness of wearing CG may be improved venous return, resulting from assisted muscle pump action by the garments [18]. The augmented muscle pump action by wearing CG would be more advantageous when Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 2 / 12 decision to publish, or preparation of the manuscript. Competing interests: Three authors (Arai M, Todoko F, Yamada E) are affiliated to DESCENTE Ltd, which made garments for the present experiment. Although they contributed to conducting the present experiment (e.g., preparing custom-made garments, evaluating pressure levels), they were not involved in manuscript writing, discussion or drawing conclusion, from the viewpoint of conflict of interest. The first author (Mizuno S) declares no conflict of interest. The present study was conducted as a part of collaborated research with DESCENTE Ltd (principal investigator: Goto K), the Ritsumeikan University received project fees for the present study. However, corresponding author (Goto K) does not have any personal relationships or competing interests to DESCENTE Ltd. exercise intensity is relatively lower, because muscle pump action is dependent on exercise intensity. In addition, positive effect of wearing CG is suggested to be observed during pro- longed exercise with accumulated fatigue [1]. However, most of these previous studies have focused on the effectiveness of wearing CG on cardiovascular and metabolic variables during maximal or intensive endurance exercise lasting less than 60 min [1,3,7,12,13,18–20]. There- fore, little information is available on whether wearing CG during prolonged running (> 60 min) affects exercise-induced changes in muscle damage markers and the inflammatory responses. The aim of this study was to investigate the effect of wearing lower-body CG at three differ- ent pressure levels (30 mmHg, 15 mmHg, below 5 mmHg) during 120 min of running on exercise performance, muscle damage markers, and the inflammatory response. Based on pre- vious studies [3,10], it was hypothesized that wearing CG exerting medium (approximately 15 mmHg) pressure would attenuate the decrease in exercise performance and increases in mus- cle damage and inflammatory markers during prolonged running. Methods Participants Eight male (mean ± standard deviation: age, 23.4 ± 2.4 years; height, 170.1 ± 2.2 cm; body mass, 62.3 ± 3.3 kg; body mass index, 21.9 ± 1.6 kg/m2; VO2max, 50.6 ± 4.1 ml/kg/min) partic- ipated in the present study. Two of ten subjects were not able to complete 120 min of exercise (exhausted at 90 min). All participants were physically active (i.e., exercising at least 1 day/ week) but not well-trained athletes. They had several years of sports participation experience. Exclusion criteria included a history of inflammatory condition and musculoskeletal disorders. Smokers and individuals taking antioxidant supplements were also excluded. Participants were instructed to maintain their normal diet and physical activity level throughout the experi- mental period. They were also asked to refrain from strenuous activity for at least 72 h prior to testing. All participants gave informed consent after being informed of the purpose and risks associated with the present study. This study was approved by the Ethics Committee of the Rit- sumeikan University, Japan. Experimental procedure All participants visited the laboratory four times over the experimental period. On the first day, they completed an incremental running test on a treadmill (Valiant; Lode B.V., Gro- ningen, Netherlands) to determine individual VO2max. The initial velocity was set as 4 km/h for 3 min, and velocity was increased by 2 km/h every 3 min. All participants were required to walk both the 4km/h and 6km/h stages, and they started running from 8km/h stage. After completing above three stages of submaximal running (9 min after exercise onset), running velocity was increased by 0.6 km/h every minute until exhaustion. The treadmill gradient was 7% throughout the test. Respiratory gases were collected and analyzed using an automatic gas analyzer (AE300S; Minato Medical Science, Tokyo, Japan). Participants also engaged in suffi- cient practice of appropriate procedures for the counter-movement jump (CMJ) test and reviewed the subjective rating scale of the measurements for familiarization. From the second to fourth visits, participants performed three exercise trials with overnight fast in a random order at the same time of the day. The exercise consisted of 120 min of uphill running (slope: 7%) on a treadmill (Elevation series E95Ta; Life Fitness Corp., Tokyo, Japan) at 60% of VO2max while wearing one of three different types of garment, separated with at least 4 weeks among trials to eliminate repeated-bout effect. The average running velocity was 6.6 ± 0.5 km/h. Uphill running was selected to induce greater metabolic response under Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 3 / 12 relatively lower running velocity. All participants maintained running throughout 120 min of exercise. The present exercise protocol was determined on the basis of a pilot experiment. Custom-made garments were utilized to match pressure levels among all participants because interindividual differences in thigh and calf circumferences were reported to produce different pressure levels for working muscles, even if the garment is the same size [5]. In the present study, appropriate size of the garment was medium for all participants. However, we prepared several types of garment (different width) for each trial (five garments for the HIGH and MED trials, respectively, three garments for the CON trial), and a garment from several types was selected individually to ensure equal pressure levels among all participants. The com- pression levels at the thigh and calf were manipulated to match about 30 mmHg for HIGH trial, about 15 mmHg for MED trial, and < 5 mmHg for CON trial, respectively. Accordingly, the exercise trials consisted of 1) wearing lower-body CG exerting high level of pressure (approximately 30 mmHg) [HIGH]; 2) wearing lower-body CG exerting medium level of pres- sure (approximately 15 mmHg) [MED]; and 3) wearing a lower-body garment without specific pressure level (< 5 mmHg) [CON]. Each garment covered the area from the waist to the ankle and was custom-made by a sportswear manufacturer (DESCENTE Ltd., Osaka, Japan). Prior to the experiment, the pressure levels were determined using an air-packed sensor (AMI3037- 2; AMI Techno, Tokyo, Japan). The sensor was placed between the skin and the garment at the thigh (50% between the greater trochanter and patellar tendon) and calf (30% between the patellar tendon and lateral malleolus). Participants were asked to maintain a standing position for 15 s while the pressure levels were recorded, and mean values were calculated during 15 s. The compression levels of each garment are presented in Table 1. Participants wore identical upper body shirts, socks, and shoes during the 120 min exercise and 60 min rest periods among trials. Time course changes in jump performance, rating of perceived exertion (RPE), HR, and blood variables were compared among the three trials. Jump performance CMJ height was evaluated before exercise (after 20 min of rest) and immediately after exercise, as an indication of maximal power output by the lower-limb muscles. We used CMJ height rather than maximal isometric strength as an indication of muscle function, as a decrease in CMJ height is reported to be related to muscle damage and decreased performance during a marathon race [14]. All participants performed the CMJ on a jump mat (Multi jump tester; DKH Corp., Tokyo, Japan) connected to a computer. They were instructed to jump as high as possible while placing their hands on the lumbar area to eliminate any upper-limb effect. The flight and contact time during the vertical jump were recorded. Each jump test was repeated Table 1. Compression applied by CG at the thigh and calf. Compression at thigh Compression at calf (mmHg) (mmHg) HIGH 26.9 ± 3.3 29.2 ± 3.8 [22.6–32.8] [24.5–37.7] MED 16.1 ± 2.0 17.9 ± 3.5 [14.3–20.1] [14.9–25.5] CON 4.4 ± 1.2 3.0 ± 1.6 [2.5–5.8] [0.7–4.8] Values are means ± SD. The values in parentheses indicate minimal and maximal pressure levels among all participants. https://doi.org/10.1371/journal.pone.0178620.t001 Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 4 / 12 twice, and the highest CMJ height value was used for analysis in each case. The intraclass cor- relation coefficient between jumps was 0.99. CMJ height was calculated from flight time using the following formula: Jump height ðcmÞ ¼ 1=8 ðflight timeÞ ^2  the gravity constant ð¼ 9:81 m=s^2Þ Heart rate and rating of perceived exertion HR, and the RPE values for respiration and legs were monitored every 10 min during 120 min of uphill running. HR was measured continuously using a wireless HR monitor (RCX5; Polar, Tokyo, Japan). The RPE values for respiration and legs were recorded using the modified Borg scale ranging from 0 (nothing at all) to 10 (maximal exertion) [21]. Participants were instructed to answer magnitude of strains for legs and respiration during running (“how do you feel the levels of fatigue for leg muscles and respiration?”) to assess RPE values. Blood sampling and analysis Blood samples were collected from antecubital vein before exercise (after 20 min of rest), 60 min during exercise, and immediately and 1 h after exercise. Blood samples were used to mea- sure blood glucose, lactate, serum creatine kinase (CK), Mb, free fatty acids (FFA) and glycerol, cortisol, and plasma interleukin-6 (IL-6) concentrations. In addition to indirect muscle damage markers (CK and Mb) and inflammatory cytokine (IL-6), serum FFA, glycerol and cortisol concentrations were assessed to evaluate metabolic responses to 120 min of running. Serum and plasma samples were obtained by centrifugation (3,000 rpm, 10 min, and 4˚C) and stored at −60˚C until analysis. Blood glucose and lactate concentrations were measured using an automatic glucose analyzer (Free style, Nipro Corp., Osaka, Japan) and a lactate analyzer (Lactate Pro; Arkray Inc., Kyoto, Japan), respectively. Serum CK, Mb, FFA, and cortisol con- centrations were measured at the SRL Clinical Laboratory (Tokyo, Japan). Serum glycerol concentrations were determined in duplicate using a commercially available kit (Cayman Chemical Co., Ann Arbor, MI, USA). Plasma IL-6 concentrations were assayed with an enzyme-linked immunosorbent assay kit (R&D Systems, Minneapolis, MN, USA). The intra- assay coefficients of variation for each measurement were: 2.7% for CK, 2.2% for Mb, 1.5% for FFA, 2.5% for glycerol, 2.9% for cortisol, and 4.4% for IL-6. Statistical analysis Data are expressed as means ± standard deviation. Time course changes in jump performance, RPE, HR, and the blood variables were initially analyzed using a two-way analysis of variance (Trial x Time) with repeated measures. When the ANOVA revealed a significant interaction or main effect, the Tukey–Kramer post-hoc test was used to assess the difference. Where appro- priate, partial eta-squared (η2) was used to quantify the effect size of ANOVA. A P-value < 0.05 was considered significant. Results The relative changes in CMJ height are presented in Fig 1. A significant interaction (trial × time, P = 0.04, η2 = 0.368) and main effect for trial (P = 0.04, η2 = 0.368) were detected, whereas no significant main effect for time was observed (P = 0.172). Immediately after exercise, CMJ height was significantly higher in the MED trial (101.1 ± 3.2%) compared with that in the HIGH trial (92.6 ± 3.4%, P = 0.04, η2 = 0.368). Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 5 / 12 Fig 2A shows mean HR for early (0–60 min) and latter (60–120 min) halves of the exercise. HR increased gradually during exercise in all trials (main effect for time: P < 0.0001, η2 = 0.844). Mean HR were significantly lower in the MED trial (0–60 min: 158 ± 4 bpm, 60–120 min: 167 ± 4 bpm) than those in the CON trial (0–60 min: 164 ± 4 bpm, 60–120 min: 177 ± 4 bpm, P = 0.01, η2 = 0.479 for mean HR for early half and η2 = 0.476 for mean HR for latter half), respectively. Significant main effect for time was observed for RPE of respiration and legs (P < 0.0001, η2 = 0.793 for RPE of respiration and η2 = 0.916 for RPE of legs). No signifi- cant differences were observed in mean RPE of respiration or legs for early (0–60 min) and lat- ter (60–120 min) halves of the exercise (P > 0.05, Fig 2B and 2C). Table 2 presents the changes in blood glucose and lactate, as well as serum CK and Mb con- centrations. Significant main effect for time was observed in all variables (P < 0.05). However, no significant interaction (trial × time) or main effect for trial were observed. Plasma IL-6 concentrations increased significantly during the exercise and post-exercise periods in all trials (main effect for time: P < 0.0001, η2 = 0.849), but no significant interaction (P > 0.05) or main effect for trial was detected (P > 0.05, η2 = 0.229, Fig 3A). However, the area under the curve (AUC) for IL-6 during 120 min exercise period was significantly lower in the MED trial (397 ± 58 pg/ml
120 min) compared with that in the CON trial (670 ± 86 pg/ ml
120 min, P = 0.04, η2 = 0.454, Fig 3B). No significant difference was observed between the HIGH and CON trials. Discussion The major finding of the present study was that the MED trial showed a significantly lower exercise-induced decrease in CMJ height compared with that of the HIGH trial and a smaller increase in HR compared with that in the CON trial. Furthermore, the increased plasma IL-6 concentration during 120 min of running was also impaired in the MED trial than in the CON trial. These findings suggest that optimal pressure (approximately 15 mmHg) exists for decrease in exercise-induced fatigue during prolonged running. A unique point of the present study was preparation of several types of custom-made gar- ments to match pressure levels (approximately 15 and 30 mmHg) among participants, because no consensus has been reached about the anti-fatiguing effect of CG, probably due to Fig 1. Changes in CMJ height. Values are mean ± standard deviation. ¶; P < 0.05 between MED and HIGH. Ex120; immediately after exercise. https://doi.org/10.1371/journal.pone.0178620.g001 Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 6 / 12 differences in the compressive pressures applied among studies [22] and participants [5]. Furthermore, we attempted to determine the influences of CG exerting different pressure lev- els on exercise-induced decrease in jump performance. As a result, an exercise-induced decrease in CMJ height was significantly greater in the HIGH trial than in the MED trial. Fig 2. Mean HR values (A), mean RPE values for respiration (B) and mean RPE values for legs (C) during 120 min of running. Values are mean ± standard deviation. †; P < 0.05 between MED and CON. https://doi.org/10.1371/journal.pone.0178620.g002 Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 7 / 12 Decreased jump height following prolonged running would be caused by several factors, including impaired neuromuscular function [23], accumulated metabolites [24] and muscle damage [14]. In particular, previous studies revealed that decreases in jump height and run- ning velocity closely related to exercise-induced increases in Mb and CK concentrations immediately after prolonged running [14]. Thus, greater jump height immediately after exer- cise, as shown in the MED trial, may reflect maintained lower-limb muscle function during prolonged running. A possible mechanism for improved jump performance may be aug- mented removal of metabolites (e.g., H+ and inorganic phosphate) in working muscle because the reduction in intramuscular pH inhibits muscle contractile function (e.g., short- ening velocity) [25]. This notion is supported by previous findings revealing that adequate external pressure assists muscle pump action and improves peripheral circulation, leading to enhanced removal of H+ and inorganic phosphate from working muscle [11,26]. Notably, the effect of lower-body CG on the exercise-induced decrease in jump height was not dependent on the pressure level applied since the post-exercise jump height was significantly lower in the HIGH trial compared with that in the MED trial. Although a high pressure CG would theoretically augment venous return from working muscle, excessive pressure exerted by CG (30–40 mmHg) significantly attenuates local blood flow [10,27]. The reduced local blood flow associated with the strong external pressure applied by CG would aggravate the accumu- lation of muscle metabolites [28,29]. Accordingly, impaired jump height immediately after the exercise period in the HIGH trial might be explained by the accumulation of intramuscu- lar metabolites under lower blood flow. Future applications evaluating local blood flow would clarify this hypothesis. Exercise-induced decrease in CMJ height was significantly correlated with increases in CK and Mb concentrations immediately after prolonged exercise (triathlon race) [14]. In the pres- ent study, no differences in CK or Mb concentrations were observed among the three trials. In contrast, the exercise-induced increase in IL-6 concentration (evaluated by AUC) was signifi- cantly lower in the MED trial compared with that in the CON trial. The finding suggests that CG with medium pressure level attenuates exercise-induced inflammation. However, because blood samples were collected during relatively initial phase (60 min) of post-exercise period, we are not able to conclude the influence of wearing CG on muscle damage and inflammation Table 2. Blood glucose, lactate and serum CK and Mb concentrations. Pre Ex60 Ex120 Post60 Glucose (mg/dL) HIGH 94 ± 6 85 ± 11 79 ± 16 * 75 ± 11 * MED 90 ± 6 84 ± 4 79 ± 10 * 77 ± 6 * CON 90 ± 8 89 ± 8 79 ± 18 * 74 ± 12 * Lactate (mmol/L) HIGH 1.8 ± 0.9 2.1 ± 1.1 2.9 ± 1.4 * 2.0 ± 0.3 * MED 1.6 ± 0.6 2.0 ± 1.1 2.3 ± 1.4 * 2.1 ± 1.0 * CON 1.2 ± 0.3 1.7 ± 0.8 2.2 ± 0.8 * 2.0 ± 0.4 * CK (U/L) HIGH 142 ± 46 170 ± 51 * 196 ± 63 * 193 ± 58 * MED 144 ± 57 172 ± 68 * 198 ± 73 * 195 ± 68 * CON 191 ± 177 225 ± 195 * 259 ± 118 * 252 ± 189 * Mb (ng/mL) HIGH 31 ± 8 55 ± 21 72 ± 31 * 106 ± 36 * MED 32 ± 13 59 ± 29 * 82 ± 30 * 106 ± 32 * CON 36 ± 21 66 ± 49 92 ± 18 * 131 ± 61 * Values are means ± SD. *; P < 0.05 vs. Pre. Ex60; 60 min during exercise. Ex120; immediately after exercise. Post60; 60 min after exercise. https://doi.org/10.1371/journal.pone.0178620.t002 Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 8 / 12 after 60 min following the exercise. Future study is necessary to determine the changes in mus- cle damage markers and inflammatory cytokines during longer period after exercise. A large amount of evidence indicates that HR during exercise is not affected by wearing CG [1–3,7,18]. However, the MED trial significantly attenuated the exercise-induced increase in HR. These inconsistent observations among studies may be associated with differences in exer- cise intensities, as previous studies required relatively high or maximal efforts during exercise, whereas the exercise intensity in the present study was moderate (60% of VO2max). Further- more, mean running velocity was low (6.6 ± 0.1 km/h) because we utilized uphill running. MacRae et al. [18] pointed out that a plausible factor for lower HR during wearing CG was Fig 3. Plasma interleukin (IL)-6 concentrations (A) and area under the curve (AUC) during 120 min of exercise. Values are mean ± standard deviation. *; P < 0.05 vs. Pre. †; P < 0.05 between MED and CON. Ex60; 60 min during exercise. Ex120; immediately after exercise. Post60; 60 min after exercise. https://doi.org/10.1371/journal.pone.0178620.g003 Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 9 / 12 augmented stroke volume associated with increased venous return. Considering that venous return would be augmented depending on exercise intensity, the influence of CG on hemody- namics is supposed to be minor during intensive exercise [3]. In fact, the wearing lower-body CG during 6 km/h treadmill running significantly attenuated the exercise-induced increase in HR, but not during 10 km/h or at 85% of VO2max running [30]. Besides, the majority of previ- ous studies used CG that provided gradually decreasing pressure levels from the distal part (ankle) to the proximal part (thigh) to enhance venous return. However, the pressures exerted between the thigh and calf were not different in the present study (Table 1), suggesting that the pressure levels were uniform across the lower-limb muscles. Therefore, wearing lower-body CG exerting medium pressure (approximately 15 mmHg) at both the thigh and calf is thought to be beneficial for reducing the exercise-induced increase in HR. Some limitations need to be considered carefully in the present study. First, the number of participants was relatively small. This is because several types of custom-made garments with different width were prepared for each size of garments (five garments for the HIGH and MED trials, respectively, three garments for the CON trial) to match pressure levels applied among participants. However, further experiment with larger sample size would be informa- tive to assist the present findings. Second, we were unable to identify changes in maximal muscular strength, hemodynamics, cardiovascular response, intramuscular metabolites in response to prolonged running. Therefore, how wearing CG attenuated exercise-induced decrease in jump performance still needs further determination. However, the present findings provide novel insight for the importance of compressive pressure in performance enhance- ment. Since prolonged running under relatively lower running velocity was selected in the cur- rent study, the findings would be specially applicable for recreational or amateur endurance runners. Conclusion Wearing lower-body CG exerting 15 mmHg on the thigh and calf attenuated decrease in jump performance compared with wearing lower-body CG exerting 30 mmHg. Furthermore, exer- cise-induced increases in HR and IL-6 concentration with CG were significantly smaller in CG exerting 15 mmHg than in garment with exerting < 5 mmHg. Acknowledgments We would like to thank all of the participants who participated in the study. Author Contributions Conceptualization: SM MA FT EY KG. Formal analysis: SM KG. Funding acquisition: KG. Investigation: SM. Methodology: SM KG. Project administration: KG. Resources: KG. Supervision: KG. Validation: KG. Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 10 / 12 Visualization: SM KG. Writing – original draft: SM. Writing – review & editing: KG. References 1. Bringard A, Perrey S, Belluye N. Aerobic Energy Cost and Sensation Responses During Submaximal Running Exercise—Positive Effects of Wearing Compression Tights. Int J Sports Med. Georg Thieme Verlag KG Stuttgart
New York; 2005; 27: 373–378. https://doi.org/10.1055/s-2005-865718 2. Kemmler W, Stengel von S, Ko¨ckritz C, Mayhew J, Wassermann A, Zapf J. Effect of compression stockings on running performance in men runners. J Strength Cond Res. 2009; 23: 101–105. https:// doi.org/10.1519/JSC.0b013e31818eaef3 PMID: 19057400 3. Ali A, Creasy RH, Edge JA. The Effect of Graduated Compression Stockings on Running Performance. J Strength Cond Res. 2011; 25: 1385–1392. https://doi.org/10.1519/JSC.0b013e3181d6848e PMID: 21293307 4. MacRae BA, Cotter JD, Laing RM. Compression garments and exercise: garment considerations, phys- iology and performance. Sports Med. 2011; 41: 815–843. https://doi.org/10.2165/11591420- 000000000-00000 PMID: 21923201 5. Born D-P, Sperlich B, Holmberg H-C. Bringing light into the dark: effects of compression clothing on per- formance and recovery. Int J Sports Physiol Perfor. 2013; 8: 4–18. 6. Engel FA, Holmberg H-C, Sperlich B. Is There Evidence that Runners can Benefit from Wearing Com- pression Clothing? Sports Med. Springer International Publishing; 2016;: 1–14. https://doi.org/10.1007/ s40279-016-0546-5 PMID: 27106555 7. Ali A, Creasy RH, Edge JA. Physiological effects of wearing graduated compression stockings during running. Eur J Appl Physiol. Springer-Verlag; 2010; 109: 1017–1025. https://doi.org/10.1007/s00421- 010-1447-1 PMID: 20354717 8. Vercruyssen F, Easthope C, Bernard T, Hausswirth C, Bieuzen F, Gruet M, et al. The influence of wear- ing compression stockings on performance indicators and physiological responses following a pro- longed trail running exercise. Eur J Sport Sci. Routledge; 2014; 14: 144–150. https://doi.org/10.1080/ 17461391.2012.730062 PMID: 24533521 9. Vecruyssen F, Gruet M, Colson S, Ehrstrom S, Brisswalter J. Compression Garments, Muscle Contrac- tile Function and Economy in Trail Runners. Int J Sports Physiol Perfor. 2016;: 1–22. https://doi.org/10. 1123/ijspp.2016-0035 PMID: 27081007 10. Sperlich B, Born D-P, Kaskinoro K, Kalliokoski KK, Laaksonen MS. Squeezing the muscle: compres- sion clothing and muscle metabolism during recovery from high intensity exercise. Mu¨ller M, editor. PLoS ONE. Public Library of Science; 2013; 8: e60923. https://doi.org/10.1371/journal.pone.0060923 PMID: 23613756 11. Miyamoto N, Kawakami Y. Effect of pressure intensity of compression short-tight on fatigue of thigh muscles. Med Sci Sports Exerc. 2014; 46: 2168–2174. https://doi.org/10.1249/MSS. 0000000000000330 PMID: 24598700 12. Barwood MJ, Corbett J, Feeney J, Hannaford P, Henderson D, Jones I, et al. Compression garments: no enhancement of high-intensity exercise in hot radiant conditions. Int J Sports Physiol Perfor. 2013; 8: 527–535. 13. Sperlich B, Haegele M, Kru¨ger M, Schiffer T, Holmberg H-C, Mester J. Cardio-respiratory and metabolic responses to different levels of compression during submaximal exercise. Phlebology. SAGE Publica- tions; 2011; 26: 102–106. https://doi.org/10.1258/phleb.2010.010017 PMID: 21228356 14. Del Coso J, Gonza´lez-Milla´n C, Salinero JJ, Abia´n-Vicen J, Soriano L, Garde S, et al. Muscle damage and its relationship with muscle fatigue during a half-iron triathlon. Wieczorek DF, editor. PLoS ONE. Public Library of Science; 2012; 7: e43280. https://doi.org/10.1371/journal.pone.0043280 PMID: 22900101 15. Nieman DC, Luo B, Dre´au D, Henson DA, Shanely RA, Dew D, et al. Immune and inflammation responses to a 3-day period of intensified running versus cycling. 2014; 39: 180–185. https://doi.org/10. 1016/j.bbi.2013.09.004 PMID: 24055861 16. Del Coso J, Ferna´ndez D, Abia´n-Vicen J, Salinero JJ, Gonza´lez-Milla´n C, Areces F, et al. Running Pace Decrease during a Marathon Is Positively Related to Blood Markers of Muscle Damage. Earnest CP, editor. PLoS ONE. 2013; 8: e57602–7. https://doi.org/10.1371/journal.pone.0057602 PMID: 23460881 Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 11 / 12 17. Doan BK, Kwon Y-H, Newton RU, Shim J, Popper EM, Rogers RA, et al. Evaluation of a lower-body com- pression garment. J Sports Sci. 2003; 21: 601–610. https://doi.org/10.1080/0264041031000101971 PMID: 12875311 18. MacRae BA, Laing RM, Niven BE, Cotter JD. Pressure and coverage effects of sporting compression garments on cardiovascular function, thermoregulatory function, and exercise performance. Eur J Appl Physiol. Springer-Verlag; 2012; 112: 1783–1795. https://doi.org/10.1007/s00421-011-2146-2 PMID: 21901265 19. Stickford ASL, Chapman RF, Johnston JD, Stager JM. Lower-Leg Compression, Running Mechanics, and Economy in Trained Distance Runners. Int J Sports Physiol Perfor. 2015; 10: 76–83. https://doi. org/10.1123/ijspp.2014-0003 PMID: 24911991 20. Wahl P, Bloch W, Mester J, Born D-P, Sperlich B. Effects of different levels of compression during sub- maximal and high-intensity exercise on erythrocyte deformability. Eur J Appl Physiol. Springer-Verlag; 2012; 112: 2163–2169. https://doi.org/10.1007/s00421-011-2186-7 PMID: 21964909 21. Borg GA. Psychophysical bases of perceived exertion. Med Sci Sports Exerc. 1982; 14: 377–381. PMID: 7154893 22. Beliard S, Chauveau M, Moscatiello T, Cros F, Ecarnot F, Becker F. Compression garments and exer- cise: no influence of pressure applied. J Sports Sci Med. Dept. of Sports Medicine, Medical Faculty of Uludag University; 2015; 14: 75–83. 23. Taylor JL, Todd G, Gandevia SC. Evidence for a supraspinal contribution to human muscle fatigue. Clin Exp Pharmacol Physiol. Blackwell Science Pty; 2006; 33: 400–405. https://doi.org/10.1111/j.1440- 1681.2006.04363.x PMID: 16620309 24. Westerblad H, Bruton JD, La¨nnergren J. The effect of intracellular pH on contractile function of intact, single fibres of mouse muscle declines with increasing temperature. J Physiol. Wiley-Blackwell; 1997; 500 (Pt 1): 193–204. 25. Rassier DE, Herzog W. Effects of pH on the length-dependent twitch potentiation in skeletal muscle. J Appl Physiol. American Physiological Society; 2002; 92: 1293–1299. https://doi.org/10.1152/ japplphysiol.00912.2001 PMID: 11842070 26. Ibegbuna V, Delis KT, Nicolaides AN, Aina O. Effect of elastic compression stockings on venous hemo- dynamics during walking. J Vasc Surg. 2003; 37: 420–425. https://doi.org/10.1067/mva.2003.104 PMID: 12563216 27. Lawrence D, Kakkar VV. Graduated, static, external compression of the lower limb: a physiological assessment. Br J Surg. 1980; 67: 119–121. PMID: 7362940 28. Nishiyasu T, Nagashima K, Nadel ER, Mack GW. Human cardiovascular and humoral responses to moderate muscle activation during dynamic exercise. J Appl Physiol. 2000; 88: 300–307. PMID: 10642393 29. Nielsen HV. External pressure—blood flow relations during limb compression in man. Acta Physiol Scand. Blackwell Publishing Ltd; 1983; 119: 253–260. https://doi.org/10.1111/j.1748-1716.1983. tb07335.x PMID: 6659990 30. Lovell DI, Mason DG, Delphinus EM, McLellan CP. Do compression garments enhance the active recovery process after high-intensity running? J Strength Cond Res. 2011; 25: 3264–3268. https://doi. org/10.1519/JSC.0b013e31821764f8 PMID: 22082795 Effects of wearing compression garment during prolonged running on exercise-induced fatigue PLOS ONE | https://doi.org/10.1371/journal.pone.0178620 May 31, 2017 12 / 12

Wearing lower-body compression garment with medium pressure impaired exercise-induced performance decrement during prolonged running.

05-31-2017

Mizuno, Sahiro,Arai, Mari,Todoko, Fumihiko,Yamada, Eri,Goto, Kazushige

eng

PMC8750590




 Citation: Parshukova, O.I.; Varlamova, N.G.; Potolitsyna, N.N.; Lyudinina, A.Y.; Bojko, E.R. Features of Metabolic Support of Physical Performance in Highly Trained Cross-Country Skiers of Different Qualifications during Physical Activity at Maximum Load. Cells 2022, 11, 39. https://doi.org/ 10.3390/cells11010039 Academic Editor: Robert Wessells Received: 1 December 2021 Accepted: 21 December 2021 Published: 23 December 2021 Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affil- iations. Copyright: © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). cells Article Features of Metabolic Support of Physical Performance in Highly Trained Cross-Country Skiers of Different Qualifications during Physical Activity at Maximum Load Olga I. Parshukova * , Nina G. Varlamova , Natalya N. Potolitsyna , Aleksandra Y. Lyudinina and Evgeny R. Bojko Institute of Physiology of Komi Science Centre of the Ural Branch of the Russian Academy of Sciences, FRC Komi SC UB RAS, 50 Pervomayskaya Str., 167982 Syktyvkar, Russia; [email protected] (N.G.V.); [email protected] (N.N.P.); [email protected] (A.Y.L.); [email protected] (E.R.B.) * Correspondence: [email protected] Abstract: The purpose of our study was to identify the features of metabolic regulation in highly trained cross-country skiers of different qualifications at different stages of the maximum load test. We examined 124 highly trained cross-country skiers (male, ages 17–24). The group consisted of two subgroups based on their competition performance: 61 nonelite athletes (Group I) and 63 elite athletes (group II), who were current members of the national team of the Komi Republic and Russia. The bicycle ergometer test was performed by using the OxyconPro system (Erich Jaeger, Hoechberg, Germany). All the examined athletes performed the exercise test on a cycle ergometer “until exhaustion”. The results of our research indicate that the studied groups of athletes with high, but different levels of sports qualifications are a convenient model for studying the molecular mechanisms of adaptation to physical loads of maximum intensity. Athletes of higher qualifications reveal additional adaptive mechanisms of metabolic regulation, which is manifested in the independence of serum lactate indicators under conditions of submaximal and maximum power from maximal oxygen uptake, and they have an NO-dependent mechanism for regulating lactate levels during aerobic exercise, including work at the anaerobic threshold. Keywords: nitric oxide; lactate; heart rate; oxygen uptake; arterial blood pressure; exercise test on a cycle ergometer; cross-country skier 1. Introduction Moderate physical activity has a positive effect on the morphology and work of the cardiovascular system of athletes due to the manifested adaptive response of the myocardium [1]. However, at the same time, a violation in the consistency of the functional state of the system associated with the work of the heart rhythm leads to an overstrain of the cardiovascular system of athletes [2]. Systematically exercising athletes usually develop myocardial hypertrophy. Pathological hypertrophy is based on dystrophic changes in the myocardium and deterioration of the microvasculature, which leads to difficulty in contraction of the left ventricular wall, which ultimately affects the decrease in the athletic performance of the body [3]. Cross-country skiers have a very high maximal oxygen uptake, and they are able to perform submaximal exercise at a rather high metabolic rate, and with cardiac output levels similar to or higher than the cardiac output levels achieved by untrained humans at maximal exercise [4]. The role of the molecular gas nitric oxide (NO) in the cardiovascular system is well established, where NO regulates a multitude of cellular processes. Endothelial cells synthesize and release NO, which mediates diverse effects, including vessel tone, haemostasis, blood pressure and vasculature remodelling [5]. The significance of NO for cardiomyocyte function is well known because it plays a role in the regulation of ion channels, Ca2 homeostasis, contractility, energetics, cell growth, and it has antioxidant effects and prevents endothelial cells from oxidative stress [6]. Cells 2022, 11, 39. https://doi.org/10.3390/cells11010039 https://www.mdpi.com/journal/cells Cells 2022, 11, 39 2 of 16 The “gold standard” of cardiorespiratory exercise testing—a test with increasing load—allows you to determine the maximum oxygen uptake, assess the level of aerobic capacity of the body and identify the reasons for limiting physical performance [7]. Cur- rently, there are many protocols for exercise testing. However, at different stages of the test, there is practically no information about the functional parameters of the cardiorespiratory system and human biochemical markers, such as heart rate, blood pressure, QRS complex, QT interval, oxygen uptake, carbon dioxide production, respiratory rate and the levels of stable nitric oxide metabolites (nitrites, nitrates) and lactate. This significantly complicates the analysis, comparison and prediction of available data. The purpose of our study was to identify the features of metabolic regulation in highly trained cross-country skiers of different qualifications at different stages of testing at physical maximum load. 2. Materials and Methods 2.1. Ethical Approval The Ethics Committee of the Institute of Physiology of the Russian Academy of Sciences, Syktyvkar approved the experimental design and protocol of the study. The study conformed to the Code of Ethics of the World Medical Association (Declaration of Helsinki). The volunteers were made aware of all the information about the experimental protocol, experimental procedures, and probable risks and inconvenience associated with performing the exercise test on a cycle ergometer “until exhaustion”. After the necessary interpretations, the volunteers gave their written informed agreement to participate in the test. Participants were aware that they were free to leave the study at any time and without consequence. 2.2. Participants The observation group included 124 highly trained cross-country skiers (male, ages 17–24). The group consisted of two subgroups based on their competition performance: Group I in- cludes 61 nonelite athletes who occupied the last ten places at official competitions, and Group II includes 63 elite athletes who occupied the first ten places at official competitions, who were current members of the national team of the Komi Republic and Russia, had no signs or history of chronic diseases. All participants had more than 5 years of cross-country skiing practice as part of their main training schedule and had extensive experience in endurance events. 2.3. Experimental Protocol The study was performed during the morning after a low-nitrate breakfast. It excluded foods and drinks that are the main sources of nitrates in human food (meat and fish products, vegetables (mainly beets, leafy green vegetables), marinades, (spirits, fruit and mineral drinks). Additionally, the night before the test all the participants consumed a standardized meal (1,674,400–1,757,420 kJcal) consisting of (in units of the percentage of the total energy supplied by the entire meal, En%) 78 En% carbohydrate, 14 En% fat and 8 En% protein. The dietary intake of the participants was assessed using a food frequency questionnaire. Calories consumed at breakfast were not standardize by body weight. The height and body weight of the athletes were measured using a medical weight growth meter (Accunig, SELVAS Healthcare, Daejeon, Korea). At rest (sitting), at the anaerobic threshold (AT) level, during peak load and during the recovery period (5th minute) in each of the athletes examined were determined by the following parameters: systolic blood pressure (SBP), diastolic pressure (DBP), heart rate (HR), Q wave, R wave, and S wave (QRS) complex, QT interval (QT), oxygen uptake (V’O2), carbon dioxide production (V’CO2), respiratory rate (Rer), and the levels of stable metabolites of NO and lactate in capillary blood samples. Blood pressure was measured by the Korotkov method on the right arm using the Microlife model BRAG-1-30 device (Widnau, Switzerland). An electrocardiogram (ECG) was recorded in 12 leads: standard according to Einthoven (I, II, III), reinforced from the extremities according to Goldberger (aVR, aVL, aVF), and thoracic according to Wilson (V1-6). Manual measurements of ECG characteristics were performed using a ruler for Cells 2022, 11, 39 3 of 16 measuring and evaluating electrocardiograms from “Heinrich Mack Nachf” (Karlsruhe, Germany). Heart rate, oxygen uptake, carbon dioxide production, and respiratory rates were obtained from test protocols. 2.4. Exercise Test on a Cycle Ergometer “until Exhaustion” On an ergometer bike (“Ergose-lect-100”, Ergoline GmbH, Hoechberg, Germany) was performed aerobic capacity (V’O2 max) testing in the “breath by breath” mode. The parameters were averaged over 15-s segments. The test included one minute of cycling without load (for adaptation of participants) followed by stepwise load increases of 40 W in 2 min increments. The first load started from 120 W. During the test, the pedalling speed was 60 rpm. The anaerobic threshold (AT) was determined by reaching a respiratory coefficient of 1 [8,9]. 2.5. Determination of NOx The NO levels in the plasma were measured using the Griess reaction, by evaluating the stable metabolites of NO, including nitrites (NO(2)(−)) and nitrates (NO(3)(−)), which were pre-sorted as an index [NOx]. As described earlier in our article [5], nitrite and nitrate are the terminal products of NO in human plasma. It is known that a strong correlation between endogenous NO production and NOx levels exist in plasma [10]. Blood samples with a volume of two ml were collected into tubes with heparin and centrifuged for 20 min at 2500× g. The separated plasma was stored at −40 ◦C until analysis. After deproteinization of plasma samples by precipitation in ethanol and centrifugation, the supernatants were incubated for 30 min at 37 ◦C with vanadium chloride to convert nitrate to nitrite. Next, the samples were mixed with Griess reagent. Samples were measured at a wavelength of 540 nm using a Spectronic Genesys-6 Spectrophotometer (Thermo Electron Scientific Instruments LLC, Madison, WI, USA). Total nitrite was measured using the Griess reagent. Samples were measured twice against a standard nitrite curve with a known concentration. The plasma nitrate concentration was calculated by subtracting the primary nitrite concentration from the total nitrite concentration. All chemicals used for NO determination were obtained from Sigma (St. Louis, MO, USA). The detection limit for NO was 0.001 µmol/L. The NO3/NO2 index was calculated as the ratio between NO3 and NO2. 2.6. Determination of Lactate Plasma lactate levels were measured using a lactate pulmonary alveolar proteinosis (PAP) enzymatic colorimetric method with “Chronolab” commercial kits (Chronolab Sys- tems, S.L. Barcelona, Spain) with an intra-assay coefficient of variance (CV) of 8%. Mea- surements were performed using a spectrophotometer at a wavelength of 546 nm. 2.7. Statistical Analysis Statistica software (STA862D175437Q, version 6.0, StatSoft Inc., 2001, Tulsa, OK, USA) was used for statistical analysis. The mean (Me) and standard deviation (SD) were calcu- lated. Differences in the dynamics of each parameter were tested by Friedman’s ANOVA. The Wilcoxon test was used to define the correlation coefficients between two variables. The Spearman rank analysis determined the correlation coefficients between two variables. A value of p < 0.05 was accepted as statistically significant. 3. Results The characteristics of the examined groups of athletes are presented in Table 1. The groups of the examined athletes did not have significant differences in age, weight or height (Table 1). At the same time, load power on the anaerobic threshold and maximal load power increased significantly (p < 0.05) with an increase in sports activity (Figure 1). Cells 2022, 11, 39 4 of 16 Table 1. Characteristics of the subjects by group, ME ± SD. Parameters Group I (n = 61) Group II (n = 63) AGE, YEARS 19.1 ± 2.1 21.0 ± 3.1 WEIGHT, KG 69.1 ± 4.8 71.1 ± 4.6 HEIGHT, CM 174.9 ± 4.7 175.4 ± 4.9 3. Results The characteristics of the examined groups of athletes are presented in Table 1. Table 1. Characteristics of the subjects by group, ME ± SD. Parameters Group I (n = 61) Group II (n = 63) AGE, YEARS 19.1 ± 2.1 21.0 ± 3.1 WEIGHT, KG 69.1 ± 4.8 71.1 ± 4.6 HEIGHT, CM 174.9 ± 4.7 175.4 ± 4.9 The groups of the examined athletes did not have significant differences in age, weight or height (Table 1). At the same time, load power on the anaerobic threshold and maximal load power increased significantly (p < 0.05) with an increase in sports activity (Figure 1). group I group II Load power on anaerobic threshold Maximal Load power 180 200 220 240 260 280 300 320 340 360 380 400 420 Load power / Watts Group I: p=0.0191 Group II: p=0.0193 * * Figure 1. Load power in cross-country skiers during the exercise test on a cycle ergometer “until exhaustion” (Mean; Box: Mean ± 2 * SD; Whisker: Min–Max). When performing the test “until exhaustion”, all cross-country skiers showed a sta- tistically significant increase in oxygen uptake during the passage of the anaerobic thresh- old (p = 0.001) compared with indicators at rest (Figure 2). At the level of the AT and peak load, group II was characterized by higher values of oxygen uptake (p < 0.01 and p < 0.001, respectively). During the recovery period, oxygen uptake values decreased significantly in both groups (p < 0.001). During rest and during the recovery period, the oxygen uptake did not differ significantly in the study groups. Figure 1. Load power in cross-country skiers during the exercise test on a cycle ergometer “until exhaustion” (Mean; Box: Mean ± 2 SD; Whisker: Min–Max). * p < 0.05. When performing the test “until exhaustion”, all cross-country skiers showed a statis- tically significant increase in oxygen uptake during the passage of the anaerobic threshold (p = 0.001) compared with indicators at rest (Figure 2). At the level of the AT and peak load, group II was characterized by higher values of oxygen uptake (p < 0.01 and p < 0.001, respectively). During the recovery period, oxygen uptake values decreased significantly in both groups (p < 0.001). During rest and during the recovery period, the oxygen uptake did not differ significantly in the study groups. Cells 2022, 11, x FOR PEER REVIEW 5 of 17 Figure 2. Oxygen uptake during the exercise test on a cycle ergometer “until exhaustion” in cross- country skiers. Statistical significance levels between groups: ** p < 0.01; *** p < 0.001.Statistical sig- nificance levels between stages of the load: ### p < 0.001. 3.1. Cardiorespiratory Parameters All cross-country skiers showed some similar dynamics of SBP during the test “until exhaustion”: an increase at AT and at peak load and a decrease at recovery (p < 0.001) (Table 2). At rest, Group I was characterized by higher SBP than Group II (p < 0.05). During the period of AT, at the peak load and at the recovery period, the DBP was Figure 2. Oxygen uptake during the exercise test on a cycle ergometer “until exhaustion” in cross- country skiers. Statistical significance levels between groups: ** p < 0.01; *** p < 0.001. Statistical significance levels between stages of the load: ### p < 0.001. Cells 2022, 11, 39 5 of 16 3.1. Cardiorespiratory Parameters All cross-country skiers showed some similar dynamics of SBP during the test “until exhaustion”: an increase at AT and at peak load and a decrease at recovery (p < 0.001) (Table 2). At rest, Group I was characterized by higher SBP than Group II (p < 0.05). Table 2. Cardiorespiratory parameters at different stages of the load in cross-country skiers, Me ± SD. Parameters Stages of the Load At Rest Anaerobic Threshold Peak Load Recovery Systolic blood pressure, mm Hg I 118.6 ± 11.7 168.5 ± 16.1 ### 188.6 ± 15.9 ### 122.2 ± 12.2 ### Ii 115.2 ± 8.9 * 163.9 ± 13.2 ### 185.8 ± 18.6 ### 125.3 ± 14.3 ### Diastolic blood pressure, mm Hg I 77.6 ± 7.6 70.6 ± 13.9 ### 75.2 ± 16.4 ### 62.6 ± 13.5 ### Ii 77.9 ± 8.8 77.1 ± 11.5 ** 83.4 ± 14.7 *,### 67.5 ± 15.4 **,### Heart rate, beats/min I 62.8 ± 13.1 166.3 ± 13.3 ### 180.2 ± 17.6 ### 108.2 ± 13.3 ### Ii 56.1 ± 10.1 ** 165.1 ± 15.0 ### 177.8 ± 17.3 ### 99.9 ± 14.6 *,### QRS complex, ms I 103.8 ± 8.9 191.8 ± 95.8 ### 216.6 ± 76.3 # 123.4 ± 41.9 ### Ii 106.4 ± 8.9 187.4 ± 83.6 ### 213.3 ± 87.3 # 113.6 ± 19.7 ### QT interval, ms I 394.2 ± 24.9 328.5 ± 93.5 ### 361.8 ± 80.4 # 316.8 ± 41.6 ### Ii 412.5 ± 29.4 *** 316.8 ± 84.3.1 ### 345.4 ± 93.6 # 311.1 ± 28.4 ## Carbon dioxide production, L/min I 0.3 ± 0.1 3.6 ± 0.5 ### 4.7 ± 0.6 ### 0.8 ± 0.2 ### Ii 0.3 ± 0.1 3.9 ± 0.6 **,### 4.8 ± 0.7 ### 0.8 ± 0.2 ### Respiratory rate, breaths per minute I 13.9 ± 3.8 35.6 ± 8.2 ### 50.2 ± 10.1 ### 25.8 ± 5.8 ### Ii 13.6 ± 3.9 36.8 ± 8.7 ### 50.8 ± 12.4 ### 25.7 ± 4.9 ### Statistical significance levels between groups: * p < 0.05; ** p < 0.01; *** p < 0.001 Statistical significance levels between stages of the load: # p < 0.05; ## p < 0.01; ### p < 0.001. During the period of AT, at the peak load and at the recovery period, the DBP was significantly higher in the athletes of Group II than in the skiers of Group I (p < 0.01; p < 0.05 and p < 0.01, respectively) (Table 2). The level of AT in Group I was decreased in DBP compared with the rest of the period (p < 0.01). Athletes in both groups showed an increase in DBP at the peak load compared with the AT Group (p < 0.001) and a decrease after five minutes of completion of the test (p < 0.001). There were no differences in DBP at rest or during the recovery period between the study groups (p > 0.05). All cross-country skiers showed some similar dynamics of heart rate during the test “until exhaustion”: an increase at the AT and at peak load and a decrease at recovery (p < 0.001) (Table 2). At rest and during the recovery period, Group I was characterized by a higher heart rate than Group II (p < 0.01 and p < 0.05, respectively). The dynamics of changes in the QRS complex in the two groups were similar; specif- ically, significant increases during the period of AT passage (p < 0.001) and at peak load (p < 0.05) and a decrease 5 minutes after the end of the test was noted (p < 0.001) (Table 2). There were no significant differences in the QRS complexation during the test “until exhaus- tion” between the study groups (p > 0.05). A significant decrease in QT interval was found in both groups of cross-country skiers during the passage of the AT and at the recovery period (p < 0.001) and an increase at the peak of the load (p < 0.05). Significant differences in QT interval between groups were found only at rest (p < 0.001). All cross-country skiers showed some similar dynamics of carbon dioxide production and respiratory rate during the test “until exhaustion”: an increase at AT and at peak load and a decrease at recovery (p < 0.001). AT Group II was characterized by higher carbon dioxide production than Group I (p < 0.01). 3.2. Biochemical Parameters At rest and during the test “until exhaustion”, the values of NOx in Group II were significantly higher than the values of NOx in Group I (p < 0.01) (Figure 3). For athletes, Cells 2022, 11, 39 6 of 16 Group I and Group II showed an increase in NOx at the AT compared with the rest (p < 0.001 and p < 0.05, respectively) and a decrease for athletes in group II at peak load (p < 0.05). During the recovery period, the NOx level in the examined groups did not change compared to the peak load. tion and respiratory rate during the test “until exhaustion”: an increase at AT and at peak load and a decrease at recovery (p < 0.001). AT Group II was characterized by higher car- bon dioxide production than Group I (p < 0.01). 3.2. Biochemical Parameters At rest and during the test “until exhaustion”, the values of NOx in Group II were significantly higher than the values of NOx in Group I (p < 0.01) (Figure 3). For athletes, Group I and Group II showed an increase in NOx at the AT compared with the rest (p < 0.001 and p < 0.05, respectively) and a decrease for athletes in group II at peak load (p < 0.05). During the recovery period, the NOx level in the examined groups did not change compared to the peak load. Figure 3. NOx level (µmol/L) during the exercise test on a cycle ergometer “until exhaustion” in cross-country skiers. Statistical significance levels between groups: ** p < 0.01. Statistical significance levels between stages of the load: # p < 0.05; ### p < 0.001. Figure 3. NOx level (µmol/L) during the exercise test on a cycle ergometer “until exhaustion” in cross-country skiers. Statistical significance levels between groups: ** p < 0.01. Statistical significance levels between stages of the load: # p < 0.05; ### p < 0.001. At rest, the NO2 value in Group II was significantly higher than the NO2 value in Group I (p < 0.05) (Table 3). The NO2 value between the studied groups of athletes at the AT, at peak load and at the recovery period did not show statistically significant changes (p > 0.05). A significant increase in the level of NO2 was detected only in Group I during the AT period compared with rest (p < 0.01). At rest and during the test “until exhaustion”, the values of NO3 in Group II were significantly higher than the values of NO3 in Group I (p < 0.01). A statistically significant (p < 0.01–0.001) increase in NO3 was observed in all study groups during the passage of AT compared with the values at rest. In Group II, during the passage of the load peak, the NO3 value decreased significantly in comparison with the AT (p < 0.01). At the same time, in Group I, a statistically significant decrease in NO3 was observed during recovery compared to the peak load (p < 0.05). Table 3. Nitric oxide and lactate levels at different stages of the load in cross-country skiers, ME ± SD. Parameters Stages of the Load At Rest Anaerobic Threshold Peak Load Recovery NO2, µmol/L I 8.1 ± 3.6 10.2 ± 4.9 ## 9.8 ± 3.9 10.7 ± 4.5 Ii 11.6 ± 5.2 * 11.6 ± 6.1 11.1 ± 5.4 11.5 ± 5.7 NO3, µmol/L I 8.2 ± 4.0 12.3 ± 6.9 ### 11.9 ± 6.2 10.2 ± 5.8 # Ii 17.2 ± 8.6 ** 20.0 ± 10.6 **,## 18.9 ± 11.3 **,# 17.3 ± 10.5 ** NO3/NO2 index I 1.5 ± 0.5 1.6 ± 0.9 1.6 ± 0.2 1.3 ± 0.8 Ii 2.3 ± 0.6 * 2.2 ± 0.6 * 2.2 ± 1.3 * 1.9 ± 0.5 * Lactate, µmol/L I 2.9 ± 0.9 6.2 ± 1.6 ### 9.6 ± 2.2 ### 9.7 ± 2.3 Ii 2.0 ± 0.8 ** 6.4 ± 1.8 ### 10.3 ± 1.8 ### 9.8 ± 2.7 Statistical significance levels between groups: * p < 0.05; ** p < 0.01 Statistical significance levels between stages of the load: # p < 0.05; ## p < 0.01; ### p < 0.001. Cells 2022, 11, 39 7 of 16 The subjects of both groups showed no changes in the dynamics of the NO3/NO2 index during the test “until exhaustion”. At the same time, Group II was characterized by higher values of the NO3/NO2 index than Group I (p < 0.05). In representatives of different sports qualifications, the level of lactate in the blood increased during the test, while it did not recover after 5 min of the end of the test (p < 0.001) (Table 3). At rest in Group I, the lactate value was higher than the lactate value in Group II (p < 0.01). 3.3. Interrelationship of Cardiorespiratory Parameters and Biochemical Parameters Correlation analysis between biochemical and cardiorespiratory parameters at differ- ent stages of the cross-country skiers in Groups I and II are presented in Tables 4 and 5, respectively. After correlation analysis, skiers of Groups I and II at rest showed a negative relationship between the values of NOx and lactate (r = −0.26, p < 0.05; r = −0.44, p < 0.001, respectively), and Group II had a negative relationship between the values of NO3 and lactate (r = −0.30, p < 0.05) (Table 4). However, in Group II, the relationship between NOx and lactate during the AT period became positive (r = 0.30, p < 0.01), and during the peak load period, the relationship between NOx and lactate again became negative (r = −0.26, p < 0.05). In Group I, during the period of the peak load, a positive relationship was found between lactate and NO2 (r = 0.39, p < 0.01), and a negative relationship was found with the NO3/NO2 index (r = −0.35, p < 0.01). During the recovery period, no correlations were found between nitric oxide and lactate in the examined groups. In general, during the AT period, most of the correlations between nitric oxide and cardiorespiratory parameters were found in Group II. Table 4. Correlations between biochemical and cardiorespiratory parameters at different stages of cross-country skiers in Group I. Stages of the Load Parameters Spearman Rank Order Correlations NOx NO2 NO3 Before load, at rest SBP −0.32 * DBP HR QRS QT Lactate −0.26 * V’O2 V’O2 max 0.30 * V’CO2 Rer Anaerobic threshold SBP −0.25 * −0.36 ** DBP HR QRS QT Lactate V’O2 0.36 ** 0.28 * V’O2 max 0.29 * V’CO2 0.34 ** Rer Cells 2022, 11, 39 8 of 16 Table 4. Cont. Stages of the Load Parameters Spearman Rank Order Correlations NOx NO2 NO3 Peak load SBP DBP HR QRS QT Lactate 0.39 ** V’O2 V’O2 max V’CO2 Rer Recovery SBP DBP HR 0.29 * QRS QT Lactate V’O2 V’O2 max V’CO2 Rer Statistical significance levels: * p < 0.05; ** p < 0.01. SBP-systolic blood pressure, DBP-diastolic blood pressure, HR-heart rate, QRS-QRS complex, QT-QT interval, V’O2-oxygen uptake, V’O2 max-maximal oxygen uptake, V’CO2-carbon dioxide production, Rer-respiratory rate. Table 5. Correlations between biochemical and cardiorespiratory parameters at different stages of cross-country skiers in Group II. Stages of the Load Parameters Spearman Rank Order Correlations NOx NO2 NO3 Before load, at rest SBP DBP HR QRS 0.37 ** QT Lactate −0.44 *** −0.30 * V’O2 V’O2 max V’CO2 −0.30 ** Rer −0.26 * −0.26 * Anaerobic threshold SBP 0.26 * DBP HR 0.31 ** 0.43 *** QRS QT Lactate 0.30 ** 0.26 * V’O2 0.33 ** 0.32 ** V’O2 max V’CO2 0.26 * 0.25 * Rer 0.26 * 0.26 * Cells 2022, 11, 39 9 of 16 Table 5. Cont. Stages of the Load Parameters Spearman Rank Order Correlations NOx NO2 NO3 Peak load SBP −0.29 * −0.26 * DBP 0.27 * HR QRS QT Lactate −0.26 * V’O2 V’O2 max 0.26 * V’CO2 Rer Recovery SBP DBP HR 0.27 * QRS QT Lactate V’O2 V’O2 max V’CO2 Rer −0.29 * −0.28 * Statistical significance levels: * p < 0.05; ** p < 0.01; *** p < 0.001. SBP-systolic blood pressure, DBP-diastolic blood pressure, HR-heart rate, QRS-QRS complex, QT-QT interval, V’O2-oxygen uptake, V’O2 max-maximal oxygen uptake, V’CO2-carbon dioxide production, Rer-respiratory rate. Correlation analysis between maximal oxygen uptake and cardiorespiratory param- eters at different stages of the cross-country skiers in Groups I and II are presented in Table 6. Table 6. Correlations between maximal oxygen uptake and cardiorespiratory parameters and lactate at different stages of cross-country skiers in Groups I and II. Stages of the Load Parameters Spearman Rank Order Correlations Groups I Groups II Before load, at rest SBP DBP HR QRS QT Lactate V’CO2 Rer Anaerobic threshold SBP −0.34 ** DBP HR QRS QT 0.26 * Lactate 0.26 * V’CO2 Rer Cells 2022, 11, 39 10 of 16 Table 6. Cont. Stages of the Load Parameters Spearman Rank Order Correlations Groups I Groups II Peak load SBP DBP HR QRS QT Lactate 0.33 ** V’CO2 Rer Recovery SBP DBP −0.26 * HR QRS QT Lactate 0.33 ** V’CO2 Rer Statistical significance levels: * p < 0.05; ** p < 0.01. SBP-systolic blood pressure, DBP-diastolic blood pressure, HR-heart rate, QRS-QRS complex, QT-QT interval, V’CO2-carbon dioxide production, Rer-respiratory rate. During the AT, the peak load and during the recovery period showed a positive relationship between the maximal oxygen uptake and lactate in Group I (r = 0.26, p < 0.05, r = 0.33, p < 0.01, r = 0.33, p < 0.01, respectively). In Group II, no correlations were found between maximal oxygen uptake and cardiorespiratory parameters. 4. Discussion The main goal of our study was to identify the features of metabolic regulation in highly trained cross-country skiers of different qualifications at different stages of testing at physical maximum load. The study showed that elite athletes with higher results at official competitions were characterized by a higher anaerobic threshold and maximal oxygen uptake. Compared to the group, the athlete was not elite, which is comparable to the literature data [11,12]. Cross-country skiers are known to have a very high maximal oxygen uptake and have equally trained upper and lower body muscles [4]. Thus, they are able to perform submaximal exercise at a rather high metabolic rate and with cardiac output levels similar to or higher than the cardiac output levels achieved by untrained humans at maximal exercise. Thus, this group of athletes is a successful model for studying metabolic effects in humans during intense physical work, especially since trained skiers can actually perform this intensive work, revealing subtle regulatory mechanisms. For example, elite Swedish skiers had a maximum oxygen uptake of 5.1 ± 0.3 L/min [13], which is 7.8% more than among the Group II skiers we examined. Our data on the maximum oxygen uptake also differ significantly in elite athletes from V’O2 max of Norwegian representatives of world- class winter sports [14,15], which are characterized by a maximum V’O2/kg from 80 to 90 mL/min/kg or 6.5 L/min, which is higher than in Group II surveyed by us by 27.6%. Most likely, this difference can be explained by different methodological approaches for determining the V’O2 max and higher anthropometric indicators of elite Swedish skiers (height 180 ± 2 cm, weight 74 ± 2 kg) [13] and Norwegian representatives of world-class winter sports [14,15]. The physical efficiency of athletes and the state of their cardiorespiratory system play leading roles among cross-country skiers in achieving high sports results. Physical aerobic exercise influences vascular remodelling, promoting angiogenesis, positively affecting the number of capillaries and therefore the gas exchange area, while improving oxygen diffusion and increasing vagus tone [14]. In the athletes examined in our study, SBP at rest Cells 2022, 11, 39 11 of 16 corresponded to the norm [16]. At the same time, in Group I at rest, the SBP was statistically higher (p < 0.05) than in Group II, but during the exercise test, significant differences in SBP between the groups disappeared. A higher SBP in skiers of the 1st group, compared with Group II, could be associated with the incompleteness of the processes of the formation of the cardiovascular system against the background of significant sports loads. At rest, DBP in all groups of cross-country skiers was above the norm, in com- parison with the data obtained from students from a physical education department (64.0 ± 4.7 mm Hg) [16]. It is known that prolonged training in open cold air can lead to an increase in peripheral vascular resistance and, as a consequence, to an increase in DBP [17]. Short-term (one-hour) cold exposure induces hypercoagulation in young healthy people, which can also cause a higher level of DBP [18]. According to the literature [19], cold air can indirectly lead to an increase in cardiovascular risks through its effect on the sympathetic and renin-angiotensin systems, blood pressure, and risk factors for atherosclerosis, such as blood viscosity, the amount of fibrinogen, lipids and uric acid. In our study, during the period of AT, at the peak load and at the recovery period, the DBP was significantly higher in the athletes of Group II than in the skiers of Group I (p < 0.01; p < 0.05 and p < 0.01, respectively) (Table 2). According to our data, the heart rate at rest in the skiers of Group I was 10.7% higher than the heart rate at rest in Group II (p < 0.01), which indicates the formation of bradycardia as a result of sports training, which was more pronounced with an increase in sports activity. This assumption is confirmed by the higher values of the QT interval at rest in Group II compared with Group I. Aerobic exercise [20] affects the parasympathetic nerve, reducing the heart rate, which has a positive effect on reducing cardiovascular diseases. With an increase in sportsmanship, lower heart rate values may indicate [3] higher functional reserves. On the one side, it was shown that low intensity exercise could improve antioxidant defences and lower lipid peroxidation levels [21]. On the other side, it is known that in the body of professional athletes under intense and strenuous physical exertion, oxidative stress [22] can occur, leading to the accumulation of lipid peroxidation products, including free radicals. Oxidative stress is the main reason for the decrease in the activity of NO syn- thase (NOS) through a decrease in the availability of the cofactor NOS-tetrahydrobiopterin and, subsequently, the inhibition of the enzymatic synthesis of NO. In our study, lower values of the NOx level during the test “to exhaustion” were observed in Group I skiers compared with athletes in Group II, which may indicate a decrease in the enzymatic synthe- sis of NO in athletes of Group I. The adaptive capacity of the body reduction with a decrease in the level of NO in the tissues, and pathological changes in metabolism are observed, leading to diseases. The primary cause of the pathogenesis of coronary heart disease and atherosclerotic vascular damage is a deficiency of NO in the vascular endothelium and myocardium [23]. There are several factors causing endothelial NO deficiency: a decrease in eNOS activity [24], destruction or capture of NO by free radicals, and/or a weakening of the effect of NO on smooth muscle [25]. In general, oxidative stress and hypoxia are believed to cause overproduction of NO, often exceeding its physiological level [26,27]. In the body under oxidative stress, NO processes take place; for example, NO covalently binds to a cysteine residue in the beta-chain of Hb to form S-nitrosohaemoglobin, as well as other proteins, which has a regulatory effect on the local tissue blood supply during hypoxic vasodilation [28]. With gradual adaptation to oxidative stress (hypoxia), the plasma level of stable NO metabolites—nitrates and nitrites—progressively increases, correlating with an increase in the vascular NO depot [29]. Perhaps this mechanism of adaptation to the gradually increasing hypoxia caused by physical exertion of maximum power was observed in the elite athletes we examined. Physiological effects [30] aimed at improving oxygen supply during hypoxia are well documented and include increased ventilation and cardiac output, erythropoiesis, and tissue vascularization [31]. Since V’O2 is determined by the interaction of several Cells 2022, 11, 39 12 of 16 factors—blood flow, blood O2 carrying capacity, diffusion of O2 from blood to tissues, ATP demands and O2 utilization by mitochondria—it is clear that the result of changes in NO production on V’O2 depends on the balance between often opposite effects at different levels of the phenomena occurring. Lactate levels in blood and tissues are assumed to increase during hypoxia. After 8 h of breathing in a chamber with hypoxic conditions of 12% O2, a moderate increase in lactate levels of 29% was observed [32]. In our study, athletes of different sports classifications showed an increase in the level of lactate in the blood during the period of maximum exercise. Compared with the indices at rest in Group I, lactate increased by 3.3 times, and in Group II by 5.2 times (p < 0.001); therefore, the athletes examined by us experienced hypoxia during physical activity. Moreover, in Group II, hypoxia had a more pronounced character and a higher increase in V’O2 during the period of maximum load. As shown earlier, including our own study, fluctuations in the level of NO in the human body trigger various adaptive reactions under conditions of acute hypoxia [33,34]. Our data indicate that the baseline lactate level before the test “until exhaustion” is also significant, and this indicator is negatively correlated with the NOx indicator. In the literature, in some clinical conditions, there is a negative correlation between lactate and NO; for example, in acute brain injury [35] and in major surgical operations [36], although with septic shock, the use of L-NMMA, an iNOS inhibitor, increases the level of lactate in the thigh muscle [37]. NO is a mediator of skeletal muscle function, especially NO, which affects cellular respiration and contractility, and in working skeletal muscle, inhibition of NOS improves the economy of muscle contraction and leads to a decrease in the outflow of lactate from the muscles by reducing oxygen cost [38]. Thiol reactions or reactive metal centres in proteins can cause NO responses for further biological events in skeletal muscle. The NO-mediated response inhibits haeme-containing proteins such as cytochrome c oxidase, thus inhibiting the function of cytochrome c oxidase and cell respiration [39]. Cytochrome c oxidase and the sarcoplasmic reticulum Ca21- ATPase in fast-twitch and slow-twitch skeletal muscle inhibits by NOS activity which in turn leads to inhibiting mitochondrial respiration in skeletal muscle [40]. Moreover, aconitase and complex I of the respiratory chain can be additional targets of NO [41]. NO is crucial for the activation and inhibition of ryanodine receptors (RyRs) [42]. For muscle contraction and excitation, the release of Ca2+ into the cytosol is necessary, the RyR play a decisive role in this process. In recent years, special properties of NO2 have been shown, which make it possible to recognize it as the most significant biologically active signalling molecule [43]. The significant vasodilatory response observed in vivo and in vitro experiments upon administration of NO2 solutions suggests that it can be an alternative source of NO [44]. NO2 was also reported to participate in adaptation to physiological conditions of hypoxia, for example, caused by physical exertion [45]. Modern concepts of nitrite-dependent mechanisms of adaptation to hypoxia are based on data on the participation of NO2 in oxygen-dependent and hypoxia-dependent nitrite reductase processes. NO released as a result of these processes is involved in the regulation of vascular tone, modulation of mitochondrial redox reactions [46], changes in the sensitivity of heart contractile proteins to oxygen [47] and calcium ions [48], and inhibition of induced NOS. Due to the cyclic metabolic transformations of NO, NO2 and NO3, the optimal level of NO is maintained, which is necessary for the normal functioning of the cardiovascular system in conditions of impaired functioning of NO synthases. However, excess NO is removed by the formation of an NO depot in the form of NO2, which protects tissues from oxidative and nitrosative stress. The source of vasoactive NO has now been established to also be NO2. It is always present in the blood and can be enzymatically reduced to NO under the action of xanthine oxidoreductase and nonenzymatic under conditions of low pH and pO2 [49]. Nitrate is reduced to nitrite and nitric oxide, which activate soluble guanylate cyclase [50]. Exercise has been shown to increase plasma nitrite levels by increasing NO synthesis in endothelial cells [51,52]. Plasma nitrite is oxidized to nitrate. This process is significantly accelerated by the presence of haeme proteins [53]. Nitrates are Cells 2022, 11, 39 13 of 16 stable in plasma until excreted in the urine. The circulation of nitrite, rather than nitrate, indicates endothelial-dependent synthesis of NO [54]. Increased NO production during exercise is likely controlled by increased nitrate excretion as a possible mechanism for controlling plasma homeostasis [55]. When oxygen pressure decreases, nitrite reduction by deoxyhaemoglobin produces NO. The generation and release of NO by erythrocytes, along with the oxygen concentration gradient, can be associated with the role of nitrite bound to erythrocytes in the processes of vasodilation in response to hypoxia. The indicators of correlation between the studied indices are of particular interest for discussion. Notably, there was a positive relationship between the parameters of V’O2max and lactate in Group 1 at all stages of testing. However, in Group II, correlations between these indicators were not revealed at all. It is also necessary to consider the correlation relationship between indicators of metabolites of nitric oxide and lactate during testing. Initially, at rest, NOx and lactate in both groups showed a significant negative correlation. In Group I, at the AT, correlation disappeared, and at the maximum load, the correlation was revealed between NO2 and lactate opposite direction. In Group II, at the AT, there was a positive relationship between NOx and lactate indicators, which inverted at maximum load. Thus, among skiers of Group II, a positive correlation of NO and lactate indicators was detected at the AT, and in Group I, the positive correlation of NO2 and lactate indicators was detected only at the maximum load. This observation suggests the existence of an adaptive mechanism for regulating the level of lactate at the AT in highly skilled skiers. The methodology of the training process in cyclic sports is known to be based on the principle of increasing the aerobic performance of the body of athletes. The phenomenon identified by us, in our opinion, may be reflects the positive effect of NO on the production of lactate, which provides an increase in aerobic performance. At the same time, at maximum load, on the contrary, it is required to maximize the activity of glycolysis and, accordingly, the production of lactate. In our opinion, activation of the reroute pyruvate away from pyruvate dehydrogenase (PDH) in an NO-dependent mechanism may be a possible mechanism explaining what is happening at the testing stage of highly skilled skiers at the AT (thereby promoting glutamine-based anaplerosis). The capability of this process was established in a recently published post [56]. To a certain extent, this hypothesis can be evidenced by our materials presented in Table 6. At rest, there was no correlation between lactate and V’O2 max in either group. However, in Group I, during exercise, a positive correlation of V’O2 max and lactate indices was manifested, which increased during the test until the recovery period. In contrast, in Group II during the entire test, there was no significant correlation between V’O2 max and lactate. In our opinion, this lack of significant correlation indicates that the level of lactate in the blood of highly qualified skiers–racers, in contrast to less qualified athletes with submaximal and maximum physical activity, does not depend on the parameters of the V’O2 max. It is likely that these athletes have developed additional adaptive mechanisms for regulating the production of lactate, one of which may be the activation of pyruvate abstraction through pyruvate dehydrogenase under conditions of aerobic work. 5. Conclusions The results of our research indicate that the studied groups of athletes with high but different levels of sports qualifications are a convenient model for studying the molecular mechanisms of adaptation to physical loads of maximum intensity. Athletes of higher qualifications reveal additional adaptive mechanisms of metabolic regulation, which is manifested in the independence of serum lactate indicators under conditions of submaximal and maximum power from V’O2 max, and they have an NO-dependent mechanism for regulating lactate levels during aerobic exercise, including work at the AT. Limitation: 1. The sample is small to suggest solid conclusions. However, our study included 124 highly trained cross-countries skiers, who, at the time of testing, were current members of the national team of the Komi Republic and Russia. All participants were exposed to hard Cells 2022, 11, 39 14 of 16 screening for a number of indicators and were unified. 2. An “until exhaustion” assessment needs further investigation due to the low degree of inter-sample analysis. 3. NO is not the only protagonist to guide the metabolic pathways. In the literature, much attention is paid to studying the role of the participation of nitric oxide interaction during inflammatory in rats [57,58]. At the same time, in our paper, we tried to reveal the mechanisms of the participation of nitric oxide in the process of adaptation to regular, hard and intense physical activity of healthy highly qualified athletes, who have developed specific mechanisms of adaptation to these loads. Author Contributions: All authors participated in designing the experiment. O.I.P. participated in the experimental procedure, carried out the biochemical studies, performed the statistical analysis and drafted the manuscript; N.G.V. participated in the experimental procedure, involved in data collection, and drafted the manuscript; N.N.P. carried out the biochemical studies, helped to draft the manuscript; A.Y.L. helped the statistical analysis; E.R.B. oversaw the experimental procedures, provided coordination, helped with drafting of the manuscript. All authors have agree with the order of presentation of the authors. All authors have read and agreed to the published version of the manuscript. Funding: The study was carried out at the expense of subsidies for the implementation of State Assignment No. GR 1021051201877-3-3.1.8. Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and approved by the Ethics Committee of Institute of Physiology of Komi Science Centre of the Ural Branch of the Russian Academy of Sciences (date of approval: 23.10.2013). Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: The data that were generated and/or analyzed during the current study are available from the corresponding author upon reasonable request. Conflicts of Interest: The authors declare no conflict of interest. References 1. Baumgartner, L.; Schulz, T.; Oberhoffer, R.; Weberruß, H. Influence of vigorous physical activity on structure and function of the cardiovascular system in young athletes—The MuCAYA-Study. Front. Cardiovasc. Med. 2019, 6, 148. [CrossRef] [PubMed] 2. Varro, A.; Baczko, I. Possible mechanisms of sudden cardiac death in top athletes: A basic cardiac electrophysiological point of view. Cardiovasc. Physiol. 2010, 460, 31–40. [CrossRef] [PubMed] 3. Turbasova, N.V.; Bulygin, A.S.; Revnivykh, I.Y.; Karpov, N.V.; Elifanov, A.V. Anxiety level and parameters of the cardiovascular system in athletes of various qualifications. Hum. Sport Med. 2019, 19, 14–19. [CrossRef] 4. Saltin, B.; Astrand, P.O. Maximal oxygen uptake in athletes. J. Appl. Physiol. 1967, 23, 353–358. [CrossRef] 5. Parshukova, O.I.; Varlamova, N.G.; Bojko, E.R. Nitric oxide production in professional skiers during physical activity at maximum load. Front. Cardiovasc. Med. Hypertens. 2020, 7, 1–8. [CrossRef] [PubMed] 6. Yol, Y.; Turgay, F.; Yigitturk, O.; Asıkovalı, S.; Durmaz, B. The effects of regular aerobic exercise training on blood nitric oxide levels and oxidized LDL and the role of eNOS intron 4a/b polymorphism. BBA Molec. Bas. Dis. 2020, 1866, 165913. [CrossRef] [PubMed] 7. Windhaber, J.; Steinbauer, M.; Holter, M.; Wieland, A.; Kogler, K.; Riedl, R.; Schober, P.; Castellani, C.; Singer, G.; Till, H. Bicycle spiroergometry: Comparison of standardized examination protocols for adolescents: Is it necessary to define own standard values for each protocol? Eur. J. Appl. Physiol. 2021, 121, 1783–1794. [CrossRef] 8. Dickstein, K.; Barvik, S.; Aarsland, T.; Snapinn, S.; Karlsson, J. A comparison of methodologies in detection of the anaerobic threshold. Circulation 1990, 81, 38–46. [PubMed] 9. Myers, J.; Ashley, E. Dangerous curves. A perspective on exercise, lactate, and the anaerobic threshold. Chest 1997, 111, 787–795. [CrossRef] [PubMed] 10. Granger, D.L.; Taintor, R.R.; Boockvar, K.S.; Hibbs, J.B., Jr. Measurement of nitrate and nitrite in biological samples using nitrate reductase and Griess reaction. Methods Enzymol. 1996, 268, 142–151. [CrossRef] 11. Lorenz, D.S.; Reiman, M.P.; Lehecka, B.J.; Naylor, A. What performance characteristics determine elite versus nonelite athletes in the same sport? Sports Health 2013, 5, 542–547. [CrossRef] 12. Mitic, P.; Nedeljkovic, J.; Bojanic, Z.; Francesko, M.; Milovanovic, I.; Bianco, A.; Drid, P. Differences in the psychological profiles of elite and non-elite athletes. Front. Psychol. 2021, 12, 635651. [CrossRef] [PubMed] 13. Calbet, J.A.L.; Jensen-Urstad, M.; Van Hall, G.; Holmberg, H.C.; Rosdahl, H.; Saltin, B. Maximal muscular vascular conductances during whole body upright exercise in humans. J. Physiology 2004, 558, 319–331. [CrossRef] Cells 2022, 11, 39 15 of 16 14. Martin, S.A.; Hadmas,, R.M. Individual adaptation in cross-country skiing based on tracking during training conditions. Sports 2019, 7, 211. [CrossRef] [PubMed] 15. Sandbakk, O.; Holmberg, H.C. Physiological capacity and training routines of elite cross-country skiers: Approaching the per limits of human endurance. Int. J. Space Physiol. Perform. 2017, 1, 1003–1011. [CrossRef] 16. Gjovaag, T.; Hjelmeland, A.K.; Oygard, J.B.; Vikne, H.; Mirtaheri, P. Acute hemodynamic and cardiovascular responses following resistance exercise to voluntary exhaustion. Effects of different loadings and exercise durations. J. Sports Med. Phys. Fit. 2016, 56, 616–623. [PubMed] 17. Varlamova, N.G.; Zenchenko, T.A.; Boyko, E.R. Annual dynamics of blood pressure and me-Teosensitivity in women. Ther. Arch. 2017, 12, 56. [CrossRef] 18. Mercer, J.B.; Osterud, B.; Tveita, T. The effect of short-term cold exposure on risk factors for cardiovascular diseases from China. Thromb. Res. 1999, 95, 93–104. [CrossRef] 19. Luo, B.; Zhang, S.; Ma, S.; Zhou, J.; Wang, B. Artificial cold air increases the cardiovascular risks in spontaneously hypertensive rats. Int. J. Environ. Res. Public Health 2012, 9, 3197–3208. [CrossRef] [PubMed] 20. Oh, D.; Hong, H.; Lee, B. The effects of strenuous exercises on resting heart rate, blood pressure, and maximal oxygen uptake. J. Exerc. Rehabil. 2016, 12, 42–46. [CrossRef] 21. de Sire, A.; Marotta, N.; Marinaro, C.; Curci, C.; Invernizzi, M.; Ammendolia, A. Role of physical exercise and nutraceuticals in modulating molecular pathways of osteoarthritis. Int. J. Mol. Sci. 2021, 22, 5722. [CrossRef] [PubMed] 22. Heitzer, T.; Krohn, K.; Albers, S.; Meinertz, T. Tetrahydrobiopterin improves endothelium-dependent vasodilation by increasing nitric oxide activity in patients with type II Diabetes mellitus. Diabetologia 2000, 43, 1435–3148. [CrossRef] 23. Besedina, A. NO-synthase activity in patients with coronary heart disease associated with hypertension of different age Groups. J. Med. Biochem. 2016, 35, 43–49. [CrossRef] 24. Chou, T.C.; Yen, M.H.; Li, C.Y.; Ding, Y.A. Alterations of nitric oxide synthase expression with aging and hypertension in rats. Hypertension 1998, 31, 643–648. [CrossRef] 25. Kizub, I.V.; Kharchenko, O.I.; Kostiuk, O.S.; Ostapchenko, L.I.; Soloviev, A.I. Protein kinase C (PKC) involved in enhancement of alpha (1)-adrenoceptor-mediated responses of the main pulmonary artery in rats with diabetes mellitus. Regul. Mech. Biosyst. 2017, 8, 287–292. [CrossRef] 26. Node, K.; Kitakaze, M.; Kosaka, H.; Komamura, K.; Minamino, T.; Inoue, M.; Tada, M.; Hori, M.; Kamada, T. Increased release of NO during ischemia reduces myocardial contractility and improves metabolic dysfunction. Circulation 1996, 93, 356–364. [CrossRef] [PubMed] 27. Hampl, V.; Herget, J. Role of nitric oxide in the pathogenesis of chronic pulmonary hypertension. Physiol. Rev. 2000, 80, 1337–1372. [CrossRef] 28. Allen, B.W.; Stamler, J.S.; Piantadosi, C.A. Hemoglobin, nitric oxide and molecular mechanisms of hypoxic vasodilation. Trends Mol. Med. 2009, 15, 452–460. [CrossRef] 29. Malyshev, I.Y.; Zenina, T.A.; Golubeva, L.Y.; Saltykova, V.A.; Manukhina, E.B.; Mikoyan, V.D.; Kubrina, L.N.; Vanin, A.F. NO-dependent mechanisms of adaptation to hypoxia. Nitric Oxide 1999, 3, 105–113. [CrossRef] 30. Murray, A.J. Metabolic adaptation of skeletal muscle to high altitude hypoxia: How new technologies could resolve the controversies. Genome Med. 2009, 1, 117. [CrossRef] 31. Sutton, J.R.; Reeves, J.T.; Wagner, P.D.; Groves, B.M.; Cymerman, A.; Malconian, M.K.; Rock, P.B.; Young, P.M.; Walter, S.D.; Houston, C.S. Operation Everest II: Oxygen transport during exercise at extreme simulated altitude. J. Appl. Physiol. 1988, 64, 1309–1321. [CrossRef] [PubMed] 32. Van Patot, M.C.T.; Serkova, N.J.; Haschke, M.; Kominsky, D.J.; Roach, R.C.; Christians, U.; Henthorn, T.K.; Honigman, B. Enhanced leukocyte HIF-1alpha and HIF-1 DNA binding in humans after rapid ascent to 4300 m. Free Radic. Biol. Med. 2009, 46, 1551–1557. [CrossRef] 33. Boiko, E.R.; Burykh, E.A. Nitric oxide metabolites level in human serum in acute normobaric hypoxia. Rossiiskii Fiziologicheskii Zhurnal Imeni IM Sechenova 2012, 98, 147–154. [PubMed] 34. MacInnis, M.J.; Carter, E.A.; Donnelly, J.; Koehle, M.S. A meta-analysis of exhaled nitric oxide in acute normobaric hypoxia. Aerosp. Med. Hum. Perform. 2015, 86, 693–697. [CrossRef] 35. Carpenter, K.L.; Timofeev, I.; Al-Rawi, P.G.; Menon, D.K.; Pickard, J.D.; Hutchinson, P.J. Nitric oxide in acute brain injury: A pilot study of NO(x) concentrations in human brain microdialysates and their relationship with energy metabolism. Acta Neurochir. 2008, 102, 207–213. [CrossRef] 36. Fujioka, S.; Noguchi, T.; Watanabe, T.; Takatsuto, S.; Yoshida, S. Biosynthesis of brassinosteroids in cultured cells of Catharanthus roseus. Phytochemistry 2000, 53, 549–553. [CrossRef] 37. Levy, B.; Valtier, M.; De Chillou, C.; Bollaert, P.E.; Cane, D.; Mallie, J.P. Beneficial effects of L-canavanine, a selective inhibitor of inducible nitric oxide synthase, on lactate metabolism and muscle high energy phosphates during endotoxic shock in rats. Shock 1999, 11, 98–103. [CrossRef] [PubMed] 38. Krause, D.S.; Van Etten, R.A. Tyrosine kinases as targets for cancer therapy. N. Engl. J. Med. 2005, 353, 172–187. [CrossRef] 39. Borutaite, V.; Brown, G.C. Rapid reduction of nitric oxide by mitochondria, and reversible inhibition of mitochondrial respiration by nitric oxide. Biochem. J. 1996, 1, 295–299. [CrossRef] [PubMed] Cells 2022, 11, 39 16 of 16 40. Klebl, B.M.; Ayoub, A.T.; Pette, D. Protein oxidation, tyrosine nitration, and inactivation of sarcoplasmic reticulum Ca21-ATPase in low-frequency stimulated rabbit muscle. FEBS Lett. 1998, 422, 381–384. [CrossRef] 41. Clementi, E.; Brown, G.C.; Feelisch, M.; Moncada, S. Persistent inhibition of cell respiration by nitric oxide: Crucial role of Snitrosylation of mitochondrial complex I and protective action of glutathione. Proc. Natl. Acad. Sci. USA 1998, 95, 7631–7636. [CrossRef] 42. Stamler, J.S.; Meissner, G. Physiology of nitric oxide in skeletal muscle. Physiol. Rev. 2001, 81, 209–237. [CrossRef] 43. Mazzone, M.; Carmeliet, P. Drug discovery: A lifeline for suffocating tissues. Nature 2008, 453, 1194–1195. [CrossRef] 44. Schulman, I.H.; Hare, J.M. Regulation cardiovascular processes by S-nitrosylation. Biochim. Biophys. Acta 2012, 1820, 752–762. [CrossRef] 45. Gladwin, M.; Shelhamer, J.; Schechter, A.; Pease-Fye, M.; Waclawei, M.; Panza, J.; Oguibene, F.; Cannon, R. Role of circulating nitrite and S-nitrosohemoglobin in the regulation of the regional blood floow in humans. Proc. Natl. Acad. Sci. USA 2000, 97, 11482–11486. [CrossRef] 46. Gladwin, M.T.; Kim-Shapiro, D.B. The functional nitrite reductase activity of the heme globins. Blood 2008, 112, 2636–2647. [CrossRef] 47. Khan, S.A.; Skaf, M.N.; Harrison, R.W.; Lee, K.; Minhas, K.M.; Kumor, A.; Fradley, M.; Shoukas, A.; Berkowitz, D.E.; Hare, J.M. Nitric oxide regulation of myocardial contractility and calcium cycling. Circ. Res. 2003, 92, 1322–1329. [CrossRef] 48. Layland, J.; Li, J.; Shah, A.M. Role cyclic GMP-dependent protein kinase in the contractile response to exogenous nitric oxide in rat cardiac myocytes. J. Physiol. 2002, 540, 457–467. [CrossRef] [PubMed] 49. Godber, B.L.; Doel, J.J.; Sapkota, G.P.; Blake, D.R.; Stevens, C.R.; Eisenthal, R.; Harrison, R. Reduction of nitrite to nitric oxide catalyzed by xanthine oxidoreductase. J. Biol. Chem. 2000, 275, 7757–7763. [CrossRef] [PubMed] 50. Zebrowska, A.; Mizia-Stec, K.; Mizia, M.; G ˛asiorb, Z.; Poprz˛ecki, S. Omega-3 fatty acids supplementation improves endothelial function and maximal oxygen uptake in endurance-trained athletes. Eur. J. Sport Sci. 2015, 15, 305–314. [CrossRef] [PubMed] 51. Wang, J.S.; Chow, S.E.; Chen, J.K. Strenuous, acute exercise a Vects reciprocal modulation of platelet and polymorphonuclear leukocyte activities under shear Xow in men. J. Thromb. Haemost. 2003, 1, 2031–2037. [CrossRef] 52. Delp, M.D.; Laughlin, M.H. Time course of enhanced endothelium mediated dilation in aorta of trained rats. Med. Sci. Sports Exerc. 1997, 29, 1454–1461. [CrossRef] [PubMed] 53. Allen, J.D.; Cobb, F.R.; Gow, A.J. Regional and whole-body markers of nitric oxide production following hyperemic stimuli. Free Radic. Biol. Med. 2005, 38, 1164–1169. [CrossRef] 54. Kleinbongard, P.; Dejam, A.; Lauer, T.; Rassaf, T.; Picker, S.A.O.; Scheeren, T.; Godecke, A.; Schrader, J.; Schulz, R.; Heusch, G.; et al. Plasma nitrite resects constitutive nitric oxide synthase activity in mammals. Free Radic. Biol. Med. 2003, 35, 790–796. [CrossRef] 55. Bode-Boger, S.M.; Boger, R.H.; Schroder, E.P.; Frolich, J.C. Exercise increases systemic nitric oxide production in men. J. Cardiovasc. Risk 1994, 1, 173–178. [CrossRef] [PubMed] 56. Palmieri, E.M.; Gonzalez-Cotto, M.; Baseler, W.A.; Davies, L.C.; Ghesquière, B.; Maio, N.; Rice, C.M.; Rouault, T.A.; Cassel, T.; Higashi, R.M.; et al. Nitric oxide orchestrates metabolic rewiring in M1 macrophages by targeting aconitase 2 and pyruvate dehydrogenase. Nat. Commun. 2020, 4, 698. [CrossRef] [PubMed] 57. Ilari, S.; Dagostino, C.; Malafoglia, V.; Lauro, F.; Giancotti, L.A.; Spila, A.; Proietti, S.; Ventrice, D.; Rizzo, M.; Gliozzi, M.; et al. Protective effect of antioxidants in nitric oxide/COX-2 interaction during inflammatory pain: The role of nitration. Antioxidants 2020, 9, 1284. [CrossRef] [PubMed] 58. Iorio, G.C.; Ammendolia, A.; Marotta, N.; Ricardi, U.; Sire, A. A bond between rheumatic diseases and cancer in the elderly: The interleukin-6 pathway. Int. J. Rheum. Dis. 2021, 24, 1317–1320. [CrossRef]

Features of Metabolic Support of Physical Performance in Highly Trained Cross-Country Skiers of Different Qualifications during Physical Activity at Maximum Load.

12-23-2021

Parshukova, Olga I,Varlamova, Nina G,Potolitsyna, Natalya N,Lyudinina, Aleksandra Y,Bojko, Evgeny R

eng

PMC10462575

Vol.:(0123456789) 1 3 Archives of Dermatological Research (2023) 315:2271–2281 https://doi.org/10.1007/s00403-023-02588-4 ORIGINAL PAPER Clinical study of a spray containing birch juice for repairing sensitive skin Xiaohong Shu1 · Shizhi Zhao3 · Wei Huo1 · Ying Tang1 · Lin Zou1 · Zhaoxia Li1 · Li Li1,2 · Xi Wang1,2 Received: 2 August 2022 / Revised: 14 December 2022 / Accepted: 16 February 2023 / Published online: 24 March 2023 © The Author(s) 2023 Abstract Sensitive skin is described as an unpleasant sensory response to a stimulus that should not cause a sensation. Sensitive skin affects an increasing proportion of the population. Sixty-seven participants who tested positive to lactic acid sting test were recruited and randomized into two groups to observe the clinical efficacy and safety of a new birch juice spray for repairing sensitive skin. One group used test spray A, while the other group used spray B as a control. Both groups were sprayed six times daily for 28 days. Noninvasive testing instruments were used to measure stratum corneum hydration, sebum content, transepidermal water loss rates, skin blood perfusion and current perception threshold before and after using spray. Facial images were captured by VISIA-CR, and the image analysis program (Image‐Pro Plus) was used to analyze these to obtain the redness value of the facial skin. Moreover, lactic acid sting test scores and participants’ self-assessments were also performed at baseline, week 2 and week 4. Both sprays A and B significantly decreased the lactic acid sting test score, transepidermal water loss rates, skin blood perfusion, and redness, while increasing the stratum corneum hydration. Compared to spray B, spray A increased sensory nerve thresholds at 5 Hz and decreased the transepidermal water loss rates, skin blood perfusion, and lactic acid sting test score. Sprays containing birch juice improved cutaneous biophysical properties in participants with sensitive skin. Keywords Birch juice · Sprays · Sensitive skin · Clinical efficacy Introduction Sensitive skin refers to a high reaction state occurring under physiological or pathological conditions, mainly in the face, manifested as the skin being prone to burning, tingling, itching, and tension when stimulated by physi- cal, chemical, physiological, and psychological factors, accompanied or not accompanied by erythema, scales, tel- angiectasia and other objective signs [1, 2]. Sensitive skin symptoms often occur repeatedly and seriously affect the appearance of a patient, causing great physical and psycho- logical pressure [3]. Factors affecting sensitive skin include external and inter- nal factors. External factors include environmental factors (humidity, temperature/climate change, environmental pol- lution, ultraviolet radiation, wind) and lifestyle (cosmet- ics, diet, and alcohol), while internal factors include sex, age, hormone level, emotion, stress, and genetic factors [4, 5]. Recently, the number of people with sensitive skin has increased due to air pollution and the diversified use of cos- metics. Approximately 50% of people claim that they have sensitive skin, and this proportion is gradually rising [6–8]. Therefore, nursing sensitive skin is increasingly important. The pathophysiology of sensitive skin includes damage to the skin barrier, an enhanced immune response, and height- ened neurosensory sensitivity. In the Muizzuddin classifi- cation, skin sensitivity can be classified into three types: skin barrier damage type, inflammatory response type, and nerve hyper-response type [4]. These types can influence and interact with each other, suggesting that in nursing sensitive skin, considering the three dimensions is crucial. At present, * Xi Wang [email protected] 1 Cosmetics Evaluation Center, West China Hospital, Sichuan University, Chengdu, Sichuan, People’s Republic of China 2 Department of Dermatology, West China Hospital, Sichuan University, No. 37 Guo Xue Xiang, Chengdu, Sichuan 610041, People’s Republic of China 3 Yoseido (Shanghai) Cosmetics R&D Co., Ltd., Shanghai, People’s Republic of China 2272 Archives of Dermatological Research (2023) 315:2271–2281 1 3 studies on sensitive skin care have been more extensive [9], although reports on the comprehensive clinical care of sensi- tive skin from multiple dimensions are few. Birch juice, a colorless and transparent or can be slightly yellow fresh juice from the birch, contains rich nutrients and biologically active material. It is often regarded as a kind of simple and quick beverage, and it also has medicinal and cosmetic effects [10–14]. To date, birch juice contains 11 kinds of fatty acids, 18 kinds of amino acids, 4 kinds of vitamins and 18 kinds of mineral elements as well as compounds of nicotinic acid, essential oil, betula bud acid, saponin, cell division elements, growth elements, sulfur ammonia elements, and pyridoxine [15–17]. Among them, amino acids, fatty acids, and vitamins play an important role in maintaining the skin barrier function, reducing inflammation, skin moisturizing, wound healing, and whitening [18–22]. The rich mineral elements in birch juice are also very valuable for skin care [23, 24]. However, the clinical application of birch juice for repairing sensitive skin has not been reported. Thus, this study used multidimensional methods to evaluate the repair effect of a moisturizing spray containing natural birch sap on sensitive skin and compared it with commercially available sprays. Materials and methods Study design This randomized, double-blind, clinical study was conducted. The research protocol was reviewed and approved by the Institutional Ethics Committee. All participants provided informed consent. Study participants Altogether, 67 people were selected, of which 33 were assigned to the Group A and 34 were in the Group B by ran- dom software distribution. The inclusion criteria were as fol- lows: aged between 18 and 60; in good health; with positive lactic acid stimulation test and experiences skin discomfort when the season changes in previous years; participated in the study voluntarily and signed the informed consent form; and able to strictly comply with the requirements of the study protocol, use the product as required, and complete follow-up. The exclusion criteria were as follows: pregnant or lactating women; with skin diseases (such as psoriasis, eczema, psoriasis, and skin cancer), evident erythema, sun- burn, wound, wear, and tattoo, in or near the test area; have participated in other clinical studies or been treated by der- matologists within the last 3 months; and used any other anti-allergy products within the last 3 months. Rejection and termination were considered when the participants requested to discontinue the test or when adverse reactions occurred, respectively. Test spray The test product (A) was a spray containing natural birch juice and birch bark extract. The control product (B) was a spray containing thermal spring water on the market. The main ingredients of spray A and spray B are shown in Table 1. Treatments All participants were asked to use the spray six times a day 15–20 cm from the face by pressing the pump head in a circular motion and spraying on the face without it. One group received spray A, and the other received spray B. To ensure dose compliance, the volume of residual spray in the container was examined at each follow-up. Assessments of skin biophysical properties were performed at the indicated times. Evaluation method The VISIA-CR 4.1 skin analysis imaging system (Can- field Imaging Systems, Fairfield, NJ, USA) equipped with a Canon EOS-5Ds Mk III SLR camera was used to cap- ture images from the front and left and right sides at 45°. The images were captured under the following lighting Table 1 Main ingredients of spray A and spray B No Name Main functional ingredient 1 Spray A Betula alba juice ≥ 88% Betula alba juice: Mineral elements (calcium, potassium, magnesium, manganese, sodium, zinc, barium, boron, strontium, ferrum, silicon, cuprum, cobalt, nickel, cadmium), amino acid (lysine, alanine, threonine, cystine, histidine, serine, valine, Isoleucine, methionine, leucine, glycine, phenylalanine, arginine, thyroxine, tryptophan, proline, aspartic acid, glutamic acid), aliphatic acid, monosaccharide, Vitamin (Vitamin C, Vitamin B1, Vitamin B2, Vitamin H) 2 Spray B Silice, trace elements (Al, Ba, Li, Sr, Zn), cations (Co2+, Mg2+, Na+), anions (HCO3 −, SO4 2−, Cl−) 2273 Archives of Dermatological Research (2023) 315:2271–2281 1 3 conditions: standards 1 and 2, cross-polarized, parallel- polarized, and UV. Tewameter® TM300 (MPA, Courage-Khazaka Electronic GmbH, Koln, Germany) was used to investigate transepidermal water loss rates (TEWL) and Corneometer® CM825 (MPA, Courage-Khazaka Electronic GmbH, Koln, Germany) was used for detecting stratum corneum (SC) hydration. TEWL was measured in triplicate on the perioral areas for each subject. SC hydration was tested thrice on the cheekbones. These values were calculated based on their average. Skin blood flow in the cheek was monitored by laser Doppler flowmetry (PeriFlux 5000; Perimed AB, Sweden). The amount of blood perfusion was used to measure skin microcirculation, and the higher the redness of the participant’s facial skin, the greater the value. The current perception threshold (CPT) test was conducted using a Neurometer® CPT/C quantitative sensory nerve detector (Neurotron Inc., Baltimore, MD, United States) via a standardized automatic double-blind test method. The lower maxillary branch of the trident meridian was tested. The electrode water was placed horizontally in the middle of one side of the mandibular bone. The Neurometer® CPT/C at three different frequencies (2000 Hz, 250 Hz, and 5 Hz) is an electric current generator that provides selective stimulation for three subpopulations of sensory nerve fibers in the skin. Typical skin sensory nerves are composed of three main subgroups of nerve fibers: Aβ fibers, which conduct skin sensation and pressure; Aδ fibers, which conduct temperature, pressure, and acute pain; and C fibers, which conduct temperature and chronic pain. The current perception threshold (CPT) of the skin can be quantitatively measured using a CPT/C neurometer which reflects the skin’s sensitivity to stimulation; the lower the CPT value, the more sensitive the sensory nerves in the skin are to stimuli, and vice versa [25]. The photographs under Visia-CR cross-polarized light were evaluated for redness value assessed using an image analysis program (Image Pro-plus 7.0; Media Cybernetic Inc., Rockville, MD, USA). The software quantified the color of the facial skin using the L, a, and b color spaces, where L was lightness, a denoted redness, and b indicated yellowness. The higher the a-value, the reddish the skin. The lactic acid sting test (LAST) was conducted as follows: 50 µl of 10% lactic acid solution was applied on one side of the nasolabial sulcus and cheek, and distilled water was applied on the other side, randomly left and right. The participants were asked about regarding their self-conscious symptoms at 2.5 min and 5 min, respectively, and scored on a 4-point scale (0 for no stinging, 1 for slight stinging, 2 for moderate stinging, and 3 for strong stinging). The two scores were then summed, and a total score ≥ 3 was classified as the LAST positive participant. In the participants’ subjective evaluation, they evaluated their skin condition at the second and fourth week of follow-up. To evaluate whether the skin discomfort caused by changing seasons is less than in previous years, the improvement standard is divided into five levels as follows: more evident discomfort, no change, slightly reduced, reduced, and significantly reduced. The number of participants with significant reduction, reduction and slight reduction is the reduction rate. The evaluation parameters also included prevented the occurrence of new sensitives, repaired sensitivity and no irritation. The score was divided into the following five levels: completely disagree, somewhat disagree, disagree, not disagree, somewhat agree, and completely agree. The total number of participants of complete agreement and some agreement is the agreed rate. In the week 4, the satisfaction evaluation was conducted, and the standard was divided into four levels: very satisfied, satisfied, general, and dissatisfied. The number of very satisfied and satisfied cases is the satisfaction rate. The parameters were evaluated before application and after 2 and 4 weeks of use. The testing environment was maintained at a constant 22 ℃ ± 1 ℃ and 50% ± 5% humidity. Before sampling, the participants sat in a temperature- controlled room quietly for 20 min. Statistical analysis The data were analyzed using SPSS version 19.0 and are expressed as the mean ± standard deviation. The data that followed approximate positive distribution were compared using mixed linear models. For participants’ self-assessment, the Wilcoxon test was used to compare W0 at different time points in the same group, and the Mann–Whitney U test was used to compare the same time difference between groups. Significance was set at p < 0.05. Results Screening results of participants Overall, 67 participants were included in the study, of which 33 were in the Group A, and 34 were in the Group B. Two participants in the Group A and one participant in the Group B could not complete the entire study due to personal reasons. Finally, 31 participants in the Group A, and 32 participants in the Group B completed the trial. The average age of the participants in the Group A was 37.9 ± 12.5 years, while that of the Group B was 39.7 ± 12.7  years. The baseline data for participants were comparable between the two groups (Table 2). 2274 Archives of Dermatological Research (2023) 315:2271–2281 1 3 Improvement of skin barrier function Both groups showed significant reductions on the TEWL value over time (F = 9.156, P < 0.05). After 4 weeks of using spray, the TEWL value was significantly lower than base- line (P < 0.05) (F = 4.151, P < 0.05), and the TEWL value in group A was significantly lower than that in group B (estimate: A:0, B:0.299). No significant interaction effects for Group and times were observed (F = 2.625, P > 0.05) (Fig. 1). As the intervention time increases, the hydration value of participants in the two groups were significantly improved (F = 24.709, P < 0.05), significant differences were found at W0 and W2 (P < 0.05), W0 and W4 (P < 0.05). No sig- nificant difference is found between the groups (F = 3.827, P > 0.05).There was no interaction effect between time points and groups (F = 1.423, P > 0.05)(Fig. 2). Improvement of facial redness A laser Doppler blood flow meter is used to measure the amount of skin microcirculation blood perfusion; the more serious the skin redness is, the higher the value of blood perfusion. The blood perfusion decreased significantly over time (F = 5.053, P < 0.05), significant differences were found at W0 and W4 (P < 0.05). Significant between- group differences were seen (F = 12.733, P < 0.05), group A had significantly lower blood perfusion compared to the group B. (estimate: A: 0, B: 37.107); there was no inter- action effect between time points and groups (F = 0.197, P > 0.05)(Fig. 3). Before and after using the spray, an image analysis program was employed to analyze the photos under the cross-polarized light of Visa CR, and obtain the value of facial redness. The facial redness of the participants in the two groups decreased significantly over time (F = 3.485, P < 0.05). There was no group significant difference (F = 1.124, P > 0.05) and group × time interaction effect (F = 0.084, P > 0.05) (Fig. 4). Figure 5 depicts the compar- ison of three participants before and after using spray A. Table 2 The baseline data Items Group A (n = 31) Group B (n = 33) P value Age, years 37.94 ± 12.50 39.76 ± 12.72 0.554 Sex, female (%) 31 (100) 33 (100) 1.000 Hydration value 60.62 ± 12.44 60.80 ± 12.84 0.748 TEWL 20.23 ± 3.31 17.57 ± 4.94 0.013 Blood perfusion 89.81 ± 62.62 112.49 ± 89.37 0.308 Redness value 6.56 ± 3.80 7.26 ± 3.71 0.803 Current perception threshold (2000 HZ) 105.18 ± 22.32 105.08 ± 27.25 0.975 Current perception threshold (250 HZ) 29.03 ± 15.24 31.12 ± 13.75 0.475 Current perception threshold (5 HZ) 16.52 ± 11.88 13.38 ± 8.08 0.857 Last 4.45 ± 0.72 4.33 ± 0.60 0.326 Fig. 1 Transepidermal water loss value. P1 = time effect; P2 = group effect; P3 = interac- tion group × time 2275 Archives of Dermatological Research (2023) 315:2271–2281 1 3 Fig. 2 Hydration value. P1 = time effect; P2 = group effect; P3 = interac- tion group × time Fig. 3 Blood perfu- sion. P1 = time effect; P2 = group effect; P3 = interac- tion group × time Fig. 4 A value. P1 = time effect; P2 = group effect; P3 = interac- tion group × time 2276 Archives of Dermatological Research (2023) 315:2271–2281 1 3 Improvement of sensory nerve sensitivity Measurement results of sensory nerve stimulation threshold At 5 Hz, there was no difference over time (F = 1.042, P > 0.05) and no effect of interaction between the time and group (F = 0.129, P > 0.05). Significant differences were observed between the groups for the sensory thresh- old at 5 Hz (F = 8.563, P < 0.05), which in group A was higher than that in group B (estimates, A: 0, B: − 4.87). At 250 Hz, there was no significant difference in time effect (F = 0.065, P > 0.05), group effect (F = 1.070, P > 0.05), and time × group effect (F = 1.414, P > 0.05). At 2000 Hz, there was no effect over time (F = 0.165, P > 0.05), no difference between groups (F = 0.404, P > 0.05), and no interaction between the two variables (F = 0.534, P > 0.05) (Fig. 6). Experimental results of LAST After 2  weeks and 4  weeks of using sprays A and B, respectively, the LAST score of the participants in the two groups decreased gradually, and a significant difference was noted compared to baseline (P < 0.05). The comparison between groups indicated that at week 4, the scores of LAST in group A were more significantly reduced than those in group B (P < 0.05) (Table 3). Subjective evaluation Subjective evaluation of the participants Before and after using the spray, the participants self- assessed their skin condition. After using spray A for 4 weeks, 83.87% of the participants believed that their skin discomfort caused by changing seasons was reduced compared with that in previous years. After 4 weeks of using spray B, 78.79% of the participants felt less skin discomfort caused by changing seasons compared with previous years. There was no significant difference between the groups (Table 4). As shown in Table 5, the participants believed that the spray prevented the occurrence of new sensitivities and repaired sensitive skin at W4 visit, the percentage of A and B for both groups was 96.78% and 87.88%, respectively. The difference between the two groups was significant. Most participants in both groups agreed that spray could repaired sensitive skin (96.77% and 93.94% in Group A and Group B, respectively). Significant differences were found between the two groups. More than 96% of the participants rated the Fig. 5 Representative images of three participants before and after the application of spray A 2277 Archives of Dermatological Research (2023) 315:2271–2281 1 3 products as non-irritating in both the groups, no significant differences between groups. As shown in Fig. 7, the satisfaction rate of the participants in group A reached 96.77% after 4 weeks of using spray A, while that of the participants in group B reached 93.94% after 4 weeks of using spray B. There was a significant dif- ference in satisfaction rate between the two groups. Discussion After the use of the birch spray by the Group A, SC hydration increased significantly, and the TEWL and redness degree were significantly reduced, while the CPT was increased, and the LAST score was significantly decreased. Thus, the spray containing natural birch juice can effectively improve the skin barrier function, relieve discomfort such as redness and tingling caused by inflammation, and reduce the sensory nerve sensitivity of the skin. The epidermis is the interface where the human body makes contact with the external environment, and one of its main functions is the barrier function. Impaired skin barrier function is a common pathological mechanism of sensitive skin [4]. The barrier function of the epidermis is closely related to the various lipids, proteins, water, inorganic salts, and other metabolites of the epidermis. The cuticle Fig. 6 Current perception threshold. P1 = time effect; P2 = group effect; P3 = interaction group × time Table 3 The results of the lactic acid sting test at week 4 Compared with week 0; *P < 0.05. Compared with Group  B; #P < 0.05 Test time Group A Group B W0 4.45 ± 0.72 4.33 ± 2.60 W2 3.68 ± 1.22* 3.79 ± 2.06* W4 2.84 ± 1.16*,# 3.42 ± 1.00* Table 4 Subjective evaluation of skin discomfort due to the change of season at week 4 a The number of participants with significant reduction, reduction, and slight reduction is defined as the reduction rate Group Compared with previous years, whether the skin discomfort caused by the change of season was reduced More obvious No change Slightly reduced Reduced Significantly reduced Reduced ratio (%) P Group A 0 (0%) 5 (16.13%) 15 (48.39%) 7 (22.58%) 4 (12.9%) 83.87 0.595 Group B 0 (0%) 7 (21.21%) 14 (42.42%) 12 (36.37%) 0 (0%) 78.79 Table 5 Participants' self-assessment at week 4 # P < 0.05 a The total number of participants expressing complete agreement and some agreement is the agreed rate Compared with Group B Group Completely disagree Somewhat disagree Not disagree Somewhat agree Completely agree Agreed rate (%) P To prevent the occurrence of a new sensitivity Group A 0 (0%) 0 (0%) 1 (3.23%) 8 (25.81%) 22 (70.97%) 96.78 0.001# Group B 0 (0%) 0 (0%) 4 (12.12%) 20 (60.61%) 9 (27.27%) 87.88 Repaired sensitive skin Group A 0 (0%) 0 (0%) 1 (3.23%) 10 (32.26%) 20 (64.52%) 96.77 0.028# Group B 0 (0%) 0 (0%) 2 (6.06%) 19 (57.58%) 20 (36.36%) 93.94 No skin irritation Group A 0 (0%) 0 (0%) 0 (0%) 1 (3.23%) 30 (96.77%) 100 0.982 Group B 0 (0%) 0 (0%) 1 (3.03%) 0 (0%) 32 (96.97%) 96.97 2278 Archives of Dermatological Research (2023) 315:2271–2281 1 3 protects the body, is an important penetration barrier of the skin, and can hinder the loss of percutaneous evaporation of water. TEWL reflects the amount of water evaporation from the skin surface and is therefore an important indica- tor for evaluating skin barrier function [26]. After 4 weeks, the TEWL value was significantly lower (P < 0.05) (Fig. 1), and the SC hydration was significantly different compared to the baseline (P < 0.05) (Fig. 2). Hence, the use of natural birch juice-containing spray significantly improved the skin barrier function in the participants. In recent years, products for sensitive skin have continuously emerged, and the common ingredients in these products are mainly minerals and plant extracts. An impaired skin barrier is one of the main mechanisms of sensitive skin, whose barrier function is mainly undertaken by the cuticle. Calcium ions are closely related to the division and differentiation of keratinocytes, as well as the barrier function of the epidermis. Yuspa et al. have reported that through keratinocyte cultures, high extracellular concentrations of calcium promote keratinocyte differentiation and stratification and glutamine transferase expression, formation of keratinocyte envelopes, and cellular differentiation indicators such as keratin 1, keratin 10 and filromerin; therefore, calcium is essential for the formation of the cuticle barrier [27]. In vitro experiments also revealed that if the isolated skin does not have sufficient calcium ions in the culture medium after the disruption of the barrier function, the calcium ion concentration and concentration gradient will not return to normal and subsequently delay the recovery of the barrier function [28, 29]. In this study, the participants with sensitive skin had significantly reduced TEWL after using a spray containing natural birch juice (Fig. 1), which may be due to the calcium ions in the birch juice that facilitate the repair of the skin barrier function. Here, using the spray containing natural birch juice 4  weeks resulted in a gradual decrease in blood flow perfusion values (Fig. 3), and each value was significantly different from its baseline value (P < 0.05). Additionally, according to the images under cross-polarized light mode, both sprays have the effect of improving the degree of redness when used alone. The difference between the two groups was not statistically significant. Only positive control was set up and no negative control was set up, which is the limitation of this study. That is, negative controls were not used to eliminate the influence of variables such as environmental, seasonal and participants’ own changes on the results. Yamasaki and Gallo have proposed that the innate immune system triggers inflammatory reactions and mediates symptoms of sensitive skin, resulting in skin redness and erythema [6, 30]. Spa therapy is an effective treatment for skin inflammation, and trace elements such as strontium and selenium may be the main effective elements of this therapy. Using a recombinant skin model, Kelerier et al. investigated the regulatory effects of strontium and selenium on inflammatory skin cytokines (IL-1α, TNF-α, and IL-6) and have reported that both strontium and selenium can effectively reduce the production of inflammatory cytokines in inflamed skin [31]. Birch juice contains strontium and selenium, which may be one of the reasons for its ability to reduce skin inflammation and redness. The current perception threshold (CPT) of skin can be quantitatively measured by a CPT/C neurometer, which Fig. 7 Percentage of participants with different satisfaction scores in groups A and B at week 4. P: Group A compared to Group B 2279 Archives of Dermatological Research (2023) 315:2271–2281 1 3 reflects the sensitivity of skin to stimulation. The lower the measured CPT value is, the more sensitive the sensory nerve of the skin is to the stimulus, and conversely, the less sensitive it is [25]. This study results revealed that after 2 and 4 weeks of spraying with the natural birch sap spray, the CPT value did not increase at 2000 Hz and 5 Hz. However, at 250 Hz, after 2 weeks and 4 weeks of spray application, the CPT value increased (Fig. 6), which indicated that the spray A reduced the sensitivity of skin sensory nerves to a certain extent. Studies have suggested that when strontium salt is used to treat faces with sensitive skin, the CPT value is significantly increased, indicating that strontium salt can improve skin sensitivity to stimulation [25]. In 1999, Hahn also reported that strontium salt can significantly reduce the sensory stimulation caused by some substances [32]. Sting sensation is considered to be one of the important characteristics of sensitive skin, so the sting sensation test is widely used in identifying sensitive skin [33]. The lactic acid sting test has been used as a method to identify facial skin sensitivity in many studies. The test is generally targeted at the nasolabial sulcus because this area has high permeability of the cuticle, high density of accessory organs, and a rich sensory neural network. The higher the LAST score is, the more sensitive the skin is; the lower the score is, the less sensitive the skin is [34, 35]. This study’s results indicated that after 2 and 4 weeks of using the spray A, the participants’ LAST scores decreased gradually (Table 3), and a significant difference was observed (P < 0.05) compared with baseline. The LAST score of people with sensitive skin decreased after using repair products have been reported in literature before [36, 37]. We found that the use of both sprays significantly reduced LAST scores, possibly because both sprays contain calcium and magnesium ions, resulting in enhanced epidermal barrier function in subjects with sensitive skin. Moreover, the birch spray can reduce the sensory nerve sensitivity of the skin. A previous work by Eunyoung Lee et al. may explain the underlying mechanism behind these benefits [31]. It has been reported in literature that LAST scores positively correlated with TEWL, a* and EI value [38], were negatively correlated with stratum corneum hydration and current perception threshold (CPT) at 250 Hz [39]. The results of spray A in this study are consistent with the correlation demonstrated in the above literature. That is, the LAST score, TEWL and redness score decreased, CPT at 250HZ and hydration value increased during the 4 weeks of experimental period. Impaired skin barrier function is t he main reason for sensitive skin [6]. When the barrier function of SC is impaired, it is less effective at preventing  water from overevaporating, resulting in TEWL increases, stratum corneum hydration decrease [40], susceptibility to irritation enhanced (the scores of lactic acid sting incread). However, other factors may also have an impact, such as changes in the nervous system and/or epidermal structure. People with sensitive skin often have less hydration, more erythema and more skin with dilated distal blood vessels. From  the  preceding  discussion, calcium ions in birch juice aids in the skin’s barrier functionality, and reducing TEWL and LAST score, increases and maintains the mois- ture content of the skin. Strontium salt of birch juice could improve skin sensitivity to stimulation and increase CPT value. And strontium and selenium could also reduce skin inflammation and redness. Spray B was selected as the control in this study because thermal spring water has been widely reported to improve sensitive skin in several studies [41–44]. Spray A has a higher satisfaction rate than spray B, possibly because of its superior benefit in improving some symptoms of sensitive skin. As mentioned above, some spray A components, such as strontium and selenium, are minerals that reduce the production of inflammatory cytokines in inflammatory skin, reducing inflammation and redness. Calcium ions promote the repair of the skin barrier function. Therefore, Spray A may be more effective than Spray B in improving these symptoms and therefore has a higher satisfaction rate. However, we initially asked about overall satisfaction, which may make these results more positive. If we asked this question at the end of the subject’s self-assessment, they would have had the opportunity to review the shortcomings of the product in detail, and the overall satisfaction would have likely reduced. And this trial is a preliminary study, more research is needed to confirm this hypothesis. There are two limitations to this experiment. First, we did not consider the impact of the environment on sensitive skin, that is, there was no negative control group. Our test site was Chengdu, Sichuan, China (102° 54′–104° 53′ E and 30° 05′–31° 26′ N), and the time was from mid-November to mid-December (average temperature was 6–12 ℃). This period just includes the transition from the end of autumn to the beginning of winter. Seasonal alternation, temperature change, sunlight, and other factors could aggravate sensitive skin [4, 45]. Cold environmental conditions could exert a negative effect on the skin. People exposed to severe weather in winter may experience dry and itchy skin, or their existing skin diseases may worsen [46, 47]. Therefore, the improvement effect of A and B may be masked by seasonal changes. The second was randomization without hierarchical grouping. The basic value distribution of TWEL in the two groups was unbalanced, and the difference was statistically significant. Thus, we adopted a mixed linear model for statistical analysis, and the differences in baselines would not affect the statistical results. In future tests, we will comprehensively consider the influence of age, redness degree, TWEL, and other factors conducting stratified grouping. In addition, considering that 2280 Archives of Dermatological Research (2023) 315:2271–2281 1 3 sensitive skin is prone to relapse, it is necessary to extend the test time, include a suitably sized negative control group, and increase the number of cases. Conclusion The birch spray is safe and effective for repairing sensitive skin, with efficacy and safety comparable to that of a widely accepted sensitive skin repair product. The results of the present study may provide a new option for the repair of sensitive skin. Author contributions Conceptualization, methodology and design: XW and SZ; Data acquisition, data analysis and interpretation: XS, WH, YT, LZ and ZL; Writing—original draft preparation: XS and YT; Writing—review and editing: LL and XW. All authors reviewed the manuscript. Funding None. Data availability The data that support the findings of this study are available from the corresponding author upon reasonable request. Declarations Conflict of interest The authors have no conflict of interest to declare. Ethics approval This study was approved by the Biomedical Ethics Committee of West China Hospital, Sichuan University (2018年审 (351)号). Consent to participate Written informed consent was obtained from all participants. Consent for publication Participants signed informed consent regard- ing publishing their data. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. References 1. Draelos ZD (1997) Sensitive skin: perceptions, evaluation, and treatment. Am J Contact Derm 8(2):67–78 2. Sulzberger M, Worthmann AC, Holtzmann U et al (2016) Effective treatment for sensitive skin: 4-t-butylcyclohexanol and licochalcone A. J Eur Acad Dermatol Venereol 30(Suppl 1):9–17 3. Misery L (2017) Neuropsychiatric factors in sensitive skin. Clin Dermatol 35(3):281–284 4. Inamadar AC, Palit A (2013) Sensitive skin: an overview. Indian J Dermatol Venereol Leprol 79(1):9–16 5. Misery L, Morisset S, Séité S et al (2021) Relationship between sensitive skin and sleep disorders, fatigue, dust, sweating, food, tobacco consumption or female hormonal changes: results from a worldwide survey of 10 743 individuals. J Eur Acad Dermatol Venereol 35(6):1371–1376 6. Berardesca E, Farage M, Maibach H (2013) Sensitive skin: an overview. Int J Cosmet Sci 35(1):2–8 7. Misery L, Weisshaar E, Brenaut E et al (2020) Pathophysiology and management of sensitive skin: position paper from the spe- cial interest group on sensitive skin of the International Forum for the Study of Itch (IFSI). J Eur Acad Dermatol Venereol 34(2):222–229 8. Misery L, Ständer S, Szepietowski JC et al (2017) Definition of sensitive skin: an expert position paper from the special interest group on sensitive skin of the international forum for the study of itch. Acta Derm Venereol 97(1):4–6 9. Achmon Y, Fishman A (2015) The antioxidant hydroxytyrosol: biotechnological production challenges and opportunities. Appl Microbiol Biotechnol 99(3):1119–1130 10. Smiljanic S, Messaraa C, Lafon-Kolb V et al (2022) Betula alba bark extract and empetrum nigrum fruit juice, a natural alter- native to niacinamide for skin barrier benefits. Int J Mol Sci 23(20):12507 11. Isidorov V, Szoka Ł, Nazaruk J (2018) Cytotoxicity of white birch bud extracts: perspectives for therapy of tumours. PLoS ONE 13(8):e0201949 12. Softa M, Percoco G, Lati E et al (2019) Birch sap (betula alba) and chaga mushroom (inonotus obliquus) extracts show anti-oxidant, anti-inflammatory and dna protection/repair activity in vitro. J Cosmet Dermatol Sci Appl 9(2):188–205 13. Wnorowski A, Bilek M, Stawarczyk K et al (2017) Metabolic activity of tree saps of diferent origin towards cultured human cells in the light of grade correspondence analysis and multiple regression modeling. Acta Soc Bot Pol 86(2):1–10 14. Svanberg I, Sõukand R, Luczaj L et al (2012) Uses of tree saps in northern and eastern parts of Europe. Acta Soc Bot Pol 81(4):343 15. Ozolinčius R, Bareika V, Met R et al (2009) Chemical composi- tion of silver birch (Betula pendula Roth) and downy birch (Betula pubescens Ehrh) Sap. Balt For 22(2):222–229 16. Ahtonen S, Kallio H (1989) Identification and seasonal variations of amino acids in birch sap used for syrup production. Food Chem 33(2):125–132 17. Kallio H, Ahtonen S, Raulo J et al (1985) Identification of the sugars and acids in birch sap. J Food Sci 50(1):266–269 18. Solano F (2020) Metabolism and functions of amino acids in the skin. Adv Exp Med Biol 1265(1):187–199 19. Yang M, Zhou M, Song L (2020) A review of fatty acids influenc- ing skin condition. J Cosmet Dermatol 19(12):3199–3204 20. Michalak M, Pierzak M, Bet K et al (2021) Bioactive compounds for skin health: a review. Nutrients 13(1):203 21. Reins RY, Hanlon SD, Magadi S et al (2016) Effects of topi- cally applied vitamin d during corneal wound healing. PLoS ONE 11(4):e0152889 22. Pullar JM, Carr AC, Vissers MCM (2017) The roles of vitamin C in skin health. Nutrients 9(8):866 23. Liu ML, Chun-Hui LV (2015) Application of vitamins and their derivatives in cosmetics. Jiangxi Chem Ind 02:7–8 24. Huang R (2012) Application of minerals in cosmetics. Deterg Cosmet 6:43–46 2281 Archives of Dermatological Research (2023) 315:2271–2281 1 3 25. Lee E, An S, Lee TR et al (2009) Development of a novel method for quantitative evaluation of sensory skin irritation inhibitors. Skin Res Tech 15(4):464–469 26. Angelova-Fischer I (2016) Irritants and skin barrier function. Curr Probl Dermatol 49:80–89 27. Yuspa SH, Kilkenny AE, Steinert PM et al (1989) Expression of murine epidermal differentiation markers is tightly regulated by restricted extracellular calcium concentrations in vitro. J Cell Biol 109(3):1207–1217 28. Elias P, Ahn S, Brown B et al (2003) Origin of the epidermal calcium gradient: regulation by barrier status and role of active vs passive mechanisms. J Invest Dermatol 119(6):1269–1274 29. Elias PM, Wu Y, Chen C et al (2010) Regulation of epidermal barrier functions by calcium. J Clin Dermatol 39(2):135–137 30. Yamasaki K, Gallo RL (2009) The molecular pathology of rosa- cea. J Dermatol Sci 55(2):77–81 31. Celerier P, Richard A, Litoux P et al (1995) Modulatory effects of selenium and strontium salts on keratinocyte-derived inflam- matory cytokines. Arch Dermatol Res 287(7):680–682 32. Hahn GS (1999) Strontium is a potent and selective inhibitor of sensory irritation. Dermatol Surg 25(9):689–694 33. Farage MA, Katsarou A, Maibach HI (2006) Sensory, clinical and physiological factors in sensitive skin: a review. Contact Derm 55(1):1–14 34. Seidenari S, Francomano M, Mantovani L (1998) Baseline bio- physical parameters in subjects with sensitive skin. Contact Derm 38(6):311–315 35. Querleux B, Dauchot K, Jourdain R et al (2008) Neural basis of sensitive skin: a fMRI study. Skin Res Tech 14(4):454–461 36. Jeong S, Lee SH, Park BD et al (2016) Comparison of the efficacy of atopalm multi-lamellar emulsion cream and physiogel intensive cream in improving epidermal permeability barrier in sensitive skin. Dermatol Ther (Heidelb) 6(1):47–56 37. Zhang Y, Jin Y, Humbert P et al (2021) An herbal cream reduces erythema of sensitive skin. J Cosmet Dermatol 20(3):792–797 38. Pan Y, Ma X, Song Y et al (2021) Questionnaire and lactic acid sting test play different role on the assessment of sensitive skin: a cross-sectional study. Clin Cosmet Inv Derm 14:1215–1225 39. Jiang W, Wang J, Zhang H et al (2022) Seasonal changes in the physiological features of healthy and sensitive skin. J Cosmet Der- matol 21(6):2581–2589 40. Berardesca E, Maibach HI (1990) Transepidermal water loss and skin surface hydration in the non invasive assessment of stratum corneum function. Derm Beruf Umwelt 38(2):50–53 41. Ferreira MO, Costa PC, Bahia MF (2010) Effect of São Pedro do Sul thermal water on skin irritation. Int J Cosmet Sci 32(3):205–210 42. Zeichner J, Seite S (2018) From probiotic to prebiotic using ther- mal spring water. J Drugs Dermatol 17(6):657–662 43. Merial-Kieny C, Castex-Rizzi N, Selas B, Mery S, Guerrero D (2011) Avène Thermal Spring Water: an active component with specific properties. J Eur Acad Dermatol Venereol 25(Suppl 1):2–5 44. Bieber T (2011) More scientific evidence for the therapeutic ben- efit of hydrotherapy in Avène. J Eur Acad Dermatol Venereol 25(Suppl 1):1 45. Berardesca E, Farage M, Maibach H (2013) Sensitive skin: an overview. Indian J Dermatol Venereol Leprol 35(1):2–8 46. Engebretsen KA, Johansen JD, Kezic S, Linneberg A, Thyssen JP (2016) The effect of environmental humidity and temperature on skin barrier function and dermatitis. J Eur Acad Dermatol Venereol 30(2):223–249 47. Brenaut E, Barnetche T, Gall-Ianotto CL et al (2020) Triggering factors in sensitive skin from the worldwide patients’ point of view: a systematic literature review and meta-analysis. J Eur Acad Dermatol Venereol 34(2):230 Publisher's Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Clinical study of a spray containing birch juice for repairing sensitive skin.

03-24-2023

Shu, Xiaohong,Zhao, Shizhi,Huo, Wei,Tang, Ying,Zou, Lin,Li, Zhaoxia,Li, Li,Wang, Xi

eng

PMC5720778

RESEARCH ARTICLE Effects of load carriage on physiological determinants in adventure racers Alex de O. Fagundes1, Elren P. Monteiro1,2, Leandro T. Franzoni1, Bruna S. Fraga1, Patrı´cia D. Pantoja1, Gabriela Fischer1,3, Leonardo A. Peyre´-Tartaruga1,4* 1 Exercise Research Laboratory, Escola de Educac¸ão Fı´sica, Fisioterapia e Danc¸a, Universidade Federal do Rio Grande do Sul, Porto Alegre, Rio Grande do Sul, Brazil, 2 Neurosciences and Rehabilitation Laboratory, Universidade Federal de Ciências da Sau´de de Porto Alegre, Porto Alegre, Rio Grande do Sul, Brazil, 3 Centro de Desportos, Universidade Federal de Santa Catarina, Floriano´polis, Brazil, 4 Post-Graduation in Pulmonology Science, Hospital de Clı´nicas de Porto Alegre, Universidade Federal do Rio Grande do Sul, Brazil * [email protected] Abstract Adventure racing athletes need run carrying loads during the race. A better understanding of how different loads influence physiological determinants in adventure racers could pro- vide useful insights to gauge training interventions to improve running performance. We compare the maximum oxygen uptake (VO2max), the cost of transport (C) and ventilatory thresholds of twelve adventure running athletes at three load conditions: unloaded, 7 and 15% of body mass. Twelve healthy men experienced athletes of Adventure Racing (age 31.3 ± 7.7 years, height 1.81 ± 0.05 m, body mass 75.5 ± 9.1 kg) carried out three maximal progressive (VO2max protocol) and three submaximal constant-load (running cost protocol) tests, defined in the following quasi-randomized conditions: unloaded, 7% and, 15% of body mass. The VO2max (unload: 59.7 ± 5.9; 7%: 61.7 ± 6.6 and 15%: 64.6 ± 5.4 ml kg-1 min-1) did not change among the conditions. While the 7% condition does neither modify the C nor the ventilatory thresholds, the 15% condition resulted in a higher C (5.2 ± 0.9 J kg-1 m-1; P = 0.001; d = 1.48) than the unloaded condition (4.0 ± 0.7 J kg-1 m-1). First ventilatory threshold was greater at 15% than control condition (+15.5%; P = 0.003; d = 1.44). Interest- ingly, the velocities on the severe-intensity domain (between second ventilatory threshold and VO2max) were reduced 1% equivalently to 1% increasing load (relative to body mass). The loading until 15% of body mass seems to affect partially the crucial metabolic and venti- latory parameters, specifically the C but not the VO2max. These findings are compatible with the concept that interventions that enhance running economy with loads may improve the running performance of adventure racing’s athletes. Introduction The adventure racing (AR) consists of a multi-sports modality involving running, mountain- bike, canoeing, vertical techniques, and others. During the event, the athletes carry back- packs of different weights (5–10 kg), including obligatory equipment, in distances varying PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 1 / 13 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Fagundes AdO, Monteiro EP, Franzoni LT, Fraga BS, Pantoja PD, Fischer G, et al. (2017) Effects of load carriage on physiological determinants in adventure racers. PLoS ONE 12 (12): e0189516. https://doi.org/10.1371/journal. pone.0189516 Editor: Luca Paolo Ardigò, Universita degli Studi di Verona, ITALY Received: October 11, 2017 Accepted: November 28, 2017 Published: December 7, 2017 Copyright: © 2017 Fagundes et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: This work was supported by the Brazilian Research Council – CNPq under Grant number 483510/2013 and 422193/2016-0; and LAPEX under Grant number 29/2015. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. about 20–100 km. Despite the increasing popularity, few studies have analyzed the crucial aspects related to training workload and running performance with loads in these athletes [1]. The load carrying induces to a higher cost of transport (C) or energy cost, i.e., the energy spent per unit distance covered [2] in comparison to unload condition. However, when the C is scaled linearly to total mass (extra + body mass) no differences, and even reductions in energy expenditure with low loads are found [3, 4]. Possibly, the elastic mechanism is opti- mized for the loaded running [3]. However, metabolic data in maximal situations or at anaero- bic threshold remain unknown. The nutritional aspects, on the other hand, are extensively studied in AR due to its very demanding characteristic from the energetic point of view. A severe negative energetic balance in AR may produce adverse effects on immunological [5], renal [6], and muscular [7] systems. In fact, progressive intensive protein depletion has been verified during adventure races [7] accompanied by a negative energetic balance [8, 9]. Thus, the relevance of knowledge on meta- bolic requirements using loads has been discussed due to the evident impact on the energetic balance of these athletes [5–9]. The determination of intensities corresponding to certain metabolic domains is useful when planning and applying interval and continuous training methods. As for speed and dis- tance, the load carriage would be a factor that affects the metabolic intensity domains. For example, when progressive load-carriage exercise is part of the training program, much larger training effects are evident than aerobic training alone [10, 11]. To the best of our knowledge, evidence-based recommendations for running training with loads are not established for AR participants. Therefore, to date, little is still known about the C and ventilatory thresholds in the specific context of trained adventure athletes. We hypothesized that the differences in the submaximal intensities would be more noticeable with 15% load than 7%, due to energy-saving mechanism acting at low loads as previously shown [3]. Using a laboratory-based measures, we addressed two main research questions: i) is there an effect of load carriage on running maximum oxygen uptake (VO2max), ventilatory thresholds and C in AR athletes? ii) If loads affect these vari- ables, is it possible to define predictive equations to estimate the crucial training intensity markers based on extra load? Materials and methods Participants Based on a minimum increase in C of 5% (~0.3 J kg-1 m-1), a coefficient of variation of 5%, an alpha error of 0.05 and a power of 90%, the minimal number of athletes required for the group was 11. Twelve healthy men, who were national level athletes of AR (age 31.3 ± 7.7 years, height 1.81 ± 0.05 m, body mass 75.5 ± 9.1 kg, training volume 39.12 ± 9.02 km per week, AR experience 64 ± 49 months), carried out three maximal progressive (VO2max protocol) and three submaximal constant-load (running cost protocol) tests, defined in the following quasi- randomized conditions: unloaded, 7% and, 15% of body mass. We choose percent loads due to inherent effects of absolute load on performance. Also, we settle these percent loads because are the loads usually used in AR [11]. All participants gave their written informed consent to participate in the study. All procedures followed were in accordance with the ethical standards and with the Helsinki Declaration of 1975, as revised in 2008, and were approved by the responsible local Ethics Committee of the Universidade Federal do Rio Grande do Sul on human experimentation. Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 2 / 13 Competing interests: The authors have declared that no competing interests exist. Experimental design During the preliminary visit, athletes were familiarized with all loads, equipment, and proto- cols. All tests were separated by about 2–4 days. Firstly, the three maximal running tests at 0, 7% and 15% of individuals’ body mass were randomized in three visits and, in the fourth visit, the submaximal tests were again randomized. The athletes used their backpacks to perform the bouts with the extra load. The backpack position was set between the first thoracic and lumbar vertebra, and it was fixed to avoid excessive oscillation (Fig 1). In all tests, the heart rate (Polar, Kempele, Finland), end-tidal partial pressure of oxygen, end-tidal partial pressure of carbon dioxide, oxygen uptake, carbon dioxide output and ventilation per minute (MED- GRAPHICS, CPX/D, Diagnostic Systems, Saint Paul, Minnesota, USA) were measured contin- uously. The gas data were registered breath-by-breath. Temperature, atmospheric pressure and humidity in the laboratory were 20 ± 2˚C, 1026 ± 10 mmHg, and 50 ± 8%, respectively. 6–20 Borg’s ratings of perceived exertion scale (RPE) was shown to the athletes during the last 30 s of each stage (maximal tests) and just after the end of submaximal tests. Each athlete received detailed instructions about the use of the scale before the beginning of the first test. The total time at each maximal test was 30 minutes, and 1 hour to the submaximal protocol. Maximal test. Before the trials, the athletes performed a warm-up walking on a treadmill (QUINTON, ST55, New York, USA) inclined at 1% for five minutes at 6.0 km h-1 [12]. During the warm-up before the tests with load, the backpack with the respective load (7 or 15% of body mass) was adjusted. They were familiarized to treadmill exercise. The initial speed of the maximal tests was 6.0 km h-1, and the increment was 1.0 km h-1 per minute until subjects reached volitional exhaustion. Submaximal tests. Initially, resting oxygen uptake was measured in orthostasis, during 5 min. The individuals were asked to carry-out a warm-up for five minutes walking on a tread- mill. Taking into account the individual outcomes from maximal tests, the athletes carried out the submaximal tests at the intensity associated with 10% below the second ventilatory thresh- old according to the respective condition (0%, 7%, and 15% of the individual’s body mass), all on the same day. The duration of each test was 6 min. On average, 10 minutes of rest between submaximal tests were enough to achieve the initial heart rate and oxygen consumption. The rates of oxygen consumption and carbon dioxide production were measured continuously during each trial. The heart rate in all submaximal tests was not greater than 80 percent of the maximal heart rate. Besides, the respiratory exchange ratio was also monitored achieving val- ues lower than one. Data analysis Maximal tests. The VO2max, the velocity associated with VO2max (vVO2max), maximal RER, maximal heart rate, first and second ventilatory thresholds, and velocity associated with first and second ventilatory thresholds (v1Tvent and v2Tvent, respectively) were determined using computerized indirect calorimetry system [13]. The highest average of five oxygen consumption values was interpreted as the VO2max [14]. The value was considered valid when, at least, one of following criteria was observed: i) estimated maximal heart rate; ii) plateau on oxygen consumption with concomitant increase in the speed (all subjects attained the true VO2max); iii) respiratory exchange ratio greater than 1.1; iv) rating of perceived exertion greater than 17 (very hard) relative to Borg scale. The first and second ventilatory thresholds were determined according to the method pro- posed by [15]. The first ventilatory threshold (also denominated as individual ventilatory threshold) was determined from the first increase in ventilation-minute with a rapid rise in the ventilatory equivalent of oxygen consumption with no concomitant increase in the ventilatory Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 3 / 13 Fig 1. The athlete and his backpack with the extra load. https://doi.org/10.1371/journal.pone.0189516.g001 Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 4 / 13 equivalent of carbon dioxide production curve. The second ventilatory threshold (also denom- inated as respiratory compensation point) was defined as follows: i) a systematic increase in the ventilatory equivalent of oxygen consumption; ii) a concomitant nonlinear increase in the ventilatory equivalent of carbon dioxide production; and iii) a reduction in the difference in the inspired and end-tidal oxygen pressure. The ventilatory thresholds were determined in a blinded way by two independent evaluators. Submaximal tests. The submaximal oxygen uptake and heart rate were averaged from the last 60 s of the test [16]. The running economy was denoted by C, expressed in J kg-1 m-1. For that, we divided the net metabolic rate (gross—stand metabolic rate) by speed and we con- verted oxygen in ml to Joules relative to combustion enthalpy of substrates resulting from oxi- dation observed indirectly from respiratory exchange ratio [17]. The metabolic rate (ECO) was also calculated and expressed in ml kg-1 min-1. The maximal and submaximal metabolic power values were normalized to body mass and expressed in ml kg-1 min-1. Recently, we showed that the relationship between metabolic parameters and performance is independent of how the parameters are relativized in runners [18–20]. All data can be seen in the supplementary material (S1 Table). Statistics The Shapiro-Wilk test was used to verify data normality. We performed the descriptive statis- tics calculating mean ± standard deviation. The Pearson product-moment correlation test was carried out in order to test the relationship among the physiological determinants of perfor- mance (VO2max, C and, ventilatory threshold) with different load conditions. The linear regression analysis was used to estimate the speeds associated empirically with ventilatory thresholds when carrying loads. Possible differences between conditions (0, 7 and 15% of body mass) were analyzed using the repeated-measures analysis of variance (ANOVA) with Bonfer- roni post hoc test. To verify the possibility of violation of the assumption of sphericity, we applied the Mauchly test using the Greenhouse-Geisser correction for all analyses. Significance was accepted at P  0.05, statistical power was 90%, and the analyses were performed in Statis- tical Package for Social Sciences version 20.0 (SPSS, Chicago, Illinois, USA). We used the Cohen’s d coefficient to determine the effect sizes [21]. We determined the differences in pro- portions using the rule of thumb criteria set out by Hopkins: trivial (< 0.2), small (0.2–0.6), moderate (0.6–1.2), or large (> 1.2). Results The physiological data are presented in Table 1. The average for 7 and 15% loads carried in the maximal and submaximal tests were 5.29, s = 0.64 kg and 11.33, s = 1.37 kg, respectively. ANOVA showed a general effect of load on vVO2max (P = 0.005, Fig 2B), which decreased 13% between 0 and 15% load. Despite no significant statistical differences among VO2max val- ues (Fig 2A), a large effect size (0.86) was observed between 0 and 15% load. First ventilatory threshold values significantly increased (P = 0.003) while the velocity asso- ciated, v1Tvent, did not differ among loads (P = 0.287, Fig 2C). On the contrary, second venti- latory threshold values were similar (P = 0.140) while the v2Tvent decreased 13.5% (P = 0.005, Fig 2D), as observed for the vVO2max. Heart rate, rating of perceived exertion at first and sec- ond ventilatory threshold, and VO2max did not differ (P > 0.05) among load conditions (Table 1). In contrast, C resulted to be significantly greater (+30%; P = 0.001; d = 1.48) at 15% load and greater (+13%; P = 0.115; d = 0.85) at 7% load compared to the unloaded running submaximal test (Fig 3). Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 5 / 13 A significant moderate correlation was observed between the second ventilatory threshold and VO2max (r = 0.74, P <0.05, Fig 4C). Interestingly, the extra-cost (the metabolic cost to transport the backpack) seems to be a function of backpack weight (r = 0.78, P <0.05, Fig 4D). We estimated the relative intensity of training (reduction on v1Tvent, v2Tvent, and vVO2- max due to load carriage) by the linear regression model. The linear regression between reduc- tion of the v1Tvent and load was weak and non significant (r2 = 0.05; P >0.05). On the other hand, the estimation of the reduction of the vVO2max (in km h-1) and v2Tvent in function of load carriage were as follows: vVO2max (load, kg) = -0.1932 x load– 0.1523 (r2 = 0.54; Table 1. Mean, standard deviation, ANOVA, post hoc (Bonferroni) and Cohen’s d effect size results from maximal (VO2max protocol) and submax- imal (running cost protocol) tests. Numbers in bold represent P < 0.05. 0% 7% 15% ANOVA Bonferroni (effect size) F P P 0–7% P 0–15% P 7–15% Load (kg) -- 5.3 ± 0.6 11.3 ± 1.4 VO2max protocol data VO2max (ml kg-1 min-1) 59.7 ± 5.9 61.7 ± 6.6 64.6 ± 5.4 2.05 0.144 0.999 (0.31) 0.156 (0.86) 0.714 (0.48) HRmax (bpm) 183 ± 9 181 ± 8 181 ± 12 0.15 0.863 0.999 (0.23) 0.999 (0.18) 0.999 (0.00) vVO2max (km h-1) 18.0 ± 1.7 16.7 ± 1.6 15.7 ± 1.6 6.36 0.005 0.151 (0.78) 0.003 (1.39) 0.412 (0.62) RERmax 1.14 ± 0.07 1.13 ± 0.09 1.15 ± 0.10 0.24 0.790 0.999 (0.12) 0.999 (0.11) 0.999 (0.21) 1Tvent (ml kg-1 min-1) 33.2 ± 3.8 37.5 ± 4.1 39.3 ± 4.6 6.75 0.003 0.050 (1.08) 0.003 (1.44) 0.904 (0.41) 2Tvent (ml kg-1 min-1) 51.8 ± 4.3 55.5 ± 6.3 56.5 ± 6.9 2.07 0.140 0.425 (0.68) 0.184 (0.81) 0.999 (0.15) v1Tvent (km h-1) 9.0 ± 0.9 8.5 ± 0.8 8.6 ± 0.7 1.30 0.287 0.427 (0.58) 0.656 (0.49) 0.999 (0.13) v2Tvent (km h-1) 14.8 ± 1.4 13.7 ± 1.4 12.8 ± 1.2 6.24 0.005 0.164 (0.78) 0.004 (1.53) 0.405 (0.69) 1Tvent% (%) 55.8 ± 5.9 61.1 ± 6.5 60.8 ± 5.9 0.68 0.514 0.861 (0.85) 0.999 (0.84) 0.999 (0.04) 2Tvent% (%) 87.1 ± 5.3 90.0 ± 5.3 87.5 ± 8.3 2.85 0.072 0.127 (0.54) 0.156 (0.05) 0.999 (0.35) v1Tvent% (%) 50.2 ± 5.4 51.3 ± 6.1 55.0 ± 4.3 2.72 0.081 0.999 (0.19) 0.098 (0.98) 0.290 (0.70) v2Tvent% (%) 82.2 ± 6.4 82.4 ± 9.1 82.3 ± 7.8 0.01 0.998 0.999 (0.02) 0.999 (0.01) 0.999 (0.01) HR at 1Tvent (bpm) 127 ± 12 130 ± 13 132 ± 15 0.54 0.585 0.999 (0.23) 0.930 (0.36) 0.999 (0.14) HR at 2Tvent (bpm) 167 ± 13 167 ± 10 166 ± 12 0.01 0.989 0.999 (0.00) 0.999 (0.07) 0.999 (0.09) RPE at 1Tvent 9.2 ± 1.2 9.1 ± 0.9 9.9 ± 1.6 1.62 0.213 0.999 (0.09) 0.453 (0.49) 0.335 (0.61) RPE at 2Tvent 14.7 ± 2.1 13.9 ± 2.5 14.4 ± 2.6 0.36 0.701 0.999 (0.34) 0.999 (0.12) 0.999 (0.19) RPE at VO2max 18.7 ±1.5 18.2 ± 1.3 18.4 ± 1.4 0.51 0.607 0.968 (0.35) 0.999 (0.20) 0.999 (0.14) Running cost protocol data ECO (ml kg-1 min-1) 42.1 ± 6.0 45.3 ± 6.7 48.1 ± 9.7 1.87 0.170 0.941 (0.50) 0.186 (0.74) 0.999 (0.33) C (J kg-1 m-1) 4.0 ± 0.7 4.6 ± 0.7 5.2 ± 0.9 7.97 0.001 0.115 (0.85) 0.001 (1.48) 0.229 (0.74) Speed (km h-1) 13.3 ± 1.2 12.3 ± 1.3 11.5 ± 1.1 1.98 0.154 0.185 (0.79) 0.530 (1.56) 0.999 (0.66) ECO% (%) 76 ± 8 80 ± 12 78 ± 11 0.32 0.725 0.999 (0.39) 0.999 (0.20) 0.99 (0.17) HR_ECO (bpm) 162 ± 15 161 ± 16 158 ± 15 0.17 0.846 0.999 (0.06) 0.999 (0.26) 0.999 (0.19) RPE_ECO 11.0 ± 2.0 11.6 ± 2.0 11.8 ± 1.9 0.66 0.525 0.999 (0.29) 0.842 (0.41) 0.999 (0.10) Note: VO2max: maximal oxygen consumption; HRmax: maximal heart rate; RERmax: maximal respiratory exchange ratio; vVO2max: velocity at VO2max; 1Tvent: first ventilatory threshold; 2Tvent: second ventilatory threshold; v1Tvent: velocity at first ventilatory threshold; v2Tvent: velocity at second ventilatory threshold; 1Tvent%: percent first ventilatory threshold; 2Tvent%: percent second ventilatory threshold; v1Tvent%: percent velocity associated with first ventilatory threshold; v2Tvent%: percent velocity associated with second ventilatory threshold; HR at 1Tvent: heart rate at first ventilatory threshold; HR at 2Tvent: heart rate at second ventilatory threshold; RPE at 1Tvent: rating of perceived exertion at first ventilatory threshold; RPE at 2Tvent: rating of perceived exertion at second ventilatory threshold; and RPE at VO2max: rating of perceived exertion at maximal oxygen consumption. ECO: metabolic rate; C: cost of transport; ECO%: percent metabolic rate; HR_ECO: heart rate during submaximal test; RPE_ECO: rating of perceived exertion during submaximal test. https://doi.org/10.1371/journal.pone.0189516.t001 Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 6 / 13 P < 0.05; SEE = 0.2186 km h-1); v2Tvent (load, kg) = -0.1676 x load– 0.1549 (r2 = 0.68; P <0.05, SEE = 0.1422 km h-1). Discussion Our first and most important research question was to determine whether there is an effect of load carriage on running VO2max, ventilatory thresholds, and C in experienced athletes of AR. We found that VO2max and second ventilatory threshold were affected similarly, that is, values reached were similar while the velocities associated were significantly reduced. We also accepted our hypothesis that adventure racers have higher differences in the C when carrying loads of 15% than 7% of body mass in comparison to the unloaded condition. Our second research question was related to testing predictive equations to estimate specific training inten- sity for adventure racers from the loads used. To the best of our knowledge, this is the first study investigating the effects of load carriage on running VO2max, C and ventilatory thresh- old in adventure racers. About the method What remains to be established is how much of the reduction in velocity is related to load car- riage and whether a similar reduction also carries out with different intensities. These are important issues given that loaded running may play a role in a substantial proportion of meta- bolic requirements during the AR. To address these questions, we chose to focus on backpack Fig 2. Mean and standard deviation of VO2max (A), velocities associated with VO2max (vVO2max, B), first (v1Tvent, C) and second (v2Tvent, D) ventilatory thresholds at different conditions (0%: unloaded; 7% of body mass; and 15% of body mass). P’s and effect sizes (Cohen’s d) are also presented. https://doi.org/10.1371/journal.pone.0189516.g002 Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 7 / 13 weights of 7 and 15 percent of body mass as these loads are frequent in adventure races [18]. As the individual speed is critical to organizing the strategy of AR teams [22], we chose to esti- mate the velocities associated with different performance threshold. Physiological meaning of findings The adventure racers’ v1Tvent did not change carrying loads until 15% of body mass, but the metabolic rate expended at that level was increased with loading condition. One possible explanation refers to the reduced running economy (C) at similar intensity found out in our study. Also, on the other hand, maintain the v1Tvent even with a higher oxygen consumption may be related to the intensity frequently used for these athletes in their races (120–130 bpm Fig 3. Cost of transport per athlete (black lines) and the average value (gray double line) at unloaded condition (0), carrying loads of 7 and 15% of body mass. https://doi.org/10.1371/journal.pone.0189516.g003 Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 8 / 13 [8]), close to the heart rate at first ventilatory threshold found here (127–132 bpm at first venti- latory threshold). The maximal heart rate, respiratory exchange ratio and rating of perceived exertion were not modified in the different conditions due to the substantial reduction in the vVO2max, which is in line with previous studies [23, 24]. At higher intensities, the vVO2max and v2Tvent were reduced as expected. We speculated that performance at the severe intensity domain was not worsened due to an optimization in the elastic bouncing of running. Since the total load is a crucial factor to explain the higher effi- ciency in larger animals due to minor hysteresis loss [25], we suggest that when using back- packs, the adventure racers have the muscle-tendon unities more loaded showing also the landing-takeoff asymmetry more elastic as recently proposed [26]. Direct evidence of animal studies [27] and indirect evidence in humans [3] support this hypothesis. Nevertheless, although we have measured the cost of carrying loads during locomotion, the function of the muscle-tendon unit cannot be ascertained in this study. Thus, the added mechanical work due to elastic bouncing with loads remains unknown, and the potential impact of loading at the level of muscle-tendon units requires further research. In our study, C increased with the extra load. However, the effects of load on the C in the scientific literature are inconclusive [3, 4, 18]. There are methodological differences in the studies mentioned above that might partly explain different findings. One important issue refers to the way of expressing the energy expenditure, especially concerning the mass Fig 4. Scatterplot between physiological variables (n = 12 subjects). A) The cost of transport in function of maximal oxygen consumption (VO2max) and, B) in function of second ventilatory threshold (2Tvent). C) The 2Tvent in function of VO2max. D) The metabolic cost of load (load cost = loaded −unloaded) in function of absolute load in kg. https://doi.org/10.1371/journal.pone.0189516.g004 Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 9 / 13 normalization. In the studies where the economy increased with load (e.g., Abe et al. [3]), the metabolic value was normalized to total mass. This procedure does not permit assessing the economy mainly related to muscles involved in the movement [2]. The ventilatory thresholds are considered good predictors of long-distance running perfor- mance [15]. In the current study, the athletes obtained moderate to high velocities [28] associ- ated with the second ventilatory threshold (0% = 14.8, s = 1.4 km h-1; 7% = 13.7, s = 1.4 km h-1; 15% = 12.8, s = 1.2 km h-1). The second ventilatory threshold expressed as a percentage of VO2max were in the range of 87–90% VO2max. These results demonstrate that the athletes are aerobically well trained. Again, we explore the effects of velocities associated with the venti- latory threshold, and these determinations may be useful to plan training programs in order not just to maximize physiological adaptations but also to reduce the probability of susceptibil- ity to overreaching and overtraining in this sport [29, 30]. From a practical point-of-view, this study provides an interesting outcome related to the one-to-one percent ratio between running velocity and extra-load (as a percentage of body mass) in the severe domain. In other words, the athlete of AR needs to pay attention that to each percent of the increase in the backpack’s load, the speed needs to be reduced at the same percentage, to maintain the same metabolic rate. These findings suggest that interventions that enhance the running economy (for exam- ple, strength training) may increase the athletes’ performance of AR. Predictive equations Specifically, our predictive equations offer a valuable tool to control the training intensity when the load is manipulated. The average reductions of the velocities associated with VO2max due to increasing load were 8 and 15% (for 7 and 15% load, respectively), and of the velocities associated with second ventilatory threshold were 10 and 15%. Interestingly, these average values indicate a constant ratio equal to one between the percent velocity reduction (due to loading) and the percent load increase. Stated in other terms, 1% of load increase (relative to body mass) is equivalent to 1% of velocity reduction (relative to running velocity without extra load). This relationship seems to be constant only on severe intensity domain, between the second ventilatory threshold and VO2max. Although it produces only empirical predictive equations, this approach is biologically meaningful and provides a useful framework for planning and developing specific training models to adventure racers. Furthermore, many attempts of estimating the ventilatory thresh- old parameters were not successful in the literature, showing an underestimation in a broad sample of endurance athletes (141 subjects, but taking part 8 adventure racers only). And, when validated, these equations are neither specific to AR [31] nor taking into account the loaded conditions [28, 32]. The experimental protocol we undertook has limitations that must be discussed. The load position is a crucial variable of C, and we used only loads on the shoulders. Although loading subjects on their shoulders had a greater negative impact on C than placing the extra load around their waists [18], backpacks are the typical way of carrying the load in AR. One important limitation is about the time-dependent effects of load on physiological parameters studied here. The adventure races are performed during 4–5 days. Therefore, the mitochon- drial function is deteriorated during the race [5], probably intensifying the negative effects of load on C. Moreover, the heart rate is qualitatively reduced in the second half of races [8]. These differ- ences show that our results are limited to regular training and race’ starting phases. Future work can use this original study as an important starting point in the quest to improve our understanding of the physiological adaptations to specific training by using loads and their Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 10 / 13 repercussions on race performance. We also suggest future studies analyzing running’s bio- mechanical alterations (stride length and frequency) under different loads in AR athletes. The results of the present study showed collectively the preservation of running primary physiological parameters of adventure racers using loads, specifically, demonstrating the main- tenance of VO2max, C (at 7% of body mass) and second ventilatory threshold, and the increase of the first ventilatory threshold (interestingly without differences in the v1Tvent). The meta- bolic cost of transporting 1 kilogram of body mass per meter of running (C) was impaired with 15% load only. Conclusion The most striking findings of this cross-sectional study are as follows: (i) The VO2max and second ventilatory threshold remain unchanged, and the responses of first ventilatory thresh- old and C were greater at 15% of body mass in comparison to unloaded condition; (ii) at severe metabolic domain (from second ventilatory threshold to VO2max), the iso-metabolic speeds were reduced 1% equivalently to 1% increasing load (relative to body mass); (iii) the C of carry- ing 7% and 15% (of body mass) loads for AR athletes are 4.6 J kg-1 m-1 (1.10 cal kg-1 m-1) and 5.2 J kg-1 m-1 (1.24 cal kg-1 m-1), respectively; and for the unloaded condition, the C is 4.0 J kg-1 m-1 (0.96 cal kg-1 m-1). Moreover, the regression model presented here is convenient for field use by adventure rac- ers, as it requires only the information of the load carried on backpacks. These findings could be further used to optimize the performance of these athletes by individualizing training inten- sities related to load carriage. Supporting information S1 Table. General dataset. (XLSX) Acknowledgments We are grateful to the Locomotion Group of the Federal University of Rio Grande do Sul for discussions and comments. L.A. Peyre´-Tartaruga is an established investigator of the Brazilian Research Council—CNPq, Brası´lia, Brazil. Author Contributions Conceptualization: Alex de O. Fagundes, Patrı´cia D. Pantoja, Leonardo A. Peyre´-Tartaruga. Data curation: Alex de O. Fagundes, Elren P. Monteiro, Leandro T. Franzoni, Bruna S. Fraga, Patrı´cia D. Pantoja, Gabriela Fischer, Leonardo A. Peyre´-Tartaruga. Formal analysis: Alex de O. Fagundes, Elren P. Monteiro, Leandro T. Franzoni, Bruna S. Fraga, Patrı´cia D. Pantoja, Gabriela Fischer, Leonardo A. Peyre´-Tartaruga. Funding acquisition: Leonardo A. Peyre´-Tartaruga. Investigation: Alex de O. Fagundes, Elren P. Monteiro, Leandro T. Franzoni, Bruna S. Fraga, Patrı´cia D. Pantoja, Gabriela Fischer, Leonardo A. Peyre´-Tartaruga. Methodology: Alex de O. Fagundes, Elren P. Monteiro, Leandro T. Franzoni, Bruna S. Fraga, Patrı´cia D. Pantoja, Gabriela Fischer, Leonardo A. Peyre´-Tartaruga. Project administration: Gabriela Fischer, Leonardo A. Peyre´-Tartaruga. Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 11 / 13 Resources: Leonardo A. Peyre´-Tartaruga. Software: Leonardo A. Peyre´-Tartaruga. Supervision: Leonardo A. Peyre´-Tartaruga. Writing – original draft: Alex de O. Fagundes, Gabriela Fischer, Leonardo A. Peyre´- Tartaruga. Writing – review & editing: Alex de O. Fagundes, Elren P. Monteiro, Leandro T. Franzoni, Bruna S. Fraga, Patrı´cia D. Pantoja, Gabriela Fischer, Leonardo A. Peyre´-Tartaruga. References 1. Simpson D, Post PG, Tashman LS. Adventure racing: The experiences of participants in the everglades challenge. J Humanist Psychol. 2015; 54, 113–128. https://doi.org/10.1177/0022167813482188 2. Saibene F, Minetti AE. Biomechanical and physiological aspects of legged locomotion in humans. Eur J Appl Physiol. 2003; 88, 297–316. https://doi.org/10.1007/s00421-002-0654-9 PMID: 12527959 3. Abe D, Fukuoka Y, Muraki S, Yasukouchi A, Sakaguchi Y, Niihata S. Effects of load and gradient on energy cost of running. J Physiol Anthropol. 2011; 30, 153–60. https://doi.org/10.2114/jpa2.30.153 PMID: 21804298 4. Bourdin M, Belli A, Arsac LM, Bosco C, Lacour JR. Effect of vertical loading on energy cost and kine- matics of running in trained male subjects. J Appl Physiol. 1995; 79, 2078–2085. PMID: 8847276 5. Levada-Pires AC, Fonseca CER, Hatanaka E, Alba-Loureiro T, Velhote FB, Curi R, et al. The effect of an adventure race on lymphocyte and neutrophil death. Eur J Appl Physiol. 2010; 109, 447–453. https:// doi.org/10.1007/s00421-010-1363-4 PMID: 20143084 6. Bowen RL, Adams JH, Myburgh KH. Nausea and high serum osmolality during a simulated ultraendur- ance adventure race: a case-control study. Int J Sports Physiol Perform. 2006; 1, 176–185. PMID: 19114751 7. Borgenvik M, Nordin M, Mattsson CM, Enqvist JK, Blomstrand E, Ekblom B. Alterations in amino acid concentrations in the plasma and muscle in human subjects during 24 h of simulated adventure racing. Eur J Appl Physiol. 2012; 112, 3679–3688. https://doi.org/10.1007/s00421-012-2350-8 PMID: 22350359 8. Enqvist JK, Mattsson CM, Johansson PH, Brink-Elfegoun T, Bakkman L, Ekblom BT. Energy turnover during 24 hours and 6 days of adventure racing. J Sports Sci. 2010; 28, 947–955. https://doi.org/10. 1080/02640411003734069 PMID: 20544486 9. Zalcman I, Guarita HV, Juzwiak CR, Crispim CA, Antunes HKM, Edwards B, et al. Nutritional status of adventure racers. Nutr. 2007; 23, 404–411. https://doi.org/10.1016/j.nut.2007.01.001 PMID: 17383160 10. Knapik JJ, Harman EA, Steelman RA, Graham BS. A systematic review of the effects of physical train- ing on load carriage performance. J Strength Cond Res. 2012; 26, 585–597. https://doi.org/10.1519/ JSC.0b013e3182429853 PMID: 22130400 11. Solomonson AA, Dicks ND, Kerr WJ, Pettitt RW. Influence of Load Carriage on High-Intensity Running Performance Estimation. J Strength Cond Res. 2016; 30, 1391–1396. https://doi.org/10.1519/JSC. 0000000000001209 PMID: 26422613 12. Jones AM, Doust JH. A 1% treadmill grade most accurately reflects the energetic cost of outdoor run- ning. J Sports Sci. 1996; 14, 321–327. https://doi.org/10.1080/02640419608727717 PMID: 8887211 13. Sun XG, Hansen JE, Garatachea N, Storer TW, Wasserman K. Ventilatory efficiency during exercise in healthy subjects. Am J Resp Crit Care Med. 2002; 166, 1443–1448. https://doi.org/10.1164/rccm. 2202033 PMID: 12450934 14. Ribeiro JP, Hughes V, Fielding RA, Holden W, Evans W, Knuttgen HG. Metabolic and ventilatory responses to steady state exercise relative to lactate thresholds. Eur J Appl Physiol Occup Physiol. 1986; 55, 215–221. PMID: 3699010 15. Wasserman K, Whipp BJ, Koyl SN, Beaver WL. Anaerobic threshold and respiratory gas exchange dur- ing exercise. J Appl Physiol. 1973; 35, 236–243. PMID: 4723033 16. Shaw AJ, Ingham SA, Folland JP. The valid measurement of running economy in runners. Med Sci Sports Exer. 2014; 46, 1968–1973. https://doi.org/10.1249/MSS.0000000000000311 PMID: 24561819 17. Pe´ronnet F, Massicotte D. Table of nonprotein respiratory quotient: an update. Can J Appl Physiol. 1991; 16, 23–29. Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 12 / 13 18. Pedersen AV, Stokke R, Mamen A. Effects of extra load position on energy expenditure in treadmill run- ning. Eur J Appl Physiol. 2007; 102, 27–31. https://doi.org/10.1007/s00421-007-0553-1 PMID: 17874122 19. Saunders PU, Pyne DB, Telford RD, Hawley JA. Factors affecting running economy in trained distance runners. Sports Med. 2004; 34, 465–485. PMID: 15233599 20. Tartaruga MP, Mota CB, Peyre´-Tartaruga LA, Brisswalter J. Scale model on performance prediction in recreational and elite endurance runners. Int J Sports Physiol Perform. 2014; 9, 650–655. https://doi. org/10.1123/ijspp.2013-0165 PMID: 24231578 21. Hopkins W, Marshall S, Batterham A, Hanin J. Progressive statistics for studies in sports medicine and exercise science. Med Sci Sports Exerc. 2009; 41, 3. https://doi.org/10.1249/MSS.0b013e31818cb278 PMID: 19092709 22. Lipman GS, Ellis MA, Lewis EJ, Waite BL, Lissoway J, Chan GK, et al. A prospective randomized blister prevention trial assessing paper tape in endurance distances (Pre-TAPED). Wilderness Environ Med. 2014; 25, 457–461. https://doi.org/10.1016/j.wem.2014.06.013 PMID: 25443754 23. Fletcher JR, Esau SP, MacIntosh BR. Economy of running: beyond the measurement of oxygen uptake. J Appl Physiol. 2009; 107, 1918–1922. https://doi.org/10.1152/japplphysiol.00307.2009 PMID: 19833811 24. Lucas SJ, Anglem N, Roberts WS, Anson JG, Palmer CD, Walker RJ, et al. Intensity and physiological strain of competitive ultra-endurance exercise in humans. J Sports Sci. 2008; 26, 477–489. https://doi. org/10.1080/02640410701552872 PMID: 18274945 25. Cavagna GA, Legramandi MA. Running, hopping and trotting: tuning step frequency to the resonant fre- quency of the bouncing system favors larger animals. JExp Biol. 2015; 218, 3276–3283. https://doi.org/ 10.1242/jeb.127142 PMID: 26347555 26. Cavagna GA. The two asymmetries of the bouncing step. Eur J Appl Physiol. 2009; 107, 739. https:// doi.org/10.1007/s00421-009-1179-2 PMID: 19727798 27. Baudinette RV, Biewener AA. Young wallabies get a free ride. Nature. 1998; 395, 653. https://doi.org/ 10.1038/27111 28. Malek MH, Housh TJ, Coburn JW, Schmidt RJ, Beck TW. Cross-validation of ventilatory threshold pre- diction equations on aerobically trained men and women. J Strength Cond Res. 2007; 21, 29–33. https://doi.org/10.1519/R-19135.1 PMID: 17313275 29. Anglem N, Lucas SJ, Rose EA, Cotter JD. Mood, illness and injury responses and recovery with adven- ture racing. Wilderness Environ Med. 2008; 19, 30–38. https://doi.org/10.1580/07-WEME-OR-091.1 PMID: 18333663 30. Fordham S, Garbutt G, Lopes P. Epidemiology of injuries in adventure racing athletes. Br J Sports Med. 2004; 38, 300–303. https://doi.org/10.1136/bjsm.2002.003350 PMID: 15155432 31. Epstein Y, Stroschein LA, Pandolf KB. Predicting metabolic cost of running with and without backpack loads. Eur J Appl Physiol Occup Physiol. 1987; 56, 495–500. PMID: 3653088 32. Malek MH, Coburn JW. A new ventilatory threshold equation for aerobically trained men and women. Clin Physiol Funct Imaging. 2009; 29, 143–150. https://doi.org/10.1111/j.1475-097X.2008.00850.x PMID: 19207417 Loaded running in adventure racing PLOS ONE | https://doi.org/10.1371/journal.pone.0189516 December 7, 2017 13 / 13

Effects of load carriage on physiological determinants in adventure racers.

12-07-2017

Fagundes, Alex de O,Monteiro, Elren P,Franzoni, Leandro T,Fraga, Bruna S,Pantoja, Patrícia D,Fischer, Gabriela,Peyré-Tartaruga, Leonardo A

eng

PMC7925537

1 Vol.:(0123456789) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports A novel device for detecting anaerobic threshold using sweat lactate during exercise Yuta Seki1,2,6, Daisuke Nakashima3,6*, Yasuyuki Shiraishi1,2, Toshinobu Ryuzaki1,2, Hidehiko Ikura1,2, Kotaro Miura1,2, Masato Suzuki1, Takatomo Watanabe4, Takeo Nagura3,5, Morio Matsumato3, Masaya Nakamura3, Kazuki Sato2, Keiichi Fukuda1 & Yoshinori Katsumata1,2* The lactate threshold (LT1), which is defined as the first rise in lactate concentration during incremental exercise, has not been non-invasively and conveniently determined in a clinical setting. We aimed to visualize changes in lactate concentration in sweat during exercise using our wearable lactate sensor and investigate the relationship between the lactate threshold (LT1) and ventilatory threshold (VT1). Twenty-three healthy subjects and 42 patients with cardiovascular diseases (CVDs) were enrolled. During exercise, the dynamic changes in lactate values in sweat were visualized in real- time with a sharp continuous increase up to volitional exhaustion and a gradual decrease during the recovery period. The LT1 in sweat was well correlated with the LT1 in blood and the VT1 (r = 0.92 and 0.71, respectively). In addition, the Bland–Altman plot described no bias between the mean values (mean differences: − 4.5 and 2.5 W, respectively). Continuous monitoring of lactate concentrations during exercise can provide additional information for detecting the VT1. Abbreviations sLT Lactate threshold in sweat bLT The lactate threshold in blood VT1 Ventilatory threshold CVD Cardiovascular disease WR Work rate NYHA New York Heart Association Functional Classification VE Ventilation VO2 Oxygen uptake VCO2 Carbon dioxide production Adequate regular physical activity is paramount to maintaining good health1,2 and preventing cardiovascular diseases (CVD). Current clinical practice guidelines and expert statements recommend aerobic exercise for patients with CVD3–5. Although an exercise test with respiratory gas analysis is the only non-invasive way to determine the ventilatory threshold (VT1)6 in clinical practice, VT1 assessment requires an expensive analyzer and expertise5,7. Additionally, it is incidentally difficult to confirm the VT1 because of oscillations in minute ventilation and inconsistencies among several factors such as the VE/VO2, the terminal exhaled O2 concentra- tion, and the VCO2/VO2 slope8. Further, careful attention is necessitated when using a respiratory gas analyzer due to possible cross-infection. An alternative method is needed to detect VT1 easily and precisely without the need for a respiratory gas analyzer. Flexible wearable sensing devices can yield important information about the underlying physiology of a human subject in a continuous, real-time, and non-invasive manner9,10. Sampling human sweat, which is rich in physiological information such as the sweat rate or sodium concentration, could enable non-invasive OPEN 1Department of Cardiology, Keio University School of Medicine, 35 Shinanomachi Shinjuku-ku, Tokyo 160-8582, Japan. 2Institute for Integrated Sports Medicine, Keio University School of Medicine, Tokyo, Japan. 3Department of Orthopaedic Surgery, Keio University School of Medicine, 35 Shinanomachi Shinjuku-ku, Tokyo 160-8582, Japan. 4Department of Clinical Laboratory, Gifu University Hospital, Gifu, Japan. 5Department of Clinical Biomechanics, Keio University School of Medicine, Tokyo, Japan. 6These authors contributed equally: Yuta Seki and Daisuke Nakashima. *email: [email protected]; [email protected] 2 Vol:.(1234567890) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ monitoring11. To date, sweat-based and non-invasive biosensors of lactate have been reported in research settings12–14 and have shown that sweat lactate is elevated in conjunction with exercise intensity. Therefore, the application of continuous lactate monitoring systems using wearable lactate sensors could influence exercise therapy in patients with CVD in clinical practice. However, these devices have not yet been applied in clinical practice, which might be due to unsuccessful miniaturization of devices and an operation that is easy to use, the inappropriate degree of accuracy as a medical device, or high cost. We have developed an innovative device in which sweat lactate may be monitored in a continuous, convenient, and non-invasive manner. We hypothesized that this device could be applied in clinical practice and successfully detected the lactate threshold (LT1) which is the first rise of lactate concentration during incremental exercise. We aimed to investigate whether a usable device in the clinical setting would enable the continuous moni- toring of sweat lactate during incremental exercise. Moreover, we elucidated the relationship among the lactate threshold in sweat (sLT), blood (bLT), and VT1 in healthy subjects and patients with CVD. Results In-vitro characterization of the lactate biosensor. Figure 1 shows the amperometric response of the lactate biosensor to increasing lactate concentrations in the physiological range of 0–10 mmol/L. The biosensor responded linearly to the lactate concentrations, especially in the ranges from 0 to 5 mmol/L, with a sensitivity of 2.4 A/mM (Fig. 1A). The range of 0–5 mmol/L was important in determining the LT1 because lactate concentra- tions around 2 mmol/L are related to LT1 / VT16. Further, the sensors responded quickly and with almost the same value to L-lactate acid three times repeatedly (Fig. 1B). Study subjects. The baseline characteristics of the healthy subjects are summarized in Table 1. The healthy subjects were predominantly male (91%), with a median age of 20 (IQR 20–21) years. Tables 1 and 2 dem- onstrates the patient background of patients with CVD. The patients were predominantly male (76%), with a median age of 63 years (interquartile [IQR], 54–71) and left ventricular ejection fraction (LVEF) of 50% (IQR, 31.7–58.5). Thirty-four (83%) patients were taking beta-blockers. Monitoring of the lactate in sweat during exercise. Figure 2 and Online Supplemental Video S1 show the lactate values in sweat during incremental exercise. Dynamic changes in sweat lactate values during the exercise tests were continuously measured and projected on the wearable device without delay in both the healthy subjects and a subset of patients with CVD. At the commencement of the cycling activity, negligible cur- rent response was measured by the lactate biosensor due to the lack of sweat. At the onset of sweating, lactate was released from the epidermis, and was selectively detected by the LOx-based biosensor. During the exercise, a drastic increase in sweat lactate values was observed as the cycling continued up to volitional exhaustion (Fig. 2). At the end of the exercise period, sweat lactate values continued to decrease relatively slowly, compared to the decrease in heart rate. Predictors associated with non-response in the lactate sensor. In patients with CVD, changes in sweat lactate values during exercise were similar to those in healthy subjects. However, 19 cases had steady low lactate values after starting the exercise until the recovery state (Online Fig. S1), suggesting a lack of sweat even with a maximum exercise load. Logistic regression analysis was performed to identify factors associated with non-response in the lactate sensor. The results of the univariate analyses are shown in Table 3. New York Heart Association functional classification (NYHA) 3 and low peak VO2 were associated with non-response in the lactate sensor among patients with CVD (odds ratio [OR], 0.06; 95% confidence interval [CI] 0.01–0.24; and OR, Figure 1. In-vitro characteristics of the sweat lactate sensor chip. (A) Amperometric response to increasing lactate concentration from 0 to 20 mM (0, 2.5, 5, 10, and 20 mM) in phosphate buffer (pH 7.0); the graph shows the corresponding calibration plots of the sensor. Applied voltage = 0.16 V versus Ag/AgCl. The data were obtained from three samples. (B) Reproducibility and long-term stability of the sweat lactate sensor; the graph shows amperometric response to l-lactic acid solution adjusted to 10 mM repeatedly three times for 90 s. Data recording was paused for 90 s for each response. The data were obtained from four samples. 3 Vol.:(0123456789) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ 1.18; 95% CI 1.02–1.42, respectively). Conversely, age, B-type natriuretic peptide, and LVEF were not associated with a non-response in the lactate sensor (OR 1.03 [95% CI 0.66–1.61], OR 0.95 [95% CI 0.84–1.05], and OR 1.17 [95% CI 0.95–1.48], respectively). Relationship among the sLT and bLT. The conversion from the steady low lactate values to the continu- ous increase easily detected in all healthy subjects and 23 patients with response in the lactate sensor (Fig. 2), was defined as sLT. Among the 23 CVD patients, the monitoring of blood lactate concentration during exercise was Table 1. Baseline characteristics of healthy subjects and patients. ACEI angiotensin-converting enzyme inhibitor, ARB angiotensin receptor blocker, BMI body mass index, BNP B-type natriuretic peptide, IQR interquartile range, LVEF left ventricular ejection fraction, NYHA New York Heart Association Functional Classification, VE/VCO2 ventilation-carbon dioxide production, VO2 oxygen uptake, VT1 ventilatory threshold. Demographic and anthropometric data Healthy subjects (n = 23) Patients (n = 42) Age, years (median, IQR) 20 (20, 21) 63 (54, 71) Male, n (%) 21 (91) 32 (76.2) Height, cm (median, IQR) 171 (165, 175) 165 (159, 172) Body weight, kg (median, IQR) 62 (58, 68) 62 (57, 71) BMI, kg/m2 (median, IQR) 22 (20, 23) 23 (21, 25) Hypertension, n (%) – 12 (28.6) Diabetes, n (%) – 9 (21.4) Dyslipidemia, n (%) – 22 (52.4) NYHA ≧3 – 19 (45.2) Device, n (%) – 4 (9.5) Laboratory data Hemoglobin, g/dL (median, IQR) – 13.7 (12.5, 14.6) Creatinine, mg/dL (median, IQR) – 0.9 (0.8, 1.1) BNP, pg/mL (median, IQR) – 146.8 (35.6, 328.0) Echocardiography data LVEF, % (median, IQR) – 49.5 (31.7, 58.5) Medications Beta-blocker, n (%) – 34 (82.9) ACEI or ARB, n (%) – 24 (58.5) Statin, n (%) – 19 (45.2) Antiplatelet drug, n (%) – 16 (38.1) Anti-arrhythmic drug, n (%) – 3 (7.1) Cardiopulmonary test data VO2 at VT1, ml/kg/min (median, IQR) – 10.5 (9.7, 12.3) VT1, sec (median, IQR) – 429.0 (391.5, 473.2) Peak VO2, mL/kg/min (median, IQR) – 15.9 (12.4, 18.9) %Peak VO2, % (median, IQR) – 67.5 (53.7, 79.7) VE/VCO2 slope (median, IQR) – 32.2 (28.6, 38.1) Table 2. Respiratory gas data during exercise in the patients. All values are presented as medians and IQRs. DBP diastolic blood pressure, HR heart rate, IQR interquartile range, RQ respiratory quotient, SBP systolic blood pressure, VE/VCO2 ventilation-carbon dioxide production, VO2 oxygen uptake, VT1 ventilatory threshold, WR work rate. Rest Warm-up VT1 Peak HR, bpm 70 (62, 82) 79 (70, 92) 96 (85, 109) 127 (113, 137) SBP, mmHg 107 (91, 123) 118 (99, 130) 127 (110, 140) 142 (116, 166) DBP, mmHg 69 (62, 80) 76 (67, 86) 75 (65, 82) 79 (72, 90) VO2, mL/kg/min 3.6 (3.4, 4.1) 6.5 (5.6, 7.1) 11.5 (9.7, 12.3) 15.9 (12.4, 18.9) RQ – – 0.89 (0.86, 0.97) 1.15 (1.08, 1.20) WR (W) – 0 46 (37, 58) 77 (62, 107) VE/VCO2 slope 32.3 (28.6, 38.1) 4 Vol:.(1234567890) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ only enabled by 13 patients. Combining these 13 patients and all healthy subjects, the relationships between the WR-sLT and WR-bLT were investigated (Fig. 3A), which described a strong relationship between each thresh- old (r = 0.92, P < 0.001). The Bland–Altman plot revealed that the mean difference between each threshold was − 4.5 W, and that there was no bias between the mean values, which displayed strong agreements between the WR-sLT and WR-bLT (Fig. 3B). Least-product regression analysis indicated no evidence of a fixed bias and a proportional bias (95% CI for y-intercept, − 9.16 to 19.1; 95% CI for the slope 0.854–1.020). Relationship among the sLT and VT1. Similarly, a good correlation was observed between the WR-sLT and WR-VT1 (r = 0.71, P < 0.001; Fig. 4A). The Bland–Altman plot described a strong agreement in the patients with CVD (Fig. 4B; the mean difference between each threshold, 2.5 W). Least-product regression analysis indi- cated a fixed bias (y-intercept, 22.7) and a proportional bias (slope, 0.57) between each threshold. Figure 2. Imaging of the lactate in the sweat during incremental exercise. Representative graphs (dots) of the lactate in sweat (LA in sweat; dark blue) and lactate in blood (LA in blood; red) during exercise with a RAMP (15 W/min) protocol ergometer are shown in the lower panel. The respiratory gas data was shown in the upper panel. HR heart rate, LA lactate, VE ventilatory equivalent, VE/VCO2 ventilation-carbon dioxide production, VE/VO2 ventilation-oxygen uptake, WR work rate. 5 Vol.:(0123456789) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ Discussion The most striking result to emerge from our findings is that the non-invasive lactate sensor enabled continuous and real-time measurement of sweat lactate values during an incremental exercise test. Furthermore, sLT strongly correlated with both bLT and VT1 in a subset of patients with CVD as well as in healthy subjects. The real-time lactate monitoring in sweat could be applicable for the detection of the LT1. Lactate has three roles including, acting as a major energy source, a gluconeogenic substrate, a cell signaling molecule, and is used as optimal fuel for working muscles15. Muscle lactate production is essential to increase exercise performance16, and lactate should be measured during exercise to track an individual’s performance and exertion level17,18. Lactate values can be conventionally acquired via clinical labs or point-of-care devices19,20; unfortunately, such approaches do not support continuous, real-time measurements, a fact that limits their utility to applications where stationary, infrequent tests are sufficient. Conversely, our devices captured the sweat lactate value during exercise in a real-time, continuous, and non-invasive manner in a subset of patients with CVD in addition to healthy subjects. Sweat lactate has been affected by the production of lactate in the body and the rate of sweating and metabolic dynamics in sweat glands17,21. In addition, lactate is secreted into sweat, mirroring the intensity of exercise, but its concentration decreases with increasing sweat volume17. Therefore, the sweat lactate Table 3. Predictors of factors associated with response in the lactate sensor. ACEI angiotensin-converting enzyme inhibitor, ARB angiotensin receptor blocker, BMI body mass index, BNP B-type natriuretic peptide, CI confidence interval, LVEF left ventricular ejection fraction, NYHA New York Heart Association Functional Classification, VE/VCO2 ventilation-carbon dioxide production, VO2 oxygen uptake, VT1 ventilatory threshold. Factor Odds ratio (95% CI) P-value Age, year (per 10-point increase) 1.03 (0.66–1.61) 0.883 Male 2.19 (0.52–10.08) 0.288 Height, cm (per 10-point increase) 1.68 (0.82–3.75) 0.173 Body weight, kg (per 10-point increase) 1.30 (0.78–2.34) 0.339 BMI, kg/m2 1.03 (0.88–1.21) 0.748 Hypertension 0.48 (0.12–1.84) 0.285 Diabetes 0.59 (0.13–2.62) 0.485 NYHA 3 (versus ≤ 2) 0.06 (0.01–0.24) < 0.001 Hemoglobin, g/dL 1.18 (0.80–1.78) 0.414 Creatinine, mg/dL 0.93 (0.01–1.02) 0.736 BNP, pg/mL (per 50-point increase) 0.95 (0.84–1.05) 0.316 LVEF, % (per 5-point increase) 1.17 (0.95–1.48) 0.153 Beta-blocker 4.04 (0.75–31.20) 0.124 ACEI or ARB 1.24 (0.35–4.41) 0.732 VO2 at VT1, mL/kg/min 1.25 (0.95–1.71) 0.136 Peak VO2, mL/kg/min 1.18 (1.02–1.42) 0.044 VE/VCO2 slope (per 5-point increase) 0.88 (0.57–1.31) 0.519 Figure 3. Validity testing of the WR at the sLT and bLT. (A) The graph shows the relationship between the work rate (WR) at the sLT and bLT. (B) The graph shows the Bland–Altman plots, which indicate the respective differences between WR at the sLT, and bLT (y-axis) for each individual against the mean of the WR at the sLT, and bLT (x-axis). Triangles indicate the data of patients and circles indicate the data of healthy subjects. r correlation coefficient, 95% CI for b’ 95% confidence interval for the slope, 95% CI for a’ 95% confidence interval for y-intercept, SD standard deviation. 6 Vol:.(1234567890) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ concentration has been reported not to have reflected the blood lactate concentration in specific circumstances, such as during vigorous exercise22. By combining the sweat lactate concentration with the sweat rate, the amount of lactate excreted from sweat may be calculated. Sweat lactate discharge may be more predictive of blood lactate levels than lactate concentrations. Further research is also warranted to examine this relationship. Our sweat lactate sensor enabled the collection of sweat immediately after discharge unlike the other devices where sweat data represented a mixed state including previously discharged sweat. Therefore, the sensor used in this study successfully captured a rise in sweat lactate without delay. Moreover, despite the discrepancy between the sweat and blood lactate concentrations during vigorous exercise, there was no fixed and proportional biases between the WT-sLT and WT-sLT, which indicated that a rise in blood lactate coincided with that in sweat lactate during an incremental exercise. An increase in lactate production from muscle cells, reflecting the LT1, may induce a simultaneous rise in sweat lactate through a change in autonomic nervous balance, hormones, acid–base equilibrium, and metabolic dynamics23–26. However, the mean difference between WR-sLT and WR-VT1 was small, but the SD was rather large, and the presence of fixed and proportional bias was also indicated, which sug- gested a poor relationship between each threshold. This may have been caused by the difficulties in confirming the VT1 in some cases due to the inconsistencies among an increase in the ventilatory equivalent, excess CO2, and modified V-slope methods. Flexibility is crucial for unobtrusive wearable devices that cause no hindrance or irritation to the wearer. Recent advances in fabrication techniques have enabled the design of wearable sensing devices in thin, confor- mal form that naturally comply with the smooth curvilinear geometry of human skin, thereby enabling close contact that is necessary for robust physiological measurements and monitoring of chemicals and electrolytes in sweat27–29. Our sensor was highly flexible and can be smoothly adjusted to curved surfaces using PET substrates. The upper arm and forehead have a high-sweat rate during physical excursion30–32 and can thus, serve as an appropriate area to measure lactate values in human sweat. Additionally, the epidermis and muscle tissues around the upper arm or forehead do not experience complex 3D strains and remain stable even during intense physical activities. Therefore, the sensor was attached to the upper arm in the healthy subjects considering easy operability for the use in outdoor sports or exercise. Conversely, in patients with CVD, who experience less sweating than healthy subjects, the sensor was attached to the forehead which has a higher-sweat rate32. However, it was not possible to continuously measure lactate in patients with NYHA3 or low peak oxygen uptake. Non-response in the sensor indicates a lack of sweat during exercise, which could be caused by intravascular dehydration by diuretics, abnormality of autonomic nervous balance, such as dominant sympathetic activity, or frailty due to heart failure. Further research is also warranted to develop wearable devices to monitor lactate values in patients without efficient sweat production. In clinical practice, exercise testing with respiratory gas analysis is the most useful way to determine VT1. However, it is often difficult to determine VT1 because of oscillations in minute ventilation and inconsisten- cies among several factors such as the VE/VO2, the terminal exhaled O2 concentration, and VCO2/VO2 slope8. Furthermore, the use of a respiratory gas analyzer has a cross-infection possibility because of the closed circuit. The determination of sLT using only sweat-based monitoring could overcome these problems, and the device developed and used here would be suitable for use in a remote patient monitoring or remote rehabilitation set- ting during isolation measures, such as that taken during the COVID pandemic. Further, real-time assessments of sweat lactate values through a wireless data transfer system can offer a rigorous aerobic exercise based on the day-to-day physical conditions of patients with CVD as well as healthy subjects (Online Fig. S2). This innova- tive system could improve persistency of cardiac rehabilitation in outpatients and relocate their therapy from hospitals to other institutions, such as commercial fitness clubs or even patients’ homes. Our findings should be interpreted with the following limitations. First, because of the observational study design, we could not deny the influence of selection bias and unmeasured confounders regarding the effect on Figure 4. Validity testing of the WR at the sLT and bLT or VT1. (A) The graph shows the relationship between the work rate (WR) at the lactate threshold in sweat (WR-sLT) and WR-ventricular threshold (WR-VT1). (B) The graph shows the Bland–Altman plots, which indicate the respective differences between WR-sLT, and WR-VT1 (y-axis) for each individual against the mean of WR-sLT, and WR-VT1 (x-axis). r correlation coefficient, 95% CI for b’ 95% confidence interval for the slope, 95% CI for a’ 95% confidence interval for y-intercept, SD standard deviation. 7 Vol.:(0123456789) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ response in the lactate sensors. Second, our study had a relatively small number of cases and included no control group in which the sensors without lactate oxidase were used for comparison. To validate that lactate and not other sweat constituents was measured, a control experiment in which an unmodified (LOx-free) amperometric biosensor should be undertaken under the same experimental conditions. Future randomized-controlled studies with different medical centers are required to overcome these limitations. Third, sweat rate was not measured during exercise because of a lack of a sweat rate sensor in our device. Therefore, it is unknown whether non- response in the lactate sensor is caused by a lack of sweat or rough contact necessary for robust physiological measurements. In addition, the exercise duration may be related to the amount of sweat. Exercise protocol improvements, such as a longer warm-up time, may overcome the lack of sweat in some cases. Further studies are needed to examine the relationship between nonresponse in the lactate sensor and a lack of sweat. Fourth, the sweat lactate sensor used in this study did not operate in an environment with a lack of sweat. It was thus not possible to measure changes in sweat lactate in the low-intensity range where there was no-sweating and in patients who did not sweat during exercise, such as with NYHA3 or low peak O2 uptake. Conclusions This was the first study to show real-time monitoring of sweat lactate values during incremental exercise in patients with CVD as well as in healthy subjects. Given the difficult situation of deciding VT1, the monitoring of lactate values in sweat could be helpful for improving the detection of VT1. Methods Lactate measurement device. L-lactic acid and hydroxymethylferrocene were obtained from Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan). Phosphate buffer solution (PBS) (0.1 mol/L, pH 7.0) was purchased from the Nacalai Tesque, Inc. (Kyoto, Japan). L-LOx (LCO-301) was purchased from the Toyobo Corp. (Osaka, Japan). Water for molecular biology (H20MB0501) was obtained from Merck KGaA (Darmstadt, Germany). Water-soluble photocurable photosensitive resin (BIOSURFINE-AWP) was obtained from Toyo Gosei Co., Ltd. (Tokyo, Japan). Methanol was obtained from FUJIFILM Wako Pure Chemical Corporation (Osaka, Japan). The original printing electrode chip (DEP-CHIP) was procured from Bio-Device Technology, Inc. (Ishikawa, Japan). Instrumentation. The original printed electrode chip (hereinafter referred to as "printed electrode"), which is the base of the lactate sensor chip, was designed using computer-aided design with a pattern shape consisting of three poles: an acting electrode, a counter electrode, and a reference electrode (Fig. 5A). Subsequently, PET substrates (Toray Industries, Inc., Tokyo, Japan) were fabricated using carbon ink, Ag/AgCl and insulating ink in a screen-printing process. These processes were outsourced to Bio-Device Technology, Inc (Ishikawa, Japan). Fabrication of lactate sensor chips. A total of 0.5 μL of hydroxymethylferrocene saturated methanol solution and 1.0 μL of L-LOX 0.5 wt% solution were applied to the working electrode of the printing electrode using a micropipette, and the working electrode was dried at room temperature (20–24 °C). The entire surface of the working electrode, counter electrode, and reference electrode was then coated with BIOSURFINE-AWP diluted to 3 wt% using pre-molecular biological water with an applicator to achieve a film thickness of 15 μm. Finally, the lactate sensor chip was fabricated by forming a protective film by exposure using a UV lamp (365 nm wavelength). The fabricated lactate sensor chips were kept refrigerated at 5 °C (Fig. 5A). When the biosensor contacts lactate, the immobilized Lox enzyme catalyzes the oxidation of lactate to generate pyruvate and H2O2. The Prussian blue transducer then selectively reduces the H2O2 to generate electrons to quantify the lactate concentration (Fig. 5B). Lactate sensor device. Lactate concentration was determined by the voltage of the working electrode (WE) on the sensor chip via a potentiostat unit, driven through the I2C interface. During measurement, elapsed time from start (in seconds), A/D converted voltage (equivalent to lactate concentration), and temperature (in Celsius) were stored as 10-byte binary strip data on flash memory using SPI. An in-house mobile application then received the data from a connected device at 1 s intervals for 10–20 min. The operating voltage was regu- lated to 3 V via an LDO regulator. The battery was charged via a USB Type-C cable and of the in-house mobile application was regularly notified its level. Power on/off and Bluetooth LE communication status had been indi- cated for the user with LEDs (Fig. 5C). In-vitro studies. The lactate concentration in human sweat depends on metabolism and level of exertion, and typically ranges from 0 to 20 mmol/L. A wide linear-detection range coupled with a fast response time is thus essential for continuous epidermal monitoring of lactate. Therefore, the electrochemical characteriza- tion of the LA sensor chip was performed using l-lactic acid solutions in 0 (pH 7.0), 2.5 (pH 7.0), 5 (pH 6.9), 10 (pH 6.8), and 20 (pH 6.6) mmol/L prepared in 0.1 mol/L phosphate buffer solution (PBS). Then, the three lactate sensor tips were evaluated using chronoamperometry at an overprinting voltage of 0.16 V (versus Ag/ AgCl). In addition, the four sensor tips were evaluated for a total of three times with L-lactate solution adjusted to 10 mmol/L, to evaluate the long-term stability of the sensor. The electrochemical characterization was per- formed at room temperature (20–24 ℃using an electrochemical analyzer from Grace Imaging. Inc. Lactate measurement in humans. Study sample and ethical approval. Twenty-three healthy subjects were recruited, and 42 consecutive patients with CVD (e.g., heart failure, cardiomyopathy, or coronary artery disease) who underwent incremental exercise testing between November 2019 and November 2020 at Keio 8 Vol:.(1234567890) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ Figure 5. Fabrication and function of the sweat lactate sensor chip. (A) Parts composition of the lactate acid sensor chip. (B) Schematic diagram of the reagent layer and processes involved in the amperometric sensing of lactate acid on the working electrode. (C) The parts composition of device. (1) Bluetooth LE System-on-Chip: Taiyo Yuden EYSHCNZWZ, (2) Potentiostat unit: Texas Instruments DAC081C085 (DAC), Microchip Technology MCP6041T-I/OT and MCP6042T-I/MS (OP AMP), (3) temperature sensor: Ablic S-5851A, (4) flash memory: Winbond Electronics W25Q32JV(4 MB), (5) power unit: Texas Instruments BQ24232RGTR (Charger IC), Synergy ScienTech AHB512229PR (Li-ion Battery: 3.7 V, 295 mAh) and ON Semiconductor LC709203F (fuel gauge). 9 Vol.:(0123456789) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ University Hospital were enrolled. The healthy subjects had a broad spectrum of aerobic capacities and fitness levels, but were not athletes, and had no comorbidities, such as hypertension, diabetes, or active lung diseases. Exclusion criteria for the patients with CVD included 2 or 3 degree-conduction block without a cardiac implant- able electronic device, severe pulmonary hypertension, decompensated heart failure, more than severe primary valvular heart diseases, and an acute phase of the acute coronary syndrome. The study protocol was approved by the Institutional Review Board of Keio University School of Medicine [permission number; 2014023, 20180357], and was conducted in accordance with the Declaration of Helsinki. All subjects provided written informed consent. Experimental procedure. The exercise tests were performed with the RAMP protocol ergometer in both healthy subjects and patients with CVD, simultaneously monitoring the changes in sweat lactate with a wear- able lactate sensor. In all healthy subjects, the sensor was attached to the upper arm, and lactates in blood were measured every 2 min. Conversely, in patients with CVD, the sensor was attached to the forehead which has a higher-sweat rate compared to the upper arm as they were less likely to sweat than healthy subjects. All the patients underwent an exercise test with respiratory gas analysis. Among the 42 patients, 17 patients refused the blood lactate test during exercise because of the invasive procedure. In only 25 patients who provided written informed consent, blood lactate concentrations were measured during the exercise test26,33. Exercise testing protocol. On the day of the exercise test, the subjects avoided heavy physical activity before the test. The subjects performed the test in the upright position on an electronically braked ergometer (STRENGTH ERGO 8, Mitsubishi Electric Engineering Company, Japan). Following a 2-min rest to stabilize the heart rate and respiratory condition, the subjects performed a 2-min warm-up pedaling at 50 W for healthy men and at 0 W for healthy women and patients, and then exercised with a progressive intensity until the sub- jects could no longer maintain the pedaling rate (volitional exhaustion). At 1-min intervals, the intensity was increased by 20 W increments for healthy subjects, and 10 or 15 W increments for CVD patients (RAMP pro- tocol). The pedaling frequency was set at 60 rev/min. The incremental exercise testing time ranged from 10 to 20 min, depending on the exercise capacities of each subject or patient. Once the exercise tests were terminated the subjects were instructed to stop pedaling and to stay on the ergometer for 3  min26,33. Respiratory gas analysis and Ventilatory threshold. The additional method is available in the sup- plemental material S1. The expired gas flows were measured using a breath-by-breath automated system (AERO- MONITOR, MINATO MedicalScience CO., LTD., Osaka, Japan). The respiratory gas exchange, including ven- tilation (VE), oxygen uptake (VO2), and carbon dioxide production (VCO2), was continuously monitored and measured using a 10-s average. VT1 was determined using the ventilatory equivalent, excess CO2, and modified V-slope methods8. Three exercise testing experts, agreed on the VT1, independently from those who determined the sLT. First, two of three experienced researchers independently and randomly evaluated the VT1 of each subject using the three methods. The researchers used all three methods to assess concurrent break point and to eliminate false breakpoint. Second, if the VO2 values determined by the independent researchers were within 3%, then the VO2 values for the two investigators were averaged. Third, if the VO2 values determined by the independent evaluators were not within 3% of one another, a third researcher then independently determined VO2. The third VO2 value was then compared to those obtained by the initial investigators. If the adjudicated VO2 value was within 3% of either of the initial investigators, then two VO2 values were averaged26,33. Lactate threshold in blood. The blood lactate values were obtained via auricular pricking and squeezing the ear lobe gently to obtain a capillary blood sample every 2 min during the exercise test. The samples were analyzed immediately for the whole blood lactate concentration (mmol/L) using a standard enzymatic method on a lactate analyzer (LACTATE PRO2, ARKRAY, Japan)34. The bLT was determined through graphical plots35. A visual interpretation was independently made of each subject by two experienced researchers to locate the first rise from baseline. If the independent determinations of the stage at LT1 differed between the two researchers, a third researcher adjudicated the difference by inde- pendently determining LT1. The three researchers then jointly agreed on the LT1 point. Lactate threshold in sweat. The sLT was defined as the first significant increase in lactate in sweat above the baseline based on the graphical plots and Change Finder scores calculated by Change Finder algorithm (Online Fig. S3). Several candidate points (change points) of sLT were extracted by applying Change Finder algorithm36 to the time-series data of the lactate values in sweat in the range from the start to the end of exer- cise. Two-step learning with a Sequentially Discounting AR (SDAR) model was used to accurately distinguish between outliers and change points in the Change Finder algorithm. Three researchers, independently of the researchers who analyzed respiratory gas exchange, jointly agreed on the point of sLT. Statistical analyses. The results are represented as median with an interquartile range (IQR) for con- tinuous variables and as percentages for categorical variables, as appropriate. A univariable logistic regression analysis was performed to estimate the adjusted odds ratios (ORs) and 95% confidence intervals (CIs) for non- response in the lactate sensor. The relationships among the work rate (WR) at the sLT, bLT, and VT1 were inves- tigated using the Pearson’s correlation coefficient test. Additionally, the Bland and Altman technique was applied to verify the similarities among the different methods37. This comparison was a graphical representation of the difference between the methods and the average of these methods. Further, ordinary least products regression 10 Vol:.(1234567890) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ analysis was used to evaluate the fixed and proportional biases between each threshold38,39. All probability values were 2-tailed with P values < 0.05 considered statistically significant. All statistical analyses were performed with R version 3.6.3 (R Core Team, 2020, R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria). Received: 2 September 2020; Accepted: 9 February 2021 References 1. Schnohr, P., O’Keefe, J. H., Marott, J. L., Lange, P. & Jensen, G. B. Dose of jogging and long-term mortality: The Copenhagen City Heart Study. J. Am. Coll. Cardiol. 65, 411–419 (2015). 2. Wen, C. P. et al. Minimum amount of physical activity for reduced mortality and extended life expectancy: A prospective cohort study. Lancet 378, 1244–1253 (2011). 3. Clark, A. M., Hartling, L., Vandermeer, B. & McAlister, F. A. Meta-analysis: secondary prevention programs for patients with coronary artery disease. Ann. Intern. Med. 143, 659–672 (2005). 4. Flynn, K. E. et al. Effects of exercise training on health status in patients with chronic heart failure: HF-ACTION randomized controlled trial. JAMA 301, 1451–1459 (2009). 5. Belardinelli, R., Georgiou, D., Cianci, G. & Purcaro, A. 10-year exercise training in chronic heart failure: A randomized controlled trial. J. Am. Coll. Cardiol. 60, 1521–1528 (2012). 6. Binder, R. K. et al. Methodological approach to the first and second lactate threshold in incremental cardiopulmonary exercise testing. Eur. J. Cardiovasc. Prev. Rehabil. 15, 726–734 (2008). 7. Group, J. C. S. J. W. Guidelines for rehabilitation in patients with cardiovascular disease (JCS 2012). Circ. J. 78, 2022–2093 (2014) 8. Gaskill, S. E. et al. Validity and reliability of combining three methods to determine ventilatory threshold. Med. Sci. Sports Exerc. 33, 1841–1848 (2001). 9. An, B. W. et al. Smart sensor systems for wearable electronic devices. Polymers (Basel). 9 (2017). 10. Kobsar, D. & Ferber, R. Wearable sensor data to track subject-specific movement patterns related to clinical outcomes using a machine learning approach. Sensors (Basel). 18 (2018). 11. Baker, L. B. Sweating rate and sweat sodium concentration in athletes: A review of methodology and intra/interindividual vari- ability. Sports Med. 47, 111–128 (2017). 12. Jia, W. et al. Electrochemical tattoo biosensors for real-time noninvasive lactate monitoring in human perspiration. Anal. Chem. 85, 6553–6560 (2013). 13. Gao, W. et al. Fully integrated wearable sensor arrays for multiplexed in situ perspiration analysis. Nature 529, 509–514 (2016). 14. Imani, S. et al. A wearable chemical-electrophysiological hybrid biosensing system for real-time health and fitness monitoring. Nat. Commun. 7, 11650 (2016). 15. Brooks, G. A. The Science and Translation of Lactate Shuttle Theory. Cell Metab. 27, 757–785 (2018). 16. Robergs, R. A. Nothing ’evil’ and no ’conundrum’ about muscle lactate production. Exp. Physiol. 96, 1097–1098; author reply 1099–1100 (2011). 17. Buono, M. J., Lee, N. V. & Miller, P. W. The relationship between exercise intensity and the sweat lactate excretion rate. J. Physiol. Sci. 60, 103–107 (2010). 18. Falk, B. et al. Sweat lactate in exercising children and adolescents of varying physical maturity. J. Appl. Physiol. 1985(71), 1735–1740 (1991). 19. Jansen, T. C. et al. Early lactate-guided therapy in intensive care unit patients: A multicenter, open-label, randomized controlled trial. Am. J. Respir. Crit. Care Med. 182, 752–761 (2010). 20. Vincent, J. L., Quintairos, E. S. A., Couto, L. Jr. & Taccone, F. S. The value of blood lactate kinetics in critically ill patients: A sys- tematic review. Crit. Care 20, 257 (2016). 21. Lamont, L. S. Sweat lactate secretion during exercise in relation to women’s aerobic capacity. J. Appl. Physiol. 1985(62), 194–198 (1987). 22. Green, J. M., Bishop, P. A., Muir, I. H., McLester, J. R. Jr. & Heath, H. E. Effects of high and low blood lactate concentrations on sweat lactate response. Int. J. Sports Med. 21, 556–560 (2000). 23. Alvear-Ordenes, I., García-López, D., De Paz, J. A. & González-Gallego, J. Sweat lactate, ammonia, and urea in rugby players. Int. J. Sports Med. 26, 632–637 (2005). 24. Benson, J. W. Jr., Buja, M. L., Thompson, R. H. & Gordon, R. S. Jr. Glucose utilization by sweat glands during fasting in man. J. Invest. Dermatol. 63, 287–291 (1974). 25. Quinton, P. M. Physiology of sweat secretion. Kidney Int. Suppl. 21, S102-108 (1987). 26. 26Shiraishi, Y. et al. Real-time analysis of the heart rate variability during incremental exercise for the detection of the ventilatory threshold. J. Am. Heart. Assoc. 7 (2018). 27. Kim, D. H., Ghaffari, R., Lu, N. & Rogers, J. A. Flexible and stretchable electronics for biointegrated devices. Annu. Rev. Biomed. Eng. 14, 113–128 (2012). 28. Kim, D. H. et al. Epidermal electronics. Science 333, 838–843 (2011). 29. Chuang, M. C. et al. Flexible thick-film glucose biosensor: influence of mechanical bending on the performance. Talanta 81, 15–19 (2010). 30. Havenith, G., Fogarty, A., Bartlett, R., Smith, C. J. & Ventenat, V. Male and female upper body sweat distribution during running measured with technical absorbents. Eur. J. Appl. Physiol. 104, 245–255 (2008). 31. Patterson, M. J., Galloway, S. D. & Nimmo, M. A. Variations in regional sweat composition in normal human males. Exp. Physiol. 85, 869–875 (2000). 32. Taylor, N. A. & Machado-Moreira, C. A. Regional variations in transepidermal water loss, eccrine sweat gland density, sweat secretion rates and electrolyte composition in resting and exercising humans. Extrem. Physiol. Med. 2, 4 (2013). 33. Miura, K. et al. Feasibility of the deep learning method for estimating the ventilatory threshold with electrocardiography data. NPJ. Digit. Med. 3, 141 (2020). 34. Bonaventura, J. M. et al. Reliability and accuracy of six hand-held blood lactate analysers. J. Sports Sci. Med. 14, 203–214 (2015). 35. Faude, O., Kindermann, W. & Meyer, T. Lactate threshold concepts: how valid are they?. Sports Med. 39, 469–490 (2009). 36. Takeuchi, J. & Yamanishi, K. A unifying framework for detecting outliers and change points from time series. Ieee T Knowl Data En 18, 482–492 (2006). 37. Bland, J. M. & Altman, D. G. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1, 307–310 (1986). 11 Vol.:(0123456789) Scientific Reports | (2021) 11:4929 | https://doi.org/10.1038/s41598-021-84381-9 www.nature.com/scientificreports/ 38. Hart, S., Drevets, K., Alford, M., Salacinski, A. & Hunt, B. E. A method-comparison study regarding the validity and reliability of the Lactate Plus analyzer. BMJ Open. 3 (2013). 39. Ludbrook, J. Statistical techniques for comparing measurers and methods of measurement: a critical review. Clin. Exp. Pharmacol. Physiol. 29, 527–536 (2002). Acknowledgements The authors thank M. Fujioka, C. Yoshida, K. Takeuchi, and R. Kendo for their technical assistance. We are grateful to Editage for editing this manuscript. Author contributions The author contributions are stated as follows; Y.S., D.N. and Y.K drew the manuscript. Y.S., D.N., M.F., M.S. and Y.K. prepared the images. Y.S., D.N., Y.S., T.R., H.I., K.M. and Y.K. collected the patient information. T.W., T.N., M.M., M.N., K.S., K.F. and Y.K. provided a critical revision of the manuscript for the key intellectual content and supervision. All of the authors have approved all aspects of our work, read, and approved the manuscript. Competing interests This study was funded by Grant-in-Aid from Scientific Research from the Japan Agency for Medical Research and Development (ID. 19ek0210130h0001) and by a grant from Kimura Memorial Heart Foundation Research Grant for 2019, Suzuken Memorial Foundation, Foundation for Total Health Promotion, and Research Grant for Public Health Science. The funders had no role in study design, data collection and analysis, decision to publish or preparation of the manuscript. D.N is a founder and shareholder of Grace imaging Inc. Y. Shiraishi is affiliated with a department endowed by Nippon Shinyaku Co., Ltd., and received a research grant from the SECOM Science and Technology Foundation and an honorarium from Otsuka Pharmaceutical Co., Ltd. M. S. is an employee of Grace imaging Inc. Additional information Supplementary Information The online version contains supplementary material available at https ://doi. org/10.1038/s4159 8-021-84381 -9. Correspondence and requests for materials should be addressed to D.N. or Y.K. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/. © The Author(s) 2021

A novel device for detecting anaerobic threshold using sweat lactate during exercise.

03-02-2021

Seki, Yuta,Nakashima, Daisuke,Shiraishi, Yasuyuki,Ryuzaki, Toshinobu,Ikura, Hidehiko,Miura, Kotaro,Suzuki, Masato,Watanabe, Takatomo,Nagura, Takeo,Matsumato, Morio,Nakamura, Masaya,Sato, Kazuki,Fukuda, Keiichi,Katsumata, Yoshinori

eng

PMC4916632

Cross-sectional study of ethnic differences in physical fitness among children of South Asian, black African– Caribbean and white European origin: the Child Heart and Health Study in England (CHASE) C M Nightingale,1,2 A S Donin,1 S R Kerry,1 C G Owen,1 A R Rudnicka,1 S Brage,3 K L Westgate,3 U Ekelund,3,4 D G Cook,1 P H Whincup1 To cite: Nightingale CM, Donin AS, Kerry SR, et al. Cross-sectional study of ethnic differences in physical fitness among children of South Asian, black African– Caribbean and white European origin: the Child Heart and Health Study in England (CHASE). BMJ Open 2016;6:e011131. doi:10.1136/bmjopen-2016- 011131 ▸ Prepublication history and additional material is available. To view please visit the journal (http://dx.doi.org/ 10.1136/bmjopen-2016- 011131). Received 12 January 2016 Revised 6 April 2016 Accepted 13 May 2016 For numbered affiliations see end of article. Correspondence to Dr C M Nightingale; [email protected] ABSTRACT Objective: Little is known about levels of physical fitness in children from different ethnic groups in the UK. We therefore studied physical fitness in UK children (aged 9–10 years) of South Asian, black African–Caribbean and white European origin. Design: Cross-sectional study. Setting: Primary schools in the UK. Participants: 1625 children (aged 9–10 years) of South Asian, black African–Caribbean and white European origin in the UK studied between 2006 and 2007. Outcome measures: A step test assessed submaximal physical fitness from which estimated VO2 max was derived. Ethnic differences in estimated VO2 max were estimated using multilevel linear regression allowing for clustering at school level and adjusting for age, sex and month as fixed effects. Results: The study response rate was 63%. In adjusted analyses, boys had higher levels of estimated VO2 max than girls (mean difference 3.06 mL O2/min/ kg, 95% CI 2.66 to 3.47, p<0.0001). Levels of estimated VO2 max were lower in South Asians than those in white Europeans (mean difference −0.79 mL O2/min/kg, 95% CI −1.41 to −0.18, p=0.01); levels of estimated VO2 max in black African–Caribbeans were higher than those in white Europeans (mean difference 0.60 mL O2/min/kg, 95% CI 0.02 to 1.17, p=0.04); these patterns were similar in boys and girls. The lower estimated VO2 max in South Asians, compared to white Europeans, was consistent among Indian, Pakistani and Bangladeshi children and was attenuated by 78% after adjustment for objectively measured physical activity (average daily steps). Conclusions: South Asian children have lower levels of physical fitness than white Europeans and black African–Caribbeans in the UK. This ethnic difference in physical fitness is at least partly explained by ethnic differences in physical activity. INTRODUCTION In the UK and in other western European countries, there are marked ethnic differ- ences in chronic disease risks. British South Asian adults have increased risks of develop- ing type 2 diabetes, coronary heart disease and stroke compared to white Europeans in the UK.1 2 British black African–Caribbeans, in contrast, have increased risks of type 2 dia- betes and stroke compared to white Europeans but lower risks of coronary heart disease.1 2 Recent evidence suggests that these ethnic differences in disease risks have their origins in childhood, with increased levels of insulin resistance, circulating glucose and body fatness in South Asian chil- dren3 4 and (to a lesser extent) black African–Caribbean children compared to Strengths and limitations of this study ▪ The submaximal fitness test used to predict VO2 max has previously been employed in a nationally representative sample of the English population. ▪ Ethnic comparisons carried out on a within- school basis to limit confounding with balanced representation of South Asians, black African– Caribbeans and white Europeans. ▪ Objective accelerometer-based levels of physical activity levels were made using a validated method and bioelectrical impedance was used to provide a valid measurement of adiposity in this multi-ethnic population. ▪ The response rate of 63% was modest and esti- mated VO2 max was obtained for 75% of the sample who completed the step test; however, characteristics of participants with and without physical fitness data were similar. Nightingale CM, et al. BMJ Open 2016;6:e011131. doi:10.1136/bmjopen-2016-011131 1 Open Access Research white Europeans.4 Marked ethnic differences in physical activity levels have been reported in the UK in adults5 and children,6 with South Asians in particular having lower physical activity levels than white Europeans. However, little is known about the extent of ethnic dif- ferences in physical fitness, a determinant of cardiovas- cular and metabolic risk in adults7–9 and children.10 An earlier report based on a small number of ethnic minor- ity children studied in the National Study of Health and Growth raised the possibility that such differences could be substantial and that they might well be explained by ethnic differences in physical activity.11 The importance of independent assessment of physical fitness is empha- sised by the results of studies, showing that physical fitness and physical activity are independently associated with metabolic risk10 and that physical fitness may be a stronger risk factor than physical activity for coronary heart disease12 and all-cause mortality.13 We have there- fore examined ethnic differences in physical fitness in a study of British school children of South Asian, black African–Caribbean and white European origin, present- ing data on the major ethnic groups and subcategories. We have particularly examined the extent to which ethnic differences in physical fitness can be explained by objectively measured physical activity (overall intensity and moderate to vigorous intensity), an important deter- minant of physical fitness,10 14 and by adiposity, which is inversely associated with physical fitness.10 RESEARCH DESIGN AND METHODS Study design The Child Heart and Health Study in England (CHASE) was a cross-sectional investigation of the health (particu- larly cardiovascular and metabolic health) of primary school children aged 9–10 years of white European, South Asian and black African–Caribbean origin in the UK carried out between October 2004 and February 2007; full details have been published elsewhere.4 6 15 Ethical approval was obtained from the relevant multi- centre research ethics committee and written, informed parental consent was obtained for all participating chil- dren. The study was based in 200 state primary schools in London, Birmingham and Leicester, half with a high prevalence of UK South Asian children (stratified by Indian, Pakistani and Bangladeshi origin) and other half with a high prevalence of UK black African–Caribbean children (stratified by black African and black Caribbean origin). Schools were drawn at random from the stratified sampling frame; schools which declined to participate were replaced by another similar school within the sam- pling frame. This report is based on the final phase of the study (81 schools studied between January 2006 and February 2007) in which additional assessments of phys- ical fitness and physical activity were made. Survey measurements A single survey team of three trained Research Nurses and two Research Assistants carried out all assessments; each observer measured approximately one-third of chil- dren in each ethnic group. Participating children had physical measurements including height and weight, and arm-to-leg bioelectrical impedance, using the Bodystat 1500 bioelectrical impedance monitor (Bodystat, Isle of Man, UK). Fat mass was derived from impedance using ethnic- and gender-specific equations derived for UK children of this age group.16 Fat mass index (FMI) [fat mass (kg)/height(m)5] was derived to be independent of height (r=−0.02)16 and was shown to be a more valid marker of body fatness than body mass index in this study population.17 Participants underwent an 8-min step test as described previously.18 19 Briefly, participants were fitted with a combined heart rate (HR) and movement sensor (Actiheart, CamNtech, Papworth, UK), then followed an audible prompt instructing them to progressively increase their step frequency, ramping from 15 to 32.5 body lifts/min (rate of change: 2.5 body lifts/min2) on a 150 mm high step. The step test was terminated if the participant was unable to maintain the prescribed step frequency, even after verbal encouragement from the investigator. After test termination, 2 min of seated recovery was measured. The combined sensor-recorded ECG and acceleration waveforms (128 and 32 Hz sam- pling, respectively), were summarised in 15 s epochs. Data were visually reviewed and noisy ECG data were excluded from analysis. Heart beats were detected using a modified Pan-Tompkins peak detection algorithm.18 Estimation of VO2 max was done in a similar manner as Health Survey for England 2008.20 Briefly, predicted workload was regressed against instantaneous HR (expressed above resting level) and 1-min recovery HR was extracted using quadratic regression against recovery time (first 90 s); these parameters were combined with resting HR and test duration to define the submaximal relationship between HR and workload, which was then extrapolated to predicted maximal HR21 to predict maximal work capacity. This was converted to VO2 max by first adding an estimate of resting metabolic rate22 and then dividing by the energetic value of oxygen to estimate.20 Physical activity was assessed by accelerometry; chil- dren were asked to wear an activity monitor (GT1M; ActiGraph LLC, Pensacola, Florida, USA), on their left hip during waking hours (apart from during water-based activities) for 7 days following measurement and then return the instrument to the school. The ActiGraph monitor was worn over the left hip on an elasticised belt. Non-wear time, defined as periods of at least 20 consecutive minutes of zero counts, was excluded and remaining data were summarised into mean daily counts, counts per minute (CPM), steps per day and time spent at moderate to vigorous intensity (moderate– vigorous physical activity, MVPA; ≥2000 CPM counts). All participants with one or more days of valid data were included in the analysis; a valid day being defined as at least 600 min of registered time. 2 Nightingale CM, et al. BMJ Open 2016;6:e011131. doi:10.1136/bmjopen-2016-011131 Open Access Ethnicity and socioeconomic status The ethnicity of the child was defined using parental infor- mation on the self-reported ethnicity of both parents where available (63%), or using the parentally defined ethnic origin of the child (36%), or using information on parental and grand-parental place of birth provided by the child, cross-checked with observer assessment of ethnic origin (1%). Children were broadly defined (as previously described4) as white European, South Asian, black African–Caribbean and other ethnicity; more detailed ethnic subcategories of South Asians (Indian, Pakistani, Bangladeshi and South Asian other) and black African– Caribbeans (black African, black Caribbean and black other) were also used in analyses. Parents and children provided information on parental occupation, which was coded using the National Statistics-Socioeconomic Classification (NS-SEC) as previously described.23 Statistical methods Statistical analyses were carried out using Stata/SE soft- ware (Stata/SE V.12 for Windows; StataCorp LP, College Station, Texas, USA). Estimated VO2 max was normally dis- tributed (see online supplementary figure S1). We used a previously published regression calibration method to allow for measurement error in the physical activity vari- ables (counts, CPM, steps and moderate to vigorous activ- ity).24 This method allows for within-child variation in physical activity across a variable number of days of recording (between 1 and 7) and by day of the week and provides an unbiased average of counts and CPM for each child. Most children (87%) had 3 or more full days of recorded physical activity data. Restricting the analyses to these children did not materially affect the results. Gender differences in estimated VO2 max were assessed using multilevel models adjusted for age, ethnicity and month of the year (fitted as fixed effects) and school fitted as a random effect to take account of clustering of children within schools; all models were fitted using the xtmixed command in Stata. Similar models were used to quantify ethnic differences in estimated VO2 max adjusted for age, sex, month and school. An interaction between ethnic group and sex was fitted and likelihood ratio tests were used to examine whether ethnic differences in phys- ical fitness were modified by sex. Associations between estimated VO2 max and physical activity counts, CPM, fat mass index and resting HR were plotted and were quanti- fied using correlation coefficients. To examine whether ethnic differences in estimated VO2 max were explained by ethnic differences in physical activity, activity counts, CPM, steps or time spent in moderate to vigorous activity was fitted as a covariate in the model; fat mass index was fitted as a covariate to examine whether adiposity accounted for ethnic differences in estimated VO2 max. RESULTS Of 3571 children invited to participate in this phase of the study, 2236 (63%) took part in the physical fitness test. Response rates were similar among South Asians and white Europeans and other ethnic groups (67%, 65% and 63%, respectively) but slightly lower among black African–Caribbeans (59%); response rates were higher in girls than boys (66% and 59%, respectively). Among 2236 children who took part in the test, 73 parti- cipants were removed from the analysis due to not main- taining prescribed step frequency during the test and 538 participants did not have adequate HR data. Estimated VO2 max values were therefore derived for 1625 participants; the proportion of those with valid physical fitness data was similar in boys and girls and unrelated to ethnicity, socioeconomic position or phys- ical characteristics. The children with estimated VO2 max values included similar numbers of boys and girls (825 and 800, respectively) and similar numbers of children of white European, South Asian, black African– Caribbean and other ethnicity (424, 407, 413 and 381, respectively). Of these children, 1215 also had objective measures of physical activity. Unadjusted means and SDs for estimated VO2 max are shown by gender and ethnic group in online supple- mentary table S1; adjusted mean levels and gender dif- ferences in VO2 max are shown in table 1. The overall level of estimated VO2 max in the study population was 39.4 mL O2/min/kg (95% reference range 30.6, 48.2). Girls had markedly lower levels of estimated VO2 max than boys; on average, the level of estimated VO2 max in girls was 3.06 mL O2/min/kg lower (95% CI 2.66 to 3.47) than that in boys. This gender difference was apparent in all individual ethnic groups and there was no strong evidence of an interaction between ethnic group and sex (p=0.33). Mean levels of estimated VO2 max for white European, South Asian, black African–Caribbean and ‘other’ ethnic groups are presented in table 1 and adjusted dif- ferences compared to white Europeans are summarised in table 2. South Asian children had lower levels and black African–Caribbeans had higher levels of estimated VO2 max than white Europeans, whereas levels of esti- mated VO2 max in ‘other ethnicity’ were similar to those in white Europeans. Within ethnic minority groups, there was no strong evidence of heterogeneity between Indian, Pakistani and Bangladeshi children, all of whom had lower estimated VO2 max values than white Europeans. Estimated VO2 max levels were higher in black African–Caribbean children, with the difference apparently concentrated in boys and black African chil- dren; the test for heterogeneity between black Africans and Caribbeans was of borderline statistical significance (p=0.07). Estimated VO2 max was positively correlated with phys- ical activity counts, CPM, steps and time spent in MVPA; correlation coefficients (r) were 0.40, 0.35, 0.34 and 0.37 respectively. Estimated VO2 max was inversely correlated with resting HR (r=−0.66) and fat mass index (r=−0.37). As reported previously6 and shown in online supplemen- tary table S2, physical activity levels (including counts, Nightingale CM, et al. BMJ Open 2016;6:e011131. doi:10.1136/bmjopen-2016-011131 3 Open Access CPM, steps and MVPA) were lower in South Asians com- pared to white Europeans; black African–Caribbeans had similar levels of physical activity (CPM and MVPA), though mean levels of counts were higher and mean levels of steps were lower. Adiposity levels (particularly fat mass index) were previously reported in the larger CHASE study to be higher among South Asians com- pared to white Europeans and similar among black African–Caribbeans compared to white Europeans.17 In this smaller subset of the CHASE study, fat mass index was higher in South Asians, though the difference was not statistically significant, and lower in black African– Caribbeans compared to white Europeans. The effect of adjustment for physical activity (counts, CPM, steps or MVPA) and adiposity (fat mass index) on the overall ethnic differences in physical fitness are pro- vided in table 3 for the subset of data with measure- ments of physical activity and adiposity. The higher level of estimated VO2 max in black African–Caribbean chil- dren compared to white Europeans was reduced by 25% when adjusting for objectively measured physical activity counts; adjustment for steps increased the difference by 34%. Adjustment for fat mass index reduced the differ- ence in estimated VO2 max by 51% on its own and by 54% in combination with adjustment for physical activity (table 3). The lower level of estimated VO2 max in South Asian children compared to white Europeans was not statistically significantly different in this subset; adjust- ment for objectively measured physical activity further reduced this difference. For South Asians, adjustment for fat mass index did not have an appreciable effect on the ethnic difference in estimated VO2 max either as a single adjustment or in combination with adjustment for physical activity (table 3). In similar analyses, the effects of adjustment for phys- ical activity and adiposity on gender differences in esti- mated VO2 max are given in table 3. For all ethnic groups combined, girls had a markedly lower level of estimated VO2 max compared to boys; this difference was reduced by 36% by adjustment for physical activity Table 2 Ethnic differences in estimated VO2 max (mL O2/min/kg): overall and by sex Difference in estimated VO2 max (95% CI), p value Compared to white Europeans Boys (n=825) Girls (n=800) All South Asian −0.69 (−1.51 to 0.14) 0.10 −0.95 (−1.79 to −0.11) 0.03 −0.79 (−1.41 to −0.18) 0.01 Indian −1.16 (−2.36 to 0.03) 0.06 −1.93 (−3.29 to −0.58) 0.01 −1.52 (−2.45 to −0.59) 0.001 Pakistani −0.85 (−1.98 to 0.29) 0.14 −0.98 (−2.15 to 0.19) 0.10 −0.90 (−1.75 to −0.05) 0.04 Bangladeshi −0.15 (−1.48 to 1.19) 0.83 −0.67 (−1.83 to 0.49) 0.26 −0.38 (−1.29 to 0.53) 0.42 Black African–Caribbean 1.03 (0.25 to 1.82) 0.01 0.13 (−0.69 to 0.94) 0.76 0.60 (0.02 to 1.17) 0.04 Black African 1.64 (0.70 to 2.58) <0.001 0.24 (−0.71 to 1.19) 0.62 0.95 (0.27 to 1.62) 0.01 Black Caribbean 0.10 (−0.99 to 1.18) 0.86 0.17 (−0.96 to 1.29) 0.77 0.14 (−0.65 to 0.93) 0.72 Other 0.31 (−0.50 to 1.11) 0.46 −0.48 (−1.31 to 0.34) 0.25 −0.07 (−0.65 to 0.51) 0.80 Adjusted for sex, age quartiles, month, ethnic group, an interaction between ethnic group and sex (except for analysis of all children combined) and school (random effect). Table 1 Adjusted means for estimated VO2 max (mL O2/min/kg) by sex and ethnic group Mean estimated VO2 max (95% CI) Ethnic group or subgroup n Boys (n=825) Girls (n=800) p (sex difference) All p (ethnicity)* All children 1625 40.9 (40.5 to 41.3) 37.8 (37.4 to 38.2) <0.0001 39.4 (39.0 to 39.7) White European 424 40.7 (40.1 to 41.3) 38.1 (37.5 to 38.8) <0.0001 39.4 (39.0 to 39.9) South Asian 407 40.0 (39.4 to 40.7) 37.2 (36.5 to 37.8) <0.0001 38.6 (38.1 to 39.2) 0.13 Indian 111 39.6 (38.5 to 40.7) 36.2 (35.0 to 37.5) <0.0001 37.9 (37.1 to 38.8) Pakistani 147 39.9 (38.9 to 40.9) 37.2 (36.1 to 38.2) <0.0001 38.6 (37.8 to 39.3) Bangladeshi 121 40.6 (39.4 to 41.8) 37.5 (36.5 to 38.5) <0.0001 39.1 (38.2 to 39.9) Black African–Caribbean 413 41.8 (41.1 to 42.4) 38.3 (37.6 to 38.9) <0.0001 40.0 (39.6 to 40.5) 0.07 Black African 230 42.4 (41.6 to 43.2) 38.4 (37.6 to 39.2) <0.0001 40.4 (39.8 to 41.0) Black Caribbean 148 40.8 (39.9 to 41.8) 38.3 (37.3 to 39.3) <0.001 39.6 (38.9 to 40.3) Other 381 41.0 (40.4 to 41.7) 37.7 (37.0 to 38.3) <0.0001 39.4 (38.9 to 39.9) Adjusted for age quartiles, month, sex, ethnic group, an interaction between ethnic group and sex (except for analysis of all ethnic groups combined) and school (random effect). South Asian other and black other subgroups are not included in the table; therefore, the numbers in the subgroups do not add up to the main ethnic group totals for South Asians and black African–Caribbeans. *p Value for heterogeneity between ethnic subgroups within broader ethnic groups, that is, South Asian and black African–Caribbean. 4 Nightingale CM, et al. BMJ Open 2016;6:e011131. doi:10.1136/bmjopen-2016-011131 Open Access counts, by 28% for CPM, by 30% for steps, by 35% for moderate to vigorous activity and by 14% for fat mass index, though the differences were still highly statistically significant. Adjustment for fat mass index in combin- ation with adjustment for counts or CPM did not appre- ciably reduce the difference further. In further analyses, estimated VO2 max did not vary appreciably by socioeconomic status (NS-SEC), either overall (p[heterogeneity]=0.08) or within specific ethnic groups; furthermore, adjustment for socioeconomic status had very little effect on ethnic differences in esti- mated VO2 max (data available from authors). DISCUSSION This study provides evidence of lower levels of physical fitness in British South Asian children compared with white Europeans; black African–Caribbean children in contrast had higher levels of physical fitness compared to white Europeans. Lower levels of physical fitness were observed in girls compared to boys; this gender differ- ence was consistent across all ethnic groups studied. The lower level of physical fitness in South Asians compared to white Europeans was at least partly explained by the lower levels of physical activity in South Asian children. The higher level of physical fitness in black African– Caribbeans compared to white Europeans was at least partly explained by the higher levels of physical activity and lower levels of adiposity in black African–Caribbean children. The lower levels of physical fitness observed in girls could be partly (though not completely) explained by the lower physical activity levels in girls. Differences in adiposity did not contribute appreciably to South Asian–white European or gender differences in physical fitness. Relation to previous studies Levels of estimated VO2 max in boys and girls in the present study were slightly lower than those in children aged 10–12 years in a recent small UK-based study25 and were appreciably lower than those in children aged 9 years in the European Youth Heart Study,26 although the method of measuring VO2 max differed between studies; physical activity levels in the present study were also markedly lower than in the European Youth Heart Study.6 10 In the present study, girls had appreciably lower levels of physical fitness compared to boys, this finding is consistent with other studies in children of a similar age group in the UK11 25 and Europe.10 The finding that South Asian boys and girls had lower levels of physical fitness compared to white European children is consistent with previous findings in UK adults27 28 and with a small UK-based study of children aged 9 years;11 that study also reported that levels of physical fitness in black African–Caribbeans were similar to those of white European children, consistent with our findings.11 The finding that physical fitness was strongly related to phys- ical activity in the present study is consistent with find- ings from other studies,10 25 though no previous study has reported the contribution of physical activity to ethnic differences in physical fitness. Strengths and limitations The submaximal fitness test used in the present study to predict VO2 max has previously been used in a nationally representative sample of the English population20 and has been shown to capture two-thirds of the between-individual variance in the submaximal HR-VO2 relationship established using a wider intensity range during a more well-controlled stress-testing protocol (treadmill walking and running).19 Average levels and gender differences in estimated VO2 max were consistent with those in UK children of a similar age group who underwent maximal fitness testing.25 The cross-sectional design of this study means that we cannot infer that the association we observed between physical fitness and physical activity is causal; however, evidence from Table 3 Ethnic and sex differences in estimated VO2 max (mL O2/min/kg): effect of adjustment for physical activity levels and fat mass index Difference in estimated VO2 max (95% CI), p value Adjustments (N=1215) South Asian–white European Black African–Caribbean– white European Girls–Boys Standard −0.63 (−1.36 to 0.11) 0.09 0.80 (0.14 to 1.46) 0.02 −3.05 (−3.52 to −2.58) <0.0001 Standard+counts −0.34 (−1.05 to 0.37) 0.35 0.60 (−0.04 to 1.24) 0.07 −1.96 (−2.47 to −1.45) <0.0001 Standard+CPM −0.19 (−0.91 to 0.53) 0.61 0.81 (0.17 to 1.46) 0.01 −2.21 (−2.71 to −1.71) <0.0001 Standard+steps −0.14 (−0.87 to 0.59) 0.71 1.07 (0.42 to 1.73) 0.001 −2.12 (−2.64 to −1.60) <0.0001 Standard+MVPA −0.22 (−0.95 to 0.50) 0.54 0.72 (0.08 to 1.37) 0.03 −1.99 (−2.51 to −1.46) <0.0001 Standard+FMI −0.57 (−1.25 to 0.12) 0.11 0.39 (−0.24 to 1.01) 0.22 −2.61 (−3.05 to −2.16) <0.0001 Standard+counts+FMI −0.36 (−1.03 to 0.31) 0.29 0.29 (−0.32 to 0.90) 0.35 −1.87 (−2.36 to −1.39) <0.0001 Standard+CPM+FMI −0.26 (−0.95 to 0.42) 0.45 0.44 (−0.18 to 1.05) 0.16 −2.06 (−2.54 to −1.59) <0.0001 Standard+steps+FMI −0.19 (−0.88 to 0.49) 0.58 0.63 (0.01 to 1.24) 0.05 −1.92 (−2.41 to −1.42) <0.0001 Standard+MVPA+FMI −0.28 (−0.96 to 0.40) 0.42 0.37 (−0.25 to 0.98) 0.24 −1.88 (−2.38 to −1.38) <0.0001 Analysis carried out in a subset of 1215 children with objective measurements of physical activity. Standard adjustment is for sex, age quartiles, ethnicity, month and school (random effect).CPM, counts per minute; FMI, fat mass index; MVPA, moderate to vigorous physical activity. Nightingale CM, et al. BMJ Open 2016;6:e011131. doi:10.1136/bmjopen-2016-011131 5 Open Access randomised controlled trials has shown that interven- tions to increase activity levels in children and adoles- cents improve physical fitness,29–31 providing support for this direction of causality. The study included balanced representation of South Asians, black African– Caribbeans and white Europeans with ethnic compari- sons carried out on a within-school basis to limit confounding. Further strengths of this study include the objective and validated measurement of physical activity levels using accelerometry32 and the measurement of adiposity using bioelectrical impedance with ethnic- and gender-specific equations for the prediction of fat mass, a method which has been shown to give more valid assessment of adiposity in a multi-ethnic population.16 Although the response rate in the study was modest (63%) and estimated VO2 max was obtained for only 75% of the study sample who completed the step test, the socio-demographic and anthropometric character- istics of those participants were similar in those with and without physical fitness data, suggesting that the observed ethnic differences in physical fitness are not likely to be biased. Implications The results presented in this study showed that South Asians had lower levels and black African–Caribbeans had higher levels of physical fitness compared to white European children, and that these ethnic differences were at least partly explained by lower levels of physical activity among South Asian children and by higher levels of physical activity and lower levels of adiposity among black African–Caribbean children. However, ethnic dif- ferences in physical activity did not appear to provide a complete explanation for these differences in physical fitness. Lower levels of physical fitness are associated with higher levels of adiposity, insulin resistance and car- diovascular risk.7 10 25 South Asian children have higher levels of adiposity, insulin resistance and hyperglycaemia, and black African–Caribbean children and adolescents have higher levels of insulin resistance and hypergly- caemia compared to white Europeans.4 The lower levels of physical fitness among South Asians observed the present study (by ∼2%) could help to explain the early development of type 2 diabetes risk among South Asians; this will be an important area for further investi- gation. In contrast, the higher levels of fitness observed in black African–Caribbean children do not help to explain increased risks of type 2 diabetes in this group. Improvements in physical fitness (potentially through increases in physical activity) could be important for chronic disease prevention in the UK South Asian population. Author affiliations 1Population Health Research Institute, St George’s, University of London, London, UK 2Centre for Primary Care and Public Health, Queen Mary, University of London, London, UK 3MRC Epidemiology Unit, University of Cambridge School of Clinical Medicine, Institute of Metabolic Science, Cambridge, UK 4Department of Sport Medicine, Norwegian School of Sport Sciences, Oslo, Norway Acknowledgements The authors are grateful to the members of the CHASE study team (Julie Belbin, Claire Brannagan, Sarah Holloway, Cathy McKay, Mary McNamara, Miranda Price, Rahat Rafiq, Chloe Runeckles, Lydia Shepherd and Andrea Wathern) and to all participating schools, pupils and parents. Contributors PHW and CMN developed the idea for this paper with help from CGO, ARR, UE and DGC. PHW conceived, raised funding for and directed CHASE, with help from DGC, CGO and ARR. ASD collected data. SB, KLW and SRK derived physical fitness estimates from the results of the submaximal fitness test with support from UE. CMN and SRK carried out the statistical analyses. CMN wrote the first draft of the paper. All authors were involved in interpreting the data, critically reviewing the scientific content and all authors approved the final version. Funding This work was supported by grants from the British Heart Foundation (PG/06/003), Wellcome Trust (068362/Z/02/Z) and the National Prevention Research Initiative (G0501295). The funding partners for this NPRI award were: British Heart Foundation; Cancer Research UK; Department of Health; Diabetes UK; Economic and Social Research Council; Medical Research Council; Research and Development Office for the Northern Ireland Health and Social Services; Chief Scientist Office, Scottish Executive Health Department; and Welsh Assembly Government. Diabetes prevention research at St George’s, University of London, is supported by the National Institute of Health Research (NIHR)—Collaboration for Leadership in Applied Health Research and Care (CLAHRC) South London. Disclaimer The views expressed in this paper are those of the authors and not necessarily those of the funding agencies, the National Health Service, the NIHR or the Department of Health. Competing interests None declared. Ethics approval Multicentre Research Ethics Committee, Wales. Provenance and peer review Not commissioned; externally peer reviewed. Data sharing statement Enquiries about use of the study data set can be made by contacting Professor Peter Whincup Open Access This is an Open Access article distributed in accordance with the terms of the Creative Commons Attribution (CC BY 4.0) license, which permits others to distribute, remix, adapt and build upon this work, for commercial use, provided the original work is properly cited. See: http:// creativecommons.org/licenses/by/4.0/ REFERENCES 1. Health Survey for England 2004: The health of ethnic minority groups. In: Sproston K, Mindell J, eds. The NHS Information Centre, 2006. http://www.hscic.gov.uk/catalogue/PUB01209/heal-surv-hea- eth-min-hea-tab-eng-2004-rep.pdf 2. Wild SH, Fischbacher C, Brock A, et al. Mortality from all causes and circulatory disease by country of birth in England and Wales 2001–2003. J Public Health (Oxf) 2007;29:191–8. 3. Whincup PH, Gilg JA, Papacosta O, et al. Early evidence of ethnic differences in cardiovascular risk: cross sectional comparison of British South Asian and white children. BMJ 2002;324:635. 4. Whincup PH, Nightingale CM, Owen CG, et al. Early emergence of ethnic differences in type 2 diabetes precursors in the UK: the Child Heart and Health Study in England (CHASE Study). PLoS Med 2010;7:e1000263. 5. Fischbacher CM, Hunt S, Alexander L. How physically active are South Asians in the United Kingdom? A literature review. J Public Health (Oxf) 2004;26:250–8. 6. Owen CG, Nightingale CM, Rudnicka AR, et al. Ethnic and gender differences in physical activity levels among 9-10-year-old children of white European, South Asian and African-Caribbean origin: the Child Heart Health Study in England (CHASE Study). Int J Epidemiol 2009;38:1082–93. 6 Nightingale CM, et al. BMJ Open 2016;6:e011131. doi:10.1136/bmjopen-2016-011131 Open Access 7. Carnethon MR, Gidding SS, Nehgme R, et al. Cardiorespiratory fitness in young adulthood and the development of cardiovascular disease risk factors. JAMA 2003;290:3092–100. 8. LaMonte MJ, Eisenman PA, Adams TD, et al. Cardiorespiratory fitness and coronary heart disease risk factors: the LDS Hospital Fitness Institute cohort. Circulation 2000;102:1623–8. 9. LaMonte MJ, Barlow CE, Jurca R, et al. Cardiorespiratory fitness is inversely associated with the incidence of metabolic syndrome: a prospective study of men and women. Circulation 2005;112:505–12. 10. Ekelund U, Anderssen SA, Froberg K, et al. Independent associations of physical activity and cardiorespiratory fitness with metabolic risk factors in children: the European youth heart study. Diabetologia 2007;50:1832–40. 11. Bettiol H, Rona RJ, Chinn S. Variation in physical fitness between ethnic groups in nine year olds. Int J Epidemiol 1999;28:281–6. 12. Williams PT. Physical fitness and activity as separate heart disease risk factors: a meta-analysis. Med Sci Sports Exerc 2001;33:754–61. 13. Myers J, Kaykha A, George S, et al. Fitness versus physical activity patterns in predicting mortality in men. Am J Med 2004;117:912–18. 14. Rizzo NS, Ruiz JR, Hurtig-Wennlöf A, et al. Relationship of physical activity, fitness, and fatness with clustered metabolic risk in children and adolescents: the European youth heart study. J Pediatr 2007;150:388–94. 15. Donin AS, Nightingale CM, Owen CG, et al. Ethnic differences in blood lipids and dietary intake between UK children of black African, black Caribbean, South Asian, and white European origin: the Child Heart and Health Study in England (CHASE). Am J Clin Nutr 2010;92:776–83. 16. Nightingale CM, Rudnicka AR, Owen CG, et al. Are ethnic and gender specific equations needed to derive fat free mass from bioelectrical impedance in children of South Asian, black African-Caribbean and white European origin? Results of the Assessment of Body Composition in Children Study. PLoS ONE 2013;8:e76426. 17. Nightingale CM, Rudnicka AR, Owen CG, et al. Patterns of body size and adiposity among UK children of South Asian, black African-Caribbean and White European origin: Child Heart And health Study in England (CHASE Study). Int J Epidemiol 2011;40:33–44. 18. Brage S, Brage N, Franks PW, et al. Reliability and validity of the combined heart rate and movement sensor Actiheart. Eur J Clin Nutr 2005;59:561–70. 19. Brage S, Ekelund U, Brage N, et al. Hierarchy of individual calibration levels for heart rate and accelerometry to measure physical activity. J Appl Physiol 2007;103:682–92. 20. Health Survey for England 2008 Volume 1: Physical activity and fitness. In: Craig R, Mindell J, Hirani V, eds. The NHS Information Centre. 2009. http://www.hscic.gov.uk/catalogue/PUB00430/heal- surv-phys-acti-fitn-eng-2008-rep-v2.pdf 21. Tanaka H, Monahan KD, Seals DR. Age-predicted maximal heart rate revisited. J Am Coll Cardiol 2001;37:153–6. 22. Henry CJ. Basal metabolic rate studies in humans: measurement and development of new equations. Public Health Nutr 2005;8:1133–52. 23. Thomas C, Nightingale CM, Donin AS, et al. Socio-economic position and type 2 diabetes risk factors: patterns in UK children of South Asian, black African-Caribbean and white European origin. PLoS ONE 2012;7:e32619. 24. Owen CG, Nightingale CM, Rudnicka AR, et al. Physical activity, obesity and cardiometabolic risk factors in 9- to 10-year-old UK children of white European, South Asian and black African-Caribbean origin: the Child Heart And health Study in England (CHASE). Diabetologia 2010;53:1620–30. 25. Boddy LM, Murphy MH, Cunningham C, et al. Physical activity, cardiorespiratory fitness, and clustered cardiometabolic risk in 10- to 12-year-old school children: the REACH Y6 study. Am J Hum Biol 2014;26:446–51. 26. Adegboye AR, Anderssen SA, Froberg K, et al. Recommended aerobic fitness level for metabolic health in children and adolescents: a study of diagnostic accuracy. Br J Sports Med 2011;45:722–8. 27. Ghouri N, Purves D, McConnachie A, et al. Lower cardiorespiratory fitness contributes to increased insulin resistance and fasting glycaemia in middle-aged South Asian compared with European men living in the UK. Diabetologia 2013;56:2238–49. 28. Hall LM, Moran CN, Milne GR, et al. Fat oxidation, fitness and skeletal muscle expression of oxidative/lipid metabolism genes in South Asians: implications for insulin resistance? PLoS ONE 2010;5:e14197. 29. Annesi JJ, Westcott WL, Faigenbaum AD, et al. Effects of a 12-week physical activity protocol delivered by YMCA after-school counselors (Youth Fit for Life) on fitness and self-efficacy changes in 5-12-year-old boys and girls. Res Q Exerc Sport 2005;76:468–76. 30. Faigenbaum AD, Westcott WL, Loud RL, et al. The effects of different resistance training protocols on muscular strength and endurance development in children. Pediatrics 1999;104:e5. 31. Manios Y, Kafatos A, Mamalakis G. The effects of a health education intervention initiated at first grade over a 3 year period: physical activity and fitness indices. Health Educ Res 1998;13:593–606. 32. Ekelund U, Sjostrom M, Yngve A, et al. Physical activity assessed by activity monitor and doubly labeled water in children. Med Sci Sports Exerc 2001;33:275–81. Nightingale CM, et al. BMJ Open 2016;6:e011131. doi:10.1136/bmjopen-2016-011131 7 Open Access

Cross-sectional study of ethnic differences in physical fitness among children of South Asian, black African-Caribbean and white European origin: the Child Heart and Health Study in England (CHASE).

06-20-2016

Nightingale, C M,Donin, A S,Kerry, S R,Owen, C G,Rudnicka, A R,Brage, S,Westgate, K L,Ekelund, U,Cook, D G,Whincup, P H

eng

PMC6603669

sensors Article Effect of IMU Design on IMU-Derived Stride Metrics for Running Michael V Potter *, Lauro V Ojeda , Noel C Perkins and Stephen M Cain Department of Mechanical Engineering, University of Michigan, Ann Arbor, MI 48109, USA; [email protected] (L.V.O.); [email protected] (N.C.P.); [email protected] (S.M.C.) * Correspondence: [email protected] Received: 18 March 2019; Accepted: 5 June 2019; Published: 7 June 2019



 Abstract: Researchers employ foot-mounted inertial measurement units (IMUs) to estimate the three-dimensional trajectory of the feet as well as a rich array of gait parameters. However, the accuracy of those estimates depends critically on the limitations of the accelerometers and angular velocity gyros embedded in the IMU design. In this study, we reveal the effects of accelerometer range, gyro range, and sampling frequency on gait parameters (e.g., distance traveled, stride length, and stride angle) estimated using the zero-velocity update (ZUPT) method. The novelty and contribution of this work are that it: (1) quantifies these effects at mean speeds commensurate with competitive distance running (up to 6.4 m/s); (2) identifies the root causes of inaccurate foot trajectory estimates obtained from the ZUPT method; and (3) offers important engineering recommendations for selecting accurate IMUs for studying human running. The results demonstrate that the accuracy of the estimated gait parameters generally degrades with increased mean running speed and with decreased accelerometer range, gyro range, and sampling frequency. In particular, the saturation of the accelerometer and/or gyro induced during running for some IMU designs may render those designs highly inaccurate for estimating gait parameters. Keywords: wearable sensors; ZUPT; sensor requirements; gait 1. Introduction Studies of running biomechanics suggest that measured kinematic parameters (e.g., joint angles, stride frequency, stride length) may lead to the insight necessary to improve running performance and reduce injury risk [1,2]. Miniature inertial measurement units (IMUs) are an attractive option for analyzing human performance outside of traditional laboratory environments due to their relatively low cost, simple setup, and portability [3]. In one application, foot-mounted IMUs provide three-dimensional foot accelerations and angular rotational velocities from which foot trajectories (and associated gait parameters) are derived during walking/running [4–10]. Doing so requires minimizing the accumulated drift error in the estimated foot velocity and position using the so-called “zero-velocity update” (ZUPT) method [4,5,11]. The ZUPT method exploits the fact that the foot is nearly stationary at some time during the stance phase and uses that condition to estimate the foot velocity drift error for each gait cycle. Prior studies confirm that the ZUPT method yields accurate foot trajectory estimates for walking gait [9,12,13] and running gait with modest speeds (up to 4.36 m/s) [14]. However, little research addresses the requirements that ensure accurate trajectory estimates, particularly at faster running speeds, such as those observed in competitive middle- and long-distance running (up to 6.5 m/s) [15–17]. One limitation is that the sensor never achieves exactly zero-velocity, even in walking [18], and this assumption becomes increasingly suspect for faster running speeds. Another limitation lies with the range and sampling frequency of the inertial sensors themselves. Bailey and Harle [19] investigated Sensors 2019, 19, 2601; doi:10.3390/s19112601 www.mdpi.com/journal/sensors Sensors 2019, 19, 2601 2 of 19 the effect of IMU sampling frequency and accelerometer range on foot trajectory estimates and found that errors increase with increased running speed and decreased sampling frequency. While the errors observed in [19] were relatively small, the study considered modest running speeds (2.3–3.4 m/s) that are well below those of typical competitive middle- and long-distance running (up to 6.5 m/s). Additionally, the experiments in [19] were conducted on a treadmill rather than running overground which may also have influenced the conclusions. For example, significant differences may arise in gait kinematics when comparing walking on a treadmill versus overground [20]. Additionally, in a pilot study [21], the authors evaluated estimates of IMU-derived running speed using a treadmill. They observed that fluctuations in the treadmill belt speed, especially at higher running speeds with accompanying larger ground reaction forces, generated significant discrepancies between the reported belt speed and IMU-derived estimates of running speed. The speed fluctuations (i.e., accelerations) of the treadmill belt render it a non-inertial frame; thus, IMU-measured accelerations and angular velocities cannot be directly integrated to yield accurate estimates of foot velocity and position relative to the belt (as assumed using the ZUPT method). These treadmill-based limitations were a primary factor in modifying the pilot study protocol to the overground-based protocol of the study presented in this paper. Recently, Mitschke, et al. [22] examined the impact of accelerometer range on IMU-derived estimates of stride length, velocity, and tibial acceleration for overground running. They found no significant degradation in stride parameter estimates when the accelerometer range was ±32 g or greater, but significant degradation with smaller accelerometer ranges. Similar to [19], the study [22] considered only modest running speeds (up to 3.6 m/s) and did not disclose the fundamental reasons for the inaccurate estimates within the ZUPT method. In addition to the limitations imposed by accelerometer range and sampling frequency considered in these prior studies, we hypothesize that gyro range may also impose limitations on achieving accurate foot trajectory estimates for running. The effect of gyro range on stride estimates has likely not been studied previously because many commercial IMUs are unlikely to experience gyro saturation at the modest speeds observed in past studies. However, we hypothesize that the effects of gyro range on these estimates will become increasingly important at higher running speeds. The objective of this study is to reveal the impact of an IMU’s accelerometer range, gyro range, and sampling frequency on estimated stride parameters (i.e., stride length, stride angle, and total distance traveled) during overground walking and running up to competitive distance running speeds (up to 6.4 m/s) and over a wide range of sensor ranges and sampling frequencies typically found in commercially-available IMUs. This study also addresses the impact of gait speed on the estimated distance traveled in the presence of no saturation of the IMU signals. The novelty and contribution of this work are that it: (1) quantifies these effects at mean speeds commensurate with competitive distance running (up to 6.4 m/s); (2) identifies the root causes of inaccurate foot trajectory estimates obtained from the ZUPT method; and (3) offers important engineering recommendations for selecting accurate IMUs for studying human running. The results of this study will aid coaches and researchers in selecting appropriate IMUs to study stride parameters in outdoor environments and at speeds up to 6.4 m/s using the ZUPT method. 2. Materials and Methods 2.1. Subjects and Experimental Protocol Six healthy subjects (3 female, 3 male; mean (standard deviation) age 24.2 (±6.0) years, height 1.71 (±0.12) m, mass 68.4 (±15.2) kg) were recruited for this study. All subjects verified that they felt capable of completing the experimental protocol. Informed consent was obtained from all subjects and the study was approved by the University of Michigan IRB. Two IMU designs were employed. The first (IMU 1) provides a high accelerometer range and the second (IMU 2) provides a high sampling frequency as reported by the IMU specifications in Table 1. Each subject wore both IMUs strapped Sensors 2019, 19, 2601 3 of 19 together and placed on the bridge of both feet as shown in Figure 1 (only right foot shown) and secured with a strap and tape to limit their movement with respect to each other and to the foot. Table 1. Specifications for IMU 1 and IMU 2. IMU 1 IMU 2 Model Opal Custom design Manufacturer APDM Insight Sports, Ltd. Sampling Frequency (Hz) 128 1000 Accelerometer Range (g) ±200 ±32 Angular Gyro Range (deg/s) ±2000 ±4000 Figure 1. Two inertial measurement unit (IMU) designs and the means of attachment to a foot. (a) IMU 1 (Opal, APDM, left) and IMU 2 (custom design, Insight Sports, Ltd., right); (b and c) Side and front views of attachment of both IMUs to the instep via a Velcro strap. Sensor axes are denoted by X, Y, and Z (X and Z largely lie in foot sagittal plane and Y largely points to subject’s left). Note that these axes are illustrated to aid interpretation of the raw data signals presented in Figures 4 and 8, and that the zero-velocity update (ZUPT) method herein does not require specific sensor alignment to anatomical axes. We employed a method similar to [4] to assess the accuracy of ZUPT-based foot trajectory estimates outside of laboratory environments as follows. Each subject completed ten straight 100-meter trials on (level) asphalt. The subjects completed the first two trials at a perceived slow walk and fast walk, respectively. For the third through to the tenth trial, subjects ran at increasing speeds from a perceived slow jog (third trial) up to maximal sprint (tenth trial). For each trial, the subjects started at rest with the front of both shoes aligned with the start line. The subjects then walked/ran 100 meters and stopped with the front of both shoes aligned with the finish line. Rest between trials was self-selected by the subjects. Since all trials started and ended with the subject at rest, the time to complete each trial was readily identified from the start and end of significant acceleration (magnitude). The known 100-meter distance traveled was divided by the trial time to yield the mean speed for each trial. We employed the mean speed as an independent variable in the analyses below. 2.2. Overview of the ZUPT Method While the ZUPT method is generally known, we provide an overview for the reader’s benefit in following the discussions offered later in this paper. The ZUPT method used here draws largely from [4] and [9]. The method begins with estimating the instantaneous orientation of the IMU relative to an inertial frame as further detailed in [23]. The orientation is described by the rotation matrix, R, that defines the orientation of the sensor’s three orthogonal axes [ˆis, ˆjs,ˆks] (corresponding to x, y, Sensors 2019, 19, 2601 4 of 19 and z sensor axes, respectively) relative to the orthogonal axes [ˆiw, ˆjw,ˆkw] of a world (i.e., inertial) frame (corresponding to x, y, and z world axes, respectively, with z pointing in the direction of gravity) per:   ˆiw ˆjw ˆkw   = R   ˆis ˆjs ˆks   (1) Following [23], the angular velocity is integrated to estimate R at each time step with a Kalman filter used to correct drift error in the tilt angle. The three components of IMU-measured acceleration (axs, ays, azs) yield the acceleration components in the world frame (axw, ayw, azw) through:   axw ayw azw   = R   axs ays azs   −   0 0 g   (2) where g is the acceleration of gravity. These acceleration components in the inertial frame are used to estimate velocity in the inertial frame using the fact that the foot-mounted IMU returns to zero-velocity during the stance phase as described below. First, zero-velocity times are identified as times of minimum angular velocity magnitude during the stance phase. A stride is defined by the time interval between two successive zero-velocity times tn−1 < t < tn. Thus, the stride time is: ts = tn − tn−1 (3) and the number of data samples, l, in each stride is: l = ts × Fs + 1 (4) where Fs is the sampling frequency of the IMU. For each stride, the initial velocity is set to zero at tn−1 and the acceleration components in the inertial frame (2) are integrated to estimate the velocity components in the inertial frame (´vxw, ´vyw, ´vzw). Because it is assumed that the velocity at tn returns to zero, the estimated velocity error (per sample) in the stride is: → Verror =   ´vxw ´vyw ´vzw   (t = tn) × 1 l − 1, (5) assuming linear drift. For each sample number, k ∈ (1, l), within the stride (e.g., k = 3 for third sample in stride), a linear velocity drift error correction is applied to each of the three world frame velocity components:   vxcorr vycorr vzcorr   =   ´vxw ´vyw ´vzw   − → Verror × (k − 1). (6) The corrected velocity for the whole trial is integrated to estimate the foot position throughout the trial. The position estimates are segmented by zero-velocity times and used to estimate the stride length and stride angle as follows. Figure 2 gives a two-dimensional illustration of the stride length and angle. Sensors 2019, 19, 2601 5 of 19 Figure 2. Illustration of stride length , ∆, and stride angle, θ. Orange dots on footprints represent consecutive zero-velocity times. Identified zero-velocity times define the start position, → Sn, and end position, → En, of the nth stride as given by: → Sn =   Sxw,n Syw,n Szw,n   , (7) → En =   Exw,n Eyw,n Ezw,n   , (8) which denote the X, Y, and Z components of the start and end positions in the world frame for the nth stride. The stride length, ∆, is calculated as the total three-dimensional displacement of the IMU (and thus the foot) between the start and end of the stride. We note that in many biomechanical studies, stride length often refers to only the anterior–posterior component of the foot displacement; however, in using this technology for biomechanical analyses, there is a precedent to use alternative definitions of stride length, such as the total horizontal displacement of the foot, such as in [14]. Other studies make assumptions about the sensor/foot orientation, such as assuming a particular sensor axis is perfectly aligned with the medial–lateral axis of the subject [22] or impose a level ground assumption to constrain vertical drift [24]. While such assumptions may be useful in simplifying calculations or in interpreting some of the other foot parameters obtainable using similar ZUPT method applications (e.g., if interested in the foot roll and pitch angles, precise alignment of sensor and anatomical axes may be helpful or even necessary), we do not impose such restrictions in our method (e.g., our method can be used for non-level walking/running and our method does not require any alignment between the sensor and anatomical axes). In particular, the stride length (as defined for this study) for the nth stride is: ∆n = r (Exw,n − Sxw,n)2 +  Eyw,n − Syw,n 2 + (Ezw,n − Szw,n)2. (9) The stride angle, θ, is calculated as the three-dimensional angle between successive strides’ vectors. First, we define the stride vector for the nth stride as: → Dn =   Exw,n − Sxw,n Eyw,n − Syw,n Ezw,n − Szw,n   . (10) The nth stride angle is thus computed as: θn = tan−1    → Dn × → Dn+1  → Dn· → Dn+1   , (11) Sensors 2019, 19, 2601 6 of 19 with × and · being the standard vector cross and dot products, respectively. We again note that the estimated stride parameters used in this study (stride length and stride angle) do not require any assumptions of a particular alignment of the sensor axes on the foot. 2.3. Data Processing and Analysis Each IMU’s acceleration and angular velocity data were used to estimate the foot’s three-dimensional trajectory throughout each trial using the ZUPT method described above. The stride lengths were added (for both the left and right foot and then averaged) to yield the estimated total distance traveled (Dcalc) during each trial. The cumulative distance error: Derr = (Dcalc/Dtruth − 1) × 100%, (12) is reported for each trial, where Dtruth is the known distance traveled (100 meters in this study). To investigate the effect of accelerometer range, the raw accelerometer data from IMU 1, that possesses an acceleration range ±200 g, were numerically truncated to seven smaller ranges; namely 100, 75, 50, 24, 16, 10, and 6 g. For example, to investigate the effect of a 16 g accelerometer, any acceleration outside the range of −16 g < a < 16 g was set to the corresponding limit (−16 g or 16 g) to simulate sensor saturation at that limit. These seven ranges were chosen because they are typical of commercial IMU designs. Note that the acceleration data never exceeded ±100 g in any trial and thus we used the data with the ±100 g accelerometer range as the baseline for this analysis. After the raw accelerometer data were modified in this manner, the ZUPT method was used (with the modified accelerometer data and raw angular velocity data as input) to estimate the foot trajectories. The cumulative distance error (12) was computed as a function of the mean speeds for each accelerometer range. These data were then fit to a linear mixed-effects model [25] to test the statistical significance of the following: (1) the effect of mean gait speed on the cumulative distance error with no accelerometer saturation (i.e., using the 100 g range accelerometer), and (2) the effect of smaller accelerometer ranges on these estimates (i.e., using the 75, 50, 24, 16, 10, and 6 g range accelerometers). The statistical model and its full results are detailed in Appendix A. Similarly, to investigate the effect of gyro range, the raw angular velocity data from IMU 1, which possesses an angular velocity range of ±2000 deg/s, were also numerically truncated to four smaller ranges; namely 1500, 1000, 750, and 500 deg/s (i.e., ranges common in commercially available IMU designs). For this analysis, we used the data with the ±2000 deg/s gyro range as the baseline. After the raw angular velocity data were modified in this manner, the ZUPT method was used (with the modified angular velocity data and raw accelerometer data as input) to estimate the foot trajectories and cumulative distance error as described above. An analogous statistical model to the one described above was also employed, but investigating gyro range instead of accelerometer range. Additionally, for both accelerometer and gyro range effects, the amount of data that was lost due to truncation was quantified. To accomplish this, we compared the integrated area under the truncated signal and that of the non-truncated signal. In particular, to quantify the data lost due to accelerometer saturation, we defined the percent data loss as: Ldata =  1 − R  →a trun dt R  →a non dt   × 100%, (13) where  →a trun  is the acceleration magnitude of the truncated signal,  →a non  is the corresponding acceleration magnitude of the non-truncated signal, and dt is the time per sample. The integration is over the length of the trial. The percent data loss due to angular velocity saturation was defined analogously. Note that IMU 1 was specifically chosen for studying the effects of accelerometer and gyro ranges and we verified that all data were within the design ranges (±200 g and ±2000 deg/s) for IMU 1 in all Sensors 2019, 19, 2601 7 of 19 trials. By contrast, accelerometer saturation arose in IMU 2 (±32 g) at higher running speeds. However, IMU 2 was specifically chosen to study the effect of sampling frequency due to the sampling frequency limitations of IMU 1 (128 Hz). IMU 2 possesses a sampling frequency of 1000 Hz, far beyond the minimum 250 Hz rate recommended in [19] for obtaining accurate foot position and velocity estimates using the ZUPT method in running at speeds up to 3.4 m/s. For the purpose of studying the effect of sampling frequency, the accelerometer and gyro data from IMU 2 (1000 Hz) were also down-sampled to four smaller sampling frequencies (500, 250, 125, and 62.5 Hz) typical of commercial IMU designs. To that end, we employed two down-sampling methods as further described in the Results section. For these analyses, we used the data with the 1000 Hz sampling frequency as the baseline. In addition to studying the cumulative distance error, we also report stride-to-stride variations in the differences of individual stride length and stride angle estimates as defined above (e.g., the standard deviation of the stride length difference over a trial) as these data also reveal important conclusions. To this end, the stride length difference for a particular stride and trial (e.g., stride 2 for subject 1 and trial 1) was defined as: ∆di f = ∆base − ∆est, (14) where the ∆base is the estimated stride length using the baseline (non-saturated/non-downsampled) IMU data and ∆est is the estimated stride length using the truncated or downsampled data generated as discussed above. 3. Results 3.1. Effect of Accelerometer Range The statistical analysis (Table A1 of Appendix A) reveals a significant effect of speed on cumulative distance error (p < 0.01) with the 100 g accelerometer range despite no observed accelerometer or gyro signal saturation in any of the trials with this range. Additionally, the statistical analysis reveals that using an accelerometer range of 24 g or below leads to significantly greater degradation of the estimated distance traveled with speed (p < 0.01 for 24 g, p < 0.001 for 16 g, 10 g, and 6 g) relative to the 100 g range. The accelerometer ranges of 75 g and 50 g revealed no statistically significant effects compared to 100 g. These findings are observable in Figure 3 which illustrates the cumulative distance error versus mean running speed for the original (100 g) accelerometer range and for each of the six truncated accelerometer ranges utilizing data from IMU 1. As illustrated, subjects achieved mean speeds up to 6.4 m/s with the upper end of this range, similar to speeds observed in elite distance running [15]. Because the peak accelerations rarely exceeded 50 g, the cumulative distance errors for the 100, 75, and 50 g accelerometer ranges are visually indistinguishable on this scale as expected from the statistical results. Additionally, note that the cumulative distance error results converge across accelerometer ranges with decreased running speeds because these speeds generally yield lower accelerations (e.g., for the slowest trial, accelerations never exceeded 6 g, yielding identical cumulative distance errors for all accelerometer ranges considered). For walking and low running speeds (i.e., <2.2 m/s), the illustrated results largely confirm the cumulative distance errors reported by others for walking [4,9]. Additionally, for these low speeds, the IMU-estimated cumulative stride distances are nearly independent of accelerometer range. This is expected since the peak accelerations rarely exceeded 6 g for walking and low running speeds. However, for high running speeds, significant portions of each stride cycle generated accelerations larger than 6 g, leading to the large observable degradations in the estimated cumulative distance at the higher running speeds with decreased accelerometer range. Importantly, Figure 3 shows that accelerometers with ranges exceeding 50 g yield cumulative distance errors no greater than 5% for all mean speeds observed in this study (up to 6.4 m/s). Thus, depending on the accuracy needs for a particular use of these estimates, the ZUPT method may yield acceptable results (i.e., errors of 5% or less) even at the highest speeds observed in this study, provided no saturation arises in the accelerometer (and gyro) signals. Sensors 2019, 19, 2601 8 of 19 By contrast, at the opposite extreme, errors exceeding 30% are observable for the 6 g accelerometer where significant saturation occurs. Figure 3. Effect of accelerometer range and mean running speed on the cumulative distance error. Note that errors arising from the 100, 75, and 50 g accelerometers are indistinguishable on this scale. The shaded region indicates estimates within ±5% error. The degradation of estimates of the cumulative distance traveled with increased speed and decreased accelerometer range traces to saturation in the accelerometer signals. Figure 4 illustrates the effect of truncating the accelerometer range for sample walking (Figure 4a) and running (Figure 4b) trials. In the sample walking trial, the three acceleration components never exceed 6 g and therefore distance estimates based on any of the accelerometer ranges considered (6 g through 100 g) yield essentially identical results. However, in the sample running trial, the acceleration components often exceed 6 g and for significant portions of the gait cycle. The data loss leads to significant foot trajectory errors largely due to how the ZUPT method corrects for velocity drift error as described in detail in the Discussion. Figure 4. Effect of accelerometer range limits on acceleration data for a sample (a) walking trial (mean speed 1.4 m/s) and a sample (b) running trial (mean speed 5.8 m/s). Acceleration data is saturated when the slope is zero as is most apparent for the X-axis acceleration of the 6 g accelerometer for the running trial. Shaded areas indicate stance phase. Sensors 2019, 19, 2601 9 of 19 We note that the presence of saturation does not necessarily lead to poor estimates. In particular, acceptable results may still be obtainable if the amount of data that is lost due to saturation remains small. To illustrate what might be “acceptable” levels of saturation for the accelerometer range, we further quantified the amount of data that is lost due to saturation for each trial and accelerometer range and present the relationship of the cumulative distance error and amount of data lost due to saturation in Figure 5. These results show that the cumulative distance errors remain below 5% when the percentage of acceleration data lost due to saturation is below 1.5%, irrespective of the accelerometer range. Figure 5. Cumulative distance error versus percentage of acceleration data lost due to saturation for all trials and for each accelerometer range. The shaded region indicates estimates within ±5% error. Beyond studying errors in the cumulative distance traveled, we also considered differences in stride length and stride angle estimates on a stride by stride basis. To this end, we compared the estimated length and angle of each stride (where stride angle refers to the angle between successive strides as defined in the Methods) using truncated accelerometer data to the same quantities estimated from untruncated accelerometer data, employing the 100 g accelerometer as the benchmark. Figure 6 illustrates the standard deviation of the resulting differences in the individual stride length (Figure 6a) and stride angle (Figure 6b) estimates. When the variation is large, the agreement between truncated and baseline estimates of the given parameter for individual strides is small. These variations increase strongly with mean speed and decreased accelerometer range. As both mean speed and accelerometer range contribute to accelerometer saturation, they also significantly impact the estimates of these metrics on an individual stride basis. Figure 6. Dependence of the standard deviation of difference in (a) stride length and (b) stride angle estimates with mean speed and accelerometer range. The differences are computed with respect to the results of the 100 g accelerometer as the benchmark. Sensors 2019, 19, 2601 10 of 19 3.2. Effect of Gyro Range The statistical analysis (Table A2 of Appendix A) reveals a significant effect of speed on cumulative distance error (p < 0.01) with the 2000 deg/s gyroscope range despite no observed accelerometer or gyro signal saturation in any of the trials with this range. Additionally, the statistical analysis reveals that using a gyro range of 750 deg/s or below leads to significantly greater degradation of the estimated distance traveled with speed (p < 0.05 for 750 deg/s, p < 0.001 for 500 deg/s) versus the 2000 deg/s range. Gyro ranges of 1500 deg/s and 1000 deg/s reveal no statistically significant effects compared to 2000 deg/s. These findings are observable in Figure 7 which illustrates the cumulative distance error versus mean running speed for the five gyro ranges considered utilizing data from IMU 1. Because the angular velocities rarely exceeded 1000 deg/s, the distance error is nearly the same for the 2000, 1500, and 1000 deg/s range gyros. The distance errors remain within 5% for all gyro ranges of at least 1000 deg/s for the entire range of mean speeds studied herein (up to 6.4 m/s). Figure 7. Effect of gyro range and mean running speed on the cumulative distance error. Note that distance errors arising from the 2000 and 1500 deg/s gyros are indistinguishable on this scale. The shaded region indicates estimates within ±5% error. As in Figure 3, the cumulative distance traveled is underestimated at faster (running) speeds and this underestimation increases with speed and decreased gyro range. However, the cumulative distance traveled is often overestimated at slower (walking) speeds and this overestimation increases with decreased gyro range; observe the slower (walking) trials with a 500 deg/s range gyro. The underestimation versus overestimation traces to saturation in distinct portions of the stride cycle as revealed in Figure 8 for sample walking (Figure 8a) and running (Figure 8b) trials. Observe in Figure 8a that the Y-axis angular velocity for the 500 deg/s gyro exhibits saturation during a modest fraction of the stance phase near toe-off (end of the stance phase) for walking. By contrast, Figure 8b reveals that for maximal sprinting, the same angular velocity component saturates during toe-off, heel-strike (beginning of the stance phase), and for a significant portion of the swing phase. The portion of the stride cycle in which data is lost leads to overestimation versus underestimation because of how it impacts the ZUPT algorithm as described in detail in the Discussion. Sensors 2019, 19, 2601 11 of 19 Figure 8. Effect of gyro range limits on angular velocity for a sample (a) walking trial (mean speed 1.4 m/s) and a sample (b) running trial (mean speed 5.8 m/s). Angular velocity data is saturated when the slope is zero as is most apparent for the Y-axis angular velocity of the 500 deg/s gyro for the running trial. Shaded areas indicate stance phases. As with the accelerometer, we note that the presence of saturation in the gyro does not necessarily lead to poor estimates; in particular, acceptable results may still be obtainable if the amount of data that is lost due to saturation remains small. To illustrate what might be “acceptable” levels of saturation for gyro signals, we further quantify the amount of data that is lost due to saturation for each trial and gyro range and present the relationship of cumulative distance error and amount of data lost due to saturation in Figure 9. These results show that cumulative distance errors remain below 5% when the percentage of angular velocity data lost due to saturation is below 2.6%, regardless of the gyro range. Figure 9. Cumulative distance error versus percentage of angular velocity data lost due to saturation for all trials and for each gyro range. The shaded region indicates estimates within ± 5% error. As in the previous section, we also considered differences in stride length and stride angle estimates on a stride by stride basis. To this end, we compared the estimated length and angle of each stride (where stride angle refers to the angle between successive strides as defined in the Methods) using truncated gyro data to the same quantities estimated from untruncated gyro data, employing the 2000 deg/s gyro as the benchmark. Figure 10 illustrates the standard deviation of the resulting differences in the individual stride length (Figure 10a) and stride angle (Figure 10b) estimates. When the Sensors 2019, 19, 2601 12 of 19 variation is large, the agreement between the truncated and baseline estimates of the given parameter for individual strides is small. These variations increase strongly with mean speed and decreased gyro range. As both mean speed and gyro range contribute to gyro saturation, they also significantly impact the estimates of these metrics on an individual stride basis. Figure 10. Dependence of the standard deviation of the difference in (a) stride length and (b) stride angle estimates with mean speed and gyro range. The differences are computed with respect to the results of the 2000 deg/s gyro as the benchmark. 3.3. Effect of Sampling Frequency Sampling methods can vary widely in commercial IMUs. In particular, one or more filters are commonly employed within the IMU hardware and/or software before data is output at the IMUs specified sampling frequency. Therefore, an IMU having a higher sampling frequency (specification) does not necessarily imply it will lead to superior estimates of stride parameters in the context of this study. Because the filters and sampling methods are generally hidden to the user and vary between manufacturers, we considered the effect of sampling frequency by studying two simple sampling methods, including both an extreme method (no filtering before down-sampling) and a common method (low pass filter before down-sampling). For both methods, data from IMU 2 was utilized as that IMU design yields data at a high (1000 Hz) sampling frequency. Neither sampling method demonstrated a statistically significant effect for the interaction of speed and sampling frequency on the cumulative distance error except for the most extreme downsampling used in this study (Method 1 at the lowest sampling frequency). See Appendix A for full statistical results. Because this one exception represents a most unrealistic scenario and because no other sampling method and sampling frequency combinations studied herein revealed statistically significant effects for this interaction, we offer no further results for this effect. However, significant differences do arise in the estimated stride lengths and stride angles of the individual strides as reported below. Method 1 constitutes simple down-sampling performed without filtering (e.g., when down-sampling from 1000 to 500 Hz, every other sample is retained). This overly simplistic approach introduces aliasing effects and hence sub-optimal results [26,27]. Figure 11 illustrates the standard deviation of differences in stride length (Figure 11a) and stride angle (Figure 11b) estimates using Method 1 as functions of both mean speed and sampling frequency. The differences are with respect to the same quantities computed using the original data (i.e., data sampled at 1000 Hz). The standard deviation of the difference from simple down-sampling quickly grows (i.e., increasing variation) with increasing mean speed and decreasing sampling frequency. For example, at the lowest sampling frequency (62.5 Hz), the standard deviation in stride length Sensors 2019, 19, 2601 13 of 19 difference becomes a significant fraction of the stride length at higher mean speeds. Thus, the reliability of these measures is significantly impacted by both mean speed and sampling frequency. Figure 11. Dependence of the standard deviation of the difference in (a) stride length and (b) stride angle estimates with mean speed and sampling frequency. Method 1: down-sampling without filtering. The differences are computed with respect to the results of the 1000 Hz sampling frequency as the benchmark. Method 2 follows a more common strategy known as decimation [28,29], which consists of low pass filtering prior to down-sampling. We used the decimate function in MATLABTM [30] which utilizes a low pass Chebyshev Type I filter (infinite impulse response, order 8) before down-sampling the data. Figure 12 illustrates the results from Method 2, analogous to those of Method 1. The results still illustrate increased differences in estimates with increased speed and decreased sampling frequency, but significantly less than that observed using Method 1 (Figure 11). For example, the variation of the stride length difference is reduced by nearly a factor of five (compare scales of Figures 11a and 12a). These results suggest that sensor hardware that employs well-designed filters can significantly mitigate the adverse impact of limited sampling frequencies. Figure 12. Dependence of the standard deviation of the difference in (a) stride length and (b) stride angle estimates with mean speed and sampling frequency. Method 2: low pass filtering before down-sampling (using MATLABTM decimate function). The differences are computed with respect to the results of the 1000 Hz sampling frequency as the benchmark. Note the differences in the y-axis scales compared to Figure 11. Sensors 2019, 19, 2601 14 of 19 4. Discussion Overall, this study highlights the importance of proper sensor selection in order to estimate accurate gait parameters from foot-mounted IMUs using the ZUPT method. Accurate estimates of the cumulative distance traveled are possible upon limiting acceleration and angular velocity saturation. Importantly, we observed that the cumulative distance error using the ZUPT method remains below 5% when acceleration saturation is limited to 1.5% and when angular velocity saturation is limited to 2.6%; refer to Figures 5 and 9. We also observed that gait parameter estimates degrade with higher mean speeds even without sensor saturation (p < 0.01); refer to Appendix A. However, the results confirm that ZUPT-based algorithms yield accurate estimates for some applications (i.e., less than 5% cumulative distance error) over the entire range of mean speeds studied herein (up to 6.4 m/s) contingent on the IMU design. Importantly, lower range inertial sensors yield significant errors in gait parameter estimates at higher mean speeds due to (increasingly larger) data saturation. Interestingly, saturation may produce both overestimates and underestimates of the cumulative distance traveled depending on which signal (acceleration or angular velocity) is saturated, in which part of the stride cycle most of the saturation occurs, and the mean speed. These errors arise from error sources within the ZUPT method (detailed in the Methods) as follows. The gyro data is employed to estimate the orientation of the IMU (via integrating angular velocity) and this is critical to accurately resolving the acceleration into the world frame. The orientation estimates are corrected for drift error using a Kalman filter based on the core assumption of zero-mean Gaussian gyro noise. However, this assumption is violated when the angular velocity saturates, leading to inaccurate estimates of orientation and thus improper resolution of acceleration in the world frame. Subsequent integration of poorly resolved (and even possibly saturated) acceleration yields inaccurate estimates of velocity and position. Additionally, even with proper IMU orientation estimates, saturation of accelerometer signals will create velocity drift errors that do not increase linearly in time between the zero-velocity update times as assumed in the ZUPT method. These error sources suggest an intuitive explanation for why the estimated total distance traveled is increasingly underestimated with increased speed and decreased accelerometer range. In this study, the majority of acceleration data lost due to saturation was acceleration directed opposite to the direction of travel (i.e., deceleration). Consequently, the uncorrected velocity in the direction of travel was overestimated at the end of a stride. However, when the (linear) velocity-drift correction was then applied, it consistently lead to underestimated velocity in the direction of travel and corresponding underestimated stride length. By contrast, there is no parallel explanation for why saturated gyro data may lead to both over and under estimates of the stride length. In particular, note that saturation of gyro data creates errors farther upstream in the ZUPT algorithm and specifically in the orientation estimation. Errors in the orientation estimation may yield both over and under estimates of the stride length. Despite these several error sources, accurate velocity and position estimates are still obtained if either the percentage of the missing sensor data remains small (as described above) or if the saturation occurs along a sensor axis that does not contribute significantly to the estimate. A naïve user may be tempted to conclude that it is always best to select an IMU with the largest ranges for acceleration and angular velocity to always avoid saturation. However, increased range often comes with the tandem penalty of reduced resolution (e.g., if range is increased, but bit resolution is not) as well as increased sensor noise, which may both defeat the apparent advantage of higher range sensors. Consequently, there could well be instances where an IMU possessing an accelerometer that admits minor saturation yields superior stride parameter estimates relative to one possessing a higher range accelerometer that admits no saturation. Of course, a superior concept is to employ multiple accelerometers and/or rate gyros with increasing (and even slightly overlapping) ranges, a concept not studied herein. We note that in estimating the total distance traveled, symmetrically distributed stride length errors (i.e., some overestimated and some underestimated) may cancel, leading to accurate estimates Sensors 2019, 19, 2601 15 of 19 of total distance traveled. Therefore, it is important to understand how individual stride length and angle estimates are affected by sensor parameters. We chose to study these effects using the reported standard deviation of the stride length (and stride angle) differences, where these differences were compared to the baseline estimates (using the maximal sensor parameters). By using both the cumulative error in the distance traveled and the standard deviations of stride length and stride angle differences, we were able to reach sound conclusions of how IMU parameters affect individual stride length estimates as reported herein despite not explicitly having stride by stride ground truth data. In particular, we note that standard deviations of stride length and stride angle differences appear to converge as sensor parameters (ranges and sampling frequency) approach the nominal parameters for that sensor (Figures 6 and 10–12). This apparent convergence suggests that (as one would expect) the baseline estimates are likely the best available estimates and thus the differences likely correspond to degradations in the estimates. However, we also acknowledge that independent ground truth estimates of individual stride lengths and angles are required to confirm this conclusion and that data was not available in the present study. The convergence in standard deviations also suggests that improvements in estimates of individual stride parameters when going beyond the ranges and sampling frequencies utilized by the sensor in this study will be minor compared to the degradation effects due to sensor limitations demonstrated in this study. For these experiments, sampling frequency showed no significant impact on estimates of the total distance traveled (except in one limited condition); however, it significantly impacted estimates of the individual stride parameters (stride length and stride angle). In particular, the variance of the individual stride parameters was significantly influenced by the filtering/sampling method employed. This effect was demonstrated using two simple down-sampling methods. While this analysis demonstrated differences in stride parameter estimates with a reduction of the sampling frequency on a single IMU, caution must be exercised when comparing sampling frequencies between IMU designs. Many factors of IMU design in addition to sampling frequency (e.g., sensor hardware, sensor placement) impact stride parameter estimates. Thus, it remains possible for an IMU with a modest sampling frequency (e.g., 128 Hz) to yield superior stride parameter estimates to another IMU design with a higher sampling frequency (e.g., 1000 Hz). Finally, we describe several limitations of this study which we also believe do not alter the core conclusions. First, we acknowledge that the sensors available for this study both had limitations (in range and sampling frequency) that may affect the accuracy of the calculated stride metrics presented herein. Despite these limitations, we demonstrated important conclusions about the effects of the sensor parameters on the selected stride parameters. We duly note additional factors not studied herein that may impact the accuracy of stride estimates (e.g., sensor noise, sensor bandwidth, sensor resolution, Kalman filter tuning, etc.) that motivate future studies. Second, these experiments were conducted over a relatively short (100 m) distance. However, the results for the cumulative distance error are expected to hold for any (i.e., longer) distance because the cumulative distance error is equivalent to a percent error in the mean stride length estimates. Third, in the statistical analyses detailed in Appendix A, we did not evaluate subject-specific effects (i.e., we removed the effect of subjects on our results by treating subject as a random effect). Future studies could investigate the effect of subject demographics (e.g., weight, height, etc.) on estimates of gait parameters using the ZUPT method. Fourth, we note that our study is not well suited for traditional statistical power analyses. Thus, we remind the reader that caution should be employed when interpreting the effects that were not found to be significant in the statistical analyses as they may be subject to type-II statistical errors (i.e., an effect not found to be significant does not guarantee that there is no effect). However, we also note that the potential presence of type-II errors (for the effects not observed to be significant) in no way diminishes the importance of the effects that were observed to be significant and their associated conclusions. Fifth, we suggest that future studies investigate the impact of sensor properties on other stride parameters obtainable from the ZUPT method (e.g., foot clearance, foot roll angle, etc.). Sensors 2019, 19, 2601 16 of 19 5. Conclusions Appropriate selection of the ranges and sampling frequencies of the inertial sensors embedded in IMUs is crucial for accurately estimating foot trajectories (hence gait parameters) from foot-mounted IMUs, and particularly for the speeds associated with competitive distance running. In this study, we investigated the effects of mean gait speed and sensor parameters on estimates of stride parameters. The novelty and contribution of this work are that it: (1) quantifies these effects at mean speeds commensurate with competitive distance running (up to 6.4 m/s); (2) identifies the root causes of inaccurate foot trajectory estimates obtained from the ZUPT method; and (3) offers important engineering recommendations for selecting accurate IMUs for studying human running. Estimates of the cumulative distance traveled (from the individual stride length estimates) degrade with speed; however, across the range of mean speeds studied here, estimates remained within 5% of ground truth if there was no or minor saturation of the accelerometer (1.5% or less) or gyro (2.6% or less) signals as defined herein. In particular, the reported experiments required accelerometer ranges of at least 50 g and gyro ranges of at least 1000 deg/s to avoid significant errors in estimates of the cumulative distance traveled for mean running speeds up to 6.4 m/s. Errors that arise due to sensor saturation trace to core assumptions that are violated in the underlying estimation procedure based on the ZUPT method (i.e., zero-mean Gaussian noise, zero-velocity, and linear velocity drift assumptions). For applications similar to the ones described in this paper, accurate results remain possible even with modest sampling frequencies (e.g., 128 Hz), provided well-designed filters are employed. Author Contributions: Conceptualization, M.V.P. and S.M.C.; Formal analysis, M.V.P.; Funding acquisition, N.C.P.; Investigation, M.V.P. and S.M.C.; Methodology, M.V.P. and S.M.C.; Supervision, L.V.O., N.C.P., and S.M.C.; Writing—original draft, M.V.P.; Writing—review and editing, L.V.O., N.C.P., and S.M.C. Funding: Parts of this research were funded by US Army Contracting Command-APG, Natick Contracting Division, Natick, MA, contract number W911QY-15-C-0053. This material is also based upon work supported by the National Science Foundation Graduate Research Fellowship Program under Grant No. DGE 1256260. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation. Acknowledgments: The authors would like to thank Corey Powell for his help with the statistical analyses for this study. Conflicts of Interest: The authors declare no conflict of interest. Appendix A —Full Statistical Results from Linear Mixed-Effects Models We report results for four statistical analyses that reveal the effects of accelerometer range, gyro range, and downsampling by two methods. For each of the four analyses, we utilized a linear mixed-effects model to test the effects of speed, sensor parameter (sensor range or sampling frequency), and the combination of speed and sensor parameter on the cumulative distance error. In each model, we removed subject effects by treating the subject as a random effect. Because we are interested in the effect of mean speed on the cumulative distance error without saturation, we treated mean speed as a continuous variable and fixed effect. We are also interested in the interaction of mean speed and the sensor parameter (i.e., how speed affects the cumulative distance error for a change in the sensor parameter versus that effect for the baseline sensor parameter) and included this interaction as a fixed effect. The model also calculated the effects of the sensor parameter alone (effect of only sensor parameter independent of speed) and a y-axis intercept (cumulative distance error with a mean speed of 0 m/s) for the baseline sensor condition; however, we did not include those additional findings as they did not add to the conclusions or the interpretation of the presented results. In each analysis, we used estimates obtained from the original (non-truncated or non-downsampled) IMU data as the baseline. All statistical analyses were run in R statistical software using lme4 and lmerTest packages [25,31,32]. The function lmer was used to fit the models and the function confint was used to compute 95% confidence intervals for all estimates. Sensors 2019, 19, 2601 17 of 19 The four tables below present results from this model for each of the four analyses. The first (shaded) row in each table reports the effect of speed only on the cumulative distance error for the baseline sensor condition. This includes an estimated linear slope for the cumulative distance error versus speed (i.e., an estimated slope of 1%/(m/s) would indicate that the cumulative distance error increases by an estimated 1% for each 1 m/s increase in mean speed) for the baseline sensor condition where a negative value means that the error is negative as defined in this paper (i.e., underprediction of cumulative distance). This slope also represents the sensitivity of the cumulative distance error to mean speed. In the remaining (unshaded) rows, we report the interactions of mean speed and the sensor parameter (i.e., how a change to sensor parameters compared to the baseline sensor condition impacts the sensitivity of the cumulative distance error to mean speed). The estimated slopes in these rows denote the estimated difference in the slope of the cumulative distance error versus speed over that for the baseline condition. Thus, these entries report the additional error sensitivity to speed for the specified change in the sensor parameter relative to the baseline. For example, in Table A1, the estimated cumulative distance speed error is negative and it grows by −0.39%/(m/s) for the baseline sensor. The error sensitivity then grows (i.e., worsens) by an additional −0.57%/(m/s) for a sensor employing an accelerometer with a 24 g range. The third and fourth columns report the significance (p-value) and the 95% confidence interval for the estimated slope/error sensitivity. Table A1. Effect of speed on the cumulative distance versus accelerometer range. First (shaded) row reports the estimated slope of the cumulative distance versus mean speed using the baseline sensor (100 g range accelerometer). Remaining (unshaded) rows report estimated differences in this slope for IMUs with indicated acceleration ranges compared to the baseline IMU (i.e., the increased error sensitivity to speed with the indicated reductions in acceleration range relative to the baseline range). Speed Effect Estimated Slope (%/(m/s)) p-Value 95% Conf. Int. Baseline (100g) −0.39 <0.01 (−0.65, −0.13) 75 g vs. Baseline 0.00 0.99 (−0.37, 0.37) 50 g vs. Baseline −0.03 0.88 (−0.39, 0.34) 24 g vs. Baseline −0.57 <0.01 (−0.94, −0.21) 16 g vs. Baseline −1.08 <0.001 (−1.45, −0.72) 10 g vs. Baseline −2.24 <0.001 (−2.60, −1.87) 6 g vs. Baseline −5.70 <0.001 (−6.06, −5.33) Table A2. Effect of speed on the cumulative distance versus gyro range. First (shaded) row reports the estimated slope of cumulative distance versus mean speed using the baseline sensor (2000 deg/s gyro). Remaining (unshaded) rows report estimated differences in this slope for IMUs with indicated angular velocity ranges compared to the baseline IMU (i.e., the increased error sensitivity to speed following the indicated reductions in angular velocity range relative to the baseline range). Speed Effect Estimated Slope (%/(m/s)) p-Value 95% Conf. Int. Baseline (2000 deg/s) −0.41 <0.01 (−0.68, −0.13) 1500 deg/s vs. Baseline 0.00 0.99 (−0.39, 0.39) 1000 deg/s vs. Baseline 0.05 0.81 (−0.34, 0.44) 750 deg/s vs. Baseline −0.40 <0.05 (−0.80, −0.01) 500 deg/s vs. Baseline −3.08 <0.001 (−3.47, −2.69) Sensors 2019, 19, 2601 18 of 19 Table A3. Effect of speed on the cumulative distance versus sampling frequency for Method 1 (downsampling). First (shaded) row reports the estimated slope of the cumulative distance traveled versus mean speed using the baseline sensor (1000 Hz). Remaining (unshaded) rows report the estimated differences in this slope for IMUs with the indicated sampling frequency compared to the baseline IMU (i.e., the increased error sensitivity to speed following the indicated reductions in sampling frequency relative to the baseline sampling frequency). Speed Effect Estimated Slope (%/(m/s)) p-Value 95% Conf. Int. Baseline (1000 Hz) −1.12 <0.001 (−1.48, −0.77) 500 Hz vs. Baseline −0.09 0.73 (−0.59, 0.41) 250 Hz vs. Baseline −0.13 0.63 (−0.63, 0.38) 125 Hz vs. Baseline −0.26 0.32 (−0.76, 0.24) 62.5 Hz vs. Baseline 0.94 <0.001 (0.44, 1.44) Table A4. Effect of speed on the cumulative distance versus sampling frequency for Method 2 (decimation). First (shaded) row reports the estimated slope of the cumulative distance traveled versus mean speed using the baseline sensor (1000 Hz). Remaining (unshaded) rows report the estimated differences in this slope for IMUs with the indicated sampling frequency compared to the baseline IMU (i.e., the increased error sensitivity to speed following the indicated reductions in sampling frequency relative to the baseline sampling frequency). Speed Effect Estimated Slope (%/(m/s)) p-Value 95% Conf. Int. Baseline (1000 Hz) −1.17 <0.001 (−1.50, −0.85) 500 Hz vs. Baseline −0.01 0.96 (−0.47, 0.45) 250 Hz vs. Baseline −0.03 0.90 (−0.49, 0.43) 125 Hz vs. Baseline 0.08 0.74 (−0.38, 0.54) 62.5 Hz vs. Baseline 0.12 0.60 (−0.33, 0.58) References 1. Cavanagh, P.R.; Kram, R. Stride length in distance running: Velocity, body dimensions, and added mass effects. Med. Sci. Sports Exerc. 1989, 21, 467–479. [CrossRef] [PubMed] 2. Novacheck, T.F. The biomechanics of running. Gait Posture 1998, 7, 77–95. [CrossRef] 3. Norris, M.; Anderson, R.; Kenny, I.C. Method analysis of accelerometers and gyroscopes in running gait: A systematic review. Proc. Inst. Mech. Eng. Part P J. Sport Eng. Technol. 2014, 228, 3–15. [CrossRef] 4. Ojeda, L.V.; Borenstein, J. Non-GPS navigation for security personnel and first responders. J. Navig. 2007, 60, 391–407. [CrossRef] 5. Foxlin, E. Pedestrian tracking with shoe-mounted inertial sensors. IEEE Comput. Graph. Appl. 2005, 25, 38–46. [CrossRef] [PubMed] 6. Reenalda, J.; Maartens, E.; Homan, L.; Buurke, J.H.J. Continuous three dimensional analysis of running mechanics during a marathon by means of inertial magnetic measurement units to objectify changes in running mechanics. J. Biomech. 2016, 49, 3362–3367. [CrossRef] [PubMed] 7. Xing, H.; Li, J.; Hou, B.; Zhang, Y.; Guo, M. Pedestrian stride length estimation from IMU measurements and ANN based algorithm. J. Sens. 2017. [CrossRef] 8. Hollman, J.H.; Watkins, M.K.; Imhoff, A.C.; Braun, C.E.; Akervik, K.A.; Ness, D.K. A comparison of variability in spatiotemporal gait parameters between treadmill and overground walking conditions. Gait Posture 2016, 43, 204–209. [CrossRef] 9. Rebula, J.R.; Ojeda, L.V.; Adamczyk, P.G.; Kuo, A.D. Measurement of foot placement and its variability with inertial sensors. Gait Posture 2013, 38, 974–980. [CrossRef] 10. Sabatini, A.M.; Martelloni, C.; Scapellato, S.; Cavallo, F. Assessment of walking features from foot inertial sensing. IEEE Trans. Biomed. Eng. 2005, 52, 486–494. [CrossRef] 11. Elwell, J. Inertial navigation for the urban warrior. In Proceedings of the SPIE; SPIE: Bellingham, WA, USA, 1999; Volume 3709, pp. 196–204. Sensors 2019, 19, 2601 19 of 19 12. Ojeda, L.V.; Rebula, J.R.; Adamczyk, P.G.; Kuo, A.D. Mobile platform for motion capture of locomotion over long distances. J. Biomech. 2013, 46, 2316–2319. [CrossRef] [PubMed] 13. Kitagawa, N.; Ogihara, N. Estimation of foot trajectory during human walking by a wearable inertial measurement unit mounted to the foot. Gait Posture 2016, 45, 110–114. [CrossRef] [PubMed] 14. Brahms, C.M.; Zhao, Y.; Gerhard, D.; Barden, J.M. Stride length determination during overground running using a single foot-mounted inertial measurement unit. J. Biomech. 2018, 71, 302–305. [CrossRef] [PubMed] 15. Leskinen, A.; Häkkinen, K.; Virmavirta, M.; Isolehto, J.; Kyröläinen, H. Comparison of running kinematics between elite and national-standard 1500-m runners. Sports Biomech. 2009, 8, 1–9. [CrossRef] [PubMed] 16. Hanley, B. Pacing, packing and sex-based differences in Olympic and IAAF World Championship marathons. J. Sports Sci. 2016, 34, 1675–1681. [CrossRef] [PubMed] 17. Hoogkamer, W.; Kram, R.; Arellano, C.J. How Biomechanical Improvements in Running Economy Could Break the 2-h Marathon Barrier. Sports Med. 2017, 47, 1739–1750. [CrossRef] [PubMed] 18. Peruzzi, A.; Della Croce, U.; Cereatti, A. Estimation of stride length in level walking using an inertial measurement unit attached to the foot: A validation of the zero velocity assumption during stance. J. Biomech. 2011, 44, 1991–1994. [CrossRef] 19. Bailey, G.P.; Harle, R.K. Investigation of sensor parameters for kinematic assessment of steady state running using foot mounted IMUs. In Proceedings of the 2nd International Congress on Sports Sciences Research and Technology Support, Rome, Italy, 16–18 October 2014; pp. 154–161. 20. Ojeda, L.V.; Rebula, J.R.; Kuo, A.D.; Adamczyk, P.G. Influence of contextual task constraints on preferred stride parameters and their variabilities during human walking. Med. Eng. Phys. 2015, 37, 929–936. [CrossRef] 21. Potter, M.V.; Ojeda, L.V.; Perkins, N.C.; Cain, S.M. Influence of accelerometer range on accuracy of foot-mounted IMU based running velocity estimation. In Proceedings of the Annual Meeting of the American Society of Biomechanics, Boulder, CO, USA, 8–11 August 2017; pp. 1133–1134. 22. Mitschke, C.; Kiesewetter, P.; Milani, T.L. The effect of the accelerometer operating range on biomechanical parameters: Stride length, velocity, and peak tibial acceleration during running. Sensors 2018, 18. [CrossRef] 23. Ojeda, L.V.; Zaferiou, A.M.; Cain, S.M.; Vitali, R.V.; Davidson, S.P.; Stirling, L.A.; Perkins, N.C. Estimating stair running performance using inertial sensors. Sensors 2017, 17. [CrossRef] 24. Bailey, G.P.; Harle, R. Assessment of foot kinematics during steady state running using a foot-mounted IMU. Procedia Eng. 2014, 72, 32–37. [CrossRef] 25. Bates, D.; Maechler, M.; Bolker, B.; Walker, S. Fitting Linear Mixed-Effects Models Using lme4. J. Stat. Softw. 2015, 67, 1–48. [CrossRef] 26. Nyquist, H. Certain topics in telegraph transmission theory. Trans. AIEE 1928, 47, 617–644. [CrossRef] 27. Shannon, C.E. Communication in the presence of noise. Proc. Inst. Radio Eng. 1949, 37, 10–21. [CrossRef] 28. Antoniou, A. Digital Signal. Processing; McGraw-Hill Education: New York, NY, USA, 2006; ISBN 0-07-145424-1. 29. Lyons, R. Understanding Digital Signal. Processing; Prentice Hall: Upper Saddle River, NJ, USA, 2001; ISBN 0-201-63467-8. 30. MATLAB Decimate. Available online: https://www.mathworks.com/help/signal/ref/decimate.html (accessed on 25 July 2018). 31. R Core Team. R: A Language and Environment for Statistical Computing; R Foundation for Statistical Computing: Vienna, Austria, 2017; Available online: https://www.R-project.org (accessed on 7 March 2019). 32. Kuznetsova, A.; Brockhoff, P.B.; Christensen, R.H.B. lmerTest Package: Tests in Linear Mixed Effects Models. J. Stat. Softw. 2017, 82, 1–26. [CrossRef] © 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Effect of IMU Design on IMU-Derived Stride Metrics for Running.

06-07-2019

Potter, Michael V,Ojeda, Lauro V,Perkins, Noel C,Cain, Stephen M

eng

PMC6195805

Wrist-worn Accelerometry for Runners: Objective Quantification of Training Load VICTORIA H. STILES1, MATTHEW PEARCE1, ISABEL S. MOORE2, JOSS LANGFORD3, and ALEX V. ROWLANDS4,5,6 1Sport and Health Sciences, College of Life and Environmental Sciences, University of Exeter, Exeter, UNITED KINGDOM; 2Cardiff School of Sport and Health Sciences, Cardiff Metropolitan University, Cardiff, UNITED KINGDOM; 3GENEActiv, Activinsights, Cambridgeshire, UNITED KINGDOM; 4Diabetes Research Centre, University of Leicester, Leicester, UNITED KINGDOM; 5National Institute for Health Research (NIHR), Leicester Biomedical Research Centre, Leicester, UNITED KINGDOM; and 6Alliance for Research in Exercise, Nutrition and Activity (ARENA), Sansom Institute for Health Research, Division of Health Sciences, University of South Australia, Adelaide, AUSTRALIA ABSTRACT STILES, V. H., M. PEARCE, I. S. MOORE, J. LANGFORD, and A. V. ROWLANDS. Wrist-worn Accelerometry for Runners: Objective Quantification of Training Load. Med. Sci. Sports Exerc., Vol. 50, No. 11, pp. 2277–2284, 2018. Purpose: This study aimed to apply open-source analysis code to raw habitual physical activity data from wrist-worn monitors to: 1) objectively, unobtrusively, and accurately discriminate between ‘‘running’’ and ‘‘nonrunning’’ days; and 2) develop and compare simple accelerometer-derived metrics of external training load with existing self-report measures. Methods: Seven-day wrist-worn accelerometer (GENEActiv; Activinsights Ltd, Kimbolton, UK) data obtained from 35 experienced runners (age, 41.9 T 11.4 yr; height, 1.72 T 0.08 m; mass, 68.5 T 9.7 kg; body mass index, 23.2 T 2.2 kgImj2; 19 [54%] women) every other week over 9 to 18 wk were date-matched with self-reported training log data. Receiver operating characteristic analyses were applied to accelerometer metrics (‘‘Average Acceleration,’’ ‘‘Most Active-30mins,’’ ‘‘MinsQ400 mg’’) to discriminate between ‘‘running’’ and ‘‘nonrunning’’ days and cross-validated (leave one out cross-validation). Variance explained in training log criterion metrics (miles, duration, training load) by accelerometer metrics (MinsQ400 mg, ‘‘workload (WL) 400-4000 mg’’) was examined using linear regression with leave one out cross-validation. Results: Most Active-30mins and MinsQ400 mg had 994% accuracy for correctly classifying ‘‘running’’ and ‘‘nonrunning’’ days, with validation indicating robustness. Variance explained in miles, duration, and training load by MinsQ400 mg (67%–76%) and WL400–4000 mg (55%–69%) was high, with validation indicating robustness. Conclusions: Wrist- worn accelerometer metrics can be used to objectively, unobtrusively, and accurately identify running training days in runners, reducing the need for training logs or user input in future prospective research or commercial activity tracking. The high percentage of variance explained in existing self-reported measures of training load by simple, accelerometer-derived metrics of external training load supports the future use of accelerometry for prospective, preventative, and prescriptive monitoring purposes in runners. Key Words: WORKLOAD, TRAINING EXPOSURE, TRAINING PROGRAMS, ATHLETE MONITORING, INJURY PREVENTION, PERFORMANCE R unners are suggested to be particularly at risk of developing a running-related injury (RRI) if they have one or a combination of the following: a history of injury, low or high running experience (high indicates that long distances have been run for many years), a low (women) or high (men) weekly training frequency, a low or high overall weekly running mileage or a sudden increase in training load (1–3). Characteristics of external training load (work done) typically described as the distance, frequency, intensity, and duration of running per day/week or month are therefore highly modifiable risk factors for RRI (1–4). Optimal patterns of training load relative to rest and sleep (recovery) are also im- portant in the prevention of RRI and illness (5–7). However, a single validated method enabling longitudinal training patterns to be objectively, accurately, and unobtrusively quantified in runners is unavailable. A more detailed understanding of the influence of training load on RRI and performance could be enhanced by an improved ability to objectively monitor sim- ple, yet meaningful characteristics of external training load in runners on a large population scale (5,8,9). Within research and applied settings, characteristics of external training load, such as miles and duration, are Address for correspondence: Victoria Stiles, Ph.D., Sport and Health Sci- ences, College of Life and Environmental Sciences, University of Exeter, Exeter, United Kingdom; E-mail: [email protected]. Submitted for publication February 2018. Accepted for publication June 2018. 0195-9131/18/5011-2277/0 MEDICINE & SCIENCE IN SPORTS & EXERCISE Copyright  2018 The Author(s). Published by Wolters Kluwer Health, Inc. on behalf of the American College of Sports Medicine. This is an open access article distributed under the Creative Commons Attribution License 4.0 (CCBY), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. DOI: 10.1249/MSS.0000000000001704 2277 EPIDEMIOLOGY typically recorded using a training log (self-reported or coach- reported), global navigational satellite system (GNSS), or prescribed within a training program. To avoid inaccuracies from self-reported data due to recall bias (overreported/ underreported training/activity), characteristics of external training load can be more accurately quantified using objec- tive measurements (9,10). For example, initial findings from the use of pedometers in military recruits to estimate dis- tances covered over consecutive weeks of training have highlighted the importance of capturing evidence of previ- ously unreported additional training and habitual physical activity (PA) associated with stress fractures (9). In addition to pedometers, there has been a vast increase in the use of more sophisticated, commercial, consumer-focused wrist- worn activity trackers that are worn 24/7 to monitor habitual PA (11). These usually incorporate accelerometers that sam- ple at various frequencies and/or GNSS. With or without additional user input to improve the accuracy of identifying training events, the external training characteristics objec- tively recorded by the majority of these devices seem only to replicate those captured in a training log, for example, dis- tance and duration. Restricting objective quantification of training characteristics to the replication of existing metrics may limit insight into the possible effects of accelerometer- derived metrics of external training load on performance and injury outcomes in runners. Accelerometer-derived measures of load, if available, tend to input accelerations to ‘‘black-box’’ on-board processors and produce manufacturer-specific, proprietary metrics that ap- pear difficult to interpret (5). For example, in team sports (12–15), there has been some development with the use of vest-/back-mounted triaxial accelerometers to provide a proprietary measure termed PlayerLoadi (modified vector magnitude in arbitrary units representing rates of change in instantaneous acceleration (16)). Proprietary metrics limit comparisons with data recorded by other devices. The wear location and limited battery life in these devices also limits their ability to monitor other important training or nontraining activity outside of training sessions. These aspects, alongside other practical issues related to access to longitudinal data, limit the use of accelerometry in running-related research that seeks to develop new measures of external training load that might help reduce RRI and improve performance. The ability to objectively, unobtrusively, and accurately quantify external training load without user input using high- resolution, triaxial, open source (nonproprietary) acceleration data from a single wearable device over weeks at a time, is therefore attractive. Wrist-worn accelerometers are now widely used in very large research cohorts to measure characteristics of habitual PA (17) including sleep without the need for a sleep diary (18). These research-grade monitors generate high- resolution raw data, which can be processed using open- source software, facilitating the development of metrics most appropriate for a specific research question. For ex- ample, outputs from these monitors have been validated with ground reaction force data (19,20) enabling metrics indicative of external mechanical loading relative to bone health to be established (21). A similar approach could therefore be developed to provide a field-based proxy mea- sure of external mechanical load (biomechanical risk factor) relevant to injury. Example metrics in PA and health re- search include ‘‘Average Acceleration,’’ ‘‘Most Active- 30mins’’ or ‘‘MinsQ400 mg’’ (Table 1) which describe the intensity of activity in different user-defined periods or time spent above user-defined intensities of activity (e.g. 400 mg is a validated vigorous activity threshold in adults (23)). Al- though these wrist-worn triaxial accelerometer-derived metrics are validated for use in large-scale population PA research, it is not yet known whether they can be used to ac- curately and unobtrusively measure external training load in runners in the field. The application of these sample metrics provides a justifiable starting point for objectively classifying and quantifying an alternative measure of external training load in runners. Further experimentation with the creation of a composite metric of workload (WL400–4000 mg; Table 1) from intensity multiplied by duration (25), may also provide a possible accelerometer-derived alternative to Foster_s (24) composite measure of training load (RPE  duration). Embedding a procedure for classifying running and nonrunning training days from accelerometer data and accurately obtaining accelerometer-derived metrics of external training load within existing, validated protocols for accurately monitoring habitual activity (26,27) including those used to derive accurate mea- sures of sleep (18), would benefit subsequent analysis of pat- terns of training relative to rest and recovery (6,7). The benefits of high wear compliance and increased measurement reliability associated with the use of wrist-worn monitors (22) would also support this future analysis. The aims of this study are to assess whether simple PA metrics derived from the application of open-source analysis code to repeated week-long raw habitual PA data from wrist- worn tri-axial accelerometers in runners can be used to 1) ob- jectively, unobtrusively and accurately discriminate between running training days and nonrunning days; and 2) quantify external training load on running training days. It was hypoth- esized that the Most Active 30mins metric (Table 1) would be the best discriminator for classifying running and nonrunning days as it focuses on a single continuous period of activity rather than an average derived from the entire day. It was also hypothesized that MinsQ400 mg and WL400–4000 mg would demonstrate at least a moderate level of correspondence (variance explained) with existing self-reported measures of training load (criterion measures) from a training log. METHODS Participants Forty-one runners (22 women) with 92 yr running expe- rience who were training for an event (e.g. 10 km, half/full marathon) were recruited. An early attrition rate (14.6%) due http://www.acsm-msse.org 2278 Official Journal of the American College of Sports Medicine EPIDEMIOLOGY to injury or withdrawal resulted in 35 runners (19 women [54%]; Table 2) monitoring their training load for at least nine consecutive weeks (mean 12.6 T 2.3 wk) between De- cember 2015 and June 2016, to obtain a range of training intensities before their target event. Variations in average weekly running mileage, duration and pace (minutes per mile) averaged over the monitoring period indicate heterogeneous characteristics of training in this sample of runners (Table 2). The Sport and Health Sciences Ethics Committee at the University of Exeter approved this study, and all participants provided written informed consent. Self-monitoring of Training Load For the duration of the monitoring period, after each ac- tivity, runners were required to record the following data as soon as possible in a training log: date of session; start/end times (session duration calculated); activity/training type (e.g., road/off-road/track/treadmill run or other (e.g., gym, swim- ming, cycling, circuits or yoga)); running miles covered; an overall session RPE after consulting a visual scale, regardless of session type. Training logs were returned every 2 wk (mail) for manual input into a database. A composite measure of running training load (session RPE  session duration) in arbitrary units was subsequently calculated (7,24). Using ac- tivity/training type data, each day was classified as either a ‘‘running’’ (all surface types) or ‘‘nonrunning’’ day, with the latter further classified as either ‘‘other training’’ (e.g. gym, swimming, cycling, circuits or yoga) or a ‘‘rest’’ day. Where running training occurred twice on 1 d, running miles and duration were summed, a mean running RPE was calculated and training load was recalculated. If different types of train- ing including running occurred on the same day, that day was labeled as a ‘‘running’’ day. Average self-reported weekly training load characteristics for the sample of runners are presented in Table 2. Accelerometer Monitoring of Training Load Runners were issued with a GENEActiv accelerometer (100 Hz, triaxial, T8g; Activinsights Ltd, Kimbolton UK) every other week to wear on their nondominant wrist to collect 7 d of data. Monitoring on alternate weeks allowed monitors to be refreshed and reduced participant burden to TABLE 2. Summary characteristics for runners including self-reported weekly training volume metrics. Average Range (Min–Max) Age (yr) 41.9 (11.4) 23–63 Height (m) 1.72 (0.08) 1.60–1.86 Mass (kg) 68.5 (9.7) 54.1–93.8 BMI (kgImj2) 23.2 (2.2) 18.8–27.7 aMiles per week (miles) 22.1 (12.4–34.1) 1–149.5 aTotal duration per week (min) 208.0 (124–323) 10–1048 aMinutes per mile (min) 9.3 (8.2–10.6) 6.2–41.6 aRunning sessions per week 3 (1–4) 0–13 All values are means (standard deviations) unless indicated otherwise. aMedian (inter quartile range). Ranges for training volume metrics represent minimum- maximum volumes from individual weeks. TABLE 1. Acceleration metrics considered* for discriminating between running and nonrunning days and used# to quantify external training load on running training days. Acceleration Metric Description Rationale Average Acceleration* Average daily dynamic acceleration in mg This metric (also known as ENMO) has previously been used to quantify levels of habitual PA (17,22). The inclusion of running activity within a day was assumed to lead to higher average acceleration for that day. This metric was not used to quantify external training load on running training days as average acceleration reflects the whole day and not just higher intensity accelerations reflective of training. Most Active-30mins* Average acceleration in mg for the most active continuous 30-min period of the day This metric identifies the single most active 30-min period of activity within a day and not the entire day. By looking at a single continuous period of activity, it has the potential to discriminate well between a day including a running training session (regardless of length of run) compared to the most active 30-min period on other training and rest (nonrunning) days. This metric was not used to quantify training load on running days as it only corresponds to 30 min of the day. MinsQ400 mg*# Time in minutes accumulated throughout the day at or greater than an intensity of 400 mg An intensity of 400 mg is a validated threshold of activity used to estimate time spent at vigorous (six times the rate of energy expenditure at rest; 6 METS [metabolic equivalents]) levels of habitual PA in population research (23). All high intensity activity is summed, wherever it occurs within a day, which means this metric may be useful for both discrimination of days and quantification of external training load. aWL400–4000 mg# Time in minutes accumulated in 50 mg bins Q400 mg was multiplied by the average intensity of the bin (e.g. 425 mg was the average intensity of the 400–450 mg bin) to create individual workload (WL) bins in mg-minutes (mgmins). WL bins between 400 and 4000 mg were summed to create a total WL metric in mgmins. Similar to methods presented by Foster (24) for calculating session rating of perceived effort as a composite measure of training load (RPE  duration), this acceleration metric multiplies intensity by duration to give value to short but potentially meaningful amounts of high-intensity activity that may be particularly relevant in new models relating accelerometer-derived metrics of external training load with RRI and performance outcomes. The lower and upper accelerations (400 and 4000 mg) border the zone where accelerations typical of running fall (23). This metric was considered for quantifying external training load only. aAs recently proposed by Hillsdon (25). OBJECTIVELY QUANTIFYING TRAINING LOAD Medicine & Science in Sports & Exercised 2279 EPIDEMIOLOGY help maximize wear compliance during test weeks. Partici- pants were requested to wear the monitor 24 hId-1. As mini- mal differences exist in accelerometer output between monitors worn on dominant and nondominant wrists during higher intensity activity (28), runners were permitted to swap the wear location of the GENEActiv to their dominant wrist for the duration of a run if the wear location clashed with the preferred placement of another personal wearable device. Raw acceleration files were extracted and processed through an open-source package (GGIR Version 1.2–8, (26)) in R (http://cran.r-project.org) for autocalibration and calculation of the dynamic acceleration in milligravitational units (mg) averaged over 5-s epochs (the resultant vector magnitude corrected for gravity, ENMO, as described pre- viously (27)). A total of 1532 d were obtained from which 1494 (97.5%) accelerometer days with at least 10 h of wear per waking day (29) were analyzed. Time accumulated in bins spanning 50-mg intervals between 50 and 4000 mg (50–99.99 mg; 100–149.99 mg; 150–199.99 mg etc) were obtained with activity G50 mg considered non–meaningful (21,30), and the incidence of time accumulated 94000 mg extremely brief and rare. Accelerometer Metrics and Statistical Analysis Discriminating between ‘‘running’’ and ‘‘non- running’’ days (aim 1). Accelerometer data were time- matched with training log data for each calendar day in STATA (version 15). Average Acceleration, Most Active 30-mins, and MinsQ400 mg, which are typical metrics used to describe characteristics of habitual PA, were considered candidates for discriminating between ‘‘running,’’ ‘‘other training’’ and ‘‘rest’’ days (Table 1). Receiver operating characteristic (ROC) analyses were carried out for these metrics to derive the optimum thresholds for discrimination between running and nonrunning training days. Perfor- mances were summarized by calculating the area under the ROC curves (AUROC). Similar to the methods by Evenson et al (31), thresholds were selected that optimized the bal- ance between sensitivity (running classified as ‘‘running’’) and specificity (nonrunning classified as ‘‘nonrunning’’). Optimal thresholds were applied to the data and the per- centage of days correctly classified as ‘‘running’’ and ‘‘nonrunning’’ calculated. The percentage of days correctly classified as ‘‘nonrunning’’ was further broken down according to whether the day was an ‘‘other training’’ day or a ‘‘rest’’ day. The percentage of misclassification for each type of ‘‘other training’’ misclassified as ‘‘running’’ was also identified. To detect a medium effect size with power of 80% and alpha of 0.05 (AUROC of 0.6 as significantly different from an AUROC of 0.5, no association), a total sample of at least 258 days (sample ratio of 1:1 with 129 positive days and 129 negative days) was required. The generalizability and performance of the ROC models on unseen data was assessed using leave-one-out-cross-validations (LOOCV) (32). Estimation of external training load on accelerometer- classified ‘‘running’’ days (aim 2). From training log data, miles, duration, and training load, which are frequently monitored to understand the influence of training load on performance, injury, and illness (1,3,6,7,24,33), were used to represent external and composite criterion measures of train- ing load (criterion measures). On running training days that were classified using cut points from accelerometer metrics which demonstrated the highest levels of accuracy for cor- rectly classifying running days (see aim 1), accelerometer- derived metrics of training load (MinsQ400 mg and WL400–4000 mg; Table 1) were examined to see how closely they corresponded to criterion measures. On each set of classified days, variances explained in training log criterion measures (miles, duration, and training load) by MinsQ400 mg and WL400–4000 mg were examined using linear regression analysis. The generalizability and performance of the model on unseen data was assessed using LOOCV. Statistical anal- yses were carried out in STATA (version 15) with an alpha level set at 0.05. RESULTS Discriminating between running and nonrunning days (aim 1). From 35 participants, a total of 1494 d with 910-h wear were analyzed, of which 694 were ‘‘running’’ days, 641 were ‘‘rest’’ days, and 159 were ‘‘other training’’ days. Each participant contributed 18 to 56 d (mean [SD] = 42.7 [8.8]). Of these, 2 to 42 (19.8 [10]) were ‘‘running’’ days, 0 to 37 (18.3 [9.0]) were ‘‘rest’’ days, and 0 to 23 (4.5 [6.0]) were ‘‘other training’’ days. Cutpoints for identifying running days from habitual PA using respective accelerometer metrics with area under curve (AUC) significant at P G 0.05 are presented in Table 3. Discrimination between ‘‘running’’ and ‘‘rest’’ days was excellent (88%–94% agreement; Table 3). Both the Most Active-30mins and MinsQ400 mg had 994% accuracy for classifying running as ‘‘running’’ and nonrunning as ‘‘nonrunning’’ and were subsequently used to separately classify running days for aim 2. ‘‘Average Acceleration’’ performed similarly for correctly classifying ‘‘nonrunning’’ days (93%), but was weaker at correctly classifying ‘‘run- ning’’ days (88%). Irrespective of the metric, the greatest inaccuracy was from misclassifying ‘‘other training’’ days as ‘‘running’’ days, ranging from 14% misclassification for Most Active-30mins to 33% misclassification for Average Acceler- ation. The LOOCV procedure indicated robustness and sta- bility as the high performance was maintained (AUC Q0.93). The rate of misclassification of other training activities as running is shown in Table 4. The most frequent other training activities undertaken were cycling (47 occurrences) and gym/exercise classes (45 occurrences). The most likely activities to be misclassified by the Average Acceleration metric were field/racket sport (95%), circuit training (57%), and then cycling (53%). A similar pattern was found when using MinsQ400 mg, except circuit training was not http://www.acsm-msse.org 2280 Official Journal of the American College of Sports Medicine EPIDEMIOLOGY misclassified. The Most Active-30mins metric performed better for field or racket sports (25% misclassification) and generally across the board, but still misclassified nearly a third of cycling occurrences. Estimation of external training load on accelerometer- classified running days (aim 2). On running days classi- fied using MinsQ400 mg or using Most Active-30mins (accelerometer metrics most successful at classifying run- ning days from aim 1), the accelerometer-derived training load metric MinsQ400 mg explained approximately 75% to 76% and 74% of the variance in miles and duration, respectively, and 67% and 71% of the variance in training load (Table 5). The variance explained by WL400–4000 mg in miles, dura- tion, and training load was slightly lower at 63% to 69% on running days classified using either MinsQ400 mg or Most Active-30mins, except for training load when running days were classified using MinsQ400 mg, which was much lower (55%). The LOOCV procedure indicated robustness and stability as the high performance was maintained in all cases. DISCUSSION Raw acceleration data from wrist-worn accelerometers widely used in research can be used to objectively, unob- trusively, and accurately identify running training days and quantify external training load in runners. Importantly, the accelerometer metrics used are embedded within existing, validated open-source software for processing and analyzing accelerometer data for accurate quantification of habitual PA (26,27). As a field-based proxy measure of external me- chanical load (19,20), use of these accelerometer-derived metrics will enhance future research that seeks to further understand the influence of objectively measured modifiable patterns of external training load relative to rest and sleep on RRI and performance outcomes (1–8). Discriminating between running and nonrunning days. The high degree of accuracy for correctly classifying running days and days with no training indicates that wrist- worn accelerometer metrics can be used to objectively and unobtrusively discriminate between running and nonrunning days. While each accelerometer metric was able to discrimi- nate between these days, the mean acceleration recorded during ‘‘Most Active-30mins’’ was the best discriminator. As running is characterized by high accelerations (23), which incorporate an impact peak (19,20), high accelerations for the most active continuous 30 min of the day likely reflect the deliberate inclusion of a running session. The length of the training session may not match 30 min, but the elevation of the acceleration alone is sufficient to simply differentiate between running and nonrunning days. In contrast, metrics that sum time spent at high accelerations across the day (MinsQ400 mg), or the average accelerations across the day, can be elevated due to short activity bursts spread across the day which may or may not be part of a training session. Even for the Most Active-30mins, a degree of mis- classification in ‘‘field or racket sports’’ and ‘‘circuits’’ is likely due to these activities, including aspects of running or lunging and jumping, which could elevate average acceleration to exceed magnitudes typically found during running (19,20). For cycling, road or track vibration also has the potential to elevate this average acceleration to similar levels found when running. Post hoc analysis of demographic and training data indicated that very short runs may be a potential source of misclassification. However, we are also cognizant that vali- dation of accelerometer data in the field is complicated by the use of potentially inaccurate self-reported training log or TABLE 3. Optimum accelerometer cutpoints for differentiation between running and nonrunning days (includes rest days [no training] and other-training days [days with a different type of training]). Accelerometer Metrics Average Acceleration MinsQ400 mg Most Active-30mins Cutpoint 40.9 mg 22.4 min 525.3 mg AUC (95% CI) a0.93 (0.92–0.95) a0.95 (0.94–0.96) a0.97 (0.96–0.98) Agreement (%) 88 92 94 ‘‘Running’’ correctly classified as ‘‘running’’ (%) 88 94 94 ‘‘Rest’’ correctly classified as ‘‘nonrunning’’ (%) 93 95 95 ‘‘Other training’’ correctly classified as ‘‘nonrunning’’ (%) 67 72 86 LOOCV AUC (95% CI) 0.93 (0.92–0.95) 0.95 (0.94–0.96) 0.97(0.96–0.98) aSignificantly different (P G 0.05) to the null hypothesis of an AUC of 0.5. CI, confidence interval. TABLE 4. Percentage of ‘‘other training’’ activities misclassified as ‘‘running’’ when using ‘‘Average Acceleration,’’ ‘‘MinsQ400 mg’’, and ‘‘Most Active-30mins’’ to discriminate between ‘‘running’’ and ‘‘nonrunning’’ days. % Misclassified as ‘‘running’’ by Accelerometer Metrics Other Training Actual Number of Occurrences Average Acceleration MinsQ400 mg Most Active-30mins Field or racket sport 20 95.0 90.0 25.0 Circuit training 7 57.1 0 14.3 Cycling 47 53.2 53.2 31.9 Walk 6 16.7 0 0 Gym/exercise class 45 4.5 4.4 2.2 Swimming 32 0 0 0 Sailing 2 0 0 0 OBJECTIVELY QUANTIFYING TRAINING LOAD Medicine & Science in Sports & Exercised 2281 EPIDEMIOLOGY training program information (underreported and overreported training activity) (9,10). In using this simple metric to identify running training days in future studies, any accepted level of misclassification will depend on the nature of the activity misclassified relative to the research question. Estimation of external training load on running days. When running days were classified using Most Active-30mins, approximately 71% to 76% of the variance in Miles, Duration or Training Load was explained by the accelerometer-derived training load metric MinsQ400 mg, which was approximately 7%–13% more than the variance explained by the composite workload metric WL400–4000 mg on these days. When running days were classified using MinsQ400 mg, similarly high levels of variance in miles and duration were explained by the accelerometer-derived training load metrics MinsQ400 mg and WL400–4000 mg compared with when days were classified using Most Active-30mins but 4% and 9% less variance was explained in training load by respective metrics on these days. Despite differences, these accelerometer-derived metrics correspond highly with crite- rion measures, especially miles and duration, which suggests a high degree of convergent validity with existing training log methods for quantifying external training load. MinsQ400 mg in particular appears to be a good measure of external training load and can be easily obtained from longitudinal monitoring of habitual PA. For comparison, fast walking at 5 kmIhj1 in adults yields approximately 170 T 56 mg from a wrist-worn accelerometer, whereas running at 8 kmIhj1 yields approximately 760 T 200 mg (23). A threshold of 400 mg, which is also validated to quantify vigorous activity equivalent to 6 METs (23) (Table 1), therefore, provides sufficient margins to avoid capturing lower-intensity walking- type activity while capturing lower accelerations introduced by large variability when running at 8 kmIhj1 and lower ac- celerations from slower speed running. For comparison, ad- ditional analysis of accelerometer-derived metrics of external training load on days classified using training log informa- tion, indicated that MinsQ400 mg (75%, 75%, and 72%, re- spectively) and WL400–4000 mg (65%, 60%, and 62%, respectively) explained similar variation in miles, duration and training load when days were classified using either MinsQ400 mg or Most Active-30mins. An ability to use a single simple accelerometer-derived metric (e.g., MinsQ400 mg) to accurately classify running days and provide a valid measure of external training load, lays the foundation for overcoming challenges, such as ease of use and data interpretation described by Bourdon and colleagues (5) for accelerometry to be used in training program prescription. Further, it would be possible to use the regression analyses to predict outcomes familiar to runners (e.g., miles), but for analytical purposes, we believe it preferable to use the directly measured metrics. Accelerometer metrics are also highly correlated with laboratory-based measures of ground reaction force (19,20), which suggests that accelerometer-derived metrics of exter- nal training load may add more value to models of RRI and performance than existing training log-based measures. Further research to determine whether MinsQ400 mg and/or WL400–4000 mg translate into meaningful measures of external training load in relation to injury and performance would be beneficial. Implications of this study. The ability to obtain ac- curate, objective training records without the need for user input removes the reliance on the creation of a subjective training log, reduces participant burden, avoids bias, and other reporting inaccuracies associated with logging or marking data on paper or a device, (9,10) and facilitates the accurate monitoring of runners_ training behavior in future prospective studies. It also removes the need to match training log data, sometimes with multiple entries, with accelerometer data across days. A high monitor wear compliance (90% of days 916 h; 76% of days 922 h) in this population also supports its inclusion in the future analysis of patterns of training relative to rest and sleep. In contrast to GNSS devices, which are reliant on tracking a physical change in position in an outdoor envi- ronment (5), accelerometers also have the advantage of being able to be used anywhere, even to monitor external load when running on the spot. Developing accelerometer-derived mea- sures of external training load provides a natural extension of an accelerometer_s existing ability to accurately measure all aspects of habitual PA including rest and sleep longitudinally. Further developments. Streamlining methods for collecting, generating, and visualizing simple accelerometer- derived training load metrics (5) could facilitate their inclu- sion in commercial activity trackers and healthcare monitors for training program monitoring, prescription, and injury prevention purposes. It would also be beneficial to examine time spent at higher intensities of acceleration (e.g. 91000 mg approximating 10 kmIhj1; 28) to separately analyze higher and lower-intensity running. In an effort to avoid bias from self-reported measures of miles and duration (10), further comparison of accelerometer-derived metrics with objectively measured criterion measures (e.g., GPS) might be beneficial, TABLE 5. Percentage of the variance explained in miles, duration and training load when using ‘‘MinsQ400 mg’’ and ‘‘WL400–4000 mg’’ to quantify external training load on ‘‘running’’ days classified using MinsQ400 mg and Most Active-30mins. Miles Duration Training Load R 2 LOOCV R 2 R 2 LOOCV R 2 R 2 LOOCV R 2 Training load on ‘‘running’’ days classified using ‘‘Mins Q 400 mg’’ MinsQ400 mg 74.8 74.6 73.7 73.4 66.9 70.0 WL400–4000 mg 69.2 68.9 63.5 63.2 54.8 64.2 Training load on ‘‘running’’ days classified using ‘‘Most Active-30mins’’ MinsQ400 mg 76.2 76.0 74.6 74.4 70.8 70.4 WL400–4000 mg 69.0 68.7 62.9 62.5 64.0 63.5 R2, coefficient of determination from linear regression; LOOCV R2, coefficient of determination for LOOCV. http://www.acsm-msse.org 2282 Official Journal of the American College of Sports Medicine EPIDEMIOLOGY however it would be important to avoid over-burdening the runner with a requirement to wear multiple monitoring de- vices. To improve the classification of running from other training activities, alternative analysis methods such as those used in the sedentary sphere (34,35), which consider the orientation of the monitor due to wrist position to estimate upright, sitting, and lying down postures, could also be ex- plored to distinguish running from other activities, such as cycling. Analysis of the frequency compositions of the raw acceleration signal from different activities may also allow the selection of suitable filters to improve classification perfor- mance or enable metrics to be developed that allow cycling and racket sports to be identified explicitly. Although the high performance from the LOOCV carried out in this study demonstrate the robustness of using accelerometer metrics to classify running days, further validation of the method in an independent sample would also be beneficial. STRENGTHS AND LIMITATIONS We used longitudinal methods for objectively monitoring habitual PA in runners every other week and obtained written self-reported training logs every week over at least a 9-wk period. The incorporation of runners leading up to the events of differing lengths ensured a wide range of training patterns for testing. The results were robust across this range. During this period, runners remained motivated to complete training logs and were familiar with wearable devices yielding a large number of matched training log and accelerometer days with the added benefit of high accelerometer wear com- pliance. A rich bank of data was therefore obtained, allowing robust statistical methods to be used with cross-validations to address each research question. However, the nature of the sample does limit the generalizability of the results. All par- ticipants were self-identified runners who were training for an event. Most did undertake some form of cross-training, but the degree of engagement in other activities may be greater in people who do not identify as runners, or runners when they are not leading up to an event. Further research should in- vestigate the degree of misclassification of ‘‘other training’’ as ‘‘running’’ in other populations. CONCLUSIONS Wrist-worn accelerometer metrics can be used to objec- tively, unobtrusively, and accurately identify running train- ing days in runners, reducing the need for training logs or user input in future prospective research or commercial ac- tivity tracking. A high percentage of the variance explained in existing metrics by new, simple, accelerometer-derived metrics of external training load supports the development and future use of accelerometry for prospective, preventa- tive, and prescriptive monitoring purposes in runners. This project was supported by Medical Research Council Prox- imity to Discover funding (Reference: MC_PC_14127) in collabora- tion with Activinsights Ltd, UK. The authors would like to thank the runners who volunteered their time and committed to having their training load monitored over multiple weeks. A. R. is with the National Institute for Health Research (NIHR) Bio- medical Research Centre based at University Hospitals of Leicester and Loughborough University, the National Institute for Health Re- search Collaboration for Leadership in Applied Health Research and Care – East Midlands (NIHR CLAHRC-EM) and the Leicester Clinical Trials Unit. The views expressed are those of the authors and not necessarily those of the NHS, the NIHR or the Department of Health. Conflicts of Interest: As a collaborative study with industry supported by MRC Proximity to Discover funding, the industry partner may potentially benefit from the outcomes from the research. However, the open-source analysis procedures employed in the current study impose no restriction for other members of the activity monitoring in- dustry to also benefit. There are no other competing interests. The results of the study are presented clearly, honestly, and without fabri- cation, falsification, or inappropriate data manipulation and do not constitute endorsement by ACSM. REFERENCES 1. van Gent RN, Siem D, van Middelkoop M, et al. Incidence and determinants of lower extremity running injuries in long distance runners: a systematic review. Br J Sports Med. 2007;41(8):469–80. 2. Nielsen RO, Buist I, SLrensen H, Lind M, Rasmussen S. Training errors and running related injuries: A systematic review. Int J Sports Phys Ther. 2012;7(1):58–75. 3. van der Worp MP, ten Haaf DS, van Cingel R, de Wijer A, Nijhuis-van der Sanden MW, Staal JB. Injuries in runners; a sys- tematic review on risk factors and sex differences. PLoS One. 2015;10(2):e0114937. 4. Hreljac A. Etiology, prevention and early intervention of overuse injuries in runners: a biomechanical perspective. Phys Med Rehabil Clin N Am. 2005;16(3):651–67. 5. Bourdon PC, Cardinale M, Murray A, et al. Monitoring athlete training loads: consensus statement. Int J Sports Physiol Perform. 2017;12(2 Suppl):S2161–70. 6. Schwellnus M, Soligard T, Alonso J, et al. How much is too much? (Part 2) International Olympic Committee consensus statement on load in sport and risk of illness. Br J Sports Med. 2016;50(17): 1043–52. 7. Soligard T, Schwellnus M, Alonso J, et al. How much is too much? (Part 1) International Olympic Committee consensus statement on load in sport and risk of injury. Br J Sports Med. 2016;50(17):1030–41. 8. Willy RW. Innovations and pitfalls in the use of wearable devices in the prevention and rehabilitation of running related injuries. Phys Ther Sport. 2018;29:26–33. 9. Moran DS, Evans R, Arbel Y, et al. Physical and psychological stressors linked with stress fractures in recruit training. Scand J Med Sci Sports. 2013;23(4):443–50. 10. Nielsen RO, Nohr EA, Rasmussen S, SLrensen H. Classifying running-related injuries based upon etiology, with emphasis on volume and pace. Int J Sports Phys Ther. 2013;8(2):172–9. 11. Geib RW, Swink PJ, Vorel AJ, Shepard CS, Gurovich AN, Waite GN. The bioengineering of changing lifestyle and wearable technology: a mini review. Biomed Sci Instrum. 2015;51:69–76. 12. Casamichana D, Castellano J, Calleja-Gonzalez J, San Roma´n J, Castagna C. Relationship between indicators of training load in soccer players. J Strength Cond Res. 2013;27(2):369–74. 13. Scott BR, Lockie RG, Knight TJ, Clark AC, Janse de Jonge XA. A comparison of methods to quantify the in-season training load of OBJECTIVELY QUANTIFYING TRAINING LOAD Medicine & Science in Sports & Exercised 2283 EPIDEMIOLOGY professional soccer players. Int J Sports Physiol Perform. 2013; 8(2):195–202. 14. Scanlan AT, Wen N, Tucker PS, Dalbo VJ. The relationships be- tween internal and external training load models during basketball training. J Strength Cond Res. 2014;28(9):2397–405. 15. Murray NB, Gabbett TJ, Townshend AD, Blanch P. Calculating acute: chronic workload ratios using exponentially weighted moving averages provides a more sensitive indicator of injury likelihood than rolling averages. Br J Sports Med. 2017;51(9):749–54. 16. Boyd LJ, Ball K, Aughey RJ. The reliability of MinimaxX accel- erometers for measuring physical activity in Australian Football. Int J Sports Physiol Perform. 2011;6(3):311–21. 17. Doherty A, Jackson D, Hammerla N, et al. Large scale population assessment of physical activity using wrist worn accelerometers: the UK Biobank Study. PLoS One. 2017;12(2):e0169649. 18. Van Hees VT, Sabia S, Jones SE, Wood AR, Anderson KN, Kivimaki M, Frayling TM, Pack AI, Bucan M, Mazzotti DR, Gehrman PR, Singh-Manoux A, Weedon MN. Estimating sleep parameters using an accelerometer without sleep diary. bioRxiv [Internet]. 2018. [cited 2018 February 1]. Available from: https://www. biorxiv.org/content/early/2018/02/01/257972. doi: 10.1101/257972. 19. Rowlands AV, Stiles VH. Accelerometer counts and raw acceler- ation output in relation to mechanical loading. J Biomech. 2012; 45(3):448–54. 20. Stiles VH, Griew PJ, Rowlands AV. Use of accelerometry to classify activity beneficial to bone in premenopausal women. Med Sci Sports Exerc. 2013;45(12):2353–61. 21. Stiles VH, Metcalf BS, Knapp KM, Rowlands AV. A small amount of precisely measured high-intensity habitual physical ac- tivity predicts bone health in pre- and post-menopausal women in UK Biobank. Int J Epidemiol. 2017;46(6):1847–56. 22. van Hees VT, RenstrO¨ m F, Wright A, et al. Estimation of daily energy expenditure in pregnant and non-pregnant women using a wrist-worn tri-axial accelerometer. PLoS One. 2011;6(7):e22922. 23. Hildebrand M, van Hees VT, Hansen BH, Ekelund U. Age group comparability of raw accelerometer output from wrist- and hip- worn monitors. Med Sci Sports Exerc. 2014;46(9):1816–24. 24. Foster C. Monitoring training in athletes with reference to overtraining syndrome. Med Sci Sports Exerc. 1998;30(7):1164–8. 25. van der Ploeg H, Hillsdon M. Is sedentary behaviour just physical inactivity by another name? Int J Behav Nutr Phys Act. 2017;14(1):142. 26. van Hees VT, Fang Z, Langford J, et al. Auto-calibration of ac- celerometer data for free-living physical activity assessment using local gravity and temperature: an evaluation on four conti- nents. J Appl Physiol. 2014;117(7):738–44. 27. van Hees VT, Gorzelniak L, Dean Leo´n EC, et al. Separating movement and gravity components in an acceleration signal and im- plications for the assessment of human daily physical activity. PLoS One. 2013;8(4):e61691. 28. Esliger DW, Rowlands AV, Hurst TL, Catt M, Murray P, Eston RG. Validation of the GENEA accelerometer. Med Sci Sports Exerc. 2011;43(6):1085–93. 29. Sieva¨nen H, Kujala UM. Accelerometry—simple but challenging. Scand J Med Sci Sports. 2017;27(6):574–8. 30. Bakrania K, Yates T, Rowlands AV, et al. Intensity thresholds on raw acceleration data: Euclidean Norm Minus One (ENMO) and Mean Amplitude Deviation (MAD) approaches. PLoS One. 2016;11(10):e0164045. 31. Evenson KR, Catellier DJ, Gill K, Ondrak KS, McMurray RG. Calibration of two objective measures of physical activity for children. J Sports Sci. 2008;26(14):1557–65. 32. Efron B, Tibshirani RJ. An Introduction to the Bootstrap. 1st ed (Hardback). Dordrecht (Netherlands): Chapman & Hall Inc, Springer Science+Business Media; 1994. pp. 240–1. 33. Gordon D, Wightman S, Basevitch I, et al. Physiological and training characteristics of recreational marathon runners. Open Access J Sports Med. 2017;8:231–41. 34. Rowlands AV, Olds TS, Hillsdon M, et al. Assessing sedentary behavior with the GENEActiv: introducing the sedentary sphere. Med Sci Sports Exerc. 2014;46(6):1235–47. 35. Rowlands AV, Yates T, Olds TS, Davies M, Khunti K, Edwardson CL. Sedentary sphere: wrist-worn accelerometer-brand independent pos- ture classification. Med Sci Sports Exerc. 2016;48(4):748–54. http://www.acsm-msse.org 2284 Official Journal of the American College of Sports Medicine EPIDEMIOLOGY

Wrist-worn Accelerometry for Runners: Objective Quantification of Training Load.

[]

Stiles, Victoria H,Pearce, Matthew,Moore, Isabel S,Langford, Joss,Rowlands, Alex V

eng

PMC3986049

Rapid Directional Change Degrades GPS Distance Measurement Validity during Intermittent Intensity Running Jonathan C. Rawstorn1, Ralph Maddison2, Ajmol Ali3, Andrew Foskett3, Nicholas Gant1* 1 Exercise Metabolism Laboratory, Department of Sport and Exercise Science, The University of Auckland, Auckland, New Zealand, 2 National Institute for Health Innovation, School of Population Health, The University of Auckland, Auckland, New Zealand, 3 School of Sport and Exercise, Massey University, Auckland, New Zealand Abstract Use of the Global Positioning System (GPS) for quantifying athletic performance is common in many team sports. The effect of running velocity on measurement validity is well established, but the influence of rapid directional change is not well understood in team sport applications. This effect was systematically evaluated using multidirectional and curvilinear adaptations of a validated soccer simulation protocol that maintained identical velocity profiles. Team sport athletes completed 90 min trials of the Loughborough Intermittent Shuttle-running Test movement pattern on curvilinear, and multidirectional shuttle running tracks while wearing a 5 Hz (with interpolated 15 Hz output) GPS device. Reference total distance (13 200 m) was systematically over- and underestimated during curvilinear (2.6160.80%) and shuttle (23.1762.46%) trials, respectively. Within-epoch measurement uncertainty dispersion was widest during the shuttle trial, particularly during the jog and run phases. Relative measurement reliability was excellent during both trials (Curvilinear r = 1.00, slope = 1.03, ICC = 1.00; Shuttle r = 0.99, slope = 0.97, ICC = 0.99). Absolute measurement reliability was superior during the curvilinear trial (Curvilinear SEM = 0 m, CV = 2.16%, LOA 6 223 m; Shuttle SEM = 119 m, CV = 2.44%, LOA 6 453 m). Rapid directional change degrades the accuracy and absolute reliability of GPS distance measurement, and caution is recommended when using GPS to quantify rapid multidirectional movement patterns. Citation: Rawstorn JC, Maddison R, Ali A, Foskett A, Gant N (2014) Rapid Directional Change Degrades GPS Distance Measurement Validity during Intermittent Intensity Running. PLoS ONE 9(4): e93693. doi:10.1371/journal.pone.0093693 Editor: Dylan Thompson, University of Bath, United Kingdom Received January 5, 2014; Accepted March 7, 2014; Published April 14, 2014 Copyright:  2014 Rawstorn et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Funding: This study was funded by The University of Auckland. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing Interests: The authors have declared that no competing interests exist. * E-mail: [email protected] Introduction Use of Global Positioning System (GPS) technology as a performance analysis tool is increasingly common in a number of team sports [1–15]. Several studies have been dedicated to assessing the validity of GPS devices for this purpose, although newer devices have also been used without prior validation [13– 15]. Despite considerable variation in experimental design previous investigations report small relative distance and velocity measurement uncertainties, and prevailing conclusions support the use of GPS devices during team sport activities [16–25]. While GPS devices appear acceptably valid for quantifying performance across entire bouts of exercise, sport scientists, coaches and governing bodies have shown particular interest in quantifying the high intensity activity demands of match-play [6,7,10,26,27]. The utility of GPS for monitoring performance is contingent on accurate and reliable measurement of these activities, which play a critical role in determining athletes’ physiological load, and competitive match outcomes [28–31]. Sprinting and rapid acceleration are consistently associated with increased GPS measurement uncertainty, particularly over short distances [16,19,22,23,32,33]. Thus it appears likely that GPS devices underestimate some movement patterns that are of critical importance during training and match-play. While the effect of running velocity on GPS measurement validity is well established little attention has been directed toward determining the effect of rapid directional change. As athletes may execute ,550–730 turning movements during match-play [28,34] it is important to determine whether rapid directional change affects GPS measurement validity. One investigation has com- pared GPS measurement validity during linear and non-linear running, and rapid directional change was associated with reduced measurement validity [18]. However, differing self-selected veloc- ity profiles between the linear and non-linear protocols make it difficult to determine whether this effect was caused by differing directional demands, velocity demands, or a combination of both. Thus there remains a need to systematically determine whether rapid directional change effects GPS distance measurement validity during exercise protocols with equivalent velocity profiles, but differing directional demands. The Loughborough Intermittent Shuttle-running Test (LIST) is a precisely controlled intermittent intensity shuttle-running proto- col designed to simulate the activity patterns of soccer [35]. The LIST is representative of the total distance, number of sprints and number of turns common to match-play, and induces similar physiological responses [35,36]. Furthermore, precise control over velocity throughout the LIST facilitates systematic evaluation of the effect rapid directional change exerts on GPS measurement PLOS ONE | www.plosone.org 1 April 2014 | Volume 9 | Issue 4 | e93693 validity by allowing modification of directional demands without altering the velocity profile. Many early investigations evaluated GPS devices featuring a 1 Hz sampling frequency, yet more recent devices with faster sampling frequencies have demonstrated improved measurement validity during linear and multidirectional running [19,22,33]. A novel device comprising a 5 Hz GPS microcontroller and an interpolation algorithm that outputs positional data at a 15 Hz frequency was recently utilised to investigate the movement demands of soccer, rugby union and rugby sevens [13–15]; however, the distance measurement validity of this device has yet to be evaluated. Therefore, the aim of this study was to systematically assess the effect of rapid directional change on the distance measurement validity of a previously untested GPS device. It was hypothesised that rapid directional change would reduce distance measurement validity. Materials and Methods Ethics statement This study received approval from the University of Auckland Human Participants Ethics Committee. All volunteers provided written informed consent prior to participation. Participants Six amateur club and provincial level team sport athletes (age = 24.161.6 y, body mass = 72.56610.33 kg, height = 1.7960.09 m, _VO2 max~54:46+4:19 ml:kg) volun- teered to participate in this study. Experimental protocol During a preliminary trial participants completed a multi-stage fitness test to estimate maximal oxygen consumption ( _VO2max), and determine running velocities for each phase of the LIST, as previously described [37]. Participants also completed a 15 min bout of the LIST while wearing the GPS device to familiarise themselves with the movement pattern, and device operation. Participants completed two experimental trials in random order, within 7 days. Participants completed <90 min of the LIST movement pattern on shuttle or curvilinear running tracks. The movement pattern comprised sequential walk (60 m, veloci- ty = 1.54 m?s21), sprint (15 m, maximal velocity; 5 m decelera- tion), run (60 m, velocity eliciting 95% _VO2max) and jog (60 m, velocity eliciting 55% _VO2max) phases, as previously described [35,38]. This cycle was repeated 11 times (<15 min) during each of six exercise blocks. Exercise blocks were separated by 3 min rest periods. Reference cycle, block and trial distances were 200 m, 2 200 m, and 13 200 m, respectively. Identical regulation of movement velocity during both trials, via standardised auditory commands, ensured movement demands differed only in the presence (shuttle) or absence (curvilinear) of rapid directional change. Moreover, velocity regulation controlled for potential disruptive effects of environment or other extraneous variables on participants’ performance and, therefore, on reference measures. Schematic and satellite representations of the shuttle and curvilinear protocols are displayed in Figure 1. Briefly, the shuttle protocol was completed on a marked 20 m shuttle-running track similar to that described by Nicholas et al. [35], and the curvilinear protocol was completed on a marked oval track on a level athletic playing surface. The curvilinear track was designed such that one lap represented one 200 m movement cycle, the 20 m sprint followed a linear path and turn radii (25.5 m) were optimised to minimise rapid directional change. Markings at 20 m intervals along the curvilinear track facilitated adherence to velocity regulation commands as per the shuttle protocol. A foam impact mat precluded excess displacement following sprint phases during both trials. The mat was temporarily withdrawn after impact during the curvilinear trial to prevent participant deviation from the marked track. Participants who chose not to use the impact mat to aid deceleration were instructed to proceed to the mat after deceleration to ensure the correct distance was covered. A researcher monitored adherence to marked running tracks in order to prevent deviation. The shuttle and curvilinear tracks were located away from large buildings to minimise multi-pathing error and ensure clear line of sight to orbiting satellites. Track lengths were measured with a calibrated surveyor’s wheel. GPS Device A non-differentially corrected GPS device (SPI Pro X, GPSports Systems, Australia) was worn in a harness between the scapulae, as per manufacturer’s instructions. The device comprises a 5 Hz GPS microcontroller and a proprietary interpolation algorithm that outputs positional data at a 15 Hz frequency. The 15 Hz interpolation is suggested to enhance measurement accuracy compared to the raw 5 Hz data; however, technical specifications regarding this algorithm were unavailable and it is not possible to determine its effect on distance measurement validity. GPS devices do not directly measure distance; however, researchers and practitioners in the team sport domain are frequently interested in assessing distance rather than positional coordinates. Therefore terms related to ‘distance measurement’ are used throughout this paper, in preference to ‘distance calculation’, as they hold intuitive relevance in this field. Raw 15 Hz GPS data were downloaded using the manufactur- er’s proprietary software (Team AMS v2.1), and exported for manual analysis. Data recorded outside the six exercise blocks were excluded from analyses. Distances were calculated within the proprietary software and no post processing was applied to raw GPS data. Precise velocity regulation allowed calculation of reference distances at the same frequency as the GPS data. The 15 Hz output frequency was expected to yield <80 000 observa- tions per trial. Consistent with the aim of this study, additional variables calculated by the proprietary software (e.g. speed, acceleration, impact, body load) were not analysed. Data Analysis Statistical analyses were performed using PASW (v18.0, SPSS Inc., USA). Measurement validity was considered to constitute accuracy and reliability, and a multifaceted statistical approach was implemented to assess these. Measurement accuracy was evaluated by calculating total, and within-epoch biases between GPS and reference distance measures. Two-tailed paired t-tests were performed to test the null hypothesis H0: distanceGPS = dis- tancereference for each protocol. These tests determined whether GPS measures differed systematically from reference distances; however, they cannot be used as the sole indicators of agreement [39]. Factorial analysis of variance was also performed to detect effects of TRIAL (i.e. Shuttle and Curvilinear) and MOVEMENT PHASE (i.e. Walk, Jog, Run and Sprint) on GPS distance measurement biases. That is, the null hypotheses H0: biascurvilinear = biasshuttle, and H0: biaswalk = biasjog = biasrun = biassprint were tested to determine whether GPS uncertainties were systematically affected by the presence of rapid directional change and/or differing movement velocities. Statistically significant interactions were explored using Bonferroni corrected paired comparisons. Previous data indicate 99 observations will provide 95% statistical power to detect a 5% difference between GPS and reference distance GPS Validity during Multidirectional Running PLOS ONE | www.plosone.org 2 April 2014 | Volume 9 | Issue 4 | e93693 measures, assuming a standard deviation (SD) of 13.6% (Cohen’s d = 0.37) [25]. Data were subjected to Levene’s test for equality of error variances. Data displayed heteroscedastic error variance (Shuttle p = 0.001; Curvilinear p,0.001) and were logarithmically transformed and reanalysed. To aid interpretation data are reported in the unit of measurement, or relative to reference measures. Relative measurement reliability was evaluated by calculating Pearson’s correlation coefficients (r), regression coefficients (slope) and two-way random-effects intraclass correlation coefficients (ICC) [39,40]. These statistics describe the reliability with which GPS distance increases as a function of reference distance, but are not indicative of agreement between measurement tools [41]. Absolute measurement reliability was evaluated by calculating the standard error of measurement SEM:SD| ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 1{ICC p   , coef- ficient of variation ðCV~100  (SD=x) Þ and 95% limits of agreement ðLOA~x+1:96  SD Þ [39,41]. These statistics indi- cate total measurement uncertainty (i.e. systematic+random error), and facilitate comparisons between experiments using different designs [42]. In-text data are reported as mean 6 SD. Statistical significance for all calculations was set at a,0.05. Results Satellite acquisition during curvilinear (8.1961.82 satellites) and shuttle (8.4461.57 satellites) trials was consistent with several previous studies evaluating GPS measurement validity [18,19,21,23,33]. Measurement accuracy Compared to the 13 200 m reference, distance was statistically significantly overestimated during the curvilinear trial (13 543.926105.45 m; T431 870 = 772.40, p,0.001) and underes- timated during the shuttle trial (12 780.686325.61 m; T460 226 = 2403.35, p,0.001). The between-trial bias was also statistically significant (F1, 892 090 = 386 116.42, p,0.001). Statistically significant effects of MOVEMENT PHASE on measure- ment bias were detected during the curvilinear (F3, 431 867 = 31.23, p,0.001) and shuttle trials (F3, 460 223 = 15.42, p,0.001), although these differences were small (Table 1). Within-epoch biases during each movement phase are depicted graphically in Figure 2. The dispersion of within-epoch biases was widest during the shuttle trial. This effect was particularly evident during the Jog and Run phases, when bias dispersion approximated that of the sprint phase. Measurement Reliability Table 2 summarises the relative and absolute reliability of GPS distance measures. GPS and reference distance were strongly correlated during curvilinear and shuttle trials (Table 2), indicating excellent relative measurement reliability. Moreover, regression coefficients approximated 1 during both trials (Table 2). Absolute reliability metrics (SEM, LOA and CV) were larger during the shuttle trial (Table 2) indicating superior absolute measurement reliability during the curvilinear protocol. Discussion This study utilised a movement pattern that is representative of key aspects of high level team sport match-play to systematically evaluate the effect of rapid directional change on the distance measurement validity of a previously untested GPS device. The main finding is that rapid directional change degrades GPS Figure 1. Schematic and satellite representations of shuttle and curvilinear running tracks. Satellite representations comprise typical positional data from one shuttle and curvilinear trial. IR = Infrared. doi:10.1371/journal.pone.0093693.g001 GPS Validity during Multidirectional Running PLOS ONE | www.plosone.org 3 April 2014 | Volume 9 | Issue 4 | e93693 distance measurement accuracy and absolute reliability, and this effect is independent of movement velocity. The systematic positive bias between GPS and reference distances during the curvilinear trial is consistent with previous investigations employing running protocols with similarly low directional demands [18,24]. Multidirectional running protocols are also associated with systematic distance measurement biases, but the magnitude and direction of uncertainty is inconsistent [16,18,25,33]. The negative distance measurement bias during the shuttle trial is consistent with uncertainties reported during simulated court sports, team sports and non-linear running [18,33]; however, positive biases have been reported during field hockey and team sport simulations [16,25]. Consistent with the present results, the only previous study to evaluate the effect of multidirectional demands on GPS distance measurement validity reported positive and negative biases during linear and multidirectional running, respectively, with the largest absolute bias during non-linear running [18]. However, the self- selected velocity profiles differed between trials and it is unclear whether measurement validity was affected by differing directional demands, velocity demands, or a combination of these. Identical velocity regulation during the curvilinear and shuttle trials in the present study controlled for any effect of movement velocity on distance measurement validity. Thus the present findings indicate rapid multidirectional movement patterns degrade GPS distance measurement validity, and that this effect is independent of movement velocity. Excellent relative measurement reliability during curvilinear and shuttle trials are consistent with previous investigations [16,25,32]. Nonetheless, larger absolute reliability metrics during the shuttle trial demonstrate reduced distance measurement validity during rapid multidirectional movement patterns. The magnitude of shuttle trial distance measurement uncer- tainty was 3.17%. Uncertainties of similar magnitudes have previously underpinned support for the use of GPS technology during team sport-related activities [16,18–25]. Indeed, when applied to a time-motion analysis of elite soccer match-play this uncertainty indicates match distance (10 627–12 027 m) may be underestimated by just 340–380 m, depending upon playing position [43]. In addition to the magnitude of uncertainty; however, it is also important to consider which movement patterns will be most affected by this inaccuracy. As rapid multidirectional movements are an important determinant of physiological load and match outcomes [28–31] the present results suggest GPS devices are likely to misrepresent some critical aspects of match- play. This has important implications for the way in which GPS technology is used in the team sport domain. As GPS measure- ment validity is also reduced during sprinting and rapid acceleration [16,19,22,23,32,33] it appears GPS may not be an appropriate tool for evaluating match-play activity profiles, or monitoring athletes’ physiological load. Given that iterative device development (i.e. newer hardware and software components) and faster sampling frequencies are proposed to improve GPS measurement validity [16,19,22,33], this is particularly pertinent when utilising devices featuring older componentry and/or slower sampling frequencies. Body lean angle is proposed to account for a substantial proportion of negative distance measurement bias during high speed nonlinear running [18,44] and the proximal anatomical position of the present device predisposes it to similar uncertainty. This should be an important consideration when attempting to evaluate the criterion validity of GPS devices; however, post-hoc correction for lean angle would contrast the aim of this study by reducing its ecological validity and misrepresenting the measure- ment accuracy likely to be realised during real-world use. The Figure 2. Within-epoch measurement uncertainty during shuttle and curvilinear trials. Solid reference lines = mean bias. Dashed reference lines = 95% limits of agreement. Differing sample sizes reflect discrepant mean sampling frequencies (Shuttle = 14.1560.20 Hz; Curvili- near = 13.2761.44 Hz), which were below the specified 15 Hz interpolation frequency. doi:10.1371/journal.pone.0093693.g002 Table 1. Distance measurement biases during shuttle and curvilinear adaptations of the Loughborough Intermittent Shuttle- running Test. Protocol Total (%) Walk (%) Jog (%) Run (%) Sprint (%) Shuttle 22.1663.84* 22.1864.23j r s 22.2063.42w s 22.1663.41w s 21.9263.63w j r Curvilinear 2.9962.96 2.9964.06j r s 2.9561.12w s 2.9561.33w s 3.1661.87w j r Table reports mean (6 s) within-epoch measurement biases relative to reference measures. *Statistically significantly different to the curvilinear trial (p,0.001). w j r sStatistically significantly different to the Walk, Jog, Run or Sprint movement phases (p,0.001). doi:10.1371/journal.pone.0093693.t001 GPS Validity during Multidirectional Running PLOS ONE | www.plosone.org 4 April 2014 | Volume 9 | Issue 4 | e93693 effect of device position on measurement validity should be addressed during product development, when it can be balanced against other design constraints such as comfort, player safety and device exposure to impact. The reduction in distance measurement validity during the shuttle trial may be explained by examining the interaction between movement demands and GPS position sampling. As intermittent GPS sampling partitions continuous movement paths into discrete linear segments, GPS devices are constrained to calculating cumulative device displacement across individual sampling epochs. While distance may be accurately quantified by displacement during linear movements, it will be underesti- mated when non-linear movements occur within a sampling epoch. Moreover, as rapid multidirectional movements will likely cause frequent separation between distance and displacement, the negative measurement uncertainty introduced by intermittent GPS sampling is likely to be largest during many of the most critical match-play activities. While uncertainty induced by intermittent position sampling will affect all GPS devices, this mechanism may also explain the superior measurement validity of higher frequency devices during multidirectional running [19,33]. Faster sampling frequencies increase the resolution with which continuous movement paths are partitioned into linear segments and, therefore, can be expected to reduce the magnitude of separation between device distance and displacement during non-linear movements. Nonetheless, sub- optimal distance measurement validity during the shuttle trial indicates the present device’s 5 Hz sampling frequency and/or 15 Hz interpolation algorithm are insufficient to accurately quantify distance during rapid directional change. Further research is required to determine the optimal sampling frequency for quantifying performance during multidirectional movement patterns, such as those common to many team sports. The proposed mechanism underlying distance measurement uncertainty cannot account for the positive distance bias recorded during curvilinear trials. This uncertainty is consistent with previous investigations [18,24] yet its source remains unclear. Participant deviation outside the marked curvilinear running track, perhaps due to intentional protocol non-adherence or the neuromuscular fatigue induced by the LIST [45], would manifest positive measurement bias. However, adherence to marked running tracks and the precisely controlled velocity profile were monitored throughout all trials to preclude these confounds. Device-specific software components are proposed to affect measurement validity [16], but proprietary protection of these components makes evaluation of their contribution to measure- ment uncertainty problematic. Indeed, the contribution of the present device’s 15 Hz interpolation algorithm to the observed measurement uncertainty remains unclear. Nonetheless, software components cannot account for the between-trial differences in measurement validity as device configuration remained identical throughout the study. It is important to note that, although the LIST movement pattern is representative of several key aspects of match-play [35] and offers advantages for validating GPS devices compared to previous methodologies, it remains difficult to accurately simulate the complexity of match-play within a controlled environment. This limitation is common to all similar investigations, and should be considered when attempting to generalise these findings to match-play. As multiple units of the present device were not simultaneously compared it remains unclear whether inter-unit variability will preclude interchangeable use of devices among multiple athletes, and caution should be taken when comparing data between devices. This study used the movement pattern of a precisely controlled exercise protocol that is representative of several key aspects of soccer match-play to systematically assess the effect of rapid multidirectional movements on GPS distance measurement validity. Rapid directional change degrades GPS distance mea- surement validity, and this effect is independent of movement velocity. Caution is recommended when relying on GPS to quantify team sport athletes’ performance as current device technology appears unable to accurately quantify movements that play a critical role in determining physiological load and competitive match outcomes. Author Contributions Conceived and designed the experiments: JR RM AA AF NG. Performed the experiments: JR NG. Analyzed the data: JR NG. Contributed reagents/materials/analysis tools: JR RM AA AF NG. Wrote the paper: JR RM AA AF NG. References 1. Coughlan GF, Green BS, Pook PT, Toolan E, O’Connor SP (2011) Physical game demands in elite rugby union: a global positioning system analysis and possible implications for rehabilitation. Journal of Orthopaedic and Sports Physical Therapy 41: 600–605. 2. Hiscock D, Dawson B, Heasman J, Peeling P (2012) Game movements and player performance in the Australian Football League. International Journal of Performance Analysis in Sport 12: 531–545. 3. Jennings DH, Cormack SJ, Coutts AJ, Aughey RJ (2012) International field hockey players perform more high-speed running than national-level counter- parts. Journal of Strength & Conditioning Research 26: 947–952. 4. Brewer C, Dawson B, Heasman J, Stewart G, Cormack S (2010) Movement pattern comparisons in elite (AFL) and sub-elite (WAFL) Australian football games using GPS. Journal of Science and Medicine in Sport 13: 618–623. 5. Coutts AJ, Quinn J, Hocking J, Castagna C, Rampinini E (2010) Match running performance in elite Australian Rules Football. J Sci Med Sport 13: 543–548. 6. Cunniffe B, Proctor W, Baker JS, Davies B (2009) An evaluation of the physiological demands of elite rugby union using Global Positioning System tracking software. J Strength Cond Res 23: 1195–1203. 7. Macutkiewicz D, Sunderland C (2011) The use of GPS to evaluate activity profiles of elite women hockey players during match-play. Journal of Sports Sciences 29: 967–973. 8. Gabbett TJ (2010) GPS analysis of elite women’s field hockey training and competition. J Strength Cond Res 24: 1321–1324. 9. Wisbey B, Montgomery PG, Pyne DB, Rattray B (2010) Quantifying movement demands of AFL football using GPS tracking. J Sci Med Sport 13: 531–536. 10. MacLeod H, Bussell C, Sunderland C (2007) Time-motion analysis of elite women’s field hockey, with particular reference to maximum intensity movement patterns. Int J Perf Analysis Sport 7: 1–12. 11. McLellan CP, Lovell DI, Gass GC (2011) Performance analysis of elite Rugby League match play using global positioning systems. Journal of Strength & Conditioning Research 25: 1703–1710. Table 2. GPS measurement reliability during shuttle and curvilinear adaptations of the Loughborough Intermittent Shuttle-running Test. Protocol r Slope ICC SEM (m) LOA (m) CV (%) Shuttle 0.99* 0.97 0.99* 119 6453 2.44 Curvilinear 1.00* 1.03 1.00* 0 6223 2.16 Table reports comparisons of GPS and reference distance measurement for Pearson’s product-moment correlation (r), linear regression coefficient (Slope) intraclass correlation coefficient (ICC), standard error of measurement (SEM), Bland Altman’s absolute limits of agreement (LOA), and the coefficient of variation (CV). *p,0.001. doi:10.1371/journal.pone.0093693.t002 GPS Validity during Multidirectional Running PLOS ONE | www.plosone.org 5 April 2014 | Volume 9 | Issue 4 | e93693 12. Waldron M, Twist C, Highton J, Worsfold P, Daniels M (2011) Movement and physiological match demands of elite rugby league using portable global positioning systems. Journal of Sports Sciences 29: 1223–1230. 13. Del Coso J, Munoz-Fernandez VE, Munoz G, Fernandez-Elias VE, Ortega JF, et al. (2012) Effects of a caffeine-containing energy drink on simulated soccer performance. PLoS ONE [Electronic Resource] 7: e31380. 14. Del Coso J, Portillo J, Mun˜oz G, Abia´n-Vice´n J, Gonzalez-Milla´n C, et al. (2013) Caffeine-containing energy drink improves sprint performance during an international rugby sevens competition. Amino Acids: 1–9. 15. Del Coso J, Ramı´rez JA, Mun˜oz G, Portillo J, Gonzalez-Milla´n C, et al. (2012) Caffeine-containing energy drink improves physical performance of elite rugby players during a simulated match. Applied Physiology, Nutrition, and Metabolism: 1–7. 16. Coutts AJ, Duffield R (2010) Validity and reliability of GPS devices for measuring movement demands of team sports. J Sci Med Sport 13: 133–135. 17. Waldron M, Worsfold P, Twist C, Lamb K (2011) Predicting 30 m timing gate speed from a 5 Hz Global Positioning System (GPS) device. International Journal of Performance Analysis in Sport 11: 575–582. 18. Gray AJ, Jenkins D, Andrews MH, Taaffe DR, Glover ML (2010) Validity and reliability of GPS for measuring distance travelled in field-based team sports. Journal of Sports Sciences 28: 1319–1325. 19. Jennings D, Cormack S, Coutts AJ, Boyd L, Aughey RJ (2010) The validity and reliability of GPS units for measuring distance in team sport specific running patterns. International journal of sports physiology & performance 5: 328–341. 20. Johnston RJ, Watsford ML, Pine MJ, Spurrs RW, Murphy AJ, et al. (2012) The validity and reliability of 5-Hz global positioning system units to measure team sport movement demands. Journal of Strength & Conditioning Research 26: 758–765. 21. Portas MD, Harley JA, Barnes CA, Rush CJ (2010) The validity and reliability of 1-Hz and 5-Hz global positioning systems for linear, multidirectional, and soccer-specific activities. International journal of sports physiology & perfor- mance 5: 448–458. 22. Varley MC, Fairweather IH, Aughey RJ (2012) Validity and reliability of GPS for measuring instantaneous velocity during acceleration, deceleration, and constant motion. Journal of Sports Sciences 30: 121–127. 23. Waldron M, Worsfold P, Twist C, Lamb K (2011) Concurrent validity and test- retest reliability of a global positioning system (GPS) and timing gates to assess sprint performance variables. Journal of Sports Sciences 29: 1613–1619. 24. Edgecomb SJ, Norton KI (2006) Comparison of global positioning and computer-based tracking systems for measuring player movement distance during Australian football. J Sci Med Sport 9: 25–32. 25. MacLeod H, Morris J, Nevill A, Sunderland C (2009) The validity of a non- differential global positioning system for assessing player movement patterns in field hockey. J Sports Sci 27: 121–128. 26. Gray AJ, Jenkins DG (2010) Match analysis and the physiological demands of Australian football. Sports Med 40: 347–360. 27. Wisbey B, Rattray B, Pyne DB (2009) Quantifying changes in AFL player game demands using GPS tracking: 2009 AFL Season. AFL Research Board Report. Melbourne: Australian Football League. 28. Carling C, Bloomfield J, Nelsen L, Reilly T (2008) The role of motion analysis in elite soccer: contemporary performance measurement techniques and work rate data. Sports Med 38: 839–862. 29. Dellal A, Keller D, Carling C, Chaouachi A, Wong del P, et al. (2010) Physiologic effects of directional changes in intermittent exercise in soccer players. Journal of Strength & Conditioning Research 24: 3219–3226. 30. Rampinini E, Bishop D, Marcora SM, Bravo DF, Sassi R, et al. (2007) Validity of simple field tests as indicators of match-related physical performance in top- level professional soccer players. Int J Sports Med 28: 228–235. 31. Spencer M, Lawrence S, Rechichi C, Bishop D, Dawson B, et al. (2004) Time- motion analysis of elite field hockey, with special reference to repeated-sprint activity. J Sports Sci 22: 843–850. 32. Barbero-Alvarez JC, Coutts A, Granda J, Barbero-Alvarez V, Castagna C (2010) The validity and reliability of a global positioning satellite system device to assess speed and repeated sprint ability (RSA) in athletes. J Sci Med Sport 13: 232–235. 33. Duffield R, Reid M, Baker J, Spratford W (2010) Accuracy and reliability of GPS devices for measurement of movement patterns in confined spaces for court-based sports. J Sci Med Sport 13: 523–525. 34. Bloomfield J, Polman R, O’Donoghue P (2007) Physical demands of different positions in FA Premier League soccer. Journal of Sports Science and Medicine 6: 63–70. 35. Nicholas CW, Nuttall FE, Williams C (2000) The Loughborough Intermittent Shuttle Test: A field test that simulates the activity pattern of soccer. Journal of Sports Sciences 18: 97–104. 36. McGregor SJ, Nicholas CW, Drawer S, Grayson A, Williams C (1999) Metabolic responses to fluid ingestion during prolonged intermittent high intensity shuttle running. Proceedings of the Fourth Science and Football Conference. Sydney University of Technology. pp. 54. 37. Ramsbottom R, Brewer J, Williams C (1988) A progressive shuttle run test to estimate maximal oxygen uptake. British Journal of Sports Medicine 22: 141– 144. 38. Nicholas CW, Williams C, Lakomy HK, Phillips G, Nowitz A (1995) Influence of ingesting a carbohydrate-electrolyte solution on endurance capacity during intermittent, high-intensity shuttle running. J Sports Sci 13: 283–290. 39. Atkinson G, Nevill AM (1998) Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Medicine 26: 217–238. 40. Gant N, Atkinson G, Williams C (2006) The validity and reliability of intestinal temperature during intermittent running. Medicine and Science in Sports and Exercise 38: 1926–1931. 41. Bland JM, Altman DG (1986) Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1: 307–310. 42. Atkinson G, Nevill AM (1998) Statistical methods for assessing measurement error (reliability) in variables relevant to sports medicine. Sports Med 26: 217– 238. 43. Di Salvo V, Baron R, Tschan H, Calderon Montero FJ, Bachl N, et al. (2007) Performance characteristics according to playing position in elite soccer. Int J Sports Med 28: 222–227. 44. Townshend AD, Worringham CJ, Stewart IB (2008) Assessment of speed and position during human locomotion using nondifferential GPS. Med Sci Sports Exerc 40: 124–132. 45. Magalha˜es J, Rebelo A, Oliveira E, Silva JR, Marques F, et al. (2010) Impact of Loughborough Intermittent Shuttle Test versus soccer match on physiological, biochemical and neuromuscular parameters. Eur J Appl Physiol 108: 39–48. GPS Validity during Multidirectional Running PLOS ONE | www.plosone.org 6 April 2014 | Volume 9 | Issue 4 | e93693

Rapid directional change degrades GPS distance measurement validity during intermittent intensity running.

04-14-2014

Rawstorn, Jonathan C,Maddison, Ralph,Ali, Ajmol,Foskett, Andrew,Gant, Nicholas

eng

PMC10192137

1 Vol.:(0123456789) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports Determination, measurement, and validation of maximal aerobic speed Govindasamy Balasekaran *, Mun Keong Loh , Peggy Boey & Yew Cheo Ng This study determined Maximal Aerobic Speed (MAS) at a speed that utilizes maximal aerobic and minimal anaerobic contributions. This method of determining MAS was compared between endurance (ET) and sprint (ST) trained athletes. Nineteen and 21 healthy participants were selected for the determination and validation of MAS respectively. All athletes completed five exercise sessions in the laboratory. Participants validating MAS also ran an all-out 5000 m at the track. Oxygen uptake at MAS was at 96.09 ± 2.51% maximal oxygen consumption ( ˙VO2max ). MAS had a significantly higher correlation with velocity at lactate threshold (vLT), critical speed, 5000 m, time-to-exhaustion velocity at delta 50 in addition to 5% velocity at ˙VO2max (TlimυΔ50 + 5%v ˙VO2max ), and Vsub%95 (υΔ50 or υΔ50 + 5%v ˙VO2max ) compared with v ˙VO2max , and predicted 5000 m speed (R2 = 0.90, p < 0.001) and vLT (R2 = 0.96, p < 0.001). ET athletes achieved significantly higher MAS (16.07 ± 1.58 km·h−1 vs. 12.77 ± 0.81 km·h−1, p ≤ 0.001) and maximal aerobic energy (EMAS) (52.87 ± 5.35 ml·kg−1·min−1 vs. 46.42 ± 3.38 ml·kg−1·min−1, p = 0.005) and significantly shorter duration at MAS (ET: 678.59 ± 165.44 s; ST: 840.28 ± 164.97 s, p = 0.039). ST athletes had significantly higher maximal speed (35.21 ± 1.90 km·h−1, p < 0.001) at a significantly longer distance (41.05 ± 3.14 m, p = 0.003) in the 50 m sprint run test. Significant differences were also observed in 50 m sprint performance (p < 0.001), and peak post-exercise blood lactate (p = 0.005). This study demonstrates that MAS is more accurate at a percentage of v ˙VO2max than at v ˙VO2max . The accurate calculation of MAS can be used to predict running performances with lower errors (Running Energy Reserve Index Paper). The measurement of Maximal Aerobic Speed (MAS) is essential for determining aerobic and anaerobic perfor- mances of various athletes. However, there is a lack of agreement on the definition and measurement of MAS in existing literature1. Terms such as maximal velocity (Vmax), velocity at maximal oxygen uptake (v ˙VO2max ), peak running velocity, and maximal aerobic velocity have been used to represent MAS. Studies have predominantly considered v ˙VO2max as MAS1,2. However, there is a high variability in the literature regarding the speeds and increments used to measure v ˙VO2max , which is reported to produce different results for the same runner3. Studies on the relative importance of aerobic and anaerobic energy during running have suggested that time to exhaustion (Tlim) at v ˙VO2max utilizes a higher amount of anaerobic energy and therefore selecting v ˙VO2max as MAS may not be accurate4–6. Since MAS should utilize maximal aerobic energy (EMAS) and minimal possible anaerobic energy contribution, MAS should be lower than v ˙VO2max at a precise speed with a corresponding lower blood lactate (BLa) response1. In addition, there is a wide range of intergroup variation in maximal oxygen uptake ( ˙VO2max) between individuals, which vary according to the athletic background and gender of the athlete7. Hence, there is currently no universal acceptance of a single standard of measure of MAS. Exercising above critical speed (CS), which is close to the velocity of lactate threshold (vLT), leads to slow additional increases of oxygen uptake ( ˙VO2)8. Lactate threshold (LT) is usually detected at the point where BLa has a nonlinear increase during exercise as it reflects net lactate production that had exceeded lactate elimi- nation. Such BLa concentrations are usually taken during graded incremental exercise tests that indicate lactate curves. Therefore, the shift in lactate curves indicate a change in aerobic capacity, also known as LT9. This slow component of ˙VO2 becomes apparent at approximately 80–110 s from the start of maximal effort exercise, where a range of speeds is estimated as EMAS 10. One of the proposed intensities at which EMAS can be determined is known as velocity of delta 50 (υΔ50), the median of v ˙VO2max and vLT11. Measurements for vLT, v ˙VO2max , and υΔ50 of 8 highly trained long distance runners found υΔ50 to be at 91% of ˙VO2max ( ˙VO2max = 59.8 ml·kg−1·min−1, v ˙VO2max = 18.5 km·h−1, vLT = 15.2 km·h−1, υΔ50 = 16.9 km·h−1)12. However, this speed did not seem to elicit EMAS OPEN Human Bioenergetics Laboratory, Physical Education and Sports Science, National Institute of Education, Nanyang Technological University, 1 Nanyang Walk, Singapore 637616, Singapore. *email: [email protected] 2 Vol:.(1234567890) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports/ in trained athletes8. Hence, a hypothetical minimum intensity of υΔ50 + 5%v ˙VO2max will be used in this study for participants who did not achieve EMAS at υΔ50. Anaerobic energy utilization is estimated as the time spent at ˙VO2max during Tlimv ˙VO2max . This is based on the assumption that anaerobic energy stores will be completely depleted during Tlim at intensities above CS13. This has been demonstrated in previous studies assuming that maximal anaerobic energy (EMAnS) was consumed during 800–5000  m14 as well as 1500–10,000  m15 runs. It is necessary to select the intensity at which the con- sumed anaerobic energy is a representative of the anaerobic energy used at any run with an aerobic speed reserve (AeSR), where AeSR represents the difference between v ˙VO2max and CS16. MAS lies at the extreme of the range between CS and v ˙VO2max . During Tlimv ˙VO2max , the athlete attains EMAS and uses EMAnS with minimal aerobic contribution. Tlim ˙VO2max determined at other intensities within this range may consume comparatively higher percentage of ˙VO2 and thus overestimate the anaerobic energy. Hence, Tlim ˙VO2maxv ˙VO2max as anaerobic energy seems logical to measure duration at MAS (MASdur) and MAS. To determine MAS and MASdur, anaerobic energy consumption at MAS has to be minimized without com- promising its criteria. MASdur can be calculated by subtracting anaerobic energy duration from ˙VO2max till exhaustion at Vsub%95 (TlimVsub%95). This method was based on the negative linear relationship between anaerobic and aerobic energy contribution during physical activity, as anaerobic energy contribution decreases with increasing exercise duration17. Therefore, subtracting anaerobic energy duration from TlimVsub%95 may provide an accurate determination of MASdur. The objectives of this study aimed to (1) determine MAS at a speed that utilizes maximal aerobic and minimal anaerobic contributions, where MAS should fulfill four criteria (a) MAS should be lower than v ˙VO2max, (b) maxi- mal aerobic energy utilization is elicited during Tlim test, (c) MAS should occur at a specific percentage fraction of v ˙VO2max , and (d) estimated anaerobic energy contribution at TlimMAS should be lower than that at Tlimv ˙VO2max . (2) To assess whether MAS can accurately differentiate between athletes of different training orientations (endur- ance or sprint trained) and if there was an association between MASdur and aerobic performance variables of run distance and best performance times. It was hypothesized that the MAS of endurance-trained athletes would be higher than that of sprint-trained athletes, and that MAS measured would significantly correlate with 5000 m run performance and aerobic performances variables. This study has been separated into two parts. The first part of this study, which this paper is based on, utilizes a new framework of calculating MAS. This validated MAS was confirmed with the prediction of running performances in a follow-up paper that examined the Running Energy Reserve Index (RERI)18. Methods Participants. Forty participants volunteered for the study. Among the 40 athletes, 19 healthy participants (age: 29.74 ± 8.31 years; height: 171.86 ± 7.65 cm; body mass index (BMI): 22.01 ± 2.12 kg·m−2; body fat percent- age (BF%): 12.96 ± 3.10%)) were selected to validate the theoretical framework criteria of MAS. The remain- ing participants consisted of 9 sprint-trained athletes (age: 26.89 ± 9.39 years; height: 174.16 ± 5.69 cm; BMI: 23.09 ± 2.07  kg·m−2; BF%: 10.59 ± 2.55%) and 12 endurance-trained athletes (age: 31.67 ± 7.24  years; height: 173.67 ± 7.59 cm; BMI: 21.34 ± 1.27 kg·m−2; BF%: 12.74 ± 2.38%) (Table 1). These 21 athletes were selected to determine whether there were significant differences between the MAS of sprint-trained and endurance-trained athletes, and the relation of MAS with aerobic performances and variables. Participants were considered trained if they were engaged in training for at least four sessions of 60 min per week in their chosen activities for the last 12 months. Among the endurance-trained athletes, 4 were triathletes Table 1. Descriptive characteristics endurance-trained and sprint-trained athletes. Values are in means ± SD. BMI Body mass index, LBM Lean body mass, FFM Fat free mass, BMD Bone mineral density, BMC Bone mineral content, BLa Blood lactate. *p ≤ 0.05, **p ≤ 0.01: Indicates significant difference between endurance- trained and sprint-trained athletes. † The data of two participants aged 14.5 ± 0.5 years were not included due to differences between age of these two participants and the total cohort and its effect on body composition19 (Boileau and Horswill 200320). Variables Endurance-trained Sprint-trained N 12 9 Age (years) 31.67 ± 7.24 26.89 ± 9.39 Height (cm) 173.67 ± 7.59 174.16 ± 5.69 BMI (kg·m−2) 21.34 ± 1.27 23.09 ± 2.07* Fat percentage (%) 12.74 ± 2.38 10.59 ± 2.55 Hematocrit (%) 43.51 ± 2.30 45.59 ± 1.34* Haemoglobin (g·dl−1) 14.79 ± 0.78 15.50 ± 0.46* Plasma volume (%) 56.49 ± 2.30 54.41 ± 1.34* LBM (kg)† 53.13 ± 5.50 58.76 ± 4.43* FFM (kg)† 55.68 ± 5.80 61.69 ± 4.65* BMD (g·cm−2)† 1.19 ± 0.07 1.28 ± 0.10* BMC (kg)† 2.61 ± 0.23 2.93 ± 0.29* Rest BLa (mmol·L−1) 0.71 ± 0.13 0.83 ± 0.16 3 Vol.:(0123456789) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports/ and had completed the ironman distance race (3.86 km swim, 180.25 km bike, and 42.195 km run) several times. The other 6 participants were training for half and full marathon, and the remaining 2 were 10 km runners. The sprint-trained athletes were specialized in soccer and 100–400 m sprint events, and they were still actively competing in their respective events. Participants who had any history of musculoskeletal injuries in the past 6 months, smokers and medical history were exempted from this study. All participants were informed of the risk and benefits of the study and gave their informed consent to participate. This study was approved by the Ethical Review Board of the Research and Graduate Studies Committee of Physical Education & Sports Science, National Institute of Education, Nanyang Technological University, Singapore. All methods were performed in accordance with the relevant guidelines, regulations and STROBE checklist. Experimental design. The experimental design and procedures in this study were derived and modified from Bundle et al.21. A within cross-sectional design was utilized in each investigation, where each partici- pant underwent a series of exercise tests to determine MAS accurately. Participants completed exercise sessions which included (1) aerobic metabolic measurement utilizing Astrand modified running (AMRMAX) continu- ous incremental maximal treadmill protocol, (2) submaximal discontinuous treadmill run (SUBMAX) protocol, (3) Tlim at v ˙VO2max , (4) Test of Tlim at Vsub%95, and (5) speed and duration test protocols. To assess the validity of MAS, participants also ran an all-out 5000 m on the track. Participants were instructed to avoid strenuous activities, alcohol, and caffeine 24 h before testing. All laboratory sessions were conducted at the Human Bioenergetics Laboratory in the Physical Education and Sports Science department of the National Institute of Education, Nanyang Technological University, Singapore, while the 5000 m track test was performed on the 400 m track located at the Sports and Recreation Centre of Nanyang Technological University, Singapore. Pretest preparations. Prior to the tests where cardiorespiratory and aerobic metabolic parameters were measured, the flow meter, sampling line and gas calibrations of ParvoMedics TrueOne 2400 (ParvoMedics Inc, UT, USA) were performed according to the procedures explained in the instruction manual (Operator’s guide, Version 4.3, ParvoMedics Inc, UT, USA 2008). Heart rate (HR) transmitters were strapped onto the participants’ chest, and participants were required to put on the head cap, mouthpiece of a two-way non-rebreathing valve. A nose-clip was used to ensure all expired air are analyzed. In addition, participants were strapped in an upper body safety harness to prevent falling while running on the treadmill belt at various speeds. The harness did not assist or impede the participants during the tests. Experimental tests and measurements. Participants were instructed to stride the belt of the treadmill before the tests, and to hold the handrail of the treadmill or give a ‘thumbs down’ signal to stop the test due to exhaustion or discomfort. All the laboratory tests were performed on a motorized treadmill (H-P Cosmos, UK). The gradient was set at 1% for all treadmill running protocols except for ˙VO2max protocol22. Participants were encouraged to deliver their maximum effort during tests. Before performing the ˙VO2max test, height and weight of participants were recorded, and a Dual-Energy X-ray absorptiometry (DEXA, QDR 4500W, Hologic Inc, Waltham, USA) scan was performed to determine body composition. Subsequently, capillary blood sample was collected via the finger prick technique to measure resting BLa. Astrand modified running continuous incremental maximal treadmill (AMRMAX) protocol. The AMRMAX protocol was employed to determine ˙VO2max of participants. The test began with an initial speed of 8–12 km·h−1 with 0% gradient. After 3 min of running, the gradient was increased by 2.5% at 2 min stages until volitional exhaustion. Thereafter, post-exercise capillary whole blood samples were taken from the finger at every minute for 5 min. BLa was analyzed via YSI 2300 STAT Plus (2300 D, YSI Incorporated, USA) to measure peak post- exercise BLa. The expired breath-by-breath gas concentrations were analyzed using ParvoMedics TrueOne 2400 (ParvoMedics, Inc, USA) and averaged at every 15 s. HR was measured via a Polar HR transmitter (Polar Electro, Singapore) which sends its signals to the receiver of ParvoMedics TrueOne 2400 metabolic system (ParvoMed- ics, Inc, USA). ˙VO2max was determined when participants satisfied three of the following five criteria23: (1) Plateau of ˙VO2 change in ˙VO2 ≤ 2.1 ml·kg−1·min−1 in spite of increasing treadmill gradient, (2) Respiratory exchange ratio (RER) at ˙VO2max ≥ 1.1, (3) BLa > 8 mmol·L−1, (4) HR ≥ 90% of the age predicted maximal HR (HRmax), and (5) volitional exhaustion9. Submaximal discontinuous treadmill (SUBMAX) protocol. Participants performed a series of six to nine dis- continuous submaximal treadmill runs. Initial speed was set at approximately 40–60% ˙VO2max with increments of 4–5% ˙VO2max at every stage depending on the ability of the participant. All running speeds were within the range of 40–90% ˙VO2max . Running sessions were fixed at 4  min23,24, with 2–4 min recovery between sessions. Capillary blood samples were obtained with the finger prick technique and were collected immediately after each submaximal running session. Steady state cardiorespiratory and aerobic metabolic measures were recorded at every 15 s during the 3rd and 4th minute of each treadmill running session. vLT was then determined using a log–log plot method25,26. The linear relation between run speeds and cor- responding ˙VO2 were determined using a linear regression analysis21,27,26. Linear relation determined through SUBMAX protocol was extrapolated to ˙VO2max , and this velocity at ˙VO2max was termed as v ˙VO2max26. The average of vLT and v ˙VO2max was calculated to determine υΔ50. 4 Vol:.(1234567890) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports/ Oxygen Consumption till Exhaustion ( ˙VO2max till exhaustion (Tlim)) tests. Tlim tests were conducted at 100% v ˙VO2max (Tlimv ˙VO2max ) and υΔ50. However, it was found that the participants could not reach ˙VO2max at υΔ50. Hence, 5%v ˙VO2max was added to υΔ50 for all participants to achieve maximal aerobic energy during the Tlim test (υΔ50 ± 5%v ˙VO2max ). The speed at which EMAS was attained during Tlim at υΔ50 and υΔ50 ± 5%v ˙VO2max was termed Vsub%95. Achieving ≥ 95% ˙VO2max was selected as the primary criterion to measure time to attain ˙VO2max (TA ˙VO2max ) during Tlimv ˙VO2max and TlimVsub%9527,19. Participants performed a warm up protocol of 8–15 min at 60% ˙VO2max followed by a rest interval of 5–10 min. During each of the Tlim test, participants ran at a fixed speed for as long as possible until volitional exhaustion. Breath-by-breath cardiorespiratory and aerobic metabolic measures were recorded during each run. BLa samples were collected after warm up and at each minute of the first five minutes after individual Tlim run to determine peak post-exercise BLa. Breath-by-breath ˙VO2 responses recorded at Tlimv ˙VO2max were interpolated per second and the time was aligned to the start of the run with an average at every five seconds via a moving average filter. Thereafter, the data was fitted to a positive exponential nonlinear regression by means of weighted least square method using SigmaPlot software (windows version 11.0.0.77, Germany) (Eq. 1). This equation was fitted to the data collected from Tlim tests and TA ˙VO2max and Tlim ˙VO2maxconverted were computed (Eqs. 2 and 3). where ˙VO2baseline is the ˙VO2 before starting the Tlim run, A is the amplitude of ˙VO2 ( ˙VO2max– ˙VO2baseline) for I, and II components, δ is the time delay before onset of each exponential component and τ is the time constant for each component of ˙VO2 28. Speed and duration curve protocol. After pretest preparations, orientation trials were conducted by allowing participants to step onto the treadmill at fast speeds. Following a 5–10min recovery, the treadmill was set at a preselected speed. Participants then stepped on the moving treadmill with the use of the handrail and started unassisted running within 4–7 steps. They were instructed to run until volitional exhaustion, and both duration and run speeds at exhaustion were recorded. Full recovery was given between the trials, and they were allowed to discontinue the test if they were unable to perform at their best. A minimum of two to three trials were per- formed at different speeds ranged from 110% v ˙VO2max to 140% v ˙VO2max. Participants were only allowed to per- form the next trial if: (1) recovery HR was equal to or more than 120 beats·min−1 approximately, (2) participant gave consent for performing the test to the best of their abilities, and (3) duration of recovery was based on the principle of work to rest ratio. Speeds in the range of 90–140% v ˙VO2max and their corresponding durations calculated during the different Tlim sessions and speed-duration curve protocol were data fitted to determine hyperbolic relation (Fig. 1). MAS was then determined using Eq. 4. (1) ˙VO2(t) = ˙VO2 baseline + A0 ×  1 − e −  t τ0  + A1 ×  1 − e −  t−δ1 τ1  (2) Tlim ˙VO2 maxv ˙VO2 max = Tlimv ˙VO2 max−TA ˙VO2 maxv ˙VO2 max (3) Tlim ˙VO2 max converted(s) =  Tlim ˙VO2 maxv ˙VO2 max × v ˙VO2 max  Vsub%95 Figure 1. Hyperbolic relationship between speed and duration. 5 Vol.:(0123456789) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports/ where CS = critical speed; ADC = anaerobic distance capacity; MASdur = duration at MAS; and B = constant. A backward validation by predicting run performances was performed whereby MASdur was calculated by adding the time representing anaerobic energy18. Since there is a negative relationship between aerobic and anaerobic energy, aerobic energy was taken to be the negative of anaerobic energy. The following equation was employed for the calculation of MASdur (Eq. 5): The linear relation between speed and ˙VO2 (measured through the SUBMAX protocol) was extrapolated to MAS and the extrapolated ˙VO2 at MAS was considered as EMAS 21. 50 m sprint run test. Participants performed a general 10–15 min warm up run at a comfortable pace fol- lowed by dynamic stretching exercises. Following the warm up, participants performed strides of 20–40 m with 3–5 min recovery between strides. The 50 m sprint run was performed with a standing start position at the start line. At the start command, the athlete accelerated and covered the distance of 50 m in the least possible time. The speed and time at the stipulated distance intervals within 50 m were automatically recorded by the five timing gates placed within 34–50 m for sprinters and middle distance runners and within 30–46 m for endurance athletes. A minimum of two trials were performed, with a 15–20 min rest interval in between the trials, and the best performance was recorded to the nearest 0.01 s. 5000 m test. Orientation trials were performed 1 week before testing to familiarize participants with the pace of their run to elicit the best effort in testing. Prior to the actual run, participants warmed up for 10–15 min at a comfortable pace followed by stretching exercises. A rest period of 3–5 min after the warm up was given before starting the test. Participants were encouraged to run at their targeted best effort based on their fitness level and ran the whole distance at their own self-regulated pace. The time taken to cover each run was recorded to the nearest 0.01 s. Statistical analysis. Statistical analyses and data fitting procedures were performed using Statistical Pack- age for Social Sciences (SPSS) version 17.0 and SigmaPlot software (version 11.0, Systat software, Inc., 2008, Ger- many) respectively. Using a power of 0.80 and α level of 0.05 with an effect size of > 1.1, it was determined that a minimum of 10 participants were required29. Linear regression was employed to calculate vLT, v ˙VO2max, and EMAS. One-way ANOVA was utilized to measure any significant differences between BLa measured during the different Tlim tests and BLa measured at ˙VO2max (BLa ˙VO2 max) . The Wilcoxon rank test (non-parametric paired t-test) and correlation technique were employed to significantly validate the criteria of MAS, and independent t-tests were employed to compare anthropometrical and body composition measures, cardiorespiratory and aerobic metabolic measures, and MAS between endurance-trained and sprint-trained athletes. Lastly, coeffi- cient of correlation technique (very strong correlation: 0.9–1.0, strong correlation: 0.7–0.9, moderate correla- tion: 0.5–0.7) was used to assess the relationship between MAS and aerobic parameters. Statistical significance was set at p ≤ 0.05 for this study. Results As shown in Table 2, anthropometrical, body composition, and hematological measures were significantly higher among sprint-trained athletes compared to endurance-trained athletes. However, the proportion of plasma volume was significantly higher among endurance-trained athletes. Figure 2 determined the steady state of the participants during the SUBMAX protocol calculated by the submaximal efficiency equation. (4) Speed  m · s−1 = CS +  ADC B + MASdur  (5) MASdur = TlimVsub%95−  −Tlim ˙VO2 max converted  Table 2. Astrand Modified Running Protocol (AMRMAX) results in endurance-trained and sprint-trained athletes. Values are in means ± SD. ˙VO2max Maximal oxygen uptake, RERmax Respiratory exchange ratio at ˙VO 2max, HR ˙VO2max Heart rate at ˙VO2max, %HRmax Percentage of maximal heart rate, BLa ˙VO2max Blood lactate at ˙VO2max. *p < 0.05, **p < 0.01: Indicates significant difference between endurance-trained and sprint-trained athletes. a Only 8 sprint trained participants were analyzed due to technical difficulties. Variables Endurance-trained Sprint-trained N 12 9 ˙VO2max (ml·kg −1·min−1) 57.62 ± 5.40 51.12 ± 3.59** RERmax 1.16 ± 0.03 1.12 ± 0.03a ** HR ˙VO2max (beats·min-1) 181.71 ± 14.31 185.06 ± 5.81 %HRmax at ˙VO2max 96.45 ± 6.02 95.89 ± 5.96 BLa ˙VO2max (mmol·L−1) 8.26 ± 1.72 8.17 ± 1.63 6 Vol:.(1234567890) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports/ MASdur calculation. MASdur was calculated by subtracting Tlim ˙VO2maxconverted from TlimVsub%95. This however resulted in MAS being higher than Vsub%95, which elicited higher anaerobic energy and thus failed to fulfil the MAS criteria. Figure 3 shows an example of a participant whose Tlim ˙VO2maxconverted and TlimVsub%95 were at 159 s and 533 s respectively. Subtracting these two would have resulted in a corresponding speed at MASdur of 16.9 km·h−1 on the speed-duration graph ((306 s = 5 min 6 s (16.9 km·h-1) → converted to Tlimconverted = 159 s (using Eq. 3). TlimVsub95 = 533 s—(− Tlimconverted) 159 s = 692 s (MASdur) (using Eq. 5), 692 s = 11 min 32 s)). This translated to 97.1%v ˙VO2max, which was close to v ˙VO2max at which EMAnS was determined. Using the same participant in Fig. 3, adding Tlim ˙VO2maxconverted and TlimVsub%95 together resulted in a cor- responding MASdur speed at 16.1 km·h−1, which was at 92.5%v ˙VO2max . It seemed that Tlim ˙VO2maxconverted and TlimVsub%95 fulfilled the criteria of achieving MAS. This suggest that accurate calculation of MAS will result in lower error of prediction of run performances with an average of 2.39 ± 2.04% (R2 = 0.99, nT (number of running trials) = 252)) for all athletes, with treadmill trials to within an average of 2.26 ± 1.89% (R2 = 0.99, nT = 203) and track trials to within an average of 2.95 ± 2.51% (R2 = 0.99, nT = 49)18. Validation of MAS. The mean MAS was 14.50 ± 1.82  km·h−1. There was no significant difference between ˙VO2 at MAS (96.09 ± 2.51% ˙VO2max) and ˙VO2 at 95% ˙VO2max among all athletes ( ˙VO2 at MAS: Figure 2. Determination of the submaximal efficiency equation between ˙VO2 and corresponding run speeds. Figure 3. Calculation of duration at MAS. (A) indicates the duration of MAS with anaerobic and aerobic energy. (B) indicates the calculation of MAS based on duration of TlimVsub%95 and Tlim ˙VO2maxconverted at Tlimv ˙VO2max. 7 Vol.:(0123456789) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports/ 50.18 ± 5.19  ml·kg−1·min−1 vs. ˙VȮ2 at 95% ˙VO2max: 50.69 ± 4.69  ml·kg−1·min−1, p = 0.134). In addition, mean BLa at MAS (BLaMAS) (7.80 ± 1.52 mmol·L−1) was significantly lower than corresponding values at v ˙VO2max (9.11 ± 2.50 mmol·L−1; p = 0.009) and ˙VO2max (8.60 ± 1.62 mmol·L−1; p = 0.037). While BLaMAS was not signifi- cantly lower than BLa at Vsub%95 (BLaVsub%95) (8.01 ± 1.39 mmol·L−1, p = 0.174). RER, ventilatory threshold and HR at MAS were 1.05 ± 0.03, 2.19 ± 0.51 L·min−1 and 176.62 ± 26.72 beats·min−1 respectively. MAS between endurance-trained and sprint-trained athletes. Endurance-trained athletes had sig- nificantly higher mean ˙VO2max (p = 0.004) and RER at ˙VO2max (RERmax) (p = 0.007) (Table 2). vLT (p < 0.001), BLa at LT (BLaLT) (p < 0.001), ˙VO2 at LT ( ˙VO2LT) (p = 0.013) were significantly higher among ET athletes, while no significant differences were observed between both cohorts for HR at LT (HRLT) (p = 0.467) and percentage of HRmax (%HRmax) (p = 0.968) (Table 3). In addition, measured υΔ50 (p < 0.001) and υΔ50 + 5%v ˙VO2max (p < 0.001) were also significantly higher in endurance-trained athletes compared to sprint-trained athletes (Table 4). All athletes attained ≥ 95% ˙VO2max to calculate TA ˙VO2max at Tlimv ˙VO2max and TlimVsub%95 (Table 5). v ˙VO2max and Vsub%95 were significantly higher among endurance-trained athletes (p ≤ 0.001). However, sprint-trained athletes ran at these speeds for longer duration and hence Tlim was significantly different compared to ET athletes (p = 0.030). No significant differences were determined between both cohorts for TA ˙VO2max Tlim ˙VO2max, and BLa at Tlimv ˙VO2max (p = 0.164) and TlimVsub%95 (p = 0.264) (Table 5). Similar results were also calculated for Tlim ˙VO2maxconverted (sprint-trained: 167.98 ± 52.28 s; endurance-trained: 125.75 ± 76.28 s, p = 0.171). Mean CS (endurance-trained: 14.95 ± 1.40 km·h−1; sprint-trained: 11.52 ± 0.80 km·h−1, p < 0.001) was signifi- cantly higher while ADC (endurance-trained: 221.60 ± 57.74 m; sprint-trained: 313.43 ± 139.74 m, p < 0.05) was significantly lower in endurance-trained athletes compared to strength-trained athletes. MAS range was between 15.37 ± 1.57 km·h−1 (~ υΔ50) and 16.25 ± 1.64 km·h−1 (~ υΔ50 + 5%v ˙VO2max) among endurance-trained athletes and between 12.42 ± 0.81 km·h−1 (~ υΔ50) and 13.12 ± 0.85 km·h−1 (~ υΔ50 + 5%v ˙VO2max) among sprint-trained athletes. Furthermore, endurance-trained athletes achieved significantly higher MAS (endurance-trained: 16.07 ± 1.58 km·h−1; sprint-trained: 12.77 ± 0.81 km·h−1, p ≤ 0.001; 95% CI [2.091, 4.515]) and EMAS (endurance- trained: 52.87 ± 5.35 ml·kg−1·min−1; sprint-trained: 46.42 ± 3.38 ml·kg−1·min−1, p = 0.005; 95% CI [2.182, 10.716]) Table 3. Submaximal discontinuous treadmill run (SUBMAX) test results in endurance-trained and sprint- trained athletes. Values are in means ± SD. vLT velocity at lactate threshold, BLaLT Blood lactate at LT, ˙VO 2LT Oxygen uptake at LT, HRLT Heart rate at LT, %HRmax  Percentage of maximal heart rate; υΔ50 median of v ˙VO2max and vLT, υΔ50 or υΔ50 + 5%v˙VO2max mean speed of v ˙VO2max and vLT or mean speed of v ˙VO2max and vLT plus 5%v ˙VO2max. *p < 0.05, **p < 0.01, ***p < 0.001: Indicates significant difference between endurance- trained and sprint-trained athletes. Variables Endurance-trained Sprint-trained N 12 9 vLT (km·h−1) 13.37 ± 1.58 10.48 ± 0.83*** BLaLT (mmol·L−1) 1.59 ± 0.59 2.60 ± 0.49*** ˙VO2LT (ml·kg−1·min−1) 43.63 ± 4.15 38.76 ± 3.92* HRLT (beats·min-1) 157.83 ± 15.42 161.96 ± 7.14 HRLT (%HRmax) 83.73 ± 6.60 83.83 ± 4.20 υΔ50 (km·h−1) 15.37 ± 1.57 12.42 ± 0.81*** υΔ50 + 5%v ˙VO2max (km·h−1) 16.25 ± 1.64 13.12 ± 0.85*** Table 4. Oxygen kinetics and blood lactate at v ˙VO2max and Vsub%95 among endurance-trained and sprint- trained athletes. Values are presented as means ± SD. v˙VO2max Velocity at ˙VO2max, Vsub%95 speed at υΔ50 or υΔ50 + 5%v ˙VO2max at which maximal aerobic energy was obtained, Tlim ˙VO2 till exhaustion, TA˙VO2max Time to achieve 95%˙VO2max, Tlim ˙VO2max time spent at ˙VO2max, BLa Blood lactate. a Only 11 participants were analyzed due to technical difficulties. *p < 0.05, **p < 0.01, ***p < 0.001: Indicates significant difference between endurance-trained and sprint-trained athletes at Tlimv ˙VO2max and TlimVsub%95. Variables Endurance- trainedv ˙VO2max Sprint- trainedv ˙VO2max Endurance-trainedVsub%95 Sprint- trainedVsub%95 Speed (km·h−1) 17.38 ± 1.62 14.33 ± 1.11*** 16.25 ± 1.64 13.05 ± 0.82*** Tlim (s) 300.02 ± 67.43 358.50 ± 36.70* 552.84 ± 105.50 672.29 ± 120.98* TA ˙VO2max (s) 182.61 ± 34.57 206.07 ± 39.85 356.89 ± 67.59a 367.31 ± 120.52 Tlim ˙VO2max (s) 117.41 ± 71.47 152.43 ± 45.17 215.77 ± 127.80a 304.98 ± 114.18 Tlim ˙VO2maxconverted (s) 125.75 ± 76.28 167.98 ± 52.28* – – BLa (mmol·L-1) 8.96 ± 1.87 7.96 ± 1.68 7.97 ± 1.59 7.27 ± 1.25 8 Vol:.(1234567890) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports/ at significantly shorter MASdur (endurance-trained: 678.59 ± 165.44 s; sprint-trained: 840.28 ± 164.97 s, p = 0.039; 95% CI [− 314.190, − 9.177]) compared to sprint-trained athletes. MAS was also significantly correlated to ˙VO2max (r = 0.78, p < 0.001), v ˙VO2max (r = 0.98, p < 0.001). In addition, MAS had comparatively higher correlations with vLT (MAS: r = 0.97, p ≤ 0.001; v ˙VO2max: r = 0.91, p < 0.01), CS (MAS: r = 0.99; v ˙VO2max: r = 0.93), 5000 m (MAS: r = − 0.95, p < 0.001; v ˙VO2max: r = − 0.92), TlimυΔ50 + 5%v ˙VO 2max (MAS: r = − 0.71, p < 0.05; v ˙VO2max: r = − 0.62) and Vsub%95 (MAS: r = 0.997, p < 0.001; v ˙VO2max: r = 0.98, p < 0.01) compared to v ˙VO2max. MAS predicted the 5000 m speed and vLT with high accuracy (5000 m speed: R2 = 0.90; vLT: R2 = 0.96, p < 0.001). Sprint-trained athletes had significantly higher Maximal Speed (MS) (p < 0.001) and achieved this speed at a significantly longer distance (p = 0.003). Significant differences were also observed in EMAnS, 50 m sprint perfor- mance (p < 0.001), and peak post-exercise BLa (p = 0.005) in the 50 m sprint run test (Table 5). Limitations In general, the present study had no gold standard technique to validate anaerobic techniques, which may be presented as one of the limitations. Although there are other anaerobic techniques, such as, cycling or jumping, these norms are activity specific and may not accurately predict the anaerobic energy of runners or athletes involved in running. The present investigation’s results could only be compared to a similar technique, Bundle’s et al.21 anaerobic speed reserve (AnSR). The comparison in results indicated a high correlation between both methods, which indicated that MAS may also predict accurate all-out run performances. However, the accuracy of MAS to categorize middle distance athletes was not reported. Also, the techniques used for MAS in this study was different from Bundle’s use of MAS and utilizing MAS in the RERI18 had a lower error for prediction. The backward validation with lower error in prediction values was the only way to validate MAS. In future, MAS could be used to validate other similar anaerobic techniques. In addition, the effect of training on MAS was not determined. Perhaps for future studies, the effect of differ- ent types of training, such as sprint or endurance or a combination of both, on MAS can be studied. Therefore, extending the accuracy of MAS in significantly differentiating middle distance athletes may increase the sensitiv- ity of the model to detect even small changes in energy. Discussion The results from this study confirmed the hypothesis that MAS is more accurate to be measured at %v ˙VO 2max than at v ˙VO2max. The determination of MAS required a subtraction of Tlim ˙VO2maxconverted at v ˙VO2max from TlimVsub%95. This equation eliminated the anaerobic energy contribution. The concept of this study is therefore unique as the MAS determination has very little anaerobic contribution and has revealed low errors in predict- ing performance timings18. Validation of MAS. MAS was obtained at 92.45 ± 1.47%v ˙VO2max and 89.27 ± 3.56%v ˙VO2max for endurance- trained and sprint-trained athletes respectively, confirming the hypothesis that MAS should be obtained at a per- centage of v ˙VO2max rather than at v ˙VO2max. Studies have determined higher anaerobic energy at Tlimv ˙VO2max that was verified by a non-significant difference between BLa at v ˙VO2max ( BLav ˙VO2max) and BLa at ˙VO2max (BLa ˙VO2max )1,2,21,26. Similarly, no significant difference was found in anaerobic energy contribution between Tlim100%v ˙VO2max (15.1 mmol·L −1), Tlim120 (15.7 mmol·L−1) and Tlim140 (15.1 mmol·L−1)16. Tlimv ˙VO2max (269 ± 77 s) was also sig- nificantly correlated (r = − 0.52, p < 0.05) to Tlim120v ˙VO2max (86 ± 25 s) and to the blood pH after Tlim120%v ˙VO2max (r = − 0.68, p < 0.05). On the contrary, EMAS in this study was obtained at MAS. ˙VO2 at MAS (50.69 ± 4.69 ml·kg−1·min−1) was found to be at 96.08 ± 2.51% ˙VO2max, which was not significantly different from 95% ˙VO2max (50.18 ± 5.19 ml·kg−1·min−1). As most athletes did not reach EMAS at speeds of 14.10 km·h−1 which was just below MAS (14.64 km·h−1), MAS seems to be the minimal intensity of the slow component of ˙VO2. Additionally, MAS in this study was obtained at 91.08 ± 2.97%v ˙VO2max for total cohort, which was similar to other studies where most athletes achieved ˙VO2max at 91%v ˙VO2max 30. It was found that endurance-trained athletes in their study ( ˙VO2max = 60.7 ml·kg−1·min−1, v ˙VO 2max = 20 km·h−1) achieved approximately 99% ˙VO2max at 90%v ˙VO2max (18.3 km·h−1)31. This is close to 92.45%v ˙VO 2max at MAS among endurance-trained athletes in the present investigation. These studies suggest that submaxi- mal speed is sufficient for achieving an increase in ˙VO2max and should be used for training32. These findings support the validity of MAS, which is the minimal speed at which EMAS is determined. Table 5. Maximal speed of endurance-trained and sprint-trained athletes. Values are in means ± SD. MS Maximal speed, DistanceMS Distance at which MS determined, BLa50 m Peak post-exercise blood lactate after 50 m sprint run test. **p < 0.01, ***p < 0.001: Indicates significant difference between endurance-trained and sprint-trained athletes. Variables Endurance-trained Sprint-trained N 12 8 MS (km·h−1) 29.26 ± 1.33 35.21 ± 1.90*** DistanceMS (m) 36.29 ± 3.08 41.05 ± 3.14** BLa50 m (mmol·L-1) 4.16 ± 0.83 5.53 ± 1.17** 50 m (s) 7.38 ± 0.45 6.38 ± 0.43*** 9 Vol.:(0123456789) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports/ In addition, BLaMAS in this study was significantly lesser than BLav ˙VO2 max and BLa ˙VO2 max. This could be due to the slow component of ˙VO2 at a slower speed, which is directly related to the recruitment of less efficient fast twitch fibers30, anaerobic energy utilization, and to the intensity of exercise33–35. The decrement of anaerobic energy with increasing duration at TlimMAS compared to Tlimv ˙VO2max could lead to lower BLaMAS compared to BLav ˙VO2max. It was also determined that there was significant correlation between the slow component of ˙VO2 with indices of anaerobic performance (WAnT’s peak power; r = 0.77, p < 0.01)36. Since there is an inverse relation- ship between TA ˙VO2max and exercise intensity37, TA ˙VO2max would have been higher at TlimMAS as compared to Tlimv ˙VO2max. EMAS would have been attained in the later part of the run, which may minimize anaerobic energy contribution. This was shown in the present study and confirmed that MAS calculated was accurate. MAS between endurance-trained and sprint-trained athletes. This study also found that sprint- trained athletes had significantly lower MAS compared to endurance-trained athletes. This was evident in their vLT, ˙VO2max, and v ˙VO2max variables. Endurance training increases ˙VO2max by increasing cardiac stroke volume, blood volume, capillary density, and mitochondrial density in trained muscles35, allowing endurance-trained athletes to have higher ˙VO2max, vLT, and v ˙VO2max compared to sprint-trained athletes. Additionally, MAS had comparatively higher significant correlations with CS, vLT, 5000 m, TlimυΔ50 + 5%v ˙VO 2max, and Vsub%95 compared to v ˙VO2max. Furthermore, MAS was a stronger predictor of 5000 m and vLT. This was similar to a study conducted by Blondel, Berthoinm Billat & Lensel (2001), who also found significant cor- relations between 90%v ˙VO2max and CS (r = 0.69, p < 0.05)16. Additional analysis found that there was a significant negative correlation with maximal speed reserve (MSR; difference between MS and CS; r = 0.79, p ≤ 0.001). This relationship is consistent with previous studies who found that endurance-trained athletes with lower MSR achieved vLT and CS at higher speeds and had lower MAnS compared to sprinters16,38. These findings support the utility of MAS in predicting performances in most running events and may suggest more accurate performance prediction at MAS rather than at v ˙VO2max. Conclusion In conclusion, this study aimed to determine the intensity at which aerobic energy contribution is at maximal. MAS in this study was found to be at 92.45 ± 1.47%v ˙VO2max for endurance trained athletes, 89.27 ± 3.56%v ˙VO 2max for sprint trained athletes, and 91.08 ± 2.97%v ˙VO2max among the total cohort. This accurately represented EMAS with minimal contribution from anaerobic energy sources, thus confirming the hypothesis that MAS is more accurate at %v ˙VO2max rather than at v ˙VO2max. MAS for endurance-trained athletes were also significantly higher compared to sprint-trained athletes, indicating that MAS can differentiate between the types of athletes. Furthermore, MAS was found to significantly correlate with aerobic performance variables, and this suggest that submaximal speed is sufficient for training athletes. Regardless the profile of the individual, recreational athletes, collegiate athletes, elite athletes, coaches, and sports practitioners may utilize this MAS calculation to accurately derive the athlete’s individual main energy contribution source (anaerobic or aerobic energy source). Coaches may use their athletes’ MAS to prescribe training workouts that are specifically catered to them, which will predict an accurate sporting performance. Therefore, this new MAS framework demonstrates that the accurate calculation of MAS can accurately predict run performances at lower errors.18 Received: 22 February 2023; Accepted: 20 March 2023 References 1. Lacour, J. R., Padilla-Magunacelaya, S., Chatard, J. C., Arsac, L. & Barthelemy, J. C. Assessment of running velocity at maximal oxygen uptake. Eur J. Appl. Physiol. Occup. Physiol. 62(2), 77–82 (1991). 2. Renoux, J. C., Petit, B., Billat, V. & Koralsztein, J. P. Calculation of times to exhaustion at 100 and 120% maximal aerobic speed. Ergonomics 43(2), 160–166 (2000). 3. Noakes, T. D. Implications of exercise testing for prediction of athletic performance: A contemporary perspective. Med. Sci. Sports Exerc. 20(4), 319–330 (1988). 4. Medbo, J. I. & Tabata, I. Relative importance of aerobic and anaerobic energy release during short: Lasting exhausting bicycle exercise. J. Appl. Physiol. 67(5), 1881–1886 (1989). 5. Houmard, J. A. et al. Peak running velocity, submaximal energy expenditure, VO2max, and 8 km distance running performance. J. Sports Med. Phys. Fit. 31(3), 345–350 (1991). 6. Craig, N. P. et al. Aerobic and anaerobic indices contributing to track endurance cycling performance. Eur. J. Appl. Physiol. Occup. Physiol. 67(2), 150–158 (1993). 7. Astrand, P. O. & Rodahl, K. Textbook of Work Physiology 2nd edn. (McGraw-Hill, 1977). 8. Billat, V., Binsse, V., Petit, B. & Koralsztein, J. P. High level runners are able to maintain a VO2 steady-state below VO2max in an all-out run over their critical velocity. Arch. Physiol. Biochem. 106(1), 38–45 (1998). 9. Balasekaran, G., Govindaswamy, V. V., Kay Peggy, B. P. & Cheo, N. Y. Applied Physiology of Exercise (World Scientific, 2022). 10. Billat, V. L., Blondel, N. & Berthoin, S. Determination of the velocity associated with the longest time to exhaustion at maximal oxygen uptake. Eur. J. Appl. Physiol. Occup. Physiol. 80(2), 159–161 (1999). 11. Billat, V. L. et al. Oxygen kinetics and modelling of time to exhaustion whilst running at various velocities at maximal oxygen uptake. Eur. J. Appl. Physiol. 82(3), 178–187 (2000). 12. Billat, V. L. et al. Intermittent runs at the velocity associated with maximal oxygen uptake enables subjects to remain at maximal oxygen uptake for a longer time than intense but submaximal runs. Eur. J. Appl. Physiol. 81(3), 188–196 (2000). 13. Busso, T. & Chatagnon, M. Modelling of aerobic and anaerobic energy production in middle-distance running. Eur. J. Appl. Physiol. 97(6), 745–754 (2006). 14. Di Prampero, P. E. et al. Energetics of best performances in middle-distance running. J. Appl. Physiol. 74(5), 2318–2324 (1993). 15. Ward-Smith, A. J. The bioenergetics of optimal performances in middle-distance and long-distance track running. J. Biomech. 32(5), 461–465 (1999). 10 Vol:.(1234567890) Scientific Reports | (2023) 13:8006 | https://doi.org/10.1038/s41598-023-31904-1 www.nature.com/scientificreports/ 16. Blondel, N., Berthoin, S., Billat, V. & Lensel, G. Relationship between run times to exhaustion at 90, 100, 120, and 140% of vVO2max and velocity expressed relatively to critical velocity and maximal velocity. Int. J. Sports Med. 22(1), 27–33 (2001). 17. Lakomy, H. K. A. Physiology and biochemstry of sprinting. In The Handbook of Sports Medicine and Science: Running (ed. Hawley, J. A.) (Wiley-Blackwell Science, 2000). 18. Balasekaran, G., Loh, M.K., Boey, P., Ng, Y.C. Running Energy Reserve Index (RERI) as a new model for assessment and prediction of world, elite, sub-elite, and collegiate running performances. Sci. Rep. https:// doi. org/ 10. 1038/ s41598- 023- 29626-5 (2023). 19. Billat, V. & Lopes, P. Indirect methods for estimation of aerobic power. In Physiological Assessment of Human Fitness (eds Maud, P. J. & Foster, C.) 19–38 (Human Kinetics, 2006). 20. Boileau, R. A., & Horswill, C. A. (2003). Composição corporal e esportes: medidas e aplicações para perda e ganho de peso. A ciência do exercício e dos esportes. Porto Alegre. Artmed, 344–65. 21. Bundle, M. W., Hoyt, R. W. & Weyand, P. G. High-speed running performance: A new approach to assessment and prediction. J. Appl. Physiol. 95(5), 1955–1962 (2003). 22. Jones, A. M. & Doust, J. H. A 1% treadmill grade most accurately reflects the energetic cost of outdoor running. J. Sports Sci. 14(4), 321–327 (1996). 23. Howley, E. T., Bassett, T. & Welch, H. G. Criteria for maximal oxygen uptake: review and commentary. Med. Sci. Sports Exerc. 27, 1292–1301 (1995). 24. Hill, D. W. & Rowell, A. Determination of running velocity at VO2max. Med. Sci. Sports Exerc. 28, 114–119 (1996). 25. Beaver, W. L., Wasserman, K. & Whipp, B. J. Improved detection of lactate threshold during exercise using a log-log transforma- tion. J. Appl. Physiol. 59(6), 1936–1940 (1985). 26. Balasekaran, G., Govindaswamy, V. V., Kay Peggy, B. P. & Cheo, N. Y. Applied Physiology of Exercise Laboratory Manual (World Scientific, 2022). 27. O’Connor, R. & O’Connor, B. Female Fitness on Foot: Walking, Jogging, Running, Orienteering (Britwell Books, 2002). 28. Dupont, G., Blondel, N. & Berthoin, S. Time spent at VO2max: A methodological issue. Int. J. Sports Med. 24(4), 291–297 (2003). 29. Robertson, R. J. et al. OMNI scale perceived exertion at ventilatory breakpoint in children: Response normalized. Med. Sci. Sports Exerc. 33(11), 1946–1952 (2001). 30. Jenkins, S. P. R. Sports Science Handbook: I-Z (Multi-Science Publishing, 2005). 31. Bernard, O. et al. Determination of the velocity associated with VO2max. Med. Sci. Sports Exerc. 32(2), 464–470 (2000). 32. Karlsson, J., Astrand, P. O. & Ekblom, B. Training of the oxygen transport system in man. J. Appl. Physiol. 22(6), 1061–1065 (1967). 33. Migita, T. & Hirakoba, K. Effect of different pedal rates on oxygen uptake slow component during constant-load cycling exercise. J. Sports Med. Phys. Fit. 46(2), 189–196 (2006). 34. Garland, S. W., Newham, D. J. & Turner, D. L. The amplitude of the slow component of oxygen uptake is related to muscle contractile properties. Eur. J. Appl. Physiol. 91(2), 192–198 (2004). 35. Yano, T., Yunoki, T., Matsuura, R. & Ogata, H. Effect of exercise intensity on the slow component of oxygen uptake in decremental work load exercise. J. Physiol. Pharmacol. 55(2), 315–324 (2004). 36. Berger, N. J. & Jones, A. M. Pulmonary O2 uptake on-kinetics in sprint- and endurance-trained athletes. Appl. Physiol. Nutr. Metab. 32(3), 383–393 (2007). 37. Hughson, R. L., O’Leary, D. D., Betik, A. C. & Hebestreit, H. Kinetics of oxygen uptake at the onset of exercise near or above peak oxygen uptake. J. Appl. Physiol. 88(5), 1812–1819 (2000). 38. Weyand, P. G. & Bundle, M. W. Energetics of high-speed running: Integrating classical theory and contemporary observations. Am. J. Physiol. Regul. Integr. Comp. Physiol. 288(4), R956–R965 (2005). Acknowledgements Authors would like to thank all participants for volunteering in this study and researchers for collecting data. The results of this study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation, and statement that results of the present study. This research was funded by the National Institute of Education, Nanyang Technological University, Singapore, Research Support for Senior Academic Administrator Grant (RS-SAA 13/17 GB; RS 13/10 GB), and National Institute of Education Academic Research Fund Grant (RI 6/11 GB). Author contributions G.B. and M.K.L. conceptualized and designed the study. G.B. and M.K.L. collected the data with help of research- ers; G.B., M.K.L., P.B., and N.Y.C. analyzed the data; and all authors interpreted the data. All authors contributed to the writing, review, and editing of the manuscript. All authors have read and agreed to the published version of the manuscript. Competing interests The authors declare no competing interests. Additional information Correspondence and requests for materials should be addressed to G.B. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. © The Author(s) 2023

Determination, measurement, and validation of maximal aerobic speed.

05-17-2023

Balasekaran, Govindasamy,Loh, Mun Keong,Boey, Peggy,Ng, Yew Cheo

eng

PMC10688325

Supplementary File 4: list of shoes used in sprint events during the 2021-2022 seasons considered to include AFT AFT was defined as per Healey et al. (2022), whereby a superspike incorporates “a combination of lightweight, compliant and resilient foams (and/or air pods) with a stiff (nylon, PEBA, carbon-fiber) plate”. § Adidas Adizero Prime SP2 § Adidas Adizero Avanti TYO § Asics Metaspeed SP 0 § New Balance FuelCell Sigma SD-X § New Balance FuelCell SuperComp PWR-X § New Balance SuperComp MDX § Nike Air Zoom Maxfly § Nike Air Zoom Victory § Puma evoSPEED Tokyo Nitro § Puma evoSPEED Tokyo Nitro 400

The potential impact of advanced footwear technology on the recent evolution of elite sprint performances.

11-27-2023

Mason, Joel,Niedziela, Dominik,Morin, Jean-Benoit,Groll, Andreas,Zech, Astrid

eng

PMC7796355

sensors Article Combining Radar and Optical Sensor Data to Measure Player Value in Baseball Glenn Healey



 Citation: Healey, G. Combining Radar and Optical Sensor Data to Measure Player Value in Baseball. Sensors 2021, 21, 64. https:// doi.org/10.3390/s21010064 Received: 11 December 2020 Accepted: 21 December 2020 Published: 24 December 2020 Publisher’s Note: MDPI stays neu- tral with regard to jurisdictional clai- ms in published maps and institutio- nal affiliations. Copyright: © 2020 by the author. Li- censee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and con- ditions of the Creative Commons At- tribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). Department of Electrical Engineering and Computer Science, University of California, Irvine, CA 92617, USA; [email protected] Abstract: Evaluating a player’s talent level based on batted balls is one of the most important and difficult tasks facing baseball analysts. An array of sensors has been installed in Major League Baseball stadiums that capture seven terabytes of data during each game. These data increase interest among spectators, but also can be used to quantify the performances of players on the field. The weighted on base average cube model has been used to generate reliable estimates of batter performance using measured batted-ball parameters, but research has shown that running speed is also a determinant of batted-ball performance. In this work, we used machine learning methods to combine a three-dimensional batted-ball vector measured by Doppler radar with running speed measurements generated by stereoscopic optical sensors. We show that this process leads to an improved model for the batted-ball performances of players. Keywords: Bayesian; baseball analytics; machine learning; radar; intrinsic values; forecasting; sensors; batted ball; statistics; wOBA cube 1. Introduction The expanded presence of sensor systems at sporting events has enhanced the enjoy- ment of fans and supported a number of new applications [1–4]. Measuring skill on batted balls is of fundamental importance in quantifying player value in baseball. Traditional measures for batted-ball skill have been based on outcomes, but these measures have a low repeatability due to the dependence of outcomes on variables such as the defense, the ball- park dimensions, and the atmospheric conditions [5,6]. The Major League Baseball (MLB) Statcast system [2] uses Doppler radar to measure parameters that include the initial speed and direction of batted balls. These parameters can be used to compute batted-ball statistics that are more repeatable than traditional statistics [7]. Research has shown that running speed is an important determinant of batter performance that is not measured by the radar sensor [8], but the Statcast system provides running speed data using stereoscopic optical sensors. This data provides the opportunity to improve the capability of batted-ball models by combining the radar measurements with the optical measurements. The objective of this study is to determine whether combining running speed measurements with batted ball measurements can be used to improve the accuracy of models for player performance. Combining data from different sensors has been done successfully for numerous applications [9–15]. In this work, we employ a Bayesian framework and machine learning methods to build a model that combines radar batted ball data and optical running speed data. The approach generalizes a previous method [7] that considered lower-dimensional vectors consisting of only batted ball descriptors derived from a single sensor system. The model uses a nonparametric kernel method [16] to estimate the probability densities in Bayes law for vectors of radar and optical measurements acquired for over one hundred thousand batted-ball observations. A cross-validation process is used to find optimal smoothing parameters for the density estimates. The model utilizes the weighted on base average (wOBA) [17] linear weights model for run value. The result is the wOBA tesseract which represents a batted-ball value as a continuous function of four variables generated by the radar and optical sensors. Separate tesseracts are built to accommodate the effects Sensors 2021, 21, 64. https://doi.org/10.3390/s21010064 https://www.mdpi.com/journal/sensors Sensors 2021, 21, 64 2 of 14 of batter handedness. We present visualizations obtained by taking slices through the tesseracts to demonstrate properties of the model. We show that by including optical measurements for running speed, the new model is significantly more accurate than previous models that only consider measurements for batted-ball parameters. 2. Radar and Optical Sensors Beginning in 2017, the Statcast system employed radar along with optical stereo video sensors to acquire data for each MLB game. The trajectories of pitched and batted balls have been measured by Trackman’s phased-array Doppler radar component of Statcast. The Trackman radar is situated behind home plate and operates in the X-band at approximately 10.5 GHz. This radar system approximates the path of each pitch using a nine-parameter model defined by the pitch’s 3D acceleration which is assumed constant over the trajectory and the 3D velocity and position at a specified point. The system also measures the pitch spin rate from the distribution of Doppler shifts. In addition, the Trackman radar provides an estimate of the initial speed s and the 3D direction of batted balls. The direction is described by the vertical launch angle v, as shown in Figure 1, and the horizontal spray angle h, as shown in Figure 2. The angle v takes on values from −90◦ (straight down) to +90◦ (straight up) while the angle h takes on values from −45◦ (third base (3B) line) to +45◦ (first base (1B) line) for balls in fair territory. The Trackman radar is well suited for tracking the ball, but the Doppler shifts from players are difficult to discern from returns from clutter due to the players’ slower speeds. For this reason, Statcast uses stereoscopic optical video from two arrays of cameras to track the movement of players. These arrays are usually positioned in the stands on the third base side of the field and are time synchronized with the radar. This allows the movement of defenders to be tracked which allows defensive skill to be quantified using measures such as reaction time, route efficiency, and speed. The combined optical and radar sensors can also be used to measure the time from batted ball contact until the batter reaches first base. The success of a batter depends on both the quality of his batted ball contact as measured by the (s, v, h) vectors as well as his running speed as measured by time to first data. In this study we use Statcast radar and optical measurements from every regular- season MLB game during 2018. The data set includes (s, v, h) data for batted balls and associated time to first running speed measurements. For each batter with at least 20 ground balls, we use the average of his three fastest times to first to represent the batter’s time to first speed r. For switch-hitters who can bat both right and left-handed, a separate r value is computed using their batted balls as a right-handed batter and as a left-handed batter. Figure 1. Vertical angle v where v = 0◦ is parallel to the ground plane. Sensors 2021, 21, 64 3 of 14 Figure 2. Horizontal angle h in the plane of the playing field where h = −45◦ is in the direction of third base (3B), h = 0◦ is in the direction of second base (2B), h = 45◦ is in the direction of first base (1B); the three rays intersect at home plate. 3. Learning the Model from Sensor Data 3.1. Bayesian Approach Let b be a d-dimensional vector that can include the (s, v, h) batted-ball parameters and the r speed parameter. A batted ball can result in one of several outcomes Oj such as an out or a home run. Bayes rule [18] can be used to compute the a posteriori probability of an outcome Oj given b as P(Oj|b) = p(b|Oj)P(Oj) p(b) (1) where p(b) and p(b|Oj) are the probability densities for b and b given Oj respectively and P(Oj) is the a priori probability of outcome Oj. We will derive a method that uses the a posteriori probabilities P(Oj|b) to estimate the value of a batted ball given the vector b of sensor measurements. 3.2. Estimating the Conditional Densities In order to compute the a posteriori probabilities P(Oj|b) in Bayes rule we need to estimate the densities p(b|Oj) and p(b). The conditional densities p(b|Oj) have a complex dependence on the measurement vector b. An outcome Oj of a single, for example, can occur for a slowly hit ground ball toward third base or a hard hit line drive to right field. Therefore we use a nonparametric technique known as kernel density estimation [19,20] to learn the densities. In this approach, we use a set of n sensor vectors bi to construct an estimate for p(b) according to bp(b) = 1 n n ∑ i=1 G(b − bi) (2) where G(·) is the Gaussian kernel G(b) = 1 (2π)d/2|Σ|1/2 exp  −1 2bTΣ−1b  (3) where Σ is a diagonal covariance matrix defined by d parameters which determine the amount of smoothing for each element of the b vector. 3.3. Optimizing the Smoothing Parameters The d diagonal elements of the matrix Σ play an important role in determining the accuracy of bp(b) in Equation (2) [18]. If these smoothing parameters are too small then bp(b) Sensors 2021, 21, 64 4 of 14 will be composed of spikes near the bi samples and if these parameters are too large then the resulting bp(b) will be overly smooth. Cross-validation techniques have been developed to optimize the smoothing parameters by maximizing the likelihood of a set of bi vectors after building the estimate using other bi vectors [21]. An example of these techniques is leave-one-out cross-validation [16] in which the likelihood of each sample is computed after using the other samples to compute the kernel density estimate. We will take a similar but more efficient approach in this work to accommodate the size of our data set. Let σ be the d-dimensional vector of diagonal elements of Σ. We partition the n measured bi vectors into an odd group and an even group depending on whether the vector was acquired in a game starting on an odd or even day of the month. Let nv be the smaller of the sizes of the two groups. The validation set SO is defined as the first nv vectors bi from the odd group and the validation set SE is defined as the first nv vectors bi from the even group. For set SO, we find bp(b) using the n − nv vectors bi that are not in SO as a function of the vector σ. The optimal σ for SO is defined as the vector σ∗ O that maximizes the pseudolikelihood [21,22] given by σ∗ O = arg max σ ∏ bi∈SO bp(bi). (4) This process is repeated to find the vector σ∗ E that maximizes the pseudolikelihood for SE. The optimized smoothing vector σ∗ is found by averaging σ∗ O and σ∗ E . 3.4. Computing Batted Ball Values Each a posteriori probability P(Oj|b) can be estimated using Bayes rule. The estimates for the densities p(b) and p(b|Oj) in Equation (1) are generated using Equations (2) and (3) where the model data for p(b) includes all n vectors bi and the model data for each p(b|Oj) is defined by the subset of the bi vectors with outcome Oj. We use the optimized σ∗ smoothing vector derived using the method in Section 3.3 for each case. The a priori probabilities P(Oj) are estimated as nj/n where nj is the number of the n vectors bi with outcome Oj. Using these estimates, P(Oj|b) is computed using Equation (1). Many statistics such as batting average, on-base percentage, slugging average, and on-base plus slugging have been defined to quantify offensive value [23]. Each of these statistics has certain deficiencies [17]. Batting average and on-base percentage, for example, assume that all hits such as singles and doubles are equally valuable. Slugging average overweights the value of extra-base hits (doubles, triples, home runs) compared to singles. On-base plus slugging places too much value on slugging average relative to on-base percentage. Weighted on base average (wOBA) [17] overcomes these deficiencies by weighting each possible outcome according to its run value. This property has made wOBA one of the most popular and useful offensive statistics [24]. Using wOBA each of the possible batted ball outcomes Oj can be assigned a numerical value which allows the P(Oj|b) probabilities to be used to compute a single expected value for b. This is implemented using wOBA by multiplying each outcome by its average run value wj. Thus, we can represent the expected value of a batted ball as wOBA(b) = 5 ∑ j=0 wjP(Oj|b) (5) where O0 = out, O1 = single, O2 = double, O3 = triple, O4 = home run, and O5 = batter reaches on error (ROE). The wj weights for MLB are compiled for each year at [25]. In this project, we process 2018 data for which the weights are w0 = 0.000, w1 = 0.880, w2 = 1.247, w3 = 1.578, w4 = 2.031, and w5 = 0.920. If b is the three-dimensional vector b = (s, v, h) of batted-ball parameters, then the wOBA(b) function in Equation (5) can be represented by the wOBA cube. If b is the four-dimensional vector b = (s, v, h, r) of batted ball and running speed parameters, then the wOBA(b) function in Equation (5) can be represented by the four-dimensional wOBA Sensors 2021, 21, 64 5 of 14 tesseract. We will provide examples of the wOBA cube in this section and will analyze the wOBA tesseract in detail in Section 4. Figures 3 and 4 examine one-dimensional slices through the wOBA cube. Figure 3 plots wOBA(b) for ground balls with a vertical angle of −5◦ that are hit at 85 and 93 miles per hour. Minima in the two curves correspond to the typical position of infielders with the minima from left to right corresponding to the third baseman, shortstop, second baseman, and first baseman respectively. Over most horizontal angles, balls hit at 93 mph have a higher value than balls hit at 85 mph since ground balls hit at a higher speed have a higher probability of eluding a defender. 0 0.2 0.4 0.6 0.8 1 -40 -20 0 20 40 wOBA h (degrees) speed 93 speed 85 Figure 3. Weighted on base average (wOBA) for a batted ball with a vertical angle v of −5◦ for speed s of 85 miles per hour and 93 miles per hour. Figure 4 plots wOBA(b) for balls hit in the air with a vertical angle of v = +16◦ at the same two speeds. Minima in these curves correspond to the typical position of outfielders with the minima near −20◦, 0◦, and 20◦ corresponding to the left fielder, center fielder, and right fielder respectively. For this vertical angle, balls hit in the direction of an outfielder have a higher value for a speed of 85 mph because these balls often fall in front of the outfielder for hits while balls hit at 93 mph more frequently carry to the outfielder for outs. For both the ground balls and fly balls, the largest wOBA values occur for balls hit near the foul lines (|h| = 45◦) which often result in extra-base hits instead of singles. Fielder positioning is dependent on whether a batter is right-handed or left-handed. For this reason, we partition the measured b vectors by batter handedness and learn two separate wOBA(b) functions: wOBAl(b) for left-handed batters and wOBAr(b) for right- handed batters. As an example, Figure 5 plots wOBAl(b) and wOBAr(b) as a function of the horizontal angle h for a batted ball with a vertical angle v of −5◦ and a speed s of 93 miles per hour. Each curve has four minima which correspond to the typical location of the four infielders. Each of these typical locations is shifted a few degrees to the left for right-handed batters due to fielder positioning. The value of wOBAl(b) or wOBAr(b) will be referred to as the intrinsic value of the batted ball. Sensors 2021, 21, 64 6 of 14 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 -40 -20 0 20 40 wOBA h (degrees) speed 93 speed 85 Figure 4. Weighted on base average (wOBA) for a batted ball with a vertical angle v of 16◦ for speed s of 85 miles per hour and 93 miles per hour. 0 0.2 0.4 0.6 0.8 1 1.2 -40 -20 0 20 40 wOBA h (degrees) LHB RHB Figure 5. Weighted on base average (wOBA) for a batted ball with a vertical angle v of −5◦ and a speed s of 93 miles per hour for left-handed batters (LHB) and right-handed batters (RHB). Sensors 2021, 21, 64 7 of 14 3.5. Player Statistics A player’s performance on batted balls is measured by statistics that are compiled over a period of time. Each batted ball can be assigned the weight wj based on its outcome as described in Section 3.4. This outcome-based value depends on variables such as the defense, the atmospheric conditions, the ballpark dimensions, and random noise which are independent of batter skill. Let O denote the average of a player’s outcome-based values on batted balls over a period of time. The statistic O is also known as wOBA on contact or wOBAcon. A player’s intrinsic values are based on parameters (s, v, h, r) that a player has direct control over. The average of these intrinsic values over time has been shown to have a significantly higher degree of repeatability than the average O of the outcome-based values [7]. We refer to the average of a batter’s intrinsic values computed using the three-dimensional vector b = (s, v, h) of batted-ball parameters as I3 and we refer to the average of a batter’s intrinsic values using the four-dimensional vector b = (s, v, h, r) that also includes his time to first estimate r as I4. 4. wOBA Tesseract In previous work [8] we showed that players who outperform their I3 wOBAcon estimate tend to be faster runners, and many players who underperform their I3 are slower runners. This motivates augmenting the wOBA cube with batter running speed to generate the wOBA tesseract. 4.1. Time to First Measurements The Statcast system generates multiple measurements of running speed. Statcast measures sprint speed, which is derived from a runner’s fastest one second window on individual plays, and time to first which measures the time from batted ball contact to when the batter touches first base. For our application we use time to first, which includes factors such as a batter’s time to recover from the swing and start initial acceleration which affects his ability to beat out a hit. As described in Section 2, we define the running speed parameter r for batters with at least 20 ground balls as the average of the player’s three fastest measured times to first. For switch-hitters a separate r value is computed for plate appearances as a right-handed and as a left-handed batter. All other things being equal, we would expect left-handed batters to have smaller r values because they start closer to first base. For the 2018 season, the average r value over 207 qualifying left-handed batters was 4.245 s and the average r value over 319 qualifying right-handed batters was 4.305 s. Tables 1 and 2 present the left-handed and right-handed batters with the fastest r values for 2018. Figure 6 plots wOBA as a function of r for right-handed and left-handed batters for all batted balls with a vertical angle of less than 10 degrees in 2018. These are ground balls for which the r value is most relevant. We see that there is a strong dependence of batted ball value on running speed as wOBA decreases as r increases. We also see that right-handed batters have a higher wOBA for a given r since a higher fraction of ground balls from RHB are hit to the left side of the infield which requires a longer throw to first base. Table 1. Fastest time to first (r) for left-handed batters (LHB) in seconds, 2018. LHB Time to First (r) Dee Gordon 3.807 Billy Hamilton 3.814 Roman Quinn 3.824 Magneuris Sierra 3.836 Cody Bellinger 3.879 Jack Shuck 3.882 Brett Gardener 3.909 Mallex Smith 3.929 Sensors 2021, 21, 64 8 of 14 Table 2. Fastest time to first (r) for right-handed batters (RHB) in seconds, 2018. RHB Time to First (r) Delino DeShields 3.855 Dansby Swanson 3.884 Trea Turner 3.896 Jose Altuve 3.896 Harrison Bader 3.899 Starling Marte 3.904 Scott Kingery 3.923 Adam Engel 3.929 3.8 3.9 4 4.1 4.2 4.3 4.4 4.5 4.6 4.7 Time to First (seconds) 0.08 0.09 0.1 0.11 0.12 0.13 0.14 0.15 0.16 0.17 0.18 wOBA RHB LHB Figure 6. Weighted on base average (wOBA) versus time to first (r) in seconds over all batted balls with a vertical angle v < 10◦ for right-handed batters (RHB) and left-handed batters (LHB) in 2018. 4.2. Tesseract Examples The wOBA tesseract defines the mapping from (s, v, h, r) to intrinsic value. A separate wOBA tesseract was generated for right-handed and left-handed batters by applying the process described in Section 3 to 63,301 batted ball and time to first measurements for right-handed batters and 44,247 measurements for left-handed batters acquired during the 2018 MLB regular season. Figures 7 and 8 provide examples of slices through the tesseract. -50 -40 -30 -20 -10 0 10 20 30 40 50 Horizontal Angle (degrees) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 wOBA Time to First 4.0 seconds Time to First 4.4 seconds Figure 7. Weighted on base average (wOBA) for right-handed batter (RHB) batted balls with a speed s of 87 miles per hour and a vertical angle v of −9◦ for two time to first (r) values. Sensors 2021, 21, 64 9 of 14 -40 -30 -20 -10 0 10 20 30 40 Horizontal Angle (degrees) 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 wOBA Time to First 4.0 seconds Time to First 4.4 seconds Figure 8. Weighted on base average (wOBA) for left-handed batter (LHB) batted balls with a speed s of 97 miles per hour and a vertical angle v of −12◦ for two time to first (r) values. Figure 7 plots wOBA(b) for right-handed batters for two different values of r as a function of the horizontal spray angle h with the initial batted ball speed and vertical launch angle fixed at s = 87 mph and v = −9◦. The red curve corresponds to a faster than average time of r = 4.0 seconds and the black curve corresponds a slower than average time of r = 4.4 seconds. The four minima in the curves correspond to the typical position of the four infielders against right-handed batters. Near these minima we have a ground ball hit directly at an infielder and the wOBA values are similar for the different values of r. As we move away from the minima we see that a faster runner (red curve) tends to produce a higher wOBA. We see that the largest wOBA values are observed for ground balls hit near the first base line as this horizontal angle is often undefended against right-handed batters and balls down the line may go for extra bases. Figure 8 plots wOBA(b) for left-handed batters for two different values of r as a function of the horizontal spray angle h with the initial batted ball speed and vertical launch angle fixed at s = 97 mph and v = −12◦. The red curve corresponds to a faster than average time of r = 4.0 seconds and the black curve corresponds a slower than average time of r = 4.4 seconds. The four minima in the curves correspond to the typical position of the four infielders against left-handed batters. We see that the minima are shifted to the right compared to the minima for right-handed batters shown in Figure 7. Near three of these minima the wOBA values are similar for the different values of r. For a ground ball hit directly at the third baseman near h = −28◦, a faster runner enjoys an advantage since the third baseman will often be playing shallower to defend against a bunt for the faster runner and a 97 mph ground ball has a better chance of resulting in a hit. As we move away from the minima we see that a faster runner (red curve) tends to produce a higher wOBA. We see that the largest wOBA values are observed for ground balls hit near the third base line as this horizontal angle is often undefended against left-handed batters and balls down the line may go for extra bases. 4.3. Comparing I3 and I4 We computed the I3 (wOBA cube) and I4 (wOBA tesseract) estimates of wOBAcon for all batters in 2018 with at least 250 balls in play. Table 3 is a list of the I3 leaders. These batters are known for their high quality of contact. Table 4 is a list of the I4 leaders which factors running speed in addition to quality of contact into the value of each batted ball. We see that several of the slower runners (Gallo, Martinez, Judge, Goldschmidt) have a lower I4 than I3 while several of the faster runners (Trout, Story, Yelich, Betts) have a higher I4 than I3. The value of I4 − I3 depends on both the batter’s running speed parameter r and his particular collection of batted balls. Sensors 2021, 21, 64 10 of 14 Table 3. Weighted on base average (wOBA) cube (I3) leaders for 2018. Batter I3 Joey Gallo 0.597 Aaron Judge 0.544 Julio Martinez 0.544 Mike Trout 0.541 Paul Goldschmidt 0.531 Matt Carpenter 0.527 Giancarlo Stanton 0.524 Christian Yelich 0.522 Table 4. Weighted on base average (wOBA) tesseract (I4) leaders for 2018, difference between wOBA cube and wOBA tesseract values (I4 − I3), and time to first (r) in seconds. Batter I4 I4 − I3 Time to First (r) Joey Gallo 0.589 −0.008 4.319 Mike Trout 0.542 +0.001 4.062 Julio Martinez 0.535 −0.009 4.340 Aaron Judge 0.534 −0.010 4.487 Trevor Story 0.529 +0.015 3.955 Christian Yelich 0.527 +0.005 4.080 Mookie Betts 0.526 +0.007 4.055 Paul Goldschmidt 0.522 −0.009 4.309 Table 5 is a list of the batters with the highest I4 − I3 for 2018. These are the batters that would be expected to have the largest gain in wOBAcon due to their running speed given their collection of batted balls. We see that all of these players have better than average values of the running speed parameter r. Note that for switch hitters two values (L/R) of r are used. Table 5. Largest differences between weighted on base average (wOBA) cube and wOBA tesseract values (I4 − I3) for 2018 and time to first (r) in seconds; two r values are given for switch-hitters. Batter I4 − I3 Time to First (r) Cody Bellinger 0.025 3.879 Ozzie Albies 0.022 3.936/3.942 Niko Goodrum 0.019 4.08/4.022 Rougned Odor 0.018 3.984 Dansby Swanson 0.018 3.884 Odubel Herrera 0.017 3.969 Scott Kingery 0.017 3.923 Brandon Nimmo 0.017 4.113 Table 6 is a list of the batters with the lowest I4 − I3 for 2018. These are the batters that would be expected to have the largest loss in wOBAcon due to their running speed parameter r given their collection of batted balls. We see that all of these players have worse than average values of r. Sensors 2021, 21, 64 11 of 14 Table 6. Smallest differences between weighted on base average (wOBA) cube and wOBA tesseract values (I4 − I3) for 2018 and time to first (r) in seconds; two r values are given for switch-hitters. Batter I4 − I3 Time to First (r) Yasmani Grandal −0.035 4.663/4.966 Victor Martinez −0.034 4.634/4.965 Kendrys Morales −0.031 4.788/4.816 Justin Bour −0.029 4.498 Chris Davis −0.027 4.491 Albert Pujols −0.025 4.839 Yangervis Solarte −0.022 4.556/4.649 Joey Votto −0.022 4.575 4.4. Variance Reduction Differences between a batter’s observed wOBAcon O and his I3 are due to several factors including running speed, susceptibility to shifts, the ballpark, the weather, and random noise. By developing the I4 statistic we improve the accuracy of the estimate by explicitly modeling the dependence of each batted ball on the running speed parameter r. Table 7 is a list of the batters with at least 250 batted balls with the highest O − I3. We see that each of these batters had a faster than average running speed r. In addition, several of these batters, such as Carlos Gonzalez and Trevor Story in Colorado, benefited from their home ballparks [6]. We see that in each case the use of the wOBA tesseract to generate I4 improved the accuracy of the model as O − I4 is less than O − I3. Table 7. Largest differences between observed weighted on base average (wOBA) on contact (O) and wOBA cube values (O − I3) for 2018; differences between O and wOBA tesseract values (O − I4); and time to first (r) in seconds; two r values are given for switch-hitters. Batter O − I3 O − I4 Time to First (r) Carlos Gonzalez 0.063 0.054 4.150 Ronald Acuna 0.051 0.039 3.945 Mallex Smith 0.050 0.039 3.929 Brandon Nimmo 0.049 0.033 4.113 Chris Taylor 0.048 0.039 4.017 Trevor Story 0.045 0.030 3.955 Eddie Rosario 0.045 0.029 3.969 Yoan Moncada 0.045 0.029 4.094/4.175 Table 8 is a list of the batters with at least 250 batted balls with the lowest O − I3. We see that each of these batters had a slower than average running speed r except Joe Panik who was slightly better than average. Several of these players (Morales, Moreland, Calhoun, Martinez, Carpenter) were shifted on during a large fraction of their plate appearances. We see that in each case the use of the wOBA tesseract to generate I4 improved the accuracy of the model as |O − I4| is less than |O − I3|. If we consider all of the players with at least 250 batted balls in 2018, the R-squared for the set of points (O, I3) is 0.79 and the R-squared for the set of points (O, I4) is 0.85. Therefore, the model that includes running speed using the r parameter has increased accuracy for representing a batter’s wOBAcon. We therefore expect that I4 is a better estimate of wOBAcon skill and provides more value for projection [7]. Sensors 2021, 21, 64 12 of 14 Table 8. Smallest differences between observed weighted on base average (wOBA) on contact (O) and wOBA cube values (O − I3) for 2018; differences between O and wOBA tesseract values (O − I4); and time to first (r) in seconds; two r values are given for switch-hitters. Batter O − I3 O − I4 Time to First (r) Kendrys Morales −0.064 −0.033 4.788/4.816 Mitch Moreland −0.063 −0.052 4.262 Kole Calhoun −0.058 −0.045 4.315 Nelson Cruz −0.055 −0.049 4.395 Albert Pujols −0.054 −0.029 4.839 Victor Martinez −0.052 −0.018 4.634/4.965 Matt Carpenter −0.048 −0.037 4.281 Joe Panik −0.047 −0.046 4.241 5. Discussion Player valuation is a critical task for professional baseball teams that operate in an environment where player contracts are frequently worth tens of millions of dollars. Many statistics have been developed to quantify the offensive value of players. During the twentieth century these statistics, for example batting average, on base average, and slugging percentage, were based on outcomes such as whether the offensive player got a hit or made an out [23]. These outcomes, however, depend on many variables that are beyond the control of the offensive player such as the opponent fielders, the ballpark dimensions, and the weather. This dependence reduces the reliability of these statistics. The use of outcomes has also made it difficult to separate the impact of the key components that contribute to offensive value: batting skill and running speed. There have been some attempts to isolate the contributions of these components. For example, researchers have attempted to quantify running speed by using metrics like the Bill James speed score [26] which is based on factors that include an offensive player’s number of triples and stolen base attempts. But such a measure depends on factors besides running speed namely a player’s power-hitting ability and how often his team’s manager calls for stolen base attempts. Starting with the PITCHf/x system [27], sensors have been available in all MLB ballparks to recover the 3D trajectory of pitched balls since 2008. The collection of sensors has evolved and expanded and the current system, Statcast [2], consists of multiple sensor types that collect seven terabytes of data during each MLB game. Large sets of sensor data provide benefits for measurement especially in the ability to reduce the variance of estimators [28]. In addition, sensor data has enabled the discovery and measurement of new skills. Pitch trajectory data, for example, uncovered the large role that a catcher plays in determining the probability that a pitch is called a strike. This led to the quantification of a new skill called pitch framing [29] that is highly valued in the sport. Sensor data has also led to advances in the quantification of defense [30] and pitch sequencing [31]. The measurement of batted ball vectors has enabled the calculation of batting statistics that are more reliable than statistics that depend on outcomes [7]. The ability to measure running speed enables new insights into how different skill components affect offensive performance. New sensor systems [32] are becoming available that measure biomechanical data for batters and pitchers which will increase understanding of how players achieve given levels of performance [33]. These measurements can also be used to improve the level of detail of models for predicting the result of matchups [34,35]. The ability to derive models from large sets of sensor data has been enhanced by recent advances in machine learning methods [36–38]. The discrete nature of baseball makes its analysis highly amenable to these methods [39]. For many applications [40,41] the use of nonparametric models enables the recovery of functions with a complex dependence on a set of variables. In this work, we use nonparametric density estimates [16] in a Bayesian framework [18] to model a player’s offensive performance using batted ball vectors and Sensors 2021, 21, 64 13 of 14 running speed measurements generated by radar and optical sensors. We show that by applying machine learning methods to a large set of measurements acquired by multiple sensors we obtain a model with significant advantages over previous models for representing a player’s offensive performance. 6. Conclusions Analytical models in baseball have proven valuable for applications involving strat- egy [17,31,34,35], player development [33], and player evaluation [42,43]. We have com- bined data acquired by radar and optical sensors to generalize the 3D wOBA cube to the 4D wOBA tesseract. The new model accounts for the impact of batter running speed and is significantly more accurate than previous models. Thus, the use of multiple sensors enables the generation of a model that is more accurate than the model that is obtained by using either sensor in isolation. This accuracy enables the computation of offensive statistics that more reliably assess talent level on batted balls and support more accurate projections of future performance. This approach also allows separation of the impact of batted-ball skill and running speed in offensive value. An important advantage of this separation is that each skill can be regressed and projected using individual reliability and aging curves before conversion to projected offensive value during forecasting [44]. The wOBA tesseract also has the potential to improve defensive metrics by quantifying the relationship between the batter’s running speed and the difficulty of a play. We have shown that the wOBA tesseract enables visualizations that provide insights into the mapping between batted-ball and running speed parameters and intrinsic value. The process of combining sensor data and machine learning techniques to generate new statistics can be readily adapted to support other areas of sports analytics. Funding: This research received no external funding. Acknowledgments: I thank Travis Petersen at MLB Advanced Media for providing Statcast data that was used in this study. Conflicts of Interest: The author declares no conflict of interest. References 1. Clark, K. The NFL’s Analytics Revolution Has Arrived. Available online: www.theringer.com/nfl/2018/12/19/18148153/nfl- analytics-revolution (accessed on 19 December 2018). 2. Healey, G. The new Moneyball: How ballpark sensors are changing baseball. Proc. IEEE 2017, 105, 1999–2002. [CrossRef] 3. Wang, S.; Xu, Y.; Zheng, Y.; Zhu, M.; Yao, H.; Xiao, Z. Tracking a golf ball with high-speed stereo vision system. IEEE Trans. Instrum. Meas. 2019, 68, 2742–2754. [CrossRef] 4. Woo, M. Artificial Intelligence in NBA Basketball. Available online: www.insidescience.org/news/artificial-intelligence-nba- basketball (accessed on 21 December 2018). 5. Adair, R. The Physics of Baseball, 3rd ed.; Perennial: New York, NY, USA, 2002. 6. Nathan, A. Baseball at High Altitude. Available online: http://baseball.physics.illinois.edu/Denver.html (accessed on 27 November 2020). 7. Healey, G. Learning, visualizing, and assessing a model for the intrinsic value of a batted ball. IEEE Access 2017, 5, 13811–13822. [CrossRef] 8. Healey, G. The Intrinsic Value of a Batted Ball. Available online: https://tht.fangraphs.com/the-intrinsic-value-of-a-batted-ball/ (accessed on 17 March 2016). 9. Bergamini, E.; Ligorio, G.; Summa, A.; Vannozzi, G.; Cappozzo, A.; Sabatini, A. Estimating orientation using magnetic and inertial sensors and different sensor fusion approaches: Accuracy assessment in manual and locomotion tasks. Sensors 2014, 14, 18625–18649. [CrossRef] [PubMed] 10. Frolik, J.; Abdelrahman, M.; Kandasamy, P. A confidence-based approach to the self-validation, fusion and reconstruction of quasi-redundant sensor data. IEEE Trans. Instrum. Meas. 2001, 50, 1761–1769. [CrossRef] 11. Marquez, D.; Felix, P.; Garcia, C.; Tejedor, J.; Fred, A.; Otero, A. Positive and negative evidence accumulation clustering for sensor fusion: An application to heartbeat clustering. Sensors 2019, 19, 4635. [CrossRef] 12. Tian, Q.; Wang, K.; Salcic, Z. A resetting approach for INS and UWB sensor fusion using particle filter for pedestrian tracking. IEEE Trans. Instrum. Meas. 2020, 69, 5914–5921. [CrossRef] 13. Vadakkepat, P.; Jing, L. Improved particle filter in sensor fusion for tracking randomly moving object. IEEE Trans. Instrum. Meas. 2006, 55, 1823–1832. [CrossRef] Sensors 2021, 21, 64 14 of 14 14. Wang, H.; Li, S.; Song, L.; Cui, L.; Wang, P. An enhanced intelligent diagnosis method based on multi-sensor image fusion via improved deep learning network. IEEE Trans. Instrum. Meas. 2020, 69, 2648–2657. [CrossRef] 15. Wang, J.; Xie, J.; Zhao, R.; Mao, K.; Zhang, L. A new probabilistic kernel factor analysis for multisensory data fusion: Application to tool condition monitoring. IEEE Trans. Instrum. Meas. 2016, 65, 2527–2537. [CrossRef] 16. Sheather, S. Density estimation. Stat. Sci. 2004, 19, 588–597. [CrossRef] 17. Tango, T.; Lichtman, M.; Dolphin, A. The Book: Playing the Percentages in Baseball; Potomac Books: Dulles, VA, USA, 2007. 18. Duda, R.; Hart, P.; Stork, D. Pattern Classification; Wiley-Interscience: New York, NY, USA, 2001. 19. Parzen, E. On estimation of a probability density function and mode. Ann. Math. Stat. 1962, 33, 1065–1076. [CrossRef] 20. Rosenblatt, M. Remarks on some nonparametric estimates of a density function. Ann. Math. Stat. 1956, 27, 832–837. [CrossRef] 21. Guidoum, A.C. Kernel Estimator and Bandwidth Selection for Density and Its Derivatives. The Kedd Package, Version 1.03, October 2015. Available online: https://cran.r-project.org/web/packages/kedd/vignettes/kedd.pdf (accessed on 27 November 2020). 22. Duin, R. On the choice of smoothing parameters for Parzen estimators of probability density functions. IEEE Trans. Comput. 1976, C-25, 1175–1179. [CrossRef] 23. Panas, L. Beyond Batting Average; Lulu Press: Morrisville, NC, USA, 2010. 24. Slowinski, S. wOBA. Available online: https://library.fangraphs.com/offense/woba/ (accessed on 15 February 2010). 25. wOBA and FIP Constants. Available online: www.fangraphs.com/guts.aspx?type=cn (accessed on 27 November 2020). 26. James, B. The Bill James Baseball Abstract 1987; Ballantine Books: New York, NY, USA, 1987. 27. Fast, M. What the heck is PITCHf/x? In The Hardball Times Baseball Annual, 2010; Distelheim, J., Tsao, B., Oshan, J., Bolado, C., Jacobs, B., Eds.; The Hardball Times; ACTA Sports: Chicago, IL, USA, 2010; pp. 153–158. 28. Papoulis, A. Probability, Random Variables, and Stochastic Processes, 3rd ed.; McGraw-Hill: New York, NY, USA, 1991. 29. Lindbergh, B. The Art of Pitch Framing. Available online: https://grantland.com/features/studying-art-pitch-framing-catchers- such-francisco-cervelli-chris-stewart-jose-molina-others/ (accessed on 16 May 2013). 30. Tango, T. Introducing Infield Outs Above Average. Available online: https://technology.mlblogs.com/introducing-infield-outs- above-average-6467e61a98dc (accessed on 13 January 2020). 31. Healey, G.; Zhao, S. Using PITCHf/x to model the dependence of strikeout rate on the predictability of pitch sequences. J. Sports Anal. 2017, 3, 93–101. [CrossRef] 32. Lemire, J. KinaTrax’s Magic Leap: A New Way to See Data. Available online: https://sporttechie.com/mlb-kinatrax-ar- biomechanics-baseball-data (accessed on 3 December 2019). 33. Lindbergh, B.; Sawchik, T. The MVP Machine: How Baseball’s New Nonconformists Are Using Data to Build Better Players; Basic Books: New York, NY, USA, 2019. 34. Healey, G. Modeling the probability of a strikeout for a batter/pitcher matchup. IEEE Trans. Knowl. Data Eng. 2015, 27, 2415–2423. [CrossRef] 35. Healey, G. Matchup models for the probability of a ground ball and a ground ball hit. J. Sports Anal. 2017, 3, 21–35. [CrossRef] 36. Chen, Z.; Zhang, H. Learning implicit fields for generative shape modeling. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, CA, USA, 16–20 June 2019; pp. 5939–5948. 37. Healey, G.; Zhao, S. Learning and applying a function over distributions. IEEE Access 2020, 8, 172196–172203. [CrossRef] 38. Mescheder, L.; Oechsle, M.; Niemeyer, M.; Nowozin, S.; Geiger, A. Occupancy networks: Learning 3D reconstruction in function space. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Long Beach, CA, USA, 16–20 June 2019; pp. 4460–4470. 39. Koseler, K.; Stephan, M. Machine learning applications in baseball: A systematic literature review. Appl. Artif. Intell. 2017, 31, 745–763. [CrossRef] 40. Khalily, M.; Brown, T.; Tafazolli, R. Machine-learning based approach for diffraction loss variation prediction by the human body. IEEE Antennas Wirel. Propag. Lett. 2019, 18, 2301–2305. [CrossRef] 41. Wang, W.; Han, W.; Na, X.; Gong, J.; Xi, J. A probabilistic approach to measuring driving behavior similarity with driving primitives. IEEE Trans. Intell. Veh. 2020, 5, 127–138. [CrossRef] 42. Healey, G. A Bayesian method for computing intrinsic pitch values using kernel density and nonparametric regression estimates. J. Quant. Anal. Sport. 2019, 15, 59–74. [CrossRef] 43. Sawchik, T. Big Data Baseball; Flatiron Books: New York, NY, USA, 2016. 44. Bradbury, J.C. Peak athletic performance and ageing: Evidence from baseball. J. Sport. Sci. 2009, 27, 599–610. [CrossRef] [PubMed]

Combining Radar and Optical Sensor Data to Measure Player Value in Baseball.

12-24-2020

Healey, Glenn

eng

PMC7729596

sensors Article Physical Demands of U10 Players in a 7-a-Side Soccer Tournament Depending on the Playing Position and Level of Opponents in Consecutive Matches Using Global Positioning Systems (GPS) Antonio Hernandez-Martin 1 , Javier Sanchez-Sanchez 2,* , Jose Luis Felipe 2 , Samuel Manzano-Carrasco 1 , Carlos Majano 1 , Leonor Gallardo 1 and Jorge Garcia-Unanue 1 1 IGOID Research Group, Physical Activity and Sport Sciences Department, University of Castilla-La Mancha, 45071 Toledo, Spain; [email protected] (A.H.-M.); [email protected] (S.M.-C.); [email protected] (C.M.); [email protected] (L.G.); [email protected] (J.G.-U.) 2 School of Sport Sciences, Universidad Europea de Madrid, 28670 Villaviciosa de Odón, Spain; [email protected] * Correspondence: [email protected] Received: 19 November 2020; Accepted: 4 December 2020; Published: 6 December 2020



 Abstract: The aim of this study was to analyse the physical demands of U10 players in a 7-a-side-soccer tournament based on the playing positions in 6 consecutive matches by global positioning systems (GPS). Variables of total distance, relative distance in different speed zones, maximum speed, time interval between accelerations, maximum speed acceleration, maximum acceleration, acceleration distance and the number of high-intensity accelerations were analysed. Differences between playing positions were found in the total distance covered by the midfielders. They covered higher total distances than the defenders (+1167 m; 95% CI: 411 to 1922 m; effect size (ES) = 1.41; p < 0.05) and forwards (+1388 m; CI 95%: 712 a 2063 m; TE = 0.85; p < 0.05). The total covered distance increased in the final rounds with respect to the group stage (p < 0.05; ES: 0.44 to 1.62), and high-intensity actions, such as the number of accelerations, were greater in the final rounds compared to the group stage (p < 0.05; ES: 0.44 to 1.62). The physical performance of young football players in a tournament with consecutive matches on a 40 × 62 m football field on the same day is influenced by the playing position and dependent on the level difference between opponents. Keywords: tracking system; U10 soccer tournament; load; match analysis; players positions 1. Introduction Football requires high-intensity actions such as jumps, changes of direction, accelerations, decelerations and shots, interspersed with short recovery periods [1–3]. Physical demands such as aerobic and anaerobic endurance, agility or speed [3] and physiological demands such as heart rate, blood lactate concentration or RPE [4] and inter-limb strength asymmetry [5] present a critical influence in the performance of football players. Over recent years, scientific literature has focused on the movement patterns and physiological requirements of football players (both adults and youths) [6–8]. Meanwhile, it has been argued that players in the developmental stages (i.e. under U13), should not be considered as miniature adults [9] and therefore, specific football training programmes should be developed in these stages. For this reason, 7-a-side football emerged in Spain [as well as 8-a-side football in other countries, such as the United Kingdom], is practised during grassroot stages and is well regulated by the Royal Spanish Sensors 2020, 20, 6968; doi:10.3390/s20236968 www.mdpi.com/journal/sensors Sensors 2020, 20, 6968 2 of 11 Football Federation with the purpose of promoting the progression of learning and improving physical and tactical skills in young football players, such as reducing the size of the pitch [10]. This kind of football practice adapted to the movement patterns and physiological demands on young football players positively affects the development of young football players [4]. On the other hand, it is important to know the physical demands in a 7-a-side football game to prepare adequate training programmes for football players in their growing and developmental stage [2]. In light of this, global positioning systems (GPS) and accelerometers have been used to describe the physical profile of the football players in terms of distance and speed variables during friendly matches[11–13], officialmatches[14]andtoquantifythephysicalperformanceineliteyouthplayers[15–17]. In addition, previous research on adult players indicates that a player’s position can influence their physical demands [14] or the degree of fatigue during a match [18]. However, there is little scientific evidence that has used GPS to analyse the physical demands during a U10-category 7-a-side football tournament. Modern elite football currently involves a large number of tournaments and matches throughout the season, and it is not unusual for a team to play two or more matches in a very short period of time [19]. Similarly, in grassroots football, there are many tournaments where usually 6 or more 7-a-side matches are played in the same day. There are reasons to believe that too many games may lead to a lack of motivation, concentration and more incremental fatigue, which can affect coordination, leading to worse performance and an increased risk of injury [20]. However, there is not enough research on this topic. Therefore, the aim of this study was to analyse the physical demands in a 7-a-side-soccer tournament based on the playing positions and level of opponents in consecutive matches. 2. Materials and Methods 2.1. Experimental Approach to the Problem This study was designed to describe and compare distances and movement patterns (measured by GPS) of U10 7-a-side football players during a tournament of 6 games played in less than 24 h with a 1-hour break between match 1 (T1) and match 2 (T2), 40 min between T2 and match 3 (T3), 2 h and 30 min between T3 and the quarter-final match (TQ), 1 h and 35 min between TQ and the semi-final match (TS) and 55 min between TS and the final match (TF). One half of 20 min was played in T1, T2, T3, TQ and TS and one half of 30 min was played in the TF. The tournament was played on the football field of the size 40 × 62 m. 2.2. Sample Six games of a 7-a-side soccer tournament in the central region of Spain played by a U10 amateur soccer team [age = 10.2 ± 0.6 years; height = 136.5 ± 7.4 cm; body mass = 33.2 ± 6.13 kg] were analysed using a GPS with sampling rates of 15 Hz (GPSport, Canberra, Australia). The analysed matches correspond to the analysis of the players with more than 10 min played per match for the same team, with a total of 48 observations of 8 players. There was no limit to the number of substitutions according to the 7-a-side football regulations (the goalkeeper was excluded) from the tournament held on April 20th of the 2018/2019 season. The outfield players were divided into forwards (FWs), midfielders (MFs) and defenders (DFs). All subjects (and their parents or guardians) were carefully informed about the study procedures and about the possible risks and benefits associated with participating in the study, and a signed informed consent was obtained before participating in any procedure related to the study. The Clinical Research Ethics Committee of the Castilla-La Mancha Health Service [Spain] approved this study based on the latest version of the Declaration of Helsinki (Ref.: 489/24022020). 2.3. Equipment Position-tracking system (Figure 1). Global positioning systems (GPS) provide data on the location and time of satellite tracking devices and have previously been used in various investigations [21–24]. Sensors 2020, 20, 6968 3 of 11 At least four satellites orbiting the Earth are required to determine the GPS position trigonometrically and the GPS devices receive information that determines the signal traffic. Together with the addition of triaxial accelerometers, magnetometers and gyroscopes, the data are more accurate. Depending on the location and environmental obstruction, the signal quality may change. GPS devices (15 Hz, Spi Pro X, GPSports, Canberra, Australia) have demonstrated better validity and reliability values than their 1 HZ and 5 HZ predecessors and similar values with 10 Hz [25]. The 15 Hz GPS has proven to be reliable during specific football movements [26]. The GPS devices were installed and placed in a custom-made child’s vest, located at the back of the torso and well-adjusted to the body. Sensors 2020, 20, x FOR PEER REVIEW 3 of 11 trigonometrically and the GPS devices receive information that determines the signal traffic. Together with the addition of triaxial accelerometers, magnetometers and gyroscopes, the data are more accurate. Depending on the location and environmental obstruction, the signal quality may change. GPS devices (15 Hz, Spi Pro X, GPSports, Canberra, Australia) have demonstrated better validity and reliability values than their 1 HZ and 5 HZ predecessors and similar values with 10 Hz [25]. The 15 Hz GPS has proven to be reliable during specific football movements [26]. The GPS devices were installed and placed in a custom-made child’s vest, located at the back of the torso and well-adjusted to the body. Figure 1. Process of capturing the signal and transmitting the global positioning system (GPS) tracking devices. 2.4. Procedures Body mass and height measurements were completed during the last week before the tournament. Body mass was determined using the DXA (Hologic Series Discovery QDR, Software Physician’s Viewer, APEX System Software Version 3.1.2. Bedford, MA, US) following the protocols described in previous research [27]; height (cm) was measured with a scientific height rod (Seca 214, Hamburg, Germany). Before the warm-up (20 min) and before each game, a GPS unit (15 Hz, Spi Pro X, GPSports, Canberra, Australia) was attached to each player’s torso, following the protocols described by Sanchez-Sanchez [28]. A total of six matches were analysed in this study, all of them completed on the same day (20 April 2019, in the central region of Spain). Before the first match, players had rested for <24 h since their last training session or game. The tournament was played on the football field (40 × 62 m) of the organising club, including a group stage with three matches (T1, at 9:10 h: final score 5-0; T2, at 10:30: final score 3-0; T3, at 11:30: final score 2-0); a second-round quarter-final game (TQ at 14:20: final score 3-0); the semi-final (TS, at 16:15: final score 2-0); and the final (TF, at 17:30: final score 0-1). One half of 20 min (T1, T2, T3, TQ and TS) and one half of 20 min plus 10 min additional time were played in the TF. Given the U10 7-a-side football rules (regulated by the Royal Spanish Football Federation), there were unlimited substitutions. The study required football players to complete ≥10 min / game during each match of the tournament to be included in the research. 2.5. Data Processing The GPS software (Team AMS R1 2019.1 software, GPSports, Canberra, Australia) provided information about the total distance (TD) covered during the game and the percentages of distance covered in each one of the six locomotor categories with speed ranges. All players participated in a 10 m sprint test with a 5 m split time and the results were used to calculate speed zones for each player [29]: standing (zone (Z)1: 0–2 km·h−1); walking (Z2: 2–4 km·h−1); easy running (Z3: 4.1–7 km·h−1); fast running (Z4: 7.1–13.0 km·h−1); high-speed running (Z5: 13.1–17 km·h−1); sprinting (Z6: ≥17.1 Figure 1. Process of capturing the signal and transmitting the global positioning system (GPS) tracking devices. 2.4. Procedures Body mass and height measurements were completed during the last week before the tournament. Body mass was determined using the DXA (Hologic Series Discovery QDR, Software Physician’s Viewer, APEX System Software Version 3.1.2. Bedford, MA, US) following the protocols described in previous research [27]; height (cm) was measured with a scientific height rod (Seca 214, Hamburg, Germany). Before the warm-up (20 min) and before each game, a GPS unit (15 Hz, Spi Pro X, GPSports, Canberra, Australia) was attached to each player’s torso, following the protocols described by Sanchez-Sanchez [28]. A total of six matches were analysed in this study, all of them completed on the same day (20 April 2019, in the central region of Spain). Before the first match, players had rested for <24 h since their last training session or game. The tournament was played on the football field (40 × 62 m) of the organising club, including a group stage with three matches (T1, at 9:10 h: final score 5-0; T2, at 10:30: final score 3-0; T3, at 11:30: final score 2-0); a second-round quarter-final game (TQ at 14:20: final score 3-0); the semi-final (TS, at 16:15: final score 2-0); and the final (TF, at 17:30: final score 0-1). One half of 20 min (T1, T2, T3, TQ and TS) and one half of 20 min plus 10 min additional time were played in the TF. Given the U10 7-a-side football rules (regulated by the Royal Spanish Football Federation), there were unlimited substitutions. The study required football players to complete ≥10 min/game during each match of the tournament to be included in the research. 2.5. Data Processing The GPS software (Team AMS R1 2019.1 software, GPSports, Canberra, Australia) provided information about the total distance (TD) covered during the game and the percentages of distance covered in each one of the six locomotor categories with speed ranges. All players participated in a 10 m sprint test with a 5 m split time and the results were used to calculate speed zones for each player [29]: Sensors 2020, 20, 6968 4 of 11 standing (zone (Z)1: 0–2 km·h−1); walking (Z2: 2–4 km·h−1); easy running (Z3: 4.1–7 km·h−1); fast running (Z4: 7.1–13.0 km·h−1); high-speed running (Z5: 13.1–17 km·h−1); sprinting (Z6: ≥17.1 km·h−1). The GPS software also provided information about the number and average distance of the sprints. Sprint time (s) is the average time that athletes’ speed is above 17.1 km·h−1 and sprint distance (m) is the distance covered with a speed above 17.1 km·h−1. In the same way, the GPS devices registered the maximum acceleration peaks and the number of accelerations of the players in different ranges of intensity. During actual match play, this study’s players showed maximal accelerations in the range of 2.7 and 3.0 m·s−2. As a consequence of this, and due to the classification proposed by Osgnach [30], we assumed 2.5 m·s−2, as variable high-intensity accelerations are the accelerations made in the maximum intensity zone (ACCMAX). Also, using the data obtained from the Team AMS software, the average maximum speed of each acceleration (VmaxACC; km·h−1) was calculated, together with the average values of the time interval between accelerations (IACC; s), the average distance travelled in each acceleration greater than 2.5 m·s−2 (TDACC; m) and the number of high-intensity accelerations (n) in each of the matches to facilitate the comparison of results. 2.6. Statistical Analysis Data encoding and data processing were carried out using the SPSS 25.0 statistical package (SPSS Inc., Chicago, IL, USA). The normality of the variables has been analysed with the Shapiro–Wilk test. After a descriptive analysis (means and standard deviations), a comparison test was performed by the analysis of variance (ANOVA) in order to compare the physical performance variables between the three positions, and the repeated measures analysis of variance (repeated measures ANOVA) to compare the physical performance variables between the six matches. A Bonferroni post hoc test was used for pairwise comparisons in the ANOVA test and DMS test for repeated measures ANOVA. Effect size (ES; Cohen’s d) was included and evaluated as follows: 0–0.2 = trivial; 0.2−0.5 = small; 0.5−0.8 = moderate; and >0.8 high. The statistical significance criterion was established at p < 0.05. 3. Results Table 1 shows the results obtained from the GPS according to the playing position. Midfielders had significantly higher values than defenders in TD, Z4, Z5, Z6, ACCMAX and HI acceleration and significantly lower values in Z2 (p < 0.05; ES: 1.01 to 2.85). In addition, midfielders also revealed significantly higher values in TD, Z4, Z5, Z6, VMAX, VMaxACC and HI acceleration and significantly lower values in Z2 and Z3 than the forwards (p < 0.05; ES: 0.85 to 2.41). In Z1, IACC and TDACC no significant differences were found (p > 0.05). In the results obtained from the GPS, differences in the relative distances covered in the matches were identified (p < 0.05; Figure 2). The players showed no significant difference in Zone 6. Players in TQ covered a shorter distance in Zone 5 than in T1 (p < 0.05; ES: 0.93). The results obtained by the players in the distances covered in Zone 4 showed a greater distance covered in T1 compared to TS (p < 0.05; ES: 1.04) and in T2 with respect to TQ, TF (p < 0.05; ES: 0.77 to 1.09) being the greater distance covered in T2 (−8.6%; CI 95%: 3.8 to 13.3%; TE = 1.14; p < 0.05) with respect to TS. The distance travelled by players in lower speed zones showed greater variability between matches. Players in Zone 3 showed higher distances in TQ than TS (p < 0.05; ES: 0.43). Furthermore, in T3 they covered significantly lower total distance than in T1, T2, TQ, TS and TF (p < 0.05; ES: 0.85 to 1.65). Players in Zone 2 travelled a lower distance in T1 than in TQ, TS and TF (p < 0.05; ES: 0.89 to 1.30). Also, in TS and TF the distance they covered in Zone 2 was greater than in T2 (p < 0.05; ES: 0.73 to 0.93); in addition, in TF the players travelled more distance in Zone 2 than in TQ (p < 0.05; ES: 0.37). However, no significant differences were found in the distances covered in Zone 1 by the players in the different matches. Sensors 2020, 20, 6968 5 of 11 Table 1. Differences between match positions in load metrics. Defenders Midfielders Forwards TD (m) 1515.78 ± 800.94 b 2683.13 ± 854.69 c 1294.94 ± 783.26 Distance Z1 (%) 1.50 ± 0.29 1.56 ± 0.36 1.85 ± 0.54 Distance Z2 (%) 30.12 ± 6.08 b 18.87 ± 1.82 c 25.25 ± 6.90 Distance Z3 (%) 43.60 ± 4.38 39.42 ± 3.99 c 43.98 ± 5.69 Distance Z4 (%) 18.89 ± 6.29 b 30.34 ± 2.67 c 23.14 ± 7.28 Distance Z5 (%) 4.77 ± 2.31 b 7.42 ± 2.54 c 4.98 ± 3.19 Distance Z6 (%) 1.12 ± 0.91 b 2.39 ± 1.61 c 0.80 ± 0.92 VMAX (km·h−1) 18.57 ± 3.39 20.85 ± 1.80 c 17.44 ± 3.19 IACC (s) 2.66 ± 0.89 2.84 ± 0.45 2.45 ± 0.63 VMaxACC (km·h−1) 14.51 ± 2.04 14.60 ± 1.26 c 13.10 ± 1.53 ACCMAX (m·s−2) 2.71 ± 0.11 b 2.88 ± 0.15 2.75 ± 0.18 TDACC (m) 9.44 ± 4.70 10.22 ± 2.18 7.91 ± 2.98 HI acceleration (n) 3.91 ± 2.12 b 12.89 ± 5.07 c 3.69 ± 2.56 b = significant differences between defenders and midfielders; c = significant differences between midfielders and forwards; m = metres; % = relative distance percentage; s = seconds; km·h−1 = kilometers per hour; m·s−2 = metres per second squared; n = number of accelerations; TD = total distance; VMAX = maximun speed; IACC = time interval between accelerations; VmaxACC = average maximum speed acceleration; MaxACC = maximum acceleration; TDACC = average distance travelled in acceleration greater than 2.5 m·s−2; HI acceleration = number of high-intensity accelerations. Figure 2. Mean and standard deviation of % distance covered in Zone 1 (standing); Zone 2 (walking); Zone 3 (easy running); Zone 4 (fast running); Zone 5 (high-speed running) and Zone 6 (sprinting) of each tournament match. Match 1 (T1), match 2 (T2) and match 3 (T3): group stage; TQ = quarter final; TS = semi-final and TF = final. b = significant differences with T3. c = significant differences with TQ. d = significant differences with TS. e = significant differences with TF. Sensors 2020, 20, 6968 6 of 11 Figure 3 shows the results of the GPS in relation to TD, VMAX, IACC, VMaxACC, ACCMAX, TDACC and HI acceleration of each match played in the tournament. The players covered a greater total distance in TF than in T3, TQ and TS. The time interval between accelerations (IACC) was greater for players in TF than in TQ. Players achieved higher VMaxACC peaks in TF than in T2 and higher HI acceleration than in T3 and TQ (p < 0.05; ES: 0.44 to 1.62). The players covered a greater total distance in TS than in T3. Players reached higher VMaxACC peaks in TS than in T2 and showed a higher number of high-intensity accelerations (HI acceleration) than in T3 (p < 0.05; ES: 0.78 to 1.50). The players covered a greater total distance in TQ than in T3 (p < 0.05; ES: 0.52). However, the time interval between accelerations (IACC) was shorter in TQ than in T3, as was the total distance travelled in acceleration p < 0.05; ES: 1.57 to 1.65). The players covered a lower total distance in T3 than T2 and T1, the peaks of VMaxACC were lower than in T2 and showed a lower number of high-intensity accelerations (HI accelerations) than in T1 (p < 0.05; ES: 0.74 to 2.01). The players in T2 performed a lower total distance than in T1 (p < 0.05; ES: 1.59). In VMAX and ACCMAX no significant differences were found. Sensors 2020, 20, x FOR PEER REVIEW 6 of 11 TS= semi-final and TF= final. b= significant differences with T3. c= significant differences with TQ. d= significant differences with TS. e= significant differences with TF. Figure 3 shows the results of the GPS in relation to TD, VMAX, IACC, VMaxACC, ACCMAX, TDACC and HI acceleration of each match played in the tournament. The players covered a greater total distance in TF than in T3, TQ and TS. The time interval between accelerations (IACC) was greater for players in TF than in TQ. Players achieved higher VMaxACC peaks in TF than in T2 and higher HI acceleration than in T3 and TQ (p < 0.05; ES: 0.44 to 1.62). The players covered a greater total distance in TS than in T3. Players reached higher VMaxACC peaks in TS than in T2 and showed a higher number of high- intensity accelerations (HI acceleration) than in T3 (p <0.05; ES: 0.78 to 1.50). The players covered a greater total distance in TQ than in T3 (p <0.05; ES: 0.52). However, the time interval between accelerations (IACC) was shorter in TQ than in T3, as was the total distance travelled in acceleration p <0.05; ES: 1.57 to 1.65). The players covered a lower total distance in T3 than T2 and T1, the peaks of VMaxACC were lower than in T2 and showed a lower number of high-intensity accelerations (HI Figure 3. Mean and standard deviation of total distance (TD; m); maximum speed (VMAX; km·h−1); time interval between accelerations (IACC; s); average maximum speed acceleration (VMaxACC; km·h−1); maximum acceleration (MaxACC; m·s−2); average distance travelled in acceleration greater than 2.5 m·s−2 (TDACC; m); n of high-intensity accelerations (HI accelerations; n); T1, T2 and T3: group stage; TQ = quarter final; TS = Semi-final and TF = final. a = significant differences with T2; b = significant differences with T3; c = differences with TQ; d = significant differences with TS; e = significant differences with TF. Sensors 2020, 20, 6968 7 of 11 4. Discussion The current study aimed to describe and analyse the physical demands of U10 7-a-side players during a tournament based on the playing positions in consecutive matches. The main findings indicated that in total distance, as high-intensity distance, midfielders covered more distance; furthermore, high-intensity actions were higher in midfielders compared to defenders and strikers. These differences in the midfielders, accumulated in six matches, can become very important in the development of the different phases of the tournament. By exchanging positions, these demands could be equalised. In addition, the distances covered at high intensity reduced as the tournament progressed; however, in the final rounds the players showed a tendency to increase the total distance and high-intensity actions compared to the initial matches of the tournament. Thus, these results suggest that physical demands during a multi-match tournament on the same day influence a decrease in the performance of the U10 players. The U10 players used 15 Hz GPS devices, which showed higher reliability and validity values than 1 Hz and 5 Hz devices in distance covered at high speed, accelerations and short distances [25]. Previous 15 Hz GPS validation studies recreated football movements and showed a commensurate degree of accuracy in measuring distance by walking, jogging, running and sprinting linearly (CV 2.95–3.16%) and curvilinearly (CV −2.20–1.92%) [26]. The 15 Hz GPS devices showed valid results at the maximum speed reached by the U10 players (< 20 km/h−1), however these devices would not be valid enough to record maximum speed in adults, as reliability decreases by values > 20 km/h−1 [31]. With regard to accelerations, they offer reliability values (CV < 10%) for accelerations of less than 3 m/s, which are those recorded by U10 players, while the reliability is lower (CV = 30%) for accelerations greater than 3 m/s [32]. All of the multi-match tournament analysis data in the current study are novel, as the match analysis of U10 players has not been previously described. In recent years, physical performance in football has been studied during training and competition in male participants [33–35]. These studies have examined different physical parameters, such as total distance covered, sprint and high-intensity movement patterns, patterns which, in central positions, are able to maintain and even increase in three consecutive matches [36,37]. The literature has shown that these physical demands differ between playing positions [38]. Thus, attacking and defending positions are characterised by high-intensity activities, producing the highest sprint distance, and a number of accelerations and decelerations [39,40]. The differences observed in TD between different positions may be explained by the different movement patterns required for each football-specific position. According to previous studies, [41,42] our results revealed that MFs produce the highest TD compared to other positions and DFs produce the lowest TD and high-speed distances. Buchheit et al. [15] observed positional differences in U13–U18 players regarding the distance covered during matches, especially in high-intensity actions. Similar results were also observed [17] in elite youth football players aged 8–18 years. Thus, although our study allows quantifying by positions, it is important to highlight the frequency of coaches interchanging players during the different multi-match tournaments to improve technical and tactical abilities in youth football players. Analysing the results of the distances covered in each zone showed that Zone 3 is where a greater distance was recorded in consecutive matches by players (35–45%) as has been shown previously for adult players [39]. Previous studies showed that jogging is the majority movement pattern in football [29]. Therefore, U10 football players were required to use a longer time performing lower-intensity exercise to recover the match effort made [29]. Similarly, walking distances were of significantly longer relative distance (30%) in TS than in the other phases of the tournament. Conversely, medium-intensity running showed less distance covered (20%) by the participants in TF compared to the other phases. This may be attributed to the physiological demands needed to maintain this kind of velocity run, as it has been demonstrated that during a 30 s all-out cycle sprint the percentage decline in power output is lower in children than in adults [43]. If we compare the muscular characteristics of adults and children, the greater resistance to fatigue shown by children compared to adults could Sensors 2020, 20, 6968 8 of 11 be related, since children have less muscle mass, thus generating less absolute power, and they have lower glycolytic activity and higher muscle oxidative activity [44]. However, the distance covered at high-intensity running was significantly higher in T1 compared to TQ, revealing how players decrease intensity in actions based on the course of the different matches of the tournament as well as increased fatigue and physical capacity [45]. No significant differences were found in standing and sprint distances between the different rounds of the tournament, despite showing a longer duration in the final match, probably due to contextual variables of the matches. This is an interesting finding, as high-intensity and sprint running distance has been described to distinguish the standard of senior football players [46]. In the present study, U10 7-a-side football players covered approximately between 1200 and 2500 m. However, this total distance drops significantly in the last group match, and then again sees an increase in the progression from this phase to the final match. These values are less than those reported by other studies, where the players of this same category covered a greater distance (4056 m) [29]. This could be explained by the longer duration of the matches and the relaxation; once the team manages to move from the group stage there is an increase in intensity as the final phase is reached. However, this does not happen in the same way with other variables—for example, VMAX and ACCMAX remain constant throughout the tournament. This may be due to the energy path required for these efforts, which depends on the capacity of each player, and so the different fatigue-related physiological mechanism appears to operate in different periods of a football game [47]. However, TDACC is greater in the first matches, especially in T3, because the next round of the tournament is at stake. Our results show that in the final rounds of the tournament, the number of accelerations is increased, since the goal of winning the tournament is about to be achieved. In contrast, the distance covered by accelerating is less than that found in the group stage, caused mainly by the high physical load and increased fatigue. The findings are similar to previous studies [28,48] because in U10 players the contribution of the anaerobic metabolism is not as developed as in adults [49], which may help immature players to reduce metabolic stress, but on the other hand may limit their capacity to perform high-intensity actions, especially when inadequate recovery time is provided [48]. There are some limitations in this study, one of which is that the sample was composed of football players belonging to the same team, so more studies are needed to confirm the results obtained in this study, and it would also be interesting to complete the results by analysing more players from several teams at the same time. Furthermore, the system of play used during matches was not taken into account, as match training has been shown to have an impact on very high-intensity running activities with and without the ball in adult players [50], and could also affect the physical performance of the match analysed in this study. The 7-a-side soccer rule on unlimited substitutions of U10 players is also a limitation, as it could induce a great variability among U10 players in terms of distance travelled and acceleration during matches. 5. Conclusions The present study indicates that the physical performance of U10 football players in a tournament with different matches on the same day is influenced by the playing position and is dependent on the level difference between opponents. Midfielders covered more distance in high-intensity zones (Zone 5 and Zone 6) and performed more high-intensity actions (VMaxACC, MaxACC, TDACC and HI accelerations) than defenders and forwards. Regarding the level difference between opponents, the distances covered at high intensity were reduced as the tournament progressed; however, the total distance and high-intensity accelerations are higher in the final rounds, probably due to the level of the opponent and the longer duration of the final match. The results of the present study offer additional information to youth football coaches, enabling them to know the physical demands that are required in each of the matches of a tournament, and thus adjust the load of the players depending on the level difference between opponents in order to increase their performance in key matches. Furthermore, it allows training and load distribution to be designed according to the demands of a congested Sensors 2020, 20, 6968 9 of 11 schedule, taking into account the possibility of not having a limit on replacements. The results support new studies related to the performance of players in different tournaments of different amateur football categories, an area with great complexities that has not been practically investigated until now. Author Contributions: Conceptualisation, J.S.-S. and J.G.-U.; data curation, A.H.-M.; formal analysis, J.G.-U. and C.M.; investigation, S.M.-C., J.L.F., and A.H.-M.; methodology, A.H.-M.; project administration, L.G.; resources, J.S.-S. and J.L.F.; software, J.G.-U.; supervision, L.G. and J.L.F.; validation, J.G.-U. and J.L.F.; writing–original draft, A.H.-M. and S.M.-C.; writing–review and editing, L.G., J.S.-S., J.L.F. and J.G.-U. All authors have read and agreed to the published version of the manuscript. Funding: No funding has been received for the development of this study. Acknowledgments: The authors would like to thank the football sport schools for their support and collaboration, as well as all the football players who contributed in this research. A.H.-M. acknowledges the Spanish Ministry of Science, Innovation and Universities for funding the development of his PhD (Grant Number: FPU18/03222). S.M.-C. acknowledges the University of Castilla-La Mancha for funding the development of his PhD (2019/5964). J.G.-U. acknowledges “Fondo Europeo de Desarrollo Regional, Programa Operativo de la Región de Castilla-La Mancha” (2018/11744) for funding the development of his research. Conflicts of Interest: The authors declare no conflict of interest. References 1. Bloomfield, J.; Polman, R.; O’Donoghue, P. Physical demands of different positions in FA Premier League soccer. J. Sports Sci. Med. 2007, 6, 63. [PubMed] 2. Carling, C.; Bloomfield, J.; Nelsen, L.; Reilly, T. The role of motion analysis in elite soccer. Sports Med. 2008, 38, 839–862. [CrossRef] [PubMed] 3. Rebelo, A.; Brito, J.; Maia, J.; Coelho-e-Silva, M.J.; Figueiredo, A.J.; Bangsbo, J.; Malina, R.; Seabra, A. Anthropometric characteristics, physical fitness and technical performance of under-19 soccer players by competitive level and field position. Int. J. Sports Med. 2013, 34, 312–317. [CrossRef] [PubMed] 4. Rampinini, E.; Impellizzeri, F.M.; Castagna, C.; Abt, G.; Chamari, K.; Sassi, A.; Marcora, S.M. Factors influencing physiological responses to small-sided soccer games. J. Sports Sci. 2007, 25, 659–666. [CrossRef] [PubMed] 5. Coratella, G.; Beato, M.; Schena, F. Correlation between quadriceps and hamstrings inter-limb strength asymmetry with change of direction and sprint in U21 elite soccer-players. Hum. Mov. Sci. 2018, 59, 81–87. [CrossRef] [PubMed] 6. Barbero-Alvarez, J.C.; Barbero-Alvarez, V.; Vera, J.G. Activity profile in young soccer player during match play. Apunts Phys. Educ. Sports 2007, 90, 33–41. 7. Bradley, P.S.; Di Mascio, M.; Peart, D.; Olsen, P.; Sheldon, B. High intensity activity profiles of elite soccer players at different performance levels. J. Strength Cond. Res. 2010, 24, 2343–2351. [CrossRef] [PubMed] 8. Di Salvo, V.; Baron, R.; Gonzalez-Haro, C.; Gormasz, C.; Pigozzi, F.; Bachl, N. Sprinting analysis of elite soccer players during European Champions League and UEFA Cup matches. J. Sports Sci. 2010, 28, 1489–1494. [CrossRef] 9. Reilly, T.; Williams, T.; Nevill, A.M.; Franks, A. A multidisciplinary approach to talent dentification in soccer. J. Sports Sci. 2000, 18, 695–702. [CrossRef] 10. Pacheco, R. Teaching and Training in Football 7. A Game to Start in Soccer 11; Paidotribo: Barcelona, Spain, 2004. 11. Casamichana, D.; Castellano, J.; Castagna, C. Comparing the physical demands of friendly matches and smallsided games in semiprofessional soccer players. J. Strength Cond. Res. 2012, 26, 837–843. [CrossRef] 12. Buchheit, M.; Allen, A.; Poon, T.K.; Modonutti, M.; Gregson, W.; Di Salvo, V. Integrating different tracking systems in football: Multiple camera semi-automatic system, local position measurement and GPS technologies. J. Sports Sci. 2014, 32, 1844–1857. [CrossRef] [PubMed] 13. Varley, M.C.; Gabbett, T.; Aughey, R.J. Activity profiles of professional soccer, rugby league and Australian football match play. J. Sports Sci. 2014, 32, 1858–1866. [CrossRef] [PubMed] 14. Suarez-Arrones, L.; Torreño, N.; Requena, B.; Saez De Villarreal, E.; Casamichana, D.; Barbero-Alvarez, J.C.; Munguia-Izquierdo, D. Match-play activity profile in professional soccer players during official games and the relationship between external and internal load. J. Sports Med. Phys. Fit. 2015, 55, 1417–1422. Sensors 2020, 20, 6968 10 of 11 15. Buchheit, M.; Mendez-Villanueva, A.; Simpson, B.M.; Bourdon, P.C. Match running performance and fitness in youth soccer. Int. J. Sports Med. 2010, 31, 818–825. [CrossRef] [PubMed] 16. Goto, H.; Morris, J.G.; Nevill, M.E. Motion analysis of U11 to U16 elite English premier league academy players. J. Sports Sci. 2015, 33, 1248–1258. [CrossRef] 17. Saward,C.; Morris,J.G.; Nevill,M.E.; Nevill,A.M.; Sunderland,C.Longitudinaldevelopmentofmatch-running performance in elite male youth soccer players. Scand. J. Med. Sci. Sports 2015, 26, 933–942. [CrossRef] 18. Randers, M.B.; Mujika, I.; Hewitt, A.; Santisteban, J.; Bischoff, R.; Solano, R.; Zubillaga, A.; Peltola, E.; Krustrup, P.; Mohr, M. Application of four different football match analysis systems: A comparative study. J. Sports Sci. 2010, 28, 171–182. [CrossRef] 19. Thorpe, R.; Sunderland, C. Muscle damage, endocrine, and immune marker response to a soccer match. J. Strength Cond. Res. 2012, 26, 2783–2790. [CrossRef] 20. Ekstrand, J.; Waldén, M.; Hägglund, M. A congested football calendar and the wellbeing of players: Correlation between match exposure of European footballers before the World Cup 2002 and their injuries and performances during that World Cup. Br. J. Sports Med. 2004, 38, 493–497. [CrossRef] 21. Bastida-Castillo, A.; Gómez-Carmona, C.D.; De La Cruz Sánchez, E.; Pino-Ortega, J. Comparing accuracy between global positioning systems and ultra-wideband-based position tracking systems used for tactical analyses in soccer. Eur. J. Sport Sci. 2019, 19, 1157–1165. [CrossRef] 22. Castillo, D.; Raya-González, J.; Manuel Clemente, F.; Yanci, J. The influence of youth soccer players’ sprint performance on the different sided games’ external load using GPS devices. Res. Sports Med. 2020, 28, 194–205. [CrossRef] [PubMed] 23. Darbellay, J.; Malatesta, D.; Meylan, C.M.P. Monitoring matches and small-sided games in elite young soccer players. Int. J. Sports Med. 2020, 41, 832–838. [PubMed] 24. López-Fernández, J.; Sánchez-Sánchez, J.; García-Unanue, J.; Felipe, J.L.; Colino, E.; Gallardo, L. Physiological and physical responses according to the game surface in a soccer simulation protocol. Int. J. Sport Physiol. 2018, 13, 612–619. [CrossRef] [PubMed] 25. Scott, M.T.; Scott, T.J.; Kelly, V.G. The validity and reliability of global positioning systems in team sport: A brief review. J. Strength Cond. Res. 2016, 30, 1470–1490. [CrossRef] [PubMed] 26. Rawstorn, J.C.; Maddison, R.; Ali, A.; Foskett, A.; Gant, N. Rapid directional change degrades GPS distance measurement validity during intermittent intensity running. PLoS ONE 2014, 9, e93693. [CrossRef] 27. Ubago-Guisado, E.; Gómez-Cabello, A.; Sánchez-Sánchez, J.; García-Unanue, J.; Gallardo, L. Influence of different sports on bone mass in growing girls. J. Sports Sci. 2015, 33, 1710–1718. [CrossRef] 28. Sanchez-Sanchez, J.; Sanchez, M.; Hernandez, D.; Ramirez-Campillo, R.; Martínez, C.; Nakamura, F.Y. Fatigue in U12 soccer-7 players during repeated 1-day tournament games: A pilot study. J. Strength Cond. Res. 2019, 33, 3092–3097. [CrossRef] 29. Goto, H.; Morris, J.G.; Nevill, M.E. Match analysis of U9 and U10 English premier league academy soccer players using a global positioning system: Relevance for talent identification and development. J. Strength Cond. Res. 2015, 29, 954–963. [CrossRef] 30. Osgnach, C.; Poser, S.; Bernardini, R.; Rinaldo, R.; Di Prampero, P.E. Energy cost and metabolic power in elite soccer: A new match analysis approach. Med. Sci. Sports Exerc. 2010, 42, 170–178. [CrossRef] 31. Johnston, R.J.; Watsford, M.L.; Kelly, S.J.; Pine, M.J.; Spurrs, R.W. Validity and interunit reliability of 10 Hz and 15 Hz GPS units for assessing athlete movement demands. J. Strength Cond. Res. 2014, 28, 1649–1655. [CrossRef] 32. Buchheit, M.; Al Haddad, H.; Simpson, B.M.; Palazzi, D.; Bourdon, P.C.; Di Salvo, V.; Mendez-Villanueva, A. Monitoring accelerations with GPS in football: Time to slow down? Int. J. Sports Physiol. Perform. 2014, 9, 442. [CrossRef] [PubMed] 33. Carling, C.; Dupont, G. Are declines in physical performance associated with a reduction in skill-related performance during professional soccer match-play? J. Sports Sci. 2011, 29, 63–71. [CrossRef] [PubMed] 34. Rampinini, E.; Impellizzeri, F.M.; Castagna, C.; Coutts, A.J.; Wisloff, U. Technical performance during soccer matches of the Italian Serie A league: Effect of fatigue and competitive level. J. Sci. Med. Sport 2009, 12, 227–233. [CrossRef] 35. Weston, M.; Batterham, A.M.; Castagna, C.; Portas, M.D.; Barnes, C.; Harley, J.; Lovell, R. Reduction in physical match performance at the start of the second half in elite soccer. Int. J. Sports Physiol. Perform. 2011, 6, 174–182. [CrossRef] [PubMed] Sensors 2020, 20, 6968 11 of 11 36. Kołodziejczyk, M.; Chmura, P.; Milanovic, L.; Konefał, M.; Chmura, J.; Rokita, A.; Andrzejewski, M. How three consecutive matches with extra time effect on physical performance? A case study at the 2018 football men’s World Cup. Biol. Sport 2021, 38, 65–70. 37. Chmura, P.; Andrzejewski, M.; Konefał, M.; Mroczek, D.; Rokita, A.; Chmura, J. Analysis of motor activities of professional soccer players during the 2014 World Cup in Brazil. J. Hum. Kinet. 2017, 56, 187–195. [CrossRef] 38. Abbott, W.; Brickley, G.; Smeeton, N.J. Positional differences in GPS outputs and perceived exertion during soccer training games and competition. J. Strength Cond. Res. 2018, 32, 3222–3231. [CrossRef] 39. Bradley, P.S.; Noakes, T.D. Match running performance fluctuations in elite soccer: Indicative of fatigue, pacing or situational influences? J. Sports Sci. 2013, 31, 1627–1638. [CrossRef] 40. Ramírez-Campillo, R.; Gallardo, F.; Henriquez-Olguín, C.; Meylan, C.M.; Martínez, C.; Álvarez, C.; Caniuqueo, A.; Cadore, E.L.; Izquierdo, M. Effect of vertical, horizontal, and combined plyometric training on explosive, balance, and endurance performance of young soccer players. J. Strength Cond. Res. 2015, 29, 1784–1795. [CrossRef] 41. Bradley, P.S.; Sheldon, W.; Wooster, B.; Olsen, P.; Boanas, P.; Krustrup, P. High-intensity running in English FA Premier League soccer matches. J Sports Sci. 2009, 27, 159–168. [CrossRef] 42. Robineau, J.; Jouaux, T.; Lacroix, M.; Babault, N. Neuromuscular fatigue induced by a 90-minute soccer game modeling. J. Strength Cond. Res. 2012, 26, 555–562. [CrossRef] [PubMed] 43. Beneke, R.; Hutler, M.; Jung, M.; Leithauser, R.M. Modelling the blood lactate kinetics at maximal short-term exercise conditions in children, adolescents, and adults. J. Appl. Physiol. 2005, 99, 499–504. [CrossRef] [PubMed] 44. Ratel, S.; Duche, P.; Williams, C.A. Muscle fatigue during high-intensity exercise in children. Sports Med. 2006, 36, 1031–1065. [CrossRef] [PubMed] 45. Aughey, R.J. Applications of GPS technologies to field sports. Int. J. Sports Physiol. Perform. 2011, 6, 295–310. [CrossRef] [PubMed] 46. Mohr, M.; Krustrup, P.; Bangsbo, J. Match performance of high-standard soccer players with special reference to development of fatigue. J. Sports Sci. 2003, 21, 519–528. [CrossRef] 47. Mohr, M.; Krustrup, P.; Bangsbo, J. Fatigue in soccer: A brief review. J. Sports Sci. 2005, 23, 593–599. [CrossRef] 48. Buchheit, M.; Mendez-Villanueva, A.; Simpson, B.M.; Bourdon, P.C. Repeated-sprint sequences during youth soccer matches. Int. J. Sports Med. 2010, 31, 709–716. [CrossRef] 49. Zafeiridis, A.; Dalamitros, A.; Dipla, K.; Manou, V.; Galanis, N.; Kellis, S. Recovery during high-intensity intermittent anaerobic exercise in boys, teens, and men. Med. Sci. Sports Exerc. 2005, 37, 505–512. [CrossRef] 50. Aquino, R.; Vieira, L.H.P.; Carling, C.; Martins, G.H.; Alves, I.S.; Puggina, E.F. Effects of competitive standard, team formation and playing position on match running performance of Brazilian professional soccer players. Int. J. Perform. Anal. Sport 2017, 17, 695–705. [CrossRef] Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations. © 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

Physical Demands of U10 Players in a 7-a-Side Soccer Tournament Depending on the Playing Position and Level of Opponents in Consecutive Matches Using Global Positioning Systems (GPS).

12-06-2020

Hernandez-Martin, Antonio,Sanchez-Sanchez, Javier,Felipe, Jose Luis,Manzano-Carrasco, Samuel,Majano, Carlos,Gallardo, Leonor,Garcia-Unanue, Jorge

eng

PMC10098635

Citation: Maeda, Y.; Okawara, H.; Sawada, T.; Nakashima, D.; Nagahara, J.; Fujitsuka, H.; Ikeda, K.; Hoshino, S.; Kobari, Y.; Katsumata, Y.; et al. Implications of the Onset of Sweating on the Sweat Lactate Threshold. Sensors 2023, 23, 3378. https://doi.org/10.3390/s23073378 Academic Editors: Abdelhamid Errachid, Maria Giovanna Trivella and Francesca G. Bellagambi Received: 8 February 2023 Revised: 5 March 2023 Accepted: 20 March 2023 Published: 23 March 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). sensors Article Implications of the Onset of Sweating on the Sweat Lactate Threshold Yuta Maeda 1 , Hiroki Okawara 1 , Tomonori Sawada 1 , Daisuke Nakashima 1,* , Joji Nagahara 1, Haruki Fujitsuka 1, Kaito Ikeda 1, Sosuke Hoshino 1, Yusuke Kobari 2, Yoshinori Katsumata 2,3,* , Masaya Nakamura 1 and Takeo Nagura 1,4 1 Department of Orthopaedic Surgery, Keio University School of Medicine, Shinjuku, Tokyo 160-8582, Japan 2 Department of Cardiology, Keio University School of Medicine, Shinjuku, Tokyo 160-8582, Japan 3 Institute for Integrated Sports Medicine, Keio University School of Medicine, Shinjuku, Tokyo 160-8582, Japan 4 Department of Clinical Biomechanics, Keio University School of Medicine, Shinjuku, Tokyo 160-8582, Japan * Correspondence: [email protected] (D.N.); [email protected] (Y.K.); Tel.: +81-3-5363-3812 (D.N.); +81-3-3353-1211 (Y.K.) Abstract: The relationship between the onset of sweating (OS) and sweat lactate threshold (sLT) assessed using a novel sweat lactate sensor remains unclear. We aimed to investigate the implications of the OS on the sLT. Forty healthy men performed an incremental cycling test. We monitored the sweat lactate, blood lactate, and local sweating rates to determine the sLT, blood LT (bLT), and OS. We defined participants with the OS during the warm-up just before the incremental test as the early perspiration (EP) group and the others as the regular perspiration (RP) group. Pearson’s correlation coefficient analysis revealed that the OS was poorly correlated with the sLT, particularly in the EP group (EP group, r = 0.12; RP group, r = 0.56). Conversely, even in the EP group, the sLT was strongly correlated with the bLT (r = 0.74); this was also the case in the RP group (r = 0.61). Bland-Altman plots showed no bias between the mean sLT and bLT (mean difference: 19.3 s). Finally, in five cases with a later OS than bLT, the sLT tended to deviate from the bLT (mean difference, 106.8 s). The sLT is a noninvasive and continuous alternative to the bLT, independent of an early OS, although a late OS may negatively affect the sLT. Keywords: lactate threshold; sweat rate; exercise testing; incremental exercise; sweating; perspiration; body temperature regulation; sports; physiology 1. Introduction The visualization of exercise tolerance to optimize daily training is encouraged by athletes and supporters. In particular, monitoring metabolic responses, such as the anaer- obic threshold (AT), during exercise enables athletes to evaluate their aerobic capacity in real-time [1–4], which leads to a defined relative workload intensity [4,5]. Previous studies have investigated the ventilatory threshold (VT), measured using an expiration gas analyzer, and the lactate threshold (LT) as good indicators of AT [3,6]. VT testing is costly and requires expertise; thus, it is not easily accessible to recreational or younger athletes without favorable facilities. LT testing requires frequent blood sampling while interrupting strenuous exercise, meaning it does not reflect usual (continuous) exercise. In addition, contamination with other substances, such as sweat, makes the assessment difficult. In recent years, wearable sensing technology has been the focus of more precise evaluations of physiological responses in the body. We have developed a method to visualize the lactate dynamics of sweat during exercise in a noninvasive, simple, and real-time manner [7,8]. Furthermore, the sweat LT (sLT), assessed using sweat lactate dynamics, is consistent with the LT calculated from blood samples (bLT) and the VT [7]. However, unlike blood lactate levels, changes in sweat lactate levels may be affected by sweating dynamics. Sensors 2023, 23, 3378. https://doi.org/10.3390/s23073378 https://www.mdpi.com/journal/sensors Sensors 2023, 23, 3378 2 of 11 During exercise, sweating occurs with a rise in body temperature that reflects the increased workload [9–11]. Previous reports have shown a negative correlation between the local sweating rate and sweat lactate level during exercise [12–14], suggesting that increased sweating dilutes the sweat lactate concentration. However, the relationship between the onset of sweating (OS) and sLT remains unclear. If the relationship between the bLT and sLT is strongly dependent on the OS, LT determination using sweat instead of blood may require careful attention to environmental conditions, such as the temperature and humidity, based on the subject’s condition. We aimed to investigate the relationship between the OS and sLT and the effect of the OS on the relationship between the sLT and bLT during incremental exercise. 2. Materials and Methods 2.1. Participants Volunteers were recruited from December 2020 to August 2021, and 40 healthy men aged 18–37 years participated in this study. The exclusion criteria were as follows: (1) the presence of comorbidities such as active cardiopulmonary disease, hypertension, or di- abetes within two weeks; (2) elite athletes; (3) smokers and those taking medication or performance-enhancing drugs. The study protocol was conducted in compliance with the ethical guidelines for medical and health research involving human subjects and was approved by the Ethics Committees of Keio University School of Medicine (Approval No. 20180357). Written informed consent was obtained from the study participants for the participation and publication of the findings before enrollment. 2.2. Procedures and Protocols All experiments were performed at sports facilities under similar conditions (24.8 ± 1.9 ◦C temperature, 42.0 ± 9.3% relative humidity). Prior to the exercise test, the body weight (kg) and height (cm) of each participant were measured. The body mass index (BMI) and body surface area (BSA) were calculated using the formula used by DuBois (body surface area (m2) = 0.007184 × height (cm)0.725 × weight (kg)0.425) [15]. The participants were instructed to avoid drinking alcohol or caffeine for 12 h [16]. They were also required to be well hydrated and refrain from vigorous exercise for 3 h before the exercise test. Each participant completed incremental exercise using an electromagnetically braked ergometer (POWER MAX V3 Pro; Konami Sports Co., Ltd., Tokyo, Japan). The ergometer seat was adjusted to a favorable height. Two minutes of rest was set to measure the resting outcomes. Immediately after the warm-up for 4 min with a load of 20 W, the exercise test was started at 50 W, increasing by 25 W every minute. The pedaling cadence range was set to 70–80 revolutions per minute (rpm). Each test was terminated owing to (a) subjective exhaustion or (b) reaching 70 rpm for 10 s. 2.3. Measurements during the Exercise Test The sweat rate and sweat lactate concentration were continuously monitored during incremental exercise. The sweat lactate was measured using a sweat lactate sensor chip (Grace Imaging Inc., Tokyo, Japan), which we developed and applied in several studies (Figure S1) [7,8,17,18]. It is a type of electrochemical sensor that detects the potential generated by the redox reaction between the lactate and lactate oxidase. The special protective membrane structure of the sensor allows the aforementioned reaction to last for 30 to 60 min. As a result, the lactate concentration in sweat can be continuously measured throughout the exercise test. Further detailed information regarding the composition and fabrication of the sensor chip is available in our previous study [7]. This sweat lactate sensor chip connected to a wearable sweat lactate sensor was placed on the upper arm [7]. The installation area was carefully cleaned using an alcohol-free cloth to prevent contamination of the sweat sample. The sensor chip was firmly fixed with tape to prevent detachment from the skin. Real-time sweat lactate values were automatically recorded in Sensors 2023, 23, 3378 3 of 11 a connected application device (Grace Imaging Inc., Tokyo, Japan) via Bluetooth at 1 Hz. The sweat lactate value was quantified as the current value because the chip reacts with sweat lactate and generates an electric current [7,8,18,19]. The sLT was defined as the first significant increase in the lactate concentration in sweat above the baseline based on graphical plots [7,8,18,19]. The sweat rate was measured using a pre-calibrated perspiration meter (SKN-200M; SKINOS Co., Ltd., Ueda, Japan) on the upper arm [17] and a Fitbit Inspire HR (Fitbit Inc., San Francisco, CA, USA) was attached to the left wrist, two-finger widths above the ulnar styloid process, to measure the heart rate. The heart rate at rest was measured once at the onset of the warm-up. The sweat rate was recorded at 1 Hz and expressed in milligrams per square centimeter per minute (mg/cm2/min). The baseline of the local sweat rate was calculated as the average of the sweat volume in the rest period. The OS was defined as the first significant increase in sweat rate above the baseline based on graphical plots (Figure 1). y measured throughout the exercise test. Further detailed information regarding the composition and fabrication of the sensor chip is available in our previous study [7]. This sweat lactate sensor chip connected to a wearable sweat lactate sensor was placed on the upper arm [7]. The installation area was carefully cleaned using an alcohol-free cloth to prevent contamination of the sweat sample. The sensor chip was firmly fixed with tape to prevent detachment from the skin. Real-time sweat lactate values were automatically recorded in a connected application device (Grace Imaging Inc., Tokyo, Japan) via Bluetooth at 1 Hz. The sweat lactate value was quantified as the current value because the chip reacts with sweat lactate and generates an electric current [7,8,18,19]. The sLT was defined as the first significant increase in the lactate concentration in sweat above the baseline based on graphical plots [7,8,18,19]. The sweat rate was measured using a pre-calibrated perspiration meter (SKN-200M; SKINOS Co., Ltd., Ueda, Japan) on the upper arm [17] and a Fitbit Inspire HR (Fitbit Inc., San Francisco, CA, USA) was attached to the left wrist, two-finger widths above the ulnar styloid process, to measure the heart rate. The heart rate at rest was measured once at the onset of the warm-up. The sweat rate was recorded at 1 Hz and expressed in milligrams per square centimeter per minute (mg/cm2/min). The baseline of the local sweat rate was calculated as the average of the sweat volume in the rest period. The OS was defined as the first significant increase in sweat rate above the baseline based on graphical plots (Figure 1). Figure 1. Representative data of the local sweat rate, blood lactate, and sweat lactate throughout the exercise protocol, including the rest and warm-up periods. The onset of sweating (OS), blood lactate threshold (bLT), and sweat lactate threshold (sLT) are indicated by the dotted lines. The OS preceded both the bLT and sLT (35 out of 40 cases, 88%). To measure blood lactate concentration, a blood sample was obtained from the auricle at rest and every minute during exercise. The lactate concentration in the blood was immediately measured using a lactate analyzer (Lactate Pro2 LT-1730, ARKRAY, Inc., Kyoto, Japan). The blood lactate values were graphically plotted in millimoles per liter (mmol/L). The bLT was determined using graphical plots [3]. Figure 1. Representative data of the local sweat rate, blood lactate, and sweat lactate throughout the exercise protocol, including the rest and warm-up periods. The onset of sweating (OS), blood lactate threshold (bLT), and sweat lactate threshold (sLT) are indicated by the dotted lines. The OS preceded both the bLT and sLT (35 out of 40 cases, 88%). To measure blood lactate concentration, a blood sample was obtained from the auricle at rest and every minute during exercise. The lactate concentration in the blood was immediately measured using a lactate analyzer (Lactate Pro2 LT-1730, ARKRAY, Inc., Kyoto, Japan). The blood lactate values were graphically plotted in millimoles per liter (mmol/L). The bLT was determined using graphical plots [3]. 2.4. Statistical Analysis The OS, bLT, and sLT were determined visually by three researchers independently in accordance with previous reports [7,17]. We first defined subjects with the OS during the resting period or warm-up as the early perspiration (EP) group, and those with the OS during incremental exercise as the regular perspiration (RP) group. The relationships between the OS, sLT, and bLT were examined using Pearson’s correlation coefficient and a linear regression analysis. For each combination, a one-sample t-test on the difference in time was used to examine the fixed error, and a linear regression analysis on the mean and Sensors 2023, 23, 3378 4 of 11 difference in time was used to rule out the proportional error. Furthermore, to investigate the effect of an early OS on the approximation between the bLT and sLT, we compared mean bLT–sLT (s) values between the EP and RP groups using an independent t-test. Based on the results of the Kolmogorov–Smirnov test, either an independent t-test or Mann–Whitney U test was applied to the age, height, weight, BMI, body water, BSA, room temperature, and relative humidity to compare each variable between the groups. All analyses were performed using IBM SPSS Statistics version 27 (IBM Corp., Armonk, NY, USA), with the statistical significance set at 0.05. 3. Results 3.1. Participants Characteristics and Physiological Results All participants completed all procedures and were eligible for the analysis. During exercise, continuous negligible responses were initially detected in the sweat rate, sweat lactate, and blood lactate (Figure 1). Subsequently, they rapidly increased and were defined as the OS, sLT, and bLT, respectively. For all participants, these time points could be clearly defined. Based on the relationship between these points, 17 participants (43%) were categorized into the EP group and 23 participants (57%) into the RP group. Furthermore, in five participants (13%), the OS was later than the bLT. The descriptive data of the participants in the EP and RP groups are presented in Table 1. No significant differences were observed in any of the characteristics between the groups (p > 0.05). Table 1. Descriptive data of the participants. Total (n = 40) Early Perspiration Group (n = 17) Regular Perspiration Group (n = 23) p-Value Age (years) 21.8 ± 4.0 22.4 ± 4.6 21.4 ± 3.5 0.71 BMI 22.4 ± 2.1 22.7 ± 1.9 22.2 ± 2.3 0.48 BSA (m2) 1.8 ± 0.1 1.8 ± 0.1 1.8 ± 0.1 0.86 Body water (%) 57.8 ± 5.4 56.8 ± 5.6 58.6 ± 5.3 0.31 Body fat ratio (%) 17.1 ± 5.0 18.6 ± 5.4 16.0 ± 4.5 0.11 Muscle mass (kg) 53.6 ± 5.8 53.0 ± 5.3 54.0 ± 6.3 0.61 RT (◦C) 24.1 ± 1.9 24.2 ± 1.8 24.0 ± 2.0 0.91 RH (%) 42.5 ± 9.3 42.3 ± 8.6 42.6 ± 9.9 0.92 All values are presented as means ± standard deviations or n (%). Based on the pre-performed Kolmogorov– Smirnov test, the independent t-test or Mann–Whitney U test was applied for intra-group comparisons. BSA, body surface area; RT, room temperature; RH, relative humidity. Table 2 shows the physiological results including the load, heart rate, sweat lactate level, and local sweat rate. As in previous reports, the heart rate increased with the exercise intensity and the sweat rate increased after the onset of the incremental exercise. The sweat lactate levels were stable at first and then rapidly increased from the sLT to the end. Table 2. Physiological data of the participants. Baseline At the Warm-Up Onset At the Incremental Exercise Onset At the Sweat Lactate Threshold At the End of Incremental Exercise Load (watt) 0.0 ± 0.0 20.0 ± 0.0 50.0 ± 0.0 131.7 ± 48.5 261.2 ± 43.6 Heart rate (bpm) - 79.3 ± 11.1 94.4 ± 14.0 137.0 ± 23.1 172.9 ± 13.8 Sweat lactate (µA) 4.0 ± 1.1 3.9 ± 1.2 3.7 ± 1.2 3.9 ± 1.4 9.6 ± 4.6 Local sweat rate (mg/cm2/min) 0.04 ± 0.11 0.01 ± 0.09 0.08 ± 0.17 0.18 ± 0.20 0.84 ± 0.46 The baseline is the average of data in the rest period. All values are presented as means ± standard deviations. Sensors 2023, 23, 3378 5 of 11 3.2. Relationship between OS and sLT Figure 2a shows the relationship between the OS and sLT. A poor correlation was observed in the total cohort (r = 0.38, p < 0.05), and no correlation was observed in the EP group (r = 0.12, p = 0.61). In contrast, in the RP group, the OS moderately correlated with the sLT (r = 0.56, p < 0.05). In the five patients with the OS later than the bLT, the OS showed a strong correlation with the sLT (r = 0.82, p = 0.09). The Bland-Altman plot revealed that that the mean difference between the OS and sLT was large in total, especially in the EP group (total, 183.7 s, EP group, 389.6 s, RP group, 85.2 s), which validates the inconsistency between these thresholds (Figure 2b). Figure 2b also shows a positive fixed error (p < 0.05, 95% CI: 155.1–274.1) and a proportional error (p < 0.05) in total. The fixed error indicated that the sLT occurred after the OS in all the cases. Heart rate (bpm) - 79.3 ± 11.1 94.4 ± 14.0 137.0 ± 23.1 172.9 ± 13.8 Sweat lactate (µA) 4.0 ± 1.1 3.9 ± 1.2 3.7 ± 1.2 3.9 ± 1.4 9.6 ± 4.6 Local sweat rate (mg/cm2/min) 0.04 ± 0.11 0.01 ± 0.09 0.08 ± 0.17 0.18 ± 0.20 0.84 ± 0.46 The baseline is the average of data in the rest period. All values are presented as means ± standard deviations. 3.2. Relationship between OS and sLT Figure 2a shows the relationship between the OS and sLT. A poor correlation was observed in the total cohort (r = 0.38, p < 0.05), and no correlation was observed in the EP group (r = 0.12, p = 0.61). In contrast, in the RP group, the OS moderately correlated with the sLT (r = 0.56, p < 0.05). In the five patients with the OS later than the bLT, the OS showed a strong correlation with the sLT (r = 0.82, p = 0.09). The Bland-Altman plot revealed that that the mean difference between the OS and sLT was large in total, especially in the EP group (total, 183.7 s, EP group, 389.6 s, RP group, 85.2 s), which validates the inconsistency between these thresholds (Figure 2b). Figure 2b also shows a positive fixed error (p < 0.05, 95% CI: 155.1–274.1) and a proportional error (p < 0.05) in total. The fixed error indicated that the sLT occurred after the OS in all the cases. Figure 2. The relationship between the onset of sweating and sweat lactate threshold. (a) A scatter plot and approximation line between the onset of sweating (OS) and the sweat lactate threshold (sLT). The correlation was minimal in the total cohort and early perspiration group (EP), while that in the regular perspiration group (RP) was moderate. (b) Validity testing using Bland-Altman plots, which indicated the respective differences between the time at the OS and sLT (y-axis) against the mean time of OS and sLT (x-axis). Red triangles indicate the EP group data and circles indicate the RP group data. Note: r, correlation coefficient; ULOA, upper limit of agreement (=mean + 1.96 × standard deviation); LLOA, lower limit of agreement (= mean − 1.96 × standard deviation). 3.3. Relationship between OS and Blood LT Figure 3a shows the relationship between the OS and bLT. There was no correlation in the total cohort (r = 0.23, p = 0.16), EP (r = −0.06, p = 0.83) or RP groups (r = 0.30, p = 0.16). The Bland-Altman plot revealed that the mean difference between the OS and sLT was large in total, especially in the EP group (total, −195.3 s; EP group, −381.6 s; RP group, Figure 2. The relationship between the onset of sweating and sweat lactate threshold. (a) A scatter plot and approximation line between the onset of sweating (OS) and the sweat lactate threshold (sLT). The correlation was minimal in the total cohort and early perspiration group (EP), while that in the regular perspiration group (RP) was moderate. (b) Validity testing using Bland-Altman plots, which indicated the respective differences between the time at the OS and sLT (y-axis) against the mean time of OS and sLT (x-axis). Red triangles indicate the EP group data and circles indicate the RP group data. Note: r, correlation coefficient; ULOA, upper limit of agreement (=mean + 1.96 × standard deviation); LLOA, lower limit of agreement (=mean − 1.96 × standard deviation). 3.3. Relationship between OS and Blood LT Figure 3a shows the relationship between the OS and bLT. There was no correlation in the total cohort (r = 0.23, p = 0.16), EP (r = −0.06, p = 0.83) or RP groups (r = 0.30, p = 0.16). The Bland-Altman plot revealed that the mean difference between the OS and sLT was large in total, especially in the EP group (total, −195.3 s; EP group, −381.6 s; RP group, −57.6 s), which validates the inconsistency between those thresholds (Figure 3b). Figure 3b also shows both a positive fixed error (p < 0.05, 95% CI: 132.1–258.6) and a proportional error (p < 0.05) in total. The negative fixed error indicated that the bLT tended to come after the OS, while the bLT preceded the OS in five cases (out of 23 cases, 21%) in the RP group. 3.4. Effect of Early OS on the Blood–Sweat Lactate Threshold Approximation As shown in Figure 4a, the sLT was strongly correlated with the bLT, regardless of early perspiration (total, r = 0.68, p < 0.01; EP group, r = 0.74, p < 0.01; RP group, r = 0.61, p < 0.01). The linear regression analysis revealed that the sLT can be a good indicator of the bLT (total, y = 252.9 + 0.54x, p < 0.01, R = 0.68; EP group, y = 217.9 + 0.60x, p < 0.01, R = 0.74; RP group, y = 289.1 + 0.49x, p < 0.01, R = 0.61). The Bland-Altman plot showed no bias Sensors 2023, 23, 3378 6 of 11 between the bLT and sLT in each group (total, 19.3 s; EP group, 8.0 s; RP group, 27.6 s), and validated a strong agreement between those thresholds (Figure 4b). Figure 4b also shows that in the five cases where the OS occurred later than the bLT (described in blue circle), the sLT was more likely to deviate from the bLT than in other cases. The mean difference between the sLT and bLT among the five was 106.8 s. There was no fixed error (p > 0.05, 95% CI: −4.8 to 43.3) or proportional error (p > 0.05) in the total group. Sensors 2023, 23, x FOR PEER REVIEW 6 of 11 −57.6 s), which validates the inconsistency between those thresholds (Figure 3b). Figure 3b also shows both a positive fixed error (p < 0.05, 95% CI: 132.1–258.6) and a proportional error (p < 0.05) in total. The negative fixed error indicated that the bLT tended to come after the OS, while the bLT preceded the OS in five cases (out of 23 cases, 21%) in the RP group. Figure 3. The relationship between the onset of sweating and blood lactate threshold. (a) Scatter plot and approximation line between the OS and bLT. No significant correlation was found in the total, early perspiration group (EP), or regular perspiration (RP) groups. (b) Validity testing using Bland- Altman plots, which indicated the respective difference between the time at the OS and bLT (y-axis) against the mean time of the OS and bLT (x-axis). Triangles indicate the EP group data and circles indicate the RP group data. Note: r, correlation coefficient; ULOA, upper limit of agreement (= mean + 1.96 × standard deviation); LLOA, lower limit of agreement (= mean − 1.96 × standard deviation). 3.4. Effect of Early OS on the Blood–Sweat Lactate Threshold Approximation As shown in Figure 4a, the sLT was strongly correlated with the bLT, regardless of early perspiration (total, r = 0.68, p < 0.01; EP group, r = 0.74, p < 0.01; RP group, r = 0.61, p < 0.01). The linear regression analysis revealed that the sLT can be a good indicator of the bLT (total, y = 252.9 + 0.54x, p < 0.01, R = 0.68; EP group, y = 217.9 + 0.60x, p < 0.01, R = 0.74; RP group, y = 289.1 + 0.49x, p < 0.01, R = 0.61). The Bland-Altman plot showed no bias between the bLT and sLT in each group (total, 19.3 s; EP group, 8.0 s; RP group, 27.6 s), and validated a strong agreement between those thresholds (Figure 4b). Figure 4b also shows that in the five cases where the OS occurred later than the bLT (described in blue circle), the sLT was more likely to deviate from the bLT than in other cases. The mean difference between the sLT and bLT among the five was 106.8 s. There was no fixed error (p > 0.05, 95% CI: −4.8 to 43.3) or proportional error (p > 0.05) in the total group. Figure 3. The relationship between the onset of sweating and blood lactate threshold. (a) Scatter plot and approximation line between the OS and bLT. No significant correlation was found in the total, early perspiration group (EP), or regular perspiration (RP) groups. (b) Validity testing using Bland- Altman plots, which indicated the respective difference between the time at the OS and bLT (y-axis) against the mean time of the OS and bLT (x-axis). Triangles indicate the EP group data and circles indicate the RP group data. Note: r, correlation coefficient; ULOA, upper limit of agreement (=mean + 1.96 × standard deviation); LLOA, lower limit of agreement (=mean − 1.96 × standard deviation). Sensors 2023, 23, x FOR PEER REVIEW 7 of 11 Figure 4. The relationship between the sweat and blood lactate thresholds. (a) Scatter plot and approximation line between the sLT and bLT. Significant correlations were found in the total, early perspiration (EP), and regular perspiration (RP) groups. The linear regression analysis also indicated a strong effect of the sLT on the bLT estimation in each group. (b) Validity testing using Bland-Altman plots, which indicated the respective difference between the time at the bLT and sLT (y-axis) against the mean time at the sLT and bLT (x-axis). Red triangles indicate the EP group data, black or blue circles indicate the RP group data. Among the RP group data, blue circles indicate the late perspiration data, where the time of the bLT preceded the time of the OS. Note: r, correlation coefficient; R; multiple correlation coefficient; ULOA, upper limit of agreement (= mean + 1.96 × standard deviation); LLOA, lower limit of agreement (= mean − 1.96 × standard deviation). Figure 5 also shows that there is no significant difference between the bLT–sLT approximations (= sLT–bLT (s)) in the EP and RP groups (p > 0 05) Figure 4. The relationship between the sweat and blood lactate thresholds. (a) Scatter plot and approximation line between the sLT and bLT. Significant correlations were found in the total, early perspiration (EP), and regular perspiration (RP) groups. The linear regression analysis also indicated a strong effect of the sLT on the bLT estimation in each group. (b) Validity testing using Bland-Altman plots, which indicated the respective difference between the time at the bLT and sLT (y-axis) against Sensors 2023, 23, 3378 7 of 11 the mean time at the sLT and bLT (x-axis). Red triangles indicate the EP group data, black or blue circles indicate the RP group data. Among the RP group data, blue circles indicate the late perspiration data, where the time of the bLT preceded the time of the OS. Note: r, correlation coefficient; R; multiple correlation coefficient; ULOA, upper limit of agreement (=mean + 1.96 × standard deviation); LLOA, lower limit of agreement (=mean − 1.96 × standard deviation). Figure 5 also shows that there is no significant difference between the bLT–sLT approx- imations (=sLT–bLT (s)) in the EP and RP groups (p > 0.05). Figure 4. The relationship between the sweat and blood lactate thresholds. (a) Scatter plot and approximation line between the sLT and bLT. Significant correlations were found in the total, early perspiration (EP), and regular perspiration (RP) groups. The linear regression analysis also indicated a strong effect of the sLT on the bLT estimation in each group. (b) Validity testing using Bland-Altman plots, which indicated the respective difference between the time at the bLT and sLT (y-axis) against the mean time at the sLT and bLT (x-axis). Red triangles indicate the EP group data, black or blue circles indicate the RP group data. Among the RP group data, blue circles indicate the late perspiration data, where the time of the bLT preceded the time of the OS. Note: r, correlation coefficient; R; multiple correlation coefficient; ULOA, upper limit of agreement (= mean + 1.96 × standard deviation); LLOA, lower limit of agreement (= mean − 1.96 × standard deviation). Figure 5 also shows that there is no significant difference between the bLT–sLT approximations (= sLT–bLT (s)) in the EP and RP groups (p > 0.05). Figure 5. Comparison of the approximations of the sweat lactate threshold to the blood lactate threshold. Box plot of the difference between the times at the blood and sweat lactate thresholds. Black circles indicate the respective data. No significant differences were observed between the early perspiration and regular perspiration groups (p = 0.42). Note: n.s., not significance (p > 0.05). Figure 5. Comparison of the approximations of the sweat lactate threshold to the blood lactate threshold. Box plot of the difference between the times at the blood and sweat lactate thresholds. Black circles indicate the respective data. No significant differences were observed between the early perspiration and regular perspiration groups (p = 0.42). Note: n.s., not significance (p > 0.05). 4. Discussion The most striking finding of this study was the poor association between the OS and sLT during incremental exercises. Moreover, the sLT was strongly correlated with the bLT, independent of the preceding OS. These findings provide further information regarding LT measurements of sweat lactate as a noninvasive, continuous, real-time analytical alternative to blood lactate testing. The analysis of sweat, which contains various types of physiological information [20], has attracted the attention of researchers and physiologists in the athletic field. In particular, the lactate in sweat is concentrated, and the sLT is reported to correlate with VT and bLT [7], which are common indicators of the aerobic exercise capacity [1–6]. However, the relation- ship between the sweat kinetics and sweat lactate kinetics is yet to be fully investigated; therefore, the interpretation of sweat lactate concentrations remains controversial. We focused on the effect of an early OS on the time of the sLT by monitoring the local sweating rate and sweat lactate concentration. Herein, no correlation was observed between the OS and sLT in the EP and RP groups. This implies that the sLT is regulated independent of the OS and that the sweat lactate dynamics are not necessarily consistent with the sweat dynamics. Sweating occurs due to an increase in core body temperature [9], and the threshold temperature does not change with exercise [21–23] or exercise intensity [11,24–27]. In contrast, the sLT is observed simultaneously with the bLT [28,29], which manifests as an increased production of the lactate in response to the intensive energy demand in the muscles. Such different mechanisms of OS and sweat lactate generation suggest that the increase in sweat lactate production (indicating sLT) is different from the OS, as shown Sensors 2023, 23, 3378 8 of 11 in this study. We assumed that the rate of increase in exercise intensity used in this study possibly induced a rapid increase in deep body temperature that allowed the OS to occur earlier than the sLT. Our previous report showed that there was a significant correlation between the bLT and sLT [7]. Herein, we showed that the sLT was strongly associated with the bLT, regardless of the OS. Some studies noted that several physiological changes associated with increased lactate production, such as changes in the autonomic nervous balance, hormones, acid–base equilibrium, and metabolic dynamics, may explain such a simultaneous increase in lactate in different biofluids [7,30]. Environmental and biological factors complicate the interpretation of studies using different subjects because they can induce an earlier OS. Hyperthermal and humid environments induce sweat production [9], and the sweat rate in men is significantly higher than that in women due to the higher sensitivity of the eccrine glands to heat and the differences in the hormonal environment [31–34]. The high capacity for sweating in well-trained athletes supports heat acclimation to maintain high aerobic tolerance [35–38]. Thus, the effect of an early OS on the validity of the sLT to estimate the bLT needed to be investigated. As shown in this study, the sLT correlated well with the bLT, even in cases of early OS, which might lead to the greater validity and wider applicability of sweat lactate measurements during exercise. This can be attributed to the irrelevancy of the OS in the sLT due to the difference in the mechanisms of sweating onset and lactate production. In addition, the OS was poorly correlated with the bLT in the EP and RP groups, which suggests that the sLT is superior to the OS as a noninvasive parameter to estimate the bLT. Insufficient sweating remains a major challenge for the use of lactate sensors. A total of five cases showed a later OS than bLT, and the mean difference between the bLT and sLT (sLT–bLT (s)) was larger than in other cases in the EP and RP groups. This indicates that the sLT is more likely to be delayed than the bLT. Therefore, the device had difficulty in adapting to the sLT measurement under ambient conditions with poor sweating during exercise, such as with low-intensity exercise and a low number of sweat glands due to the genetic background [39]. Adjusting the exercise environment (e.g., humidity and temperature) and duration (e.g., long warm-up and total exercise time) may be required for subjects with delayed perspiration. In addition, the use of hermetic sealing or high local temperatures to promote sweating would overcome insufficient sweating in normal environments. Otherwise, improved sweat lactate sensing devices may be warranted. A sweat sampling patch that operates under novel osmotic extraction principles succeeded in withdrawing sweat without rigorous exertion; however, the data were not continuous [40]. Moreover, further studies are required to examine subjects with less perspiration. In the five cases with delayed OS, the sLT strongly correlated with the OS and tended to be more delayed. Naturally, the sLT was later than the OS because the sLT can be measured only after the sensor detects sweat. Thus, we assumed that the deviation of the sLT from the bLT was attributable to the delay in the OS. Our findings should be interpreted with the following limitations. First, our results were validated using a single type of exercise. Generally, different exercise protocols require different motor functions and associated metabolic pathways. For example, eccentric exercise leads to lower fatigue and lactate responses than concentric exercise with an equivalent load [41,42]. The OS may be delayed during exercise with a significantly short warm-up time. Second, because of the observational study design, the influence of selection bias cannot be completely excluded. The current study mainly included healthy college- aged men and had a relatively small number of cases, especially untrained participants. Further research is warranted, including with untrained subjects and women, because they may differ in their sweat kinetics, muscle mass, and lactate kinetics. Third, the sweat lactate values were obtained from the upper arms. Regional differences in sweat kinetics during cycling exercises have already been reported, and some have revealed higher sweating rates in the forehead and chest [11,43]. The relationship between the sweat kinetics and sweat lactate kinetics at different sites should be analyzed. Fourth, other parameters of Sensors 2023, 23, 3378 9 of 11 sweat dynamics, such as the local sweat rate, were not examined. The OS is an observed phenomenon at the same deep body temperature, even at different relative intensities of exercise. In contrast, the sweat rate increases significantly with increasing exercise intensity [11]. The effects of different local sweat rates on the sLT should also be examined. 5. Conclusions The results of this study revealed a poor correlation between the OS and sLT, which supports the notion that the sLT is strongly correlated with the bLT, independent of the preceding OS during incremental exercise in healthy men. These findings provide further information regarding LT measurements of sweat lactate as a noninvasive, continuous, real-time analytical alternative to blood lactate testing. Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/s23073378/s1. Figure S1: Sweat lactate sensing device. Author Contributions: Conceptualization, Y.M., H.O., T.S., D.N. and Y.K. (Yoshinori Katsumata); methodology, Y.M., H.O. and T.S.; formal analysis, Y.M., H.O., and T.S.; investigation, Y.M., H.O., T.S., J.N., H.F., K.I., S.H. and Y.K. (Yusuke Kobari); resources, D.N. and Y.K. (Yoshinori Katsumata); data curation, Y.M., H.O. and T.S.; writing—original draft preparation, Y.M.; writing—review and editing, H.O., T.S., D.N., Y.K. (Yoshinori Katsumata), J.N., H.F., K.I., S.H. and Y.K. (Yusuke Kobari); visualization, Y.M., H.O., T.S. and Y.K. (Yoshinori Katsumata); supervision, M.N. and T.N.; project administration, D.N. and Y.K. (Yoshinori Katsumata); funding acquisition, D.N. and Y.K. (Yoshinori Katsumata). All authors have read and agreed to the published version of the manuscript. Funding: This study was funded by the Japan Agency for Medical Research and Development (award numbers: 19ek0210130h0001, 20ek0210130h0002, and 21ek0210130h0003) and the Keio University Global Research Institute IoT Healthcare Research Consortium (grant number: 02-066-0008). Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki and ethical guidelines for medical and health research involving human participants and was approved by the Institutional Review Board of our institution (approval no. 20180357). Informed Consent Statement: Each participant provided written consent after being fully informed of the study purpose, possible risks, or discomfort associated with the experimental protocol, as well as the publication of the findings before enrollment. Written informed consent was obtained from the subjects for the publication of this paper. Data Availability Statement: The datasets supporting the conclusions of this study are available from the corresponding author upon reasonable request. Acknowledgments: We are grateful to Daichi Nishiumi and Yoshikazu Kikuchi for their assistance with this study. Conflicts of Interest: Daisuke Nakashima is the president of Grace Imaging, Inc., and holds shares in this company, which sells lactic-acid-sensing equipment. Daisuke Nakashima was not involved in the data acquisition and analysis. References 1. Poole, D.C.; Rossiter, H.B.; Brooks, G.A.; Gladden, L.B. The anaerobic threshold: 50+ years of controversy. J. Physiol. 2021, 599, 737–767. [CrossRef] 2. Svedahl, K.; MacIntosh, B.R. Anaerobic threshold: The concept and methods of measurement. Can. J. Appl. Physiol. 2003, 28, 299–323. [CrossRef] [PubMed] 3. Faude, O.; Kindermann, W.; Meyer, T. Lactate threshold concepts: How valid are they? Sports Med. 2009, 39, 469–490. [CrossRef] [PubMed] 4. Beneke, R.; Leithäuser, R.M.; Ochentel, O. Blood lactate diagnostics in exercise testing and training. Int. J. Sports Physiol. Perform. 2011, 6, 8–24. [CrossRef] [PubMed] 5. Kindermann, W.; Simon, G.; Keul, J. The significance of the aerobic-anaerobic transition for the determination of work load intensities during endurance training. Eur. J. Appl. Physiol. Occup. Physiol. 1979, 42, 25–34. [CrossRef] [PubMed] 6. Hollmann, W. 42 years ago—Development of the concepts of ventilatory and lactate threshold. Sports Med. 2001, 31, 315–320. [CrossRef] Sensors 2023, 23, 3378 10 of 11 7. Seki, Y.; Nakashima, D.; Shiraishi, Y.; Ryuzaki, T.; Ikura, H.; Miura, K.; Suzuki, M.; Watanabe, T.; Nagura, T.; Matsumato, M.; et al. A novel device for detecting anaerobic threshold using sweat lactate during exercise. Sci. Rep. 2021, 11, 4929. [CrossRef] 8. Okawara, H.; Sawada, T.; Nakashima, D.; Maeda, Y.; Minoji, S.; Morisue, T.; Katsumata, Y.; Matsumoto, M.; Nakamura, M.; Nagura, T. Kinetic changes in sweat lactate following fatigue during constant workload exercise. Physiol. Rep. 2022, 10, e15169. [CrossRef] 9. Gisolfi, C.V.; Wenger, C.B. Temperature regulation during exercise: Old concepts, new ideas. Exerc. Sport Sci. Rev. 1984, 12, 339–372. [CrossRef] 10. Araki, T.; Inoue, M.; Fujiwara, H. Experiment studies on sweating for exercise prescription: Total body sweat rate in relation to work load in physically trained adult males. J. Hum. Ergol. 1979, 8, 91–99. 11. Kondo, N.; Takano, S.; Aoki, K.; Shibasaki, M.; Tominaga, H.; Inoue, Y. Regional differences in the effect of exercise intensity on thermoregulatory sweating and cutaneous vasodilation. Acta Physiol. Scand. 1998, 164, 71–78. [CrossRef] [PubMed] 12. Buono, M.J.; Lee, N.V.; Miller, P.W. The relationship between exercise intensity and the sweat lactate excretion rate. J. Physiol. Sci. 2010, 60, 103–107. [CrossRef] [PubMed] 13. Green, J.M.; Bishop, P.A.; Muir, I.H.; Lomax, R.G. Gender differences in sweat lactate. Eur. J. Appl. Physiol. 2000, 82, 230–235. [CrossRef] 14. Green, J.M.; Bishop, P.A.; Muir, I.H.; Lomax, R.G. Lactate-sweat relationships in younger and middle-aged men. J. Aging Phys. Act. 2001, 9, 67–77. [CrossRef] 15. Du Bois, D. Clinical calorimetry: Tenth paper a formula to estimate the approximate surface area if height and weight be known. Arch. Intern. Med. 1916, 17, 863–871. [CrossRef] 16. Wiles, J.D.; Bird, S.R.; Hopkins, J.; Riley, M. Effect of caffeinated coffee on running speed, respiratory factors, blood lactate and perceived exertion during 1500-m treadmill running. Br. J. Sports Med. 1992, 26, 116–120. [CrossRef] 17. Okawara, H.; Sawada, T.; Nakashima, D.; Maeda, Y.; Minoji, S.; Morisue, T.; Katsumata, Y.; Matsumoto, M.; Nakamura, M.; Nagura, T. Realtime monitoring of local sweat rate kinetics during constant-load exercise using perspiration-meter with airflow compensation system. Sensors 2022, 22, 5473. [CrossRef] 18. Sawada, T.; Okawara, H.; Nakashima, D.; Ikeda, K.; Nagahara, J.; Fujitsuka, H.; Hoshino, S.; Maeda, Y.; Katsumata, Y.; Nakamura, M.; et al. Constant load pedaling exercise combined with electrical muscle stimulation leads to an early increase in sweat lactate levels. Sensors 2022, 22, 9585. [CrossRef] 19. Katsumata, Y.; Sano, M.; Okawara, H.; Sawada, T.; Nakashima, D.; Ichihara, G.; Fukuda, K.; Sato, K.; Kobayashi, E. Laminar flow ventilation system to prevent airborne infection during exercise in the COVID-19 crisis: A single-center observational study. PLoS ONE 2021, 16, e0257549. [CrossRef] 20. Baker, L.B.; Wolfe, A.S. Physiological mechanisms determining eccrine sweat composition. Eur. J. Appl. Physiol. 2020, 120, 719–752. [CrossRef] 21. Benzinger, T.H. On Physical Heat Regulation and the Sense of Temperature in Man. Proc. Natl. Acad. Sci. USA 1959, 45, 645–659. [CrossRef] [PubMed] 22. Kondo, N.; Shibasaki, M.; Aoki, K.; Koga, S.; Inoue, Y.; Crandall, C.G. Function of human eccrine sweat glands during dynamic exercise and passive heat stress. J. Appl. Physiol. 2001, 90, 1877–1881. [CrossRef] [PubMed] 23. Nadel, E.R.; Bullard, R.W.; Stolwijk, J.A. Importance of skin temperature in the regulation of sweating. J. Appl. Physiol. 1971, 31, 80–87. [CrossRef] [PubMed] 24. Chappuis, P.; Pittet, P.; Jequier, E. Heat storage regulation in exercise during thermal transients. J. Appl. Physiol. 1976, 40, 384–392. [CrossRef] [PubMed] 25. Montain, S.J.; Latzka, W.A.; Sawka, M.N. Control of thermoregulatory sweating is altered by hydration level and exercise intensity. J. Appl. Physiol. 1995, 79, 1434–1439. [CrossRef] 26. Saltin, B.; Gagge, A.P. Sweating and body temperatures during exercise. Int. J. Biometeorol. 1971, 15, 189–194. [CrossRef] 27. Taylor, W.F.; Johnson, J.M.; Kosiba, W.A.; Kwan, C.M. Graded cutaneous vascular responses to dynamic leg exercise. J. Appl. Physiol. 1988, 64, 1803–1809. [CrossRef] 28. Weiner, J.S.; Van Heyningen, R.E. Observations on lactate content of sweat. J. Appl. Physiol. 1952, 4, 734–744. [CrossRef] 29. Astrand, I. Lactate Content in Sweat. Acta Physiol. Scand. 1963, 58, 359–367. [CrossRef] 30. Brooks, G.A. The science and translation of lactate shuttle theory. Cell Metab. 2018, 27, 757–785. [CrossRef] 31. Inoue, Y.; Tanaka, Y.; Omori, K.; Kuwahara, T.; Ogura, Y.; Ueda, H. Sex- and menstrual cycle-related differences in sweating and cutaneous blood flow in response to passive heat exposure. Eur. J. Appl. Physiol. 2005, 94, 323–332. [CrossRef] 32. Frye, A.J.; Kamon, E. Sweating efficiency in acclimated men and women exercising in humid and dry heat. J. Appl. Physiol. Respir. Environ. Exerc. Physiol. 1983, 54, 972–977. [CrossRef] 33. Madeira, L.G.; da Fonseca, M.A.; Fonseca, I.A.; de Oliveira, K.P.; Passos, R.L.; Machado-Moreira, C.A.; Rodrigues, L.O. Sex-related differences in sweat gland cholinergic sensitivity exist irrespective of differences in aerobic capacity. Eur. J. Appl. Physiol. 2010, 109, 93–100. [CrossRef] [PubMed] 34. Morimoto, T.; Slabochova, Z.; Naman, R.K.; Sargent, F., 2nd. Sex differences in physiological reactions to thermal stress. J. Appl. Physiol. 1967, 22, 526–532. [CrossRef] [PubMed] 35. Henane, R.; Flandrois, R.; Charbonnier, J.P. Increase in sweating sensitivity by endurance conditioning in man. J. Appl. Physiol. Respir. Environ. Exerc. Physiol. 1977, 43, 822–828. [CrossRef] Sensors 2023, 23, 3378 11 of 11 36. Yamazaki, F.; Fujii, N.; Sone, R.; Ikegami, H. Mechanisms of potentiation in sweating induced by long-term physical training. Eur. J. Appl. Physiol. Occup. Physiol. 1994, 69, 228–232. [CrossRef] 37. Ichinose-Kuwahara, T.; Inoue, Y.; Iseki, Y.; Hara, S.; Ogura, Y.; Kondo, N. Sex differences in the effects of physical training on sweat gland responses during a graded exercise. Exp. Physiol. 2010, 95, 1026–1032. [CrossRef] [PubMed] 38. Baum, E.; Brück, K.; Schwennicke, H.P. Adaptive modifications in the thermoregulatory system of long-distance runners. J. Appl. Physiol. 1976, 40, 404–410. [CrossRef] 39. Scobbie, R.B.; Sofaer, J.A. Sweat pore count, hair density and tooth size: Heritability and genetic correlation. Hum. Hered. 1987, 37, 349–353. [CrossRef] 40. Saha, T.; Fang, J.; Mukherjee, S.; Knisely, C.T.; Dickey, M.D.; Velev, O.D. Osmotically Enabled Wearable Patch for Sweat Harvesting and Lactate Quantification. Micromachines 2021, 12, 1513. [CrossRef] 41. Horstmann, T.; Mayer, F.; Maschmann, J.; Niess, A.; Roecker, K.; Dickhuth, H.H. Metabolic reaction after concentric and eccentric endurance-exercise of the knee and ankle. Med. Sci. Sports Exerc. 2001, 33, 791–795. [CrossRef] [PubMed] 42. Goto, K.; Ishii, N.; Kizuka, T.; Kraemer, R.R.; Honda, Y.; Takamatsu, K. Hormonal and metabolic responses to slow movement resistance exercise with different durations of concentric and eccentric actions. Eur. J. Appl. Physiol. 2009, 106, 731–739. [CrossRef] [PubMed] 43. Cotter, J.D.; Patterson, M.J.; Taylor, N.A. The topography of eccrine sweating in humans during exercise. Eur. J. Appl. Physiol. Occup. Physiol. 1995, 71, 549–554. [CrossRef] [PubMed] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Implications of the Onset of Sweating on the Sweat Lactate Threshold.

03-23-2023

Maeda, Yuta,Okawara, Hiroki,Sawada, Tomonori,Nakashima, Daisuke,Nagahara, Joji,Fujitsuka, Haruki,Ikeda, Kaito,Hoshino, Sosuke,Kobari, Yusuke,Katsumata, Yoshinori,Nakamura, Masaya,Nagura, Takeo

eng

PMC9820022

Citation: Yu, Y.; Wang, R.; Li, D.; Lu, Y. Monitoring Physiological Performance over 4 Weeks Moderate Altitude Training in Elite Chinese Cross-Country Skiers: An Observational Study. Int. J. Environ. Res. Public Health 2023, 20, 266. https://doi.org/10.3390/ ijerph20010266 Academic Editor: Roberto Cejuela Anta Received: 20 November 2022 Revised: 18 December 2022 Accepted: 20 December 2022 Published: 24 December 2022 Copyright: © 2022 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). International Journal of Environmental Research and Public Health Article Monitoring Physiological Performance over 4 Weeks Moderate Altitude Training in Elite Chinese Cross-Country Skiers: An Observational Study Yichao Yu 1,2, Ruolin Wang 3, Dongye Li 1,2 and Yifan Lu 1,4,* 1 The School of Sports Medicine and Rehabilitation, Beijing Sports University, Beijing 100084, China 2 The Graduate School, Beijing Sport University, Beijing 100084, China 3 Melbourne School of Population and Global Health, The University of Melbourne, Carlton, VIC 3053, Australia 4 Key Laboratory of Sports and Physical Fitness of the Ministry of Education, Beijing Sport University, Beijing 100084, China * Correspondence: [email protected]; Tel.: +86-137-0110-2889 Abstract: The current observational study aimed to monitor the physiological performance over 4 weeks of living and training at a moderate altitude in elite Chinese cross-country skiers (8 males, mean age 20.83 ± 1.08 years). Lactate threshold, maximal oxygen uptake, blood, and body com- position tests were performed at different time points to investigate the changes in physiological performance. The data were analysed by a one-way repeated measures ANOVA and a paired sample T-test between the test results. During the training camp, systematic load monitoring was carried out. Lactate threshold velocity, lactate threshold heart rate, and upper body muscle mass increased significantly (p < 0.01) after moderate altitude training. Maximum oxygen uptake was reduced compared to pre-tests (p < 0.05). Aerobic capacity parameters (maximal oxygen uptake, haemoglobin, red blood cell count) did not significantly increase after athletes returned to sea level (p > 0.05). These findings suggest that 4 weeks of moderate altitude training can significantly improve athletes’ lactate threshold and upper body muscle mass; no significant improvement in other aerobic capacity was seen. Exposure time, training load, and nutritional strategies should be thoroughly planned for optimal training of skiers at moderate altitudes. Keywords: cross-country skiing; physiological performance; moderate altitude training 1. Introduction One of the most demanding endurance sports, cross-country skiing (XC-skiing), re- quires athletes to compete on courses between 1.5 km and 50 km long over various terrains. The different techniques of classical and skating styles require a high level of technical and physiological performance in athletes [1]. In sports training, physiological performance refers to the physiological adaptation of the athletes’ bodies for training or competition. Us- ing physiological and biochemical indicators to monitor athletes’ performance has become an essential part of training in endurance sports [2]. Altitude training is widely used in endurance sports, which can induce different responses depending on the modality used [3,4]. It can generally be divided into three main types of training: living high training high (LH-TH), living high training low (LH- TL), and living low training high (LL-TH). Coaches and sports researchers use LH-TH to increase red blood cell (RBC) count, haemoglobin (Hb), maximal oxygen uptake (VO2max), and other performance measures at sea level [5]. However, athletes cannot maintain the same intensities as those at sea level. To avoid the limitations of LH-TH, LH-TL is used to improve athletes’ physiological performance while maintaining training intensity [6,7]. Some studies have found that LH-TL can enhance haematological and neuromuscular Int. J. Environ. Res. Public Health 2023, 20, 266. https://doi.org/10.3390/ijerph20010266 https://www.mdpi.com/journal/ijerph Int. J. Environ. Res. Public Health 2023, 20, 266 2 of 11 adaptations [8,9]. In LL-TH, athletes live at sea level, with hypoxic exposure lasting for several seconds to several hours during training and repeated over several days to weeks. LL-TH has been proven to improve athletes’ erythropoietin (EPO) and skeletal muscle mitochondrial density of hypoxia-inducible factor 1α (HIF-1α) [10,11]. The core aim of altitude training is to enhance the athletes’ physiological performance through hypoxic exposure in different ways. Meanwhile, unreasonable altitude training can lead to overreaching and even overtraining of athletes [12]. In recent years, many sports researchers have explored the effects of living and training at moderate altitudes (1500–3000 m), which is closer to the traditional LH-TH modality but lower in altitude [7,13,14]. From the point of view of combining exercise physiology and sports training, such an environment provides a certain level of hypoxic stimulation to the body while approaching the intensity of basic training, thus improving physiological performance [15,16]. Research has demonstrated that three weeks of training at 1800 m significantly increased haemoglobin levels in well-trained runners [17]. Czuba et al. [18] found that prolonged exposure to hypoxic conditions (simulated moderate altitude) would continue to promote the synthesis of erythropoietin, improving RBCs’ ability to carry more oxygen. Karlsson et al. reported that 17–21 days of training at 1800 m increased elite XC-skiers and biathletes’ lactate threshold velocity [19]. Four weeks of moderate altitude training (2200 m) increased the resting metabolic rate and haemoglobin in highly trained middle-distance runners [20]. The benefits of training at moderate altitudes include increased haemoglobin mass, body metabolic efficiency, and enhanced lactate threshold [21]. In the past two decades, most Olympic XC-skiing events have been held at 1500–1800 metres, and training at moderate altitudes has become common [22,23]. However, little systematic research has focused on changes in the physiological performance of XC-skiers at moderate altitudes. This observational study aims to measure changes in physiological performance after four weeks of living and training at moderate altitude, meanwhile providing referenceable physiological evidence to help XC-skiing coaches and practitioners when using moderate altitude training. 2. Materials and Methods 2.1. Subjects and Study Design This research is an observational study of eight elite Chinese cross-country skiers over four weeks living and training at the Chinese national snow sports training base in BaShang, Chengde, Hebei Province (average sea level at 1510–1700 m, latitude at 44.5◦ N). A ‘polarised training’ plan based on a traditional Nordic XC-skiing training programme was used, and the daily training data were compiled by the scientists accompanying the team [23,24]. Throughout the design of the study, three tests were conducted (Figure 1): the pre-test (55 m above sea level) was completed five days before the moderate altitude training, the mid-test (1550 m above sea level) was performed on day 15 of the moderate altitude training, and the post-test (55 m above sea level) was conducted five days after the moderate altitude training. The participants in this study included eight male athletes involved in prepara- tion for the 2022 Winter Olympics in Beijing. The mean age was 20.83 ± 1.08 years; weight was 69.73 ± 5.12 kg; height was 179.63 ± 5.93 cm, and mean training years was 4.18 ± 1.92 years. All the athletes involved in the experiment were at the elite-level [25], were in excellent physical condition, and were uninjured. Prior to data collection, all athletes were given an informed permission form and a description of the experiment’s objective and potential dangers. Signing both agreements demonstrated a willingness to participate voluntarily. The Sports Science Experimental Ethics Committee of Beijing Sport University approved the research protocol (Grant No. 201906711). Int. J. Environ. Res. Public Health 2023, 20, 266 3 of 11 Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 3 of 12 Figure 1. Study Design of the Research. 2.2. Physiological Performance Test Physiological performance tests in treadmill running were conducted using protocols developed by the Norwegian Top Sport Centre. First, an “intermittent” incremental test was used to obtain the lactate threshold. After a 5 min recovery, the athletes conducted an incremental test to determine VO2max [26,27]. All athletes implemented a standardised warm-up process prior to the lactate threshold test under the supervision of a professional fitness coach. The warm-up routine consisted of 10 min of low-intensity jogging on a treadmill with an athlete rating of perceived exertion (RPE) of 2, followed by 10 push-ups and five squat jumps. After the athletes warmed up, the lactate threshold test was implemented on the treadmill (RL2500E, Rodby, Södertalje, Sweden) using the incremental load test method. The incline angle of the treadmill was set at 10.5% and maintained throughout the test, with the starting speed of the treadmill set at 7 km/h. Athletes ran at a constant speed for five minutes at each level of speed, with a 30 s rest interval at the end of the run. The treadmill’s speed was increased by 1 km/h between runs [26]. The heart rate level of each athlete was recorded in the last 30 s of each stage. The athlete’s blood lactate level was tested immediately after each level of the running platform test. Blood lactate concentration was measured immediately after exercise using an EKF benchtop blood lactate metre (Boisen, EKF Industrial Electronics, Magdeburg, Germany). Furthermore, the athlete’s RPE was recorded using a 0–10 scale. The lactate threshold was defined as a blood lactate level of 4 mmol/L−1 [28]; when the threshold exceeded 4 mmol/L−1, the test was stopped. Treadmill speed at 4 mmol/L−1 was calculated using linear interpolation [28]. Following the lactate threshold test, the athletes rested for 5 min before assessing their maximum oxygen uptake using a portable gas metabolism analyser (MetaMax 3B, Cortex, Leipzig, Germany) [26,27]. The treadmill incline angle for the VO2max test was 10.5%, and the treadmill start speed was 1 km/h below the end speed of the lactate threshold test. The treadmill’s speed was increased by 1 km/h every minute from the beginning of the test until the participant was exhausted. Throughout the test, the athlete wore a ventilation mask to evaluate his oxygen uptake volume. The VO2max was defined as the average of the two highest and consecutive 30 s measurements. A heart rate belt (H10, Polar, Finland) was used to monitor the athlete’s heart rate. Maximum HR was defined as the highest 5 s heart rate measurement during the VO2max test. The blood lactate concentration was measured 1 min after completing the test, and the RPE values were recorded, with the RPE counted on a 0–10 scale. The athlete’s final treadmill speed, maximum oxygen uptake, and respiratory exchange ratio (RER) were recorded. Blood tests were performed between 6:00 am and 7:00 am on each test day, with venous blood drawn by medical staff in the morning while the athlete was awake and Figure 1. Study Design of the Research. 2.2. Physiological Performance Test Physiological performance tests in treadmill running were conducted using protocols developed by the Norwegian Top Sport Centre. First, an “intermittent” incremental test was used to obtain the lactate threshold. After a 5 min recovery, the athletes conducted an incremental test to determine VO2max [26,27]. All athletes implemented a standardised warm-up process prior to the lactate threshold test under the supervision of a professional fitness coach. The warm-up routine consisted of 10 min of low-intensity jogging on a treadmill with an athlete rating of perceived exertion (RPE) of 2, followed by 10 push- ups and five squat jumps. After the athletes warmed up, the lactate threshold test was implemented on the treadmill (RL2500E, Rodby, Södertalje, Sweden) using the incremental load test method. The incline angle of the treadmill was set at 10.5% and maintained throughout the test, with the starting speed of the treadmill set at 7 km/h. Athletes ran at a constant speed for five minutes at each level of speed, with a 30 s rest interval at the end of the run. The treadmill’s speed was increased by 1 km/h between runs [26]. The heart rate level of each athlete was recorded in the last 30 s of each stage. The athlete’s blood lactate level was tested immediately after each level of the running platform test. Blood lactate concentration was measured immediately after exercise using an EKF benchtop blood lactate metre (Boisen, EKF Industrial Electronics, Magdeburg, Germany). Furthermore, the athlete’s RPE was recorded using a 0–10 scale. The lactate threshold was defined as a blood lactate level of 4 mmol/L−1 [28]; when the threshold exceeded 4 mmol/L−1, the test was stopped. Treadmill speed at 4 mmol/L−1 was calculated using linear interpolation [28]. Following the lactate threshold test, the athletes rested for 5 min before assessing their maximum oxygen uptake using a portable gas metabolism analyser (MetaMax 3B, Cortex, Leipzig, Germany) [26,27]. The treadmill incline angle for the VO2max test was 10.5%, and the treadmill start speed was 1 km/h below the end speed of the lactate threshold test. The treadmill’s speed was increased by 1 km/h every minute from the beginning of the test until the participant was exhausted. Throughout the test, the athlete wore a ventilation mask to evaluate his oxygen uptake volume. The VO2max was defined as the average of the two highest and consecutive 30 s measurements. A heart rate belt (H10, Polar, Finland) was used to monitor the athlete’s heart rate. Maximum HR was defined as the highest 5 s heart rate measurement during the VO2max test. The blood lactate concentration was measured 1 min after completing the test, and the RPE values were recorded, with the RPE counted on a 0–10 scale. The athlete’s final treadmill speed, maximum oxygen uptake, and respiratory exchange ratio (RER) were recorded. Blood tests were performed between 6:00 a.m. and 7:00 a.m. on each test day, with venous blood drawn by medical staff in the morning while the athlete was awake and fasting. Routine blood tests were analysed using a fully automated haematology analyser (BC-5180CRP Automatic Haematology Analyser, Myriad, Shenzhen, China). Blood urea Int. J. Environ. Res. Public Health 2023, 20, 266 4 of 11 (BUN) and creatine kinase (CK) were analysed using a fully automated biochemistry analyser (AU680 Automatic Biochemical Analyser, Beckman Coulter, Brea, CA, USA). All instruments were standardised using the original and matching reagents. Body composition testing was conducted only on the morning of the pre-tests and post-tests. The fat, muscle, and bone mass of the athlete’s body and all body segments (upper body, trunk, and lower body) were measured using a dual-energy X-ray bone density analyser (Luna iDXA, General Electric Company, Schenectady, NY, USA) after the athlete had finished the venous blood draw. 2.3. Training Load Monitoring This study’s four-week moderate altitude training programme involved six days of training per week, with Monday for rest and recovery. Low-intensity aerobic training was conducted five days before and after four weeks of moderate altitude training. All training loads, including volumes and intensities, were developed and implemented by the heart rate zones obtained from the athletes’ baseline physiological tests. Load monitoring used a 5-zone load intensity model (Table 1) designed by the Norwegian National Olympic Committee based on a combination of laboratory test results and actual training [24,29]. Table 1. The 5-zone model was used in the current study. Intensity Zone Blood Lactate (mmol/L) Heart Rate (% Max) RPE 5 HIT 6.0–10.0 92–97 8–10 4 4.0–6.0 87–92 7–9 3 MIT 2.5–4.0 82–87 4–7 2 LIT 1.5–2.5 72–82 3–5 1 0.8–1.5 55–72 ≤4 Abbreviations: LIT: low-intensity training; MIT, moderate-intensity training, HIT, high-intensity training. The training statistics format adhered to the training recording format suggested by the Norwegian National Olympic Committee, using exercise forms, training forms, and exercise intensity for load recording for all workouts (Figure 2) [24]. Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 4 of 12 fasting. Routine blood tests were analysed using a fully automated haematology analyser (BC-5180CRP Automatic Haematology Analyser, Myriad, Shenzhen, China). Blood urea (BUN) and creatine kinase (CK) were analysed using a fully automated biochemistry analyser (AU680 Automatic Biochemical Analyser, Beckman Coulter, Brea, CA, USA). All instruments were standardised using the original and matching reagents. Body composition testing was conducted only on the morning of the pre-tests and post-tests. The fat, muscle, and bone mass of the athlete’s body and all body segments (upper body, trunk, and lower body) were measured using a dual-energy X-ray bone den- sity analyser (Luna iDXA, General Electric Company, Schenectady, NY, USA) after the athlete had finished the venous blood draw. 2.3. Training Load Monitoring This study’s four-week moderate altitude training programme involved six days of training per week, with Monday for rest and recovery. Low-intensity aerobic training was conducted five days before and after four weeks of moderate altitude training. All training loads, including volumes and intensities, were developed and implemented by the heart rate zones obtained from the athletes’ baseline physiological tests. Load monitoring used a 5-zone load intensity model (Table 1) designed by the Norwegian National Olympic Committee based on a combination of laboratory test results and actual training [24,29]. Table 1. The 5-zone model was used in the current study. Intensity Zone Blood Lactate (mmol/L) Heart Rate (% Max) RPE 5 HIT 6.0–10.0 92–97 8–10 4 4.0–6.0 87–92 7–9 3 MIT 2.5–4.0 82–87 4–7 2 LIT 1.5–2.5 72–82 3–5 1 0.8–1.5 55–72 ≤4 Abbreviations: LIT: low-intensity training; MIT, moderate-intensity training, HIT, high-intensity training. The training statistics format adhered to the training recording format suggested by the Norwegian National Olympic Committee, using exercise forms, training forms, and exercise intensity for load recording for all workouts (Figure 2) [24]. Figure 2. Training distribution methods. abbreviations: ACT. FORMS = activity forms. Figure 2. Training distribution methods. abbreviations: ACT. FORMS = activity forms. 2.4. Statistical Analyses Data statistics and analysis were performed using IBM SPSS 25.0 and Excel 2019 software. All data were presented as mean ± SD and were tested for normality using the Shapiro–Wilk test before processing. Differences in the lactate threshold test, maximum oxygen uptake test, and blood test between the three tests were assessed using repeated measures ANOVA. Mauchly’s Test of Sphericity was used, and in the multiple comparisons Int. J. Environ. Res. Public Health 2023, 20, 266 5 of 11 were used the Bonferroni post hoc test with significant changes at p < 0.05 and highly significant differences at p < 0.01. Differences in body composition between pre- and post- tests were using the paired sample T-test with significant changes at p < 0.05 and highly significant differences at p < 0.01. 3. Results During four weeks of moderate altitude training, the eight elite Chinese male XC- skiers recorded a total training time of 67 h, with an average of 16.7 h per week. The average percentage of time spent in the i1-i5 intensity interval over four weeks was 63.6%, 28.5%, 5.8%, 1.4%, and 0.7%. The average LIT (i1–i2), MIT (i3), and HIT (i4–i5) intensities were 92.1%, 5.8%, and 2.1%, respectively. Endurance training throughout the cycle included roller skiing, skiing, running, and Nordic walking (Table 2). Table 2. Training Characteristics of 4 weeks Moderate Altitude Training. Weekly Training Patterns Week1 Week2 Week3 Week4 Total Training Total Training Time (h·wk−1) 13.0 ± 0.6 17.4 ± 1.2 16.9 ± 2.7 19.7 ± 1.4 Training Sessions·wk−1 10.4 ± 1.1 12 ± 2.3 10.5 ± 2.4 12.6 ± 1.4 Endurance Distance (km·wk−1) 155.4 ± 19.5 205.3 ± 35.1 208.9 ± 38.3 237.6 ± 17.6 Training forms Endurance Training Time (h·wk−1) 11.8 ± 1.0 14.3 ± 1.2 14.2 ± 1.6 16.9 ± 0.7 Strength Training Time (h·wk−1) 1.1 ± 0.9 3 ± 0.5 2.5 ± 1.3 2.7 ± 1.1 Sprint Training Time (h·wk−1) 0.1 ± 0.1 0.2 ± 0.1 0.2 ± 0.1 0.1 ± 0.1 Endurance Intensity distribution Zone 1 (h·wk−1) 7.5 ± 1.6 9.2 ± 3.4 9.9 ± 2.4 9.8 ± 3.2 Zone 2 (h·wk−1) 3.3 ± 1.5 3.9 ± 1.8 3.7 ± 2.5 5.4 ± 2.6 Zone 3 (h·wk−1) 0.6 ± 0.4 0.8 ± 1 0.5 ± 0.4 1.4 ± 0.9 Zone 4 (h·wk−1) 0.2 ± 0.2 0.3 ± 0.3 0.1 ± 0.1 0.2 ± 0.2 Zone 5 (h·wk−1) 0.2 ± 0.1 0.1 ± 0.1 0 ± 0.1 0.1 ± 0.1 According to Mauchly’s spherical hypothesis test, the variance–covariance matrix of all dependent variables was p < 0.01. The findings in Table 3 show the following. The athletes’ lactate threshold velocity, lactate threshold heart rate, VO2max, and maximal heart rate at VO2max were statistically significant before and after the three tests (p < 0.01). The lactate threshold velocity was 0.24 m·s−1 higher in the post-test compared to the pre-test (95% CI: 0.046–0.436, p < 0.05) and 0.31 m·s−1 higher in the post-test compared to the mid-test (95% CI: 0.020–0.056, p < 0.05). The lactate threshold heart rate was 9.0 beats·min−1 higher in the post-test compared to the pre-test (95% CI: 5.079–12.921, p < 0.05) and 9.4 beats·min−1 higher in the post-test compared to the mid-test (95% CI: 0.020–0.056, p < 0.05). The VO2max were 6.11 L·min−1 (95% CI: 2.293–9.932, p < 0.05) and 0.57 L·min−1 (95% CI: 0.213–0.917, p < 0.05) higher in the pre-test compared to the mid-test. The post-test result of VO2max increased separately 3.50 mL·min−1·kg−1 (95% CI: 0.545–6.455, p < 0.05) and 0.39 L·min−1 (95% CI: 0.108–0.674, p < 0.05) compared to the mid-test. The maximum heart rate of the athletes tested was 9.0 beats·min−1 higher in the post-test than in the mid-test (95% CI: 2.763–9.23, p < 0.01). The heart rate–velocity and lactate–velocity curves plotted from the lactate threshold test data are shown in Figure 3. After training at moderate altitude (post-test), the lactate threshold heart rate was significantly higher than pre-test and mid-test at all speed levels (p < 0.05). Int. J. Environ. Res. Public Health 2023, 20, 266 6 of 11 Table 3. Changes in Lactate Threshold and Maximum Oxygen Uptake after 4 weeks Moderate Altitude Training. Pre-Test Mid-Test Post-Test Time Effect Lactate threshold velocity(m·s−1) 3.02 ± 0.18 2.96 ± 0.14 3.27 ± 0.24 *$ F(2,14) = 10.40, p = 0.002 Lactate threshold HR(beats·min−1) 179.5 ± 9.8 179.1 ± 10.3 188.5 ± 9.2 **$$ F(2,14) = 26.46, p < 0.001 VO2max (mL·min−1·kg−1) 73.74 ± 3.63 67.63 ± 2.13 ** 71.12 ± 3.14 $ F(2,14) = 15.48, p < 0.001 VO2max (L·min−1) 4.82 ± 0.36 4.25 ± 0.36 ** 4.64 ± 0.43 $ F(2,14) = 19.70, p < 0.001 RER 1.23 ± 0.09 1.23 ± 0.05 1.11 ± 0.03 **$ F(2,14) = 14.53, p < 0.001 Maximum HR(beats·min−1) 197.4 ± 11.6 192.9 ± 8.6 198.9 ± 9.1 $$ F(2,14) = 9.41, p = 0.003 Maximum Lactate(mmol·L−1) 13.31 ± 1.42 11.69 ± 2.34 11.33 ± 2.79 F(2,14) = 1.84, p = 0.196 * indicates a significant difference compared to pre-test (* p < 0.05, ** p < 0.01). $ indicates a significant difference compared to mid-test ($ p < 0.05, $$ p < 0.01). Abbreviations: HR = heart rate, VO2max = maximal oxygen uptake, RER = respiratory exchange ratio. Int. J. Environ. Res. Public Health 2023, 20, x FOR PEER REVIEW 6 of 12 The heart rate–velocity and lactate–velocity curves plotted from the lactate threshold test data are shown in Figure 3. After training at moderate altitude (post-test), the lactate threshold heart rate was significantly higher than pre-test and mid-test at all speed levels (p < 0.05). (a) (b) Figure 3. Eight elite Chinese cross-country skiers’ lactate-velocity curve (a) and heart rate-curve (b) from lactate threshold test. * indicates a significant difference compared to pre-test (* p < 0.05). Ab- breviations: La = lactate, HR = heart rate. As demonstrated in Table 4, there were significant changes (p < 0.05 or p < 0.01) in RBC, BUN, and %MXD in the three tests. The RBC increased by 0.29 × 106·μL−1 in the pre- test compared to the mid-test (95% CI: 0.041–0.534, p < 0.05). The BUN increased 1.16 mmol·L−1 after two weeks of exposure to the moderate altitude (mid-test) compared to the pre-test (95% CI: 0.195–2.130, p < 0.05). The % MXD increased separately by 3.26% and 3.36% in the pre-test (95% CI: 1.246–5.279, p < 0.01) and mid-test (95% CI: 1.720–5.005, p < 0.01) compared to the post-test. For HCT and WBC, the F-test achieved statistical signifi- cance. There was no significant difference between the measurements. This could happen if the sample size is small, leading to unstable statistical results A paired sample T-test of body composition data during four weeks at moderate al- titude (Table 5) showed that the percentage of upper-body muscle mass increased from 11.74% to 12.03% after altitude training, with a significant increase in upper-body muscle mass before and after training (95% CI: 0.093–0.307, p < 0.01). Table 3. Changes in Lactate Threshold and Maximum Oxygen Uptake after 4 weeks Moderate Al- titude Training. Pre-Test Mid-Test Post-Test Time Effect Lactate threshold velocity(m·s−1) 3.02 ± 0.18 2.96 ± 0.14 3.27 ± 0.24 *$ F(2,14) = 10.40, p = 0.002 Lactate threshold HR(beats·min−1) 179.5 ± 9.8 179.1 ± 10.3 188.5 ± 9.2 **$$ F(2,14) = 26.46, p < 0.001 VO2max (mL·min−1·kg−1) 73.74 ± 3.63 67.63 ± 2.13 ** 71.12 ± 3.14 $ F(2,14) = 15.48, p < 0.001 VO2max (L·min−1) 4.82 ± 0.36 4.25 ± 0.36 ** 4.64 ± 0.43 $ F(2,14) = 19.70, p < 0.001 RER 1.23 ± 0.09 1.23 ± 0.05 1.11 ± 0.03 **$ F(2,14) = 14.53, p < 0.001 Maximum HR(beats·min−1) 197.4 ± 11.6 192.9 ± 8.6 198.9 ± 9.1 $$ F(2,14) = 9.41, p = 0.003 Maximum Lactate(mmol·L−1) 13.31 ± 1.42 11.69 ± 2.34 11.33 ± 2.79 F(2,14) = 1.84, p = 0.196 * indicates a significant difference compared to pre-test (* p < 0.05, ** p < 0.01). $ indicates a significant difference compared to mid-test ($ p < 0.05, $$ p < 0.01). Abbreviations: HR = heart rate, VO2max = maximal oxygen uptake, RER = respiratory exchange ratio. Figure 3. Eight elite Chinese cross-country skiers’ lactate-velocity curve (a) and heart rate-curve (b) from lactate threshold test. * indicates a significant difference compared to pre-test (* p < 0.05). Abbreviations: La = lactate, HR = heart rate. As demonstrated in Table 4, there were significant changes (p < 0.05 or p < 0.01) in RBC, BUN, and %MXD in the three tests. The RBC increased by 0.29 × 106·µL−1 in the pre-test compared to the mid-test (95% CI: 0.041–0.534, p < 0.05). The BUN increased 1.16 mmol·L−1 after two weeks of exposure to the moderate altitude (mid-test) compared to the pre-test (95% CI: 0.195–2.130, p < 0.05). The % MXD increased separately by 3.26% and 3.36% in the pre-test (95% CI: 1.246–5.279, p < 0.01) and mid-test (95% CI: 1.720–5.005, p < 0.01) compared to the post-test. For HCT and WBC, the F-test achieved statistical significance. There was no significant difference between the measurements. This could happen if the sample size is small, leading to unstable statistical results Table 4. Changes in Blood Indicators after 4-weeks Moderate Altitude Training. Pre-Test Mid-Test Post-Test Time Effect HCT 0.48 ± 0.02 0.46 ± 0.02 0.46 ± 0.02 F(2,14) = 6.85, p = 0.008 HGB/(g·L−1) 160.00 ± 10.90 156.00 ± 7.25 160.88 ± 10.48 F(2,14) = 2.71, p = 0.101 RBC/(×106·µL−1) 5.40 ± 0.32 5.11 ± 0.29 * 5.23 ± 0.27 F(2,14) = 8.70, p = 0.004 CK/(U·L−1) 207.63 ± 122.47 233.38 ± 101.77 181.38 ± 80.15 F(2,14) = 1.46, p = 0.265 BUN/(mmol·L−1) 6.65 ± 1.10 7.81 ± 0.74 * 6.71 ± 0.96 F(2,14) = 5.96, p = 0.013 %LYM 38.6 ± 6.3 39.3 ± 8.2 40.6 ± 7.8 F(2,14) = 0.39, p = 0.683 %MXD 7.8 ± 2.7 7.9 ± 1.8 4.6 ± 1.4 **$$ F(2,14) = 16.26, p < 0.001 %NEUT 53.6 ± 6.3 50.8 ± 9.0 53.1 ± 9.3 F(2,14) = 18.41, p = 0.518 WBC/(×103·µL−1) 5.84 ± 1.30 4.93 ± 0.84 5.35 ± 0.63 F(2,14) = 4.86, p = 0.025 * indicates a significant difference compared to pre-test (* p < 0.05, ** p < 0.01). $ indicates a significant difference compared to mid-test ($$ p < 0.01). Abbreviations: HCT = hematocrit, HGB = hemoglobin, RBC= red blood cell, CK = creatine kinase, BUN = blood urea nitrogen, %LYM = percentage of lymphocyte, %MXD = percentage of mononucleosi, %NEUT = percentage of neutrophil, WBC = white blood cell. Int. J. Environ. Res. Public Health 2023, 20, 266 7 of 11 A paired sample T-test of body composition data during four weeks at moderate altitude (Table 5) showed that the percentage of upper-body muscle mass increased from 11.74% to 12.03% after altitude training, with a significant increase in upper-body muscle mass before and after training (95% CI: 0.093–0.307, p < 0.01). Table 5. Changes in Body Composition after 4-weeks Moderate Altitude Training. Pre-Test Post-Test p Values Muscle mass (kg) Whole body 55.66 ± 4.01 56.01 ± 4.02 p = 0.071 Upper body 6.54 ± 0.42 6.74 ± 0.36 ## p = 0.004 Trunk 27.20 ± 2.18 27.47 ± 2.00 p = 0.485 Lower body 18.60 ± 1.82 18.57 ± 1.68 p = 0.924 Fat mass (kg) Whole body 7.81 ± 1.51 7.91 ± 1.14 p = 0.193 Upper body 0.91 ± 0.16 0.97 ± 0.18 p = 0.172 Trunk 3.13 ± 0.84 3.06 ± 0.60 p = 0.649 Lower body 2.91 ± 0.56 3.03 ± 0.47 p = 0.139 Bone mass (kg) Whole body 2.83 ± 0.16 2.84 ± 0.13 p = 0.356 Upper body 0.40 ± 0.01 0.41 ± 0.04 p = 0.172 Trunk 0.84 ± 0.05 0.81 ± 0.04 p = 0.356 Lower body 1.16 ± 0.08 1.16 ± 0.08 p > 0.05 # indicates a significant difference compared to pre-test (## p < 0.01). 4. Discussion The primary contribution of our study is to provide relevant and new data regarding the effects of four weeks of moderate altitude training on the physiological performance of elite Chinese XC-skiers. This study indicates that four weeks of living and training at a moderate altitude leads to elevated lactate threshold and upper-body muscle mass. In the case of lactic acid build-up, the athletes’ cells’ removal of lactic acid somewhat reflects the athletes’ aerobic and lactate metabolism capacity [30]. This study found that lactate threshold velocity and heart rate were significantly higher after four weeks of moderate altitude training, reflecting an increase in the athletes’ lactate threshold and aerobic capacity. The lactate–velocity curve shows that the blood lactate concentration after four weeks of moderate altitude training was consistently lower than the first two tests at the velocity level prior to reaching the lactate threshold. This improvement is consistent with Ingjer et al. [31], who discovered that excellent XC-skiers had significantly reduced lactate levels in the submaximal test after three weeks of moderate altitude training at 1900 m. In addition, a study of a group of British national team distance runners showed a 12% increase in lactate threshold velocity after four weeks of 1500–2000 m endurance training [32]. High biochemical blood indicators are the primary reason for applying altitude train- ing to increase lactate metabolism [33]. However, no similar situation occurred in this study. Unlike the traditional LH-TH plan (over 3000 m), 1550 m may be too low to increase biochemical blood indicators, such as Hb [19]. We speculate that the increase in lactate metabolism capacity has resulted from non-haematological hypoxia-induced changes. In LH-TL, elite athletes living at high altitudes of 2000–3000 m while simultaneously training below 1500 m can enhance muscle buffering capacity [34]. In this study, the living altitude of athletes was much lower, which is more conducive to recovery and may further improve muscle buffering capacity and lactate metabolism. Meanwhile, with the continuous promo- tion of training, the further improvement of the athletes’ economy of action may also have some impacts. Based on the current experimental design, we could not determine whether the effect was due to hypoxic exposure or a training plan. Maximal oxygen uptake is one of the most important indicators of XC-skiers [1,35,36]. We found that the maximum oxygen uptake test on the plateaus was significantly lower Int. J. Environ. Res. Public Health 2023, 20, 266 8 of 11 than on the plains. This is common in altitude training; as the partial pressure of oxygen decreases, oxygen from arterial blood is difficult to deliver to tissue cells, with negative effects on muscle metabolism and contraction, leading to increased peripheral fatigue [12]. Contrary to the traditional LH-TH plan, maximal oxygen uptake levels do not change significantly after returning to sea level. Ingjer et al. [31] and Chapman et al. [37] also found the same results in elite endurance athletes when they were training at 2000–3000 m. Compared with the altitude used in the traditional altitude training modality, this may be because a moderate altitude environment does not cause a strong enough stimulus to the athletes’ bodies [19]. In addition, insufficient training intensity may lead to this phenomenon. According to the training load monitoring results, the lack of HIT training may have contributed to the lack of cardiorespiratory and physiological stimulation. In addition to no changes in physiological indicators, training at 1550 m may not suffer any change in lung diffusing, which was proved in 1800 m swimming training [38]. A detailed examination applied in a meta-analysis about altitude training reveals a lag in the change in maximal oxygen uptake after altitude training. The longer the time after returning to sea level, the more pronounced the increase in maximal oxygen uptake due to altitude training [3]. Whether similar mechanisms exist at this altitude could be further verified by future experiments. RBCs are the body’s carriers of oxygen and play an important role in endurance training. This study found that after two weeks of moderate altitude training stimulation, athletes’ RBC counts and haemoglobin levels tended to decrease compared to the pre- test. After returning to sea level, neither increased relative to pre-training levels. The same result was reported in a study with XC-skiers under long-term moderate altitude training (1500–1800 m) [39]. While no EPO could be measured in the present study, the study of altitude training indicated that higher altitude had a more positive effect on RBC production in athletes [9,40]. Meanwhile, several prior studies have reported a correlation between maximal oxygen uptake values and haemoglobin concentrations in athletes [31]. Maximal oxygen uptake and haemoglobin change trends in this case study are also relevant. Contrary to previous research in other sports, our results indicate that training at 1550 m may not stimulate red cell production with concurrent amelioration of aerobic performance like VO2max. The insufficient hypoxia exposure and some hypoxia-induced disturbances in physiological function will cause it [40]. At the same time, attention should be paid to the exposure time to hypoxia and the environment; the terrain, wind speed and load could have altered this outcome. The training camp environment is windy all year round, which may lead to the loss of athletes’ body fluids, resulting in negative effects. Typically, the duration of moderate altitude training is generally 7 to 21 days [23], and 28 days of hypoxic exposure may have a negative impact. Moreover, we did not control plasma and blood volume, which may have affected the results. BUN is more sensitive to changes in training volume, as the higher the training load, the greater the rise in BUN and the slower the recovery rate early the following morning [12]. In the mid-test of this study, BUN was significantly higher than in the pre-test, which is also consistent with the accumulation of fatigue in athletes as the load increases. After moderate altitude training, the percentage of monocytes in the athletes in this study was much lower than before, and throughout training, the athletes’ bodies likely showed signs of infection-like conditions at the beginning, and the accumulated load of altitude training contributed to this phenomenon [32]. There is evidence that athletes exposed to hypoxic environments for training may experience significant changes in body composition [41]. MacDougall J et al. [42] found that long-term exposure to chronic hypoxia at high altitudes (over 5000 m) can disrupt the body’s protein synthesis, which in turn can result in lower body weight, skeletal muscle mass, and fat mass. In contrast, we did not find significant body mass, bone, or muscle loss in this case study, which is likely to be related to insufficient exposure time. The upper-body muscle mass change showed an upward trend after four weeks of moderate altitude training. It is worth noting that during this period, the Chinese XC-ski team coaching team specifically strengthened the athletes’ strength base. The training load Int. J. Environ. Res. Public Health 2023, 20, 266 9 of 11 intervention created a certain intervention mechanism on the body, resulting in the athletes’ body muscle mass being well maintained, which has been proven in elite winter sports athletes [43,44]. Another possible explanation is that specialist nutrition supplies from the national team may have led to this result. It is dangerous for endurance athletes to lose muscle and body weight during altitude training. Nevertheless, a solid nutritional approach prevents significant changes in body composition [45,46]. The study of Kayser et al. [47] has demonstrated that it is possible to maintain body composition at altitudes below 5000 metres if people intake sufficient energy. These results suggest that the athletes could maintain energy balance throughout the moderate altitude camp. Both training load and nutritional strategy may have influenced the results of this study, which can be verified in the future through well-controlled studies. Some limitations of this study should be considered. Firstly, Due to the limitations of the design, this study lacked a control group, and the observations in this study were relatively small. Nevertheless, this issue is a standard limitation of observational studies in real-world competitive sports settings. Secondly, there was a high degree of uncertainty in training during the field follow-up observations, and we needed to fully account for these confounding factors’ effects. Thirdly, due to realistic conditions, we have only discussed the physiological performance of athletes. Competitive sports are a results-driven business, and more data on sports performance are needed. Regarding further research, it would be interesting to replicate the present study with more athletes, different genders, different exposure times at moderate altitudes, or even different kinds of altitude training modalities with a control group. 5. Conclusions Four weeks of training at a moderate altitude positively affected the athletes’ lactate threshold and upper-body muscle mass. Maximum oxygen uptake was reduced in athletes tested at high altitudes compared to sea level due to being in a hypoxic environment. The aerobic capacity indicators (maximal oxygen uptake, haemoglobin, and RBC count) did not improve significantly after the athletes returned to sea level compared to pre-moderate altitude training, which may be related to altitudes that were too low, the duration of exposure to the hypoxic environment, and the training load design. When imposing moderate altitude training on athletes, exposure time, training load characteristics, and nutritional strategies must be meticulously designed to optimise training outcomes. Author Contributions: Data curation, Y.Y., R.W. and D.L.; formal analysis, Y.Y. and R.W.; investi- gation, Y.Y., R.W. and D.L.; methodology, Y.Y. and Y.L.; project administration, Y.L.; supervision, Y.L.; writing—original draft, Y.Y. and R.W; writing—review and editing, Y.Y., R.W., D.L. and Y.L. All authors have read and agreed to the published version of the manuscript. Funding: This research received no external funding. Institutional Review Board Statement: The study was conducted in accordance with the Declaration of Helsinki, and approved by the Sports Science Experimental Ethics Committee of Beijing Sport University (Grant No. 201906711, approved on 5 July 2019). Informed Consent Statement: Informed consent was obtained from all subjects involved in the study. Data Availability Statement: Not applicable. Acknowledgments: Thanks to the athletes who participated in the study and the Chinese National Olympic Committee. Conflicts of Interest: The authors declare no conflict of interest. Abbreviations LIT: low-intensity training; MIT, moderate-intensity training, HIT, high-intensity training, HR = heart rate, VO2max = maximal oxygen uptake, RER = respiratory exchange ratio, HCT = hematocrit, HGB = hemoglobin, RBC = red blood cell, CK = creatine kinase, BUN = blood urea, %LYM = percentage Int. J. Environ. Res. Public Health 2023, 20, 266 10 of 11 of lymphocyte, %MXD = percentage of mononucleosi, %NEUT = percentage of neutrophil, WBC = white blood cell. References 1. Sandbakk, O.; Holmberg, H.C. Physiological Capacity and Training Routines of Elite Cross-Country Skiers: Approaching the Upper Limits of Human Endurance. Int. J. Sports Physiol. Perform. 2017, 12, 1003–1011. [CrossRef] [PubMed] 2. Lian-Shi, F. The Functional Diagnosis Methods in Elite Athlete and Problems. Sport. Sci. Res. 2003, 24, 49–54. 3. Bonetti, D.L.; Hopkins, W.G. Sea-level exercise performance following adaptation to hypoxia: A meta-analysis. Sports Med. 2009, 39, 107–127. [CrossRef] [PubMed] 4. Flaherty, G.; O’Connor, R.; Johnston, N. Altitude training for elite endurance athletes: A review for the travel medicine practitioner. Travel Med. Infect. Dis. 2016, 14, 200–211. [CrossRef] 5. Wilber, R.L. Application of altitude/hypoxic training by elite athletes. Med. Sci. Sports Exerc. 2007, 39, 1610–1624. [CrossRef] 6. Buskirk, E.R.; Kollias, J.; Akers, R.F.; Prokop, E.K.; Reategui, E.P. Maximal performance at altitude and on return from altitude in conditioned runners. J. Appl. Physiol. 1967, 23, 259–266. [CrossRef] 7. Levine, B.D.; Stray-Gundersen, J. “Living high-training low”: Effect of moderate-altitude acclimatization with low-altitude training on performance. J. Appl. Physiol. 1997, 83, 102–112. [CrossRef] 8. Levine, B.D.; Stray-Gundersen, J. A practical approach to altitude training: Where to live and train for optimal performance enhancement. Int. J. Sports Med. 1992, 13 (Suppl. 1), S209–S212. [CrossRef] 9. Robach, P.; Schmitt, L.; Brugniaux, J.V.; Roels, B.; Millet, G.; Hellard, P.; Nicolet, G.; Duvallet, A.; Fouillot, J.P.; Moutereau, S.; et al. Living high-training low: Effect on erythropoiesis and aerobic performance in highly-trained swimmers. Eur. J. Appl. Physiol. 2006, 96, 423–433. [CrossRef] 10. Schmidt, W. Effects of intermittent exposure to high altitude on blood volume and erythropoietic activity. High Alt. Med. Biol. 2002, 3, 167–176. [CrossRef] 11. Vogt, M.; Puntschart, A.; Geiser, J.; Zuleger, C.; Billeter, R.; Hoppeler, H. Molecular adaptations in human skeletal muscle to endurance training under simulated hypoxic conditions. J. Appl. Physiol. 2001, 91, 173–182. [CrossRef] [PubMed] 12. Halson, S.L. Monitoring training load to understand fatigue in athletes. Sports Med. 2014, 44 (Suppl. 2), S139–S147. [CrossRef] [PubMed] 13. Saunders, P.U.; Telford, R.D.; Pyne, D.B.; Cunningham, R.B.; Gore, C.J.; Hahn, A.G.; Hawley, J.A. Improved running economy in elite runners after 20 days of simulated moderate-altitude exposure. J. Appl. Physiol. 2004, 96, 931–937. [CrossRef] [PubMed] 14. Schobersberger, W.; Leichtfried, V.; Mueck-Weymann, M.; Humpeler, E. Austrian Moderate Altitude Studies (AMAS): Benefits of exposure to moderate altitudes (1500–2500 m). Sleep Breath 2010, 14, 14–201. [CrossRef] [PubMed] 15. Turner, G.; Fudge, B.W.; Pringle, J.; Maxwell, N.S.; Richardson, A.J. Altitude training in endurance running: Perceptions of elite athletes and support staff. J. Sports Sci. 2019, 37, 163–172. [CrossRef] 16. Jin, Z.; Chuihui, K. The Influence of Subaltitude on Sports Coaching. J. Beijing Sport Univ. 2005, 78–79. [CrossRef] 17. Garvican-Lewis, L.A.; Halliday, I.; Abbiss, C.R.; Saunders, P.U.; Gore, C.J. Altitude Exposure at 1800 m Increases Haemoglobin Mass in Distance Runners. J. Sports Sci. Med. 2015, 14, 413–417. 18. Czuba, M.; Maszczyk, A.; Gerasimuk, D.; Roczniok, R.; Fidos-Czuba, O.; Zaj ˛ac, A.; Goła´s, A.; Mostowik, A.; Langfort, J. The effects of hypobaric hypoxia on erythropoiesis, maximal oxygen uptake and energy cost of exercise under normoxia in elite biathletes. J. Sports Sci. Med. 2014, 13, 912–920. 19. Karlsson, O.; Laaksonen, M.S.; McGawley, K. Monitoring Acclimatization and Training Responses Over 17-21 Days at 1,800 m in Elite Cross-Country Skiers and Biathletes. Front. Sports Act. Living 2022, 4, 852108. [CrossRef] 20. Woods, A.L.; Sharma, A.P.; Garvican-Lewis, L.A.; Saunders, P.U.; Rice, A.J.; Thompson, K.G. Four Weeks of Classical Altitude Training Increases Resting Metabolic Rate in Highly Trained Middle-Distance Runners. Int. J. Sport Nutr. Exerc. Metab. 2017, 27, 83–90. [CrossRef] 21. Friedmann-Bette, B. Classical altitude training. Scand. J. Med. Sci. Sports 2008, 18 (Suppl. 1), 11–20. [CrossRef] [PubMed] 22. Chapman, R.F.; Stickford, J.L.; Levine, B.D. Altitude training considerations for the winter sport athlete. Exp. Physiol. 2010, 95, 411–421. [CrossRef] [PubMed] 23. Grushin, A.A.; Nageykina, S.V. Elite female cross-Country skiers1 training for major international sport competitions in mid- altitude areas. Teor. Prakt. Fiz. Kult 2016, 5, 66–69. 24. Tønnessen, E.; Sylta, Ø.; Haugen, T.A.; Hem, E.; Svendsen, I.S.; Seiler, S. The road to gold: Training and peaking characteristics in the year prior to a gold medal endurance performance. PLoS ONE 2014, 9, e101796. [CrossRef] [PubMed] 25. McKay, A.; Stellingwerff, T.; Smith, E.S.; Martin, D.T.; Mujika, I.; Goosey-Tolfrey, V.L.; Sheppard, J.; Burke, L.M. Defining Training and Performance Caliber: A Participant Classification Framework. Int. J. Sports Physiol. Perform. 2022, 17, 317–331. [CrossRef] [PubMed] 26. Talsnes, R.K.; Hetland, T.A.; Cai, X.; Sandbakk, Ø. Development of Performance, Physiological and Technical Capacities During a Six-Month Cross-Country Skiing Talent Transfer Program in Endurance Athletes. Front. Sports Act. Living 2020, 2, 103. [CrossRef] 27. Xudan, C.; Lijuan, M.; Bei, Z.; Chungming, L.; Liangshi, F.; Xiaoping, C. The Development of Physical Ability of Talent Transferring Athletes from Different Sports in Long Term Cross Country Skiing Training—Based on Motor Function Monitorings. China Sport Sci. Technol. 2020, 56, 44–55. Int. J. Environ. Res. Public Health 2023, 20, 266 11 of 11 28. Sjödin, B.; Jacobs, I.; Svedenhag, J. Changes in onset of blood lactate accumulation (OBLA) and muscle enzymes after training at OBLA. Eur. J. Appl. Physiol. Occup. Physiol. 1982, 49, 45–57. [CrossRef] 29. Sylta, O.; Tonnessen, E.; Seiler, S. From heart-rate data to training quantification: A comparison of 3 methods of training-intensity analysis. Int. J. Sports Physiol. Perform. 2014, 9, 100–107. [CrossRef] 30. Eisenman, P.A.; Johnson, S.C.; Bainbridge, C.N.; Zupan, M.F. Applied physiology of cross-country skiing. Sports Med. 1989, 8, 67–79. [CrossRef] 31. Ingjer, F.; Myhre, K. Physiological effects of altitude training on elite male cross-country skiers. J. Sports Sci. 1992, 10, 37–47. [CrossRef] [PubMed] 32. Bailey, D.M.; Davies, B.; Romer, L.; Castell, L.; Newsholme, E.; Gandy, G. Implications of moderate altitude training for sea-level endurance in elite distance runners. Eur. J. Appl. Physiol. Occup. Physiol. 1998, 78, 360–368. [CrossRef] [PubMed] 33. Stellingwerff, T.; Peeling, P.; Garvican-Lewis, L.A.; Hall, R.; Koivisto, A.E.; Heikura, I.A.; Burke, L.M. Nutrition and Altitude: Strategies to Enhance Adaptation, Improve Performance and Maintain Health: A Narrative Review. Sports Med. 2019, 49, 169–184. [CrossRef] [PubMed] 34. Gore, C.J.; Hahn, A.G.; Aughey, R.J.; Martin, D.T.; Ashenden, M.J.; Clark, S.A.; Garnham, A.P.; Roberts, A.D.; Slater, G.J.; McKenna, M.J. Live high:train low increases muscle buffer capacity and submaximal cycling efficiency. Acta Physiol. Scand. 2001, 173, 275–286. [CrossRef] 35. Rankovic, G.; Mutavdzic, V.; Toskic, D.; Preljevic, A.; Kocic, M.; Nedin, R.G.; Damjanovic, N. Aerobic capacity as an indicator in different kinds of sports. Bosn J. Basic Med. Sci. 2010, 10, 44–48. [CrossRef] 36. Sandbakk, O.; Hegge, A.M.; Losnegard, T.; Skattebo, O.; Tonnessen, E.; Holmberg, H.C. The Physiological Capacity of the World’s Highest Ranked Female Cross-country Skiers. Med. Sci. Sports Exerc. 2016, 48, 1091–1100. [CrossRef] 37. Chapman, R.F.; Karlsen, T.; Ge, R.L.; Stray-Gundersen, J.; Levine, B.D. Living altitude influences endurance exercise performance change over time at altitude. J. Appl. Physiol. 2016, 120, 1151–1158. [CrossRef] 38. García, I.; Drobnic, F.; Galera, T.; Pons, V.; Viscor, G. Lung Diffusion in a 14-Day Swimming Altitude Training Camp at 1850 Meters. Int. J. Environ. Res. Public Health 2020, 17, 3501. [CrossRef] 39. Sun, Z.; Zhang, Y.; Xu, D.; Fei, Y.; Qiu, Q.; Gu, Y. The Effects of Six-Month Subalpine Training on the Physical Functions and Athletic Performance of Elite Chinese Cross-Country Skiers. Appl. Sci. 2022, 12, 421. [CrossRef] 40. Robach, P.; Schmitt, L.; Brugniaux, J.V.; Nicolet, G.; Duvallet, A.; Fouillot, J.P.; Moutereau, S.; Lasne, F.; Pialoux, V.; Olsen, N.V.; et al. Living high-training low: Effect on erythropoiesis and maximal aerobic performance in elite Nordic skiers. Eur. J. Appl. Physiol. 2006, 97, 695–705. [CrossRef] 41. Macdonald, J.H.; Oliver, S.J.; Hillyer, K.; Sanders, S.; Smith, Z.; Williams, C.; Yates, D.; Ginnever, H.; Scanlon, E.; Roberts, E.; et al. Body composition at high altitude: A randomized placebo-controlled trial of dietary carbohydrate supplementation. Am. J. Clin. Nutr. 2009, 90, 1193–1202. [CrossRef] [PubMed] 42. MacDougall, J.D.; Green, H.J.; Sutton, J.R.; Coates, G.; Cymerman, A.; Young, P.; Houston, C.S. Operation Everest II: Structural adaptations in skeletal muscle in response to extreme simulated altitude. Acta Physiol. Scand. 1991, 142, 421–427. [CrossRef] [PubMed] 43. Kim, T.H.; Han, J.K.; Lee, J.Y.; Choi, Y.C. The Effect of Polarized Training on the Athletic Performance of Male and Female Cross-Country Skiers during the General Preparation Period. Healthcare 2021, 9, 851. [CrossRef] [PubMed] 44. Terzis, G.; Stattin, B.; Holmberg, H.C. Upper body training and the triceps brachii muscle of elite cross country skiers. Scand. J. Med. Sci. Sports 2006, 16, 121–126. [CrossRef] [PubMed] 45. Svedenhag, J.; Saltin, B.; Johansson, C.; Kaijser, L. Aerobic and anaerobic exercise capacities of elite middle-distance runners after two weeks of training at moderate altitude. Scand. J. Med. Sci. Sports 2007, 1, 205–214. [CrossRef] 46. Gore, C.J.; Hahn, A.; Rice, A.; Bourdon, P.; Lawrence, S.; Walsh, C.; Stanef, T.; Barnes, P.; Parisotto, R.; Martin, D.; et al. Altitude training at 2690m does not increase total haemoglobin mass or sea level VO2max in world champion track cyclists. J. Sci. Med. Sport 1998, 1, 156–170. [CrossRef] 47. Kayser, B. Nutrition and high altitude exposure. Int. J. Sports Med. 1992, 13 (Suppl. 1), S129–S132. [CrossRef] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Monitoring Physiological Performance over 4 Weeks Moderate Altitude Training in Elite Chinese Cross-Country Skiers: An Observational Study.

12-24-2022

Yu, Yichao,Wang, Ruolin,Li, Dongye,Lu, Yifan

eng

PMC5053172

LETTER TO THE EDITOR Open Access Cardiorespiratory fitness: a comparison between children with renal transplantation and children with congenital solitary functioning kidney Riccardo Lubrano1,4* , Giancarlo Tancredi1, Raffaele Falsaperla2 and Marco Elli3 Abstract Children with end-stage renal disease are known to have a cardiorespiratory fitness significantly reduced. This is considered to be an independent index predictive of mortality mainly due to cardiovascular accidents. The effects of renal transplantation on cardiorespiratory fitness are incompletely known. We compared the maximal oxygen uptake (VO2 max) of children with a functioning renal transplant with that of children with congenital solitary functioning kidney, taking into consideration also the amount of weekly sport activity. Keywords: VO2 max, Cardiorespiratory fitness, Chronic renal failure, Child, Physical activity, Renal transplant, Solitary functioning kidney Dear Editor, Patients with chronic renal failure (CRF) tend to reduce their weekly amount of physical activity, with negative ef- fects on cardiorespiratory fitness and quality of life. After renal transplant the metabolic deficits induced by CRF are partially recovered and cardiorespiratory fitness improves. As we previously reported cardiorespiratory fitness of transplanted children practicing sports for more than 3 h per week is similar to normal controls exercising less that 3 h [1]. On the contrary, cardiorespiratory fitness of chil- dren with a congenital solitary functioning kidney is simi- lar to normal controls exercising for a comparable number of hours [2]. We measured the aerobic capacity in relation with weekly amount of physical activity and glomerular filtra- tion rate (GFR), comparing a group of children with a congenital solitary functioning kidney (cSFK) and a group of children with a functioning renal transplant (Tx). A standardized pediatric questionnaire was adminis- tered to all children for investigating the time dedicated weekly to physical activity [3]. On the basis of the ques- tionnaire, the children were divided into inadequately active (<3 h of physical activity per week) and adequately active (>3 h of physical activity per week). In the cSFK group we enrolled 30 patients: 15 exercis- ing more than 3 h/week (cSFK>3) and 15 less than 3 h/ week (cSFK<3). The Tx group was formed with 20 chil- dren, 10 exercising more than 3 h/week (Tx>3) and 10 less than 3 h/week (Tx<3). In all patients, transplant had been performed 6 or more years previously, following a dialysis treatment never exceeding on year. All received triple immunosuppressive therapy: 12 with tacrolimus and 8 with cyclosporine. Maximal oxygen uptake (VO2 max) was measured during a maximal incremental exercise on a treadmill (Bruce protocol) consisting of sequential increase in speed and slope every 3 min until exhaustion (breath- lessness and leg muscle pain) and/or heart rate ≥85 % of maximum (calculated with the formula 220 – age in years). During the exercise the subjects were con- nected by face mask to a breath-by-breath analyser of O2 to measure the oxygen consumption (VO2). Max- imal oxygen uptake (VO2 max) was defined as the * Correspondence: [email protected] 1Pediatric Department, Pediatric Nephrology Unit, Sapienza University of Rome, Rome, Italy 4Servizio di Nefrologia Pediatrica, Dipartimento di Pediatria, Sapienza Università di Roma, Viale Regina Elena 324, 00161 Roma, Italia Full list of author information is available at the end of the article © 2016 The Author(s). Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated. Lubrano et al. Italian Journal of Pediatrics (2016) 42:90 DOI 10.1186/s13052-016-0299-7 highest level of VO2 reached during the maximal ex- ercise test expressed as VO2 max/kg (ml/min/kg). The glomerular filtration rate (ml/min/1.73mq) was calculated with the creatinine clearance. Informed consent was obtained from both parents. The protocol conforms to the guidelines of the Declaration of Helsinki and was approved by the ethical committee of the involved institution. The children in all groups were comparable for age (years: Tx>3 12.67 ± 3.56; Tx<3 13.90 ± 1.20; cSFK>3 14.18 ± 5.29; cSFK<3 13,5 ± 4.76; p NS). GFR was also similar in all groups (GFR ml/min/1.73 mq: Tx>3 90.65 ± 22.52; Tx<3 92.02 ± 21.18; cSFK>3 99.15 ± 30.63; cSFK<3 101,02 ± 40.12; p NS). VO2 max in Tx and cSFK was significantly higher in those practicing sport for more than 3 h per week (Table 1). Children with a congenital solitary function- ing kidney had level of VO2 max consistently and significantly higher than transplanted patients (Table 1). There was no significant correlation be- tween VO2 max and GFR (VO2 max = 26.08 + 0.006; GFR R^2 = − 0.07). Our findings show that not only congenital solitary functioning kidney (cSFK), but also transplanted chil- dren with regular physical activity exceeding three hours weekly achieve higher levels of VO2 max. Adequate and regular physical exercise proves therefore beneficial in transplanted children improving their ability to cope with the increased metabolic request of physical stress and therefore reducing the risk of mortality from cardio- vascular disease [4]. VO2 max in transplanted children is consistently lower than single kidney patients with comparable physical activity. This may be due in part to the neuromuscular, metabolic, and cardiopulmonary defi- cits acquired during the exposure to uremic intoxica- tion before transplant [5], that a functioning graft can improve but not reverse completely. A combination of early transplant and prompt resumption of con- trolled adequate physical exercise post-transplant is likely to improve further the cardiorespiratory fitness in these patients, with the known benefits on cardio- vascular risk and mortality. Abbreviations 3 h/week: 3 hours a week; CRF: Chronic renal failure; cSFK: Congenital solitary functioning kidney; GFR: Glomerular filtration rate; Tx: Transplanted children; VO2 max: Maximal oxygen uptake Funding Nothing to declare. Authors’ contributions RL participated in the design of the study, performed the statistical analysis and drafted the manuscript. RF carried out data collection and helped in performing the statistical analysis. GT carried out data collection and data analysis. ME participated in the coordination and helped to draft the manuscript. All authors read and approved the final manuscript. Competing interest The authors declare that they have no competing interests. Consent for publication Not applicable. Ethics approval and consent to participate For each child informed consent was obtained from both parents and the study protocol conformed to the ethical guidelines of the 1975 Declaration of Helsinki as revised in 2000. The study was approved by the ethical committee in our institutions. Author details 1Pediatric Department, Pediatric Nephrology Unit, Sapienza University of Rome, Rome, Italy. 2General Pediatrics and Pediatric Acute and Emergency Unit, Policlinico-Vittorio-Emanuele University Hospital, Catania, Italy. 3DIBIC-Biomedical and Clinic Science Department, “Luigi Sacco”, the University of Milan-VMS Vialba Medical School, Milan, Italy. 4Servizio di Nefrologia Pediatrica, Dipartimento di Pediatria, Sapienza Università di Roma, Viale Regina Elena 324, 00161 Roma, Italia. Received: 8 September 2016 Accepted: 30 September 2016 References 1. Lubrano R, Tancredi G, Bellelli E, Gentile I, Scateni S, Masciangelo R, De Castro G, Versacci P, Elli M. Influence of physical activity on cardiorespiratory fitness in children after renal transplantation. Nephrol Dial Transplant. 2012; 27:1677–81. 2. Tancredi G, Lambiase C, Favoriti A, Ricupito F, Paoli S, Duse M, De Castro G, Zicari AM, Vitaliti G, Falsaperla R, Lubrano R. Cardiorespiratory fitness and sports activities in children and adolescents with solitary functioning kidney. Ital J Pediatr. 2016;42:43. 3. Crocker PR, Bailey DA, Faulkner RA. Measuring general levels of physical activity: preliminary evidence for the physical activity ques- tionnaire for older children. Med Sci Sports Exerc. 1997;29:1344–9. 4. Laukkannen JA, Makikallio TH, Raurammaa R, et al. Cardiorespira- tory fitness is related to the risk of sudden cardiac death. A population based follow-up study. J Am Coll Cardiol. 2010;56:1476–83. 5. Bar-Or O, Rowlands TW. Hematologic, oncologic and renal disease. In: Bar-Or O, Rowlands TW, editors. Pediatric exercise medi- cine: from physiologic principles to health care application. Champaign: Human Kinetics; 2004. p. 305–22. • We accept pre-submission inquiries • Our selector tool helps you to find the most relevant journal • We provide round the clock customer support • Convenient online submission • Thorough peer review • Inclusion in PubMed and all major indexing services • Maximum visibility for your research Submit your manuscript at www.biomedcentral.com/submit Submit your next manuscript to BioMed Central and we will help you at every step: Table 1 VO2 max/kg in the four groups of the study Tx>3 Tx<3 cSFK>3 cSFK<3 VO2 max/kg ml/ min/kg 28.99 ± 1.15 23.22 ± 1.23 46.12 ± 1.09 38.55 ± 1.97 Tx>3 vs Tx<3 p < 0.003; cSFK>3 vs cSFK<3 p < 0.01; Tx>3 vs cSFK>3 p < 0.016; Tx<3 vs cSFK<3 p < 0.001; Tx<3 vs cSFK>3 p < 0.004; Tx>3 vs cSFK<3 p < 0.001 Lubrano et al. Italian Journal of Pediatrics (2016) 42:90 Page 2 of 2

Cardiorespiratory fitness: a comparison between children with renal transplantation and children with congenital solitary functioning kidney.

10-06-2016

Lubrano, Riccardo,Tancredi, Giancarlo,Falsaperla, Raffaele,Elli, Marco

eng

PMC7379642

Supplement Table 2. Change in VO2max (L·min-1 and ml·min-1·kg-1) from 1995-1997 to 2016-2017 in the total population and by gender. L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 L·min-1 ml·min-1·kg-1 Year Mean (SD) Change Mean (SD) Change Mean (SD) Change Mean (SD) Change Mean (SD) Change Mean (SD) Change 95-97 2.47 (0.07) Ref 38.1 (1.27) Ref 3.16 (0.09) Ref 39.0 (1.22) Ref 2.80 (0.08) Ref 38.5 (0.90) Ref 98-99 2.42 (0.09) -2,1% 36.8 (1.40) -3,4% 3.05 (0.10) -3,4% 37.5 (1.45) -3,9% 2.74 (0.09) -2,2% 37.1 (1.00) -3,6% 00-01 2.45 (0.08) -1,0% 36.9 (1.33) -3,2% 3.04 (0.10) -3,9% 36.7 (1.49) -5,9% 2.75 (0.08) -1,8% 36.8 (0.99) -4,4% 02-03 2.33 (0.08) -5,8% 35.2 (1.28) -7,5% 2.94 (0.10) -6,9% 35.9 (1.39) -8,0% 2.64 (0.08) -5,8% 35.6 (0.94) -7,6% 04-05 2.35 (0.07) -4,9% 35.2 (1.32) -7,7% 2.97 (0.09) -6,1% 36.4 (1.16) -6,6% 2.66 (0.08) -4,9% 35.8 (0.87) -7,0% 06-07 2.37 (0.07) -4,1% 35.4 (1.22) -7,1% 2.97 (0.09) -6,1% 35.9 (1.15) -8,0% 2.67 (0.07) -4,6% 35.6 (0.83) -7,4% 08-09 2.39 (0.07) -3,1% 35.5 (1.19) -6,8% 2.97 (0.09) -6,1% 35.7 (1.21) -8,5% 2.68 (0.07) -4,2% 35.6 (0.84) -7,5% 10-11 2.38 (0.07) -3,5% 35.1 (1.25) -7,8% 2.99 (0.08) -5,5% 35.7 (1.11) -8,5% 2.69 (0.07) -4,0% 35.4 (0.83) -8,1% 12-13 2.38 (0.07) -3,8% 35.0 (1.21) -8,2% 2.92 (0.09) -7,5% 35.0 (1.18) -10,3% 2.65 (0.07) -5,3% 35.0 (0.83) -9,1% 14-15 2.34 (0.07) -5,2% 34.4 (1.17) -9,7% 2.90 (0.08) -8,4% 34.4 (1.08) -11,7% 2.62 (0.07) -6,4% 34.4 (0.79) -10,6% 16-17 2.34 (0.07) -5,3% 34.5 (1.12) -9,4% 2.88 (0.08) -8,7% 34.2 (1.09) -12,4% 2.61 (0.07) -6,7% 34.3 (0.77) -10,8% Women Men Total

Decline in cardiorespiratory fitness in the Swedish working force between 1995 and 2017.

11-15-2018

Ekblom-Bak, Elin,Ekblom, Örjan,Andersson, Gunnar,Wallin, Peter,Söderling, Jonas,Hemmingsson, Erik,Ekblom, Björn

eng

PMC9794057

RESEARCH ARTICLE Factors associated with high-level endurance performance: An expert consensus derived via the Delphi technique Magdalena J. KonopkaID1,2☯*, Maurice P. Zeegers1,2,3☯, Paul A. Solberg4☯, Louis Delhaije2☯, Romain Meeusen5,6☯, Geert Ruigrok2☯, Gerard Rietjens5☯, Billy Sperlich7☯ 1 Care and Public Health Research Institute, Maastricht University, Maastricht, Limburg, Netherlands, 2 Department of Epidemiology, Maastricht University Medical Centre, Maastricht, Limburg, Netherlands, 3 School of Nutrition and Translational Research in Metabolism, Maastricht University, Maastricht, Limburg, Netherlands, 4 Norwegian Olympic and Paralympic Committee and Confederation of Sports, Oslo, Norway, 5 Human Physiology and Sports Physiotherapy Research Group, Vrije Universiteit Brussel, Brussels, Brussels-Capital Region, Belgium, 6 Brussels Human Robotics Research Center (BruBotics), Vrije Universiteit Brussel, Brussels, Brussels-Capital Region, Belgium, 7 Integrative & Experimental Exercise Science & Training, Institute of Sport Science, University of Wu¨rzburg, Bavaria, Germany ☯ These authors contributed equally to this work. * [email protected] Abstract There is little agreement on the factors influencing endurance performance. Endurance perfor- mance often is described by surrogate variables such as maximum oxygen consumption, lac- tate threshold, and running economy. However, other factors also determine success and progression of high-level endurance athletes. Therefore, the aim was to identify the relevant factors for endurance performance assessed by international experts by adhering to a struc- tured communication method (i.e., Delphi technique). Three anonymous evaluation rounds were conducted initiated by a list of candidate factors (n = 120) serving as baseline input vari- ables. The items that achieved 70% of agreement in round 1 were re-evaluated in a second round. Items with a level of agreement of 70% in round 2 reached consensus and items with a level of agreement of 40–69% in round 2 were re-rated in a third round followed by a consen- sus meeting. Round 1 comprised of 27 panellists (n = 24 male) and in round 2 and 3 18 (n = 15 male) of the 27 panellists remained. Thus, the final endurance expert panel comprised of 18 international experts (n = 15 male) with 20 years of experience on average. The consensus report identified the following 26 factors: endurance capacity, running economy, maximal oxy- gen consumption, recovery speed, carbohydrate metabolism, glycolysis capacity, lactate threshold, fat metabolism, number of erythrocytes, iron deficiency, muscle fibre type, mitochon- drial biogenesis, hydrogen ion buffering, testosterone, erythropoietin, cortisol, hydration status, vitamin D deficiency, risk of non-functional overreaching and stress fracture, healing function of skeletal tissue, motivation, stress resistance, confidence, sleep quality, and fatigue. This study provides an expert-derived summary including 26 key factors for endurance performance, the “FENDLE” factors (FENDLE = Factors for ENDurance Level). This consensus report may assist to optimize sophisticated diagnostics, personalized training strategies and technology. PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 1 / 15 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Konopka MJ, Zeegers MP, Solberg PA, Delhaije L, Meeusen R, Ruigrok G, et al. (2022) Factors associated with high-level endurance performance: An expert consensus derived via the Delphi technique. PLoS ONE 17(12): e0279492. https://doi.org/10.1371/journal.pone.0279492 Editor: Daniel Boullosa, Universidade Federal de Mato Grosso do Sul, BRAZIL Received: July 8, 2022 Accepted: December 8, 2022 Published: December 27, 2022 Copyright: © 2022 Konopka et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: The author(s) received no specific funding for this work. Competing interests: The authors have declared that no competing interests exist. Introduction High performance in endurance sports is the result of the interplay of various factors including optimal training [1], recovery [2, 3], nutritional strategies [4–6], the use and handling of envi- ronmental [7] and psycho-social factors [8, 9] as well as high-tech equipment [10–12]. In addi- tion, optimal biological factors are crucial to achieve world class competition level. For instance, recent scientific evidence suggests that a complex network of genetic and other bio- logical mechanisms (e.g., transcriptomics, epigenomics, proteomics, metabolomics) contribute to the performance level of an athlete [13]. By integrating such novel research disciplines (omics) and advanced technology (e.g., machine learning) exercise science is shifting towards “precision exercise” offering training and recovery strategies tailored to the individual needs of an athlete [14–16]. The main idea of precision exercise is to base training and recovery deci- sions on individual characteristics including various biomarkers, ultimately attempting to per- sonalize training programs, minimize injury risks, optimize performance outcomes, and to identify talents in near future [16]. These attempts, e.g., in the field of precision medicine, are based on prediction models and the predictive ability of these models tend to increase with the amount and quality of data input [17]. Thus, from a “precision endurance exercise” perspec- tive, the more factors associated with high-level endurance performance are integrated, the better the prediction of the model should become as evidenced in medical research [18]. However, from a practical and a scientific point of view little agreement exists on the most important factors that influence high-level endurance performance. A first step towards devel- oping precision exercise for endurance athletes would be to identify the relevant key factors for further modelling. In addition, endurance athletes, their coaches, and researchers could be informed and guided how to set priorities regarding training strategies and future research. For example, the identified key factors could be assessed with innovative wearable technology. Therefore, the aim of this study was to reach consensus about the key factors that are consid- ered important for high-level endurance performance among international experts. The Del- phi technique allows a comprehensive and structured group communication process aiming to achieve convergence of expert opinions by employing iterative data collection [19–24]. Materials and methods Study design and consensus threshold The study protocol is available at the open science framework (doi:10.17605/OSF.IO/YH5V4). We conducted a consensus study by employing the Delphi technique according to the descrip- tion by Hsu et al., [19] the checklist by Sinha et al., [25] and Hasson et al., [26]. We also fol- lowed the recommendations for Conducting and REporting DElphi Studies (CREDES) [27]. The 12-item CREDES checklist is enclosed in S1 Table. The study was implemented in three phases: 1) preparation, 2) conduction, and 3) analysis [28]. The Delphi process (i.e., conduc- tion phase) consisted of three iterative rounds of web-based questionnaires, participant response, and controlled feedback. We employed a dichotomous scale (relevant vs. not relevant) to determine the relevance of the items. A cut-off value of 70% level of agreement was a priori chosen as the consensus threshold [27]. Level of agreement was categorized into low (0–39%), moderate (40–69%), and high (70–100%). Items with a low or moderate level of agreement were considered not relevant whereas items with a high level of agreement (70%) represented relevant factors and/or con- sensus. The steering committee advised on study as well as survey design, methodology, and content, but did not interact with or act on behalf of the panellists. Ethical approval for con- ducting the study was obtained from the Ethical Review Committee Health, Medicine, and PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 2 / 15 Life Science of Maastricht University (FHML-REC/2019/021) and conducted in accordance with the Declaration of Helsinki for human research. Study flow An initial list of candidate factors (n = 120) established by the steering committee served as input for the first round. Before distribution, the survey was pilot-tested and subsequently sent to the panellists of round 1. The items receiving a high level of agreement in round 1 were re- evaluated in a second round. All high level of agreement items in round 2 were included in the ‘consensus report’ presenting the level of agreement for each item. Moderate level of agree- ment items after round 2 were resubmitted to round 3 and in case the 70% threshold was attained, these factors were also added to the consensus report. Finally, a ‘consensus decision’ with the steering committee took place in which the remaining items of round 3 were consid- ered for additional inclusion into the consensus report. The consensus report thus contains the factors that achieved a high level of agreement (i.e., the “FENDLE” factors). The study flow is illustrated in Fig 1. Panellists In accordance with the CREDES recommendations [27] we involved experts with diverse backgrounds (e.g., high-level coaches, exercise scientists, physicians, etc.) and geographical locations. Inclusion criteria to participate in this study were: 1) age of 18 years or older; 2) good command of English language; 3) competence in or experience with (elite) endurance training. The panellists were recruited through non-probability purposive sampling to ensure that the invited experts represent sufficient experience, knowledge, and interest in the topic. Eligible panellists were professional athletes or coaches, exercise-scientists, physiotherapists, exercise-psychologists, or medical physicians. The panellists had to possess extensive experi- ence and knowledge of elite endurance performance. The qualifications and responsibilities of the expert panel are displayed in S2 Table. Invitations to take part in this study were sent through email by the steering committee members using purposive sampling. The invitation letter included the participant information leaflet containing information regarding the aim of the study, the Delphi process, what was expected from the panellists, the projected timeline, and the amount of time, effort, and commitment required to finish the study. The panellists remained anonymous except for the main investigator (MK) for communication purposes. Each survey took approximately 15 minutes to complete. Data collection The web-based surveys were conducted using Qualtrics (Qualtrics, Provo, UT). The panellists provided written informed consent for the study at the start of the first survey. In all three rounds of questionnaires, the panellists were asked to rate the proposed items as either ‘rele- vant’ or ‘not relevant’: “Which of the following factors, if any, are relevant for endurance per- formance?” Following each round the responses were analysed and anonymously reported back to the experts (e.g., number of panellists, level of agreement for each item) allowing them to reconsider pervious decisions. Preparation (PHASE I) The purpose of the preparation phase was to develop a comprehensive list of candidate factors that are possibly associated with high-level endurance performance and could potentially be incorporated into the consensus report. First, a literature search was performed in the Medline PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 3 / 15 Fig 1. Study flow. https://doi.org/10.1371/journal.pone.0279492.g001 PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 4 / 15 database to identify candidate factors (last search date: 05. November 2019). All retrieved arti- cles were read by the main investigator (MK). The potentially relevant factors were extracted using Excel and compiled into a list. Risk of bias of individual studies was not assessed. The list was subsequently revised (e.g., wording of factors, adding/ removing factors) based on recom- mendations from the steering committee. Finally, the list comprised of 120 factors (S3 Table) and served as input for the first round of the Delphi process [19]. Before distribution, the sur- vey was pilot tested by five experts with varying backgrounds (exercise scientist, athletes, coaches) and 14 years of experience on average, different from those recruited for the main study, to verify the clarity of questions. The testers were instructed not to add or remove fac- tors from the list. Suggestions from the testers were used to revise the survey. Conduction (PHASE II) Round 1. For reason of efficiency and proposed by the pilot testers, the panellists were asked to ‘uncheck’ the non-relevant factors from the list in the first round. We also invited the experts to add missing factors and to provide free text comments. Finally, we requested rele- vant demographic data, such as name, age-group, email address, profession, place of employ- ment, and quantifiable level of expertise. Round 2. Based on the results from round 1, the steering committee agreed to eliminate the candidate factors that did not reach the threshold of 70%. Thus, items that did not reach the 70% threshold in round 1 were discarded from the survey and not resubmitted in round 2. In round 2, the items that attained a level of agreement of 70% in round 1 were distributed to the panellists who completed round 1 together with the average level of agreement, the newly proposed items, and the anonymized free text comments from round 1. This time, the panellists were asked to ‘check’ the relevant factors. After round 2, the items that achieved a level of agreement of 70% reached consensus and were incorporated in the consensus report. Round 3 and consensus decision of the steering committee In round 3, we provided the anonymized results of the second survey to the experts who com- pleted both rounds (i.e., the “FENDLE PANEL”) including their own rating and group sum- mary statistics. In this round, the experts had the option to re-rate the items that reached a moderate level of agreement (40–69%) in round 2. In this ‘consensus decision’ of the FENDLE PANEL, the items that achieved a level of agreement of 70% after round 3 were also added to the consensus report. The remaining factors from round 3 were set aside for the ‘consensus decision’ amongst the steering committee members. Each member of the steering committee could propose items which were consequently discussed by the group. In case each member agreed (level of agreement = 100%), the proposed factor was added to the consensus report. Lastly, in round 2 and 3, there was no option for free text responses. S4 Table contains a link to the web-based questionnaire of round 1 and 2 as well as an illustration of round 3. Analysis (PHASE III) After each round the raw data was downloaded from Qualtrics. We checked for incomplete submissions and duplicates. In case duplicate panellists were identified, the first survey submit- ted was analysed. Incomplete submissions were excluded. In case a valid email address was provided, the research team invited the participant to finish incomplete surveys. For the data analysis, we randomly assigned identification numbers to the experts so that they remained pseudonymous. Descriptive (frequency, percentage) and inferential statistics (mean, median, interquartile range) were used to describe the demographics and to analyse the extent of con- sensus [26]. The main outcome variable was binary. Secondary outcome variables were both PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 5 / 15 categorical (gender, age-group, country, and occupation) and continuous (years of experi- ence). All analysis were performed with R (version 4.0.3). Results Survey distribution and demographics of panellists. Twenty-seven panellists completed the first round of questionnaire (July–December 2020) and were consequently invited to par- ticipate in round 2. Eighteen panellists (66.7%) from nine countries finished round 2 (Janu- ary–February 2021) and 3 (March–April 2021) and thus completed all three rounds of questionnaires. Of the 18 panellists, 15 (83.3%) were male and 72.2% were aged between 31 and 60 years. The median time in years (interquartile range) of practical experience with endurance athletes was 20.0 (10.0–33.5). Table 1 demonstrates the demographics of the panel- lists. Finally, the consensus decision of the steering committee took place in May 2021. Survey results Fig 1 illustrates the flow of the factors through the Delphi process. In the first round, 99 of the 120 candidate items achieved the 70% threshold and consequently were resubmitted in round 2. In addition, one item (sedentary lifestyle) was proposed as new factor and therefore added to round 2. In the second round, 24 of 100 items achieved consensus 70% and thus were included in the consensus report. Further, 22 items were rated with a moderate level of agree- ment and therefore resubmitted in round 3. During the ‘consensus decision’ of the FENDLE PANEL in round 3, one item “endurance capacity” was rated as relevant (level of agreement 72%) and hence integrated into the consensus report. S5–S7 Tables displays the results of round 1–3, respectively. After round 3, the ‘consensus decisions’ of the steering committee took place, in which 13 of the 21 remaining items from round 3 were subject for discussion. The steering committee agreed (100%) on one item, “recovery speed”, to be additionally included in the consensus report. S8 Table presents the results of the consensus decision of the steering committee. The final consensus report contained 26 factors (Table 2) considered rele- vant for high-level endurance training and/or performance. Lastly, 20 factors were rated with a moderate (S9 Table) and 54 with a low level of agreement, representing the factors rated as not relevant (S10 Table). Discussion Based on an international expert consensus process (i.e., Delphi technique) the aim of this study was to identify key factors that are considered important for high-level endurance per- formance. In total, 26 factors achieved consensus. The 26 FENDLE factors comprised of five different clusters: i) physiology ii) nutrition, iii) injuries, iv) psychological traits and v) fatigue. Physiology Traditionally, three physiological factors have been identified explaining endurance perfor- mance: i) ‘maximum oxygen consumption’ (i.e., the maximal capacity to take up, transport, and utilize oxygen), ii) the ability to maintain high velocity without accumulating blood lactate (‘lactate threshold’), and iii) ‘running economy’ often expressed as the oxygen utilized while running at a given constant speed [29]. Since these three physiological factors have been exten- sively investigated and linked to elite endurance performance, it seems reasonable that the expert panel identified these factors as relevant for high-level of endurance performance. Inter- estingly, this consensus report also identified the ability to quickly recover during and after endurance events (‘recovery speed’) as important factor for high-level endurance performance. PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 6 / 15 Replenishing energy storage efficiently and quickly reduces time between training and compe- tition and simultaneously can prevent injuries when training again with full energy storages [30]. The fuelling of the different energy pathways is a key factor for many endurance sports [4–6] and since most endurance disciplines involve either long-duration tasks (e.g., marathon running) and/or high-intensity exercise (e.g., 800m running) it seems plausible that the experts reached consensus about ‘glycolytic capacity’, ‘hydrogen iron buffering’, as well as ‘fat and car- bohydrate metabolism’ as key factors influencing endurance performance. Furthermore, the Table 1. Demographics of the panellists. ROUND 1 ROUND 2 & 3 i.e., FENDLE PANEL Total (n, %) 27 (100) 18 (100) Gender (n, %) Males 24 (88.9) 15 (83.3) Females 2 (7.4) 2 (11.1) Missing 1 (3.7) 1 (5.6) Age group in years (n, %) 18–30 5 (18.5) 4 (22.3) 31–60 20 (74.1) 13 (72.1) 60+ 1 (3.7) 1 (5.6) Missing 1 (3.7) 0 Country (n, %) Canada 1 (3.7) 1 (5.6) USA 1 (3.7) 1 (5.6) United Kingdom 2 (7.4) 1 (5.6) Belgium 4 (14.8) 3 (16.7) Germany 3 (11.1) 2 (11.1) Italy 3 (11.1) 2 (11.1) Netherlands 8 (29.7) 5 (27.6) Norway 1 (3.7) 1 (5.6) Sweden 2 (7.4) 2 (11.1) Finland 1 (3.7) 0 Spain 1 (3.7) 0 Occupation (n, %) Chief Executive Officer (triathlon) 1 (3.7) 1 (5.6) PhD student (sport science) 5 (18.5) 4 (22.2) Physician (sports and rehabilitation) 2 (7.4) 1 (5.6) Physiotherapist 1 (3.7) 1 (5.6) Professor (sport science) 6 (22.3) 6 (33.1) Director (triathlon) 1 (3.7) 1 (5.6) Head Coach Road cycling 1 (3.7) 1 (5.6) Soccer 1 (3.7) 1 (5.6) Triathlon 2 (7.4) 2 (11.1) Unknown 4 (14.8) 0 Missing 3 (11.1) 0 Practical experience in years (median, IQRa) 20.0 (9.0–33.0) 20.0 (10.0–33.5) FENDLE: acronym for Factors for ENDurance Level. aIQR = interquartile range. https://doi.org/10.1371/journal.pone.0279492.t001 PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 7 / 15 muscle fibre spectrum [31, 32] including the density and efficiency of mitochondria [33] have substantial impact on endurance performance. It is well known that ‘fibre type distribution’ and ‘mitochondrial biogenesis’ are important for increasing oxygen utilization for energy pro- duction stimulated by various training methods [34, 35]. Oxygen delivery and utilization are critical components for the energy turnover in the muscle cells [36]. Substantial evidence indi- cates that improving oxygen transportation by the augmentation of haemoglobin levels (e.g., through high-altitude training, medication, or blood transfusion) significantly improves endurance performance [37]. Thus, it seems reasonable that the experts rated the ‘number of red blood cells’ and ‘iron deficiency’ as key factors for high-level endurance performance. Finally, three hormones reached consensus as key physiological factors for endurance perfor- mance: ‘testosterone’, ‘erythropoietin’, and ‘cortisol’. Testosterone is known to stimulate mus- cle mass and to reduce body fat [38]. Erythropoietin induces erythropoiesis, the maturation and proliferation of oxygen-delivering red blood cells [39]. Finally, in the skeletal muscle the level of cortisol plays a fundamental role in regulating energy homeostasis [40]. During exer- cise, the high level of cortisol increases the availability of metabolic substrates, protects from immune cell activity, and maintains vascular integrity [41]. Table 2. Consensus report describing the FENDLE factors, the 26 factors considered the most important to influence endurance training and/or performance. Cluster Factor Level of agreement (%) Physiology General Endurance capacitya 72.2 Economy of movement 88.9 Maximal oxygen consumption 94.4 Recovery speedb 66.7 Metabolism Carbohydrate metabolism 100.0 Glycolysis capacity 100.0 Lactate threshold 88.9 Fat metabolism 88.9 Blood Number of red blood cells 100.0 Iron deficiency 94.4 Muscle Muscle fibres—type 1 vs. type 2a/x 94.4 Mitochondrial biogenesis 88.9 Hydrogen ion buffering 88.9 Hormones Testosterone 94.4 Erythropoietin 83.3 Cortisol 77.8 Nutrition Electrolyte balance/ hydration status 77.8 Vitamin D deficiency 72.2 Injuries Risk of non-functional overreaching 88.9 Risk of stress fracture 77.8 Healing function of skeletal muscle tissue 88.8 Psychology Motivation capacity 94.4 Stress resistance 88.9 Self-confidence 72.2 Fatigue Sleep quality 94.4 Level of fatigue 77.8 FENDLE: acronym for Factors for ENDurance Level. aAdded after the consensus decision of the FENDLE PANEL (round 3). bAdded after the consensus decision of the steering committee. https://doi.org/10.1371/journal.pone.0279492.t002 PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 8 / 15 Injury and nutrition Good health is essential for achieving high training volumes necessary for maximal endurance performances. Since endurance athletes engage in numerous low- to high intensity trainings and thus are exposed to constant muscle and tissue damage, it seems reasonable that the experts identified the ‘healing function of skeletal muscle tissue’, ‘risk of stress fracture’, and ‘risk of non-functional overreaching’ as key factors for high-level endurance performance. Moreover, tissue repair is supported by various nutritional factors. For example, a low level of vitamin D has been associated with maladaptation of skeletal muscle and bone tissue [42]. Proper muscle renewal in response to exercise is required for optimal hypertrophic effects [43]. Therefore, it is plausible that ‘vitamin D deficiency’ has reached consensus in the present analysis. Further, ‘hydration status’ and ‘electrolyte balance’ were rated as key endurance fac- tors by the panellists. The loss of body fluids during exercise is mostly due to sweating and the replacement of sodium loss is widely recommended [44]. The hydration of athletes tends to vary according to individual factors (e.g., thirst response, acclimation, gut training) and envi- ronmental factors (e.g., ambient temperature, provision of drinks at aid stations) as well as the type and intensity of exercise, making individualized fluid replacement strategies necessary for high-level endurance performance [5, 45]. Psychology and fatigue This report contains three psychological key factors relevant for high-level endurance perfor- mance: ‘motivation’, ‘stress resistance’, and ‘self-confidence’. Motivation is a key factor when describing (long-term) success in endurance athletes. Intrinsic motivation determines the high engagement in disciplined training and therefore is crucial for any athlete [46]. Professional athletes must be able to deal with stress during training and competitions [47]. In literature, self-confidence is one of the most cited factors thought to affect athletic performance [48, 49]. Finally, it is well known that engaging in extensive endurance training induce central and peripheral fatigue [50]. Therefore, the ‘sleep quality’ and ‘level of fatigue’ are not surprising to be identified as key factors influencing high-level endurance performance. Areas of disagreement The number of relevant endurance factors declined from 99 factors in the first round to 24 fac- tors in the second round. For instance, the volume of heart (85%) and lungs (78%) were rated as relevant in the first round but have been discarded after the second round (level of agree- ment round 2: 33% and 17%, respectively). In addition, although myoglobin plays an impor- tant role in the oxidative capacity of endurance trained runners [51], ‘myoglobin storage capacity’ was rated by 89% of the experts as relevant in round 1, but only by 33% in round 2. Furthermore, zinc (level of agreement: 85% round 1, 28% round 2), magnesium (level of agree- ment: 85% round 1; 39% round 2), vitamin C (level of agreement: 78% round 1; 22% round 2), and Vitamin E (level of agreement: 74% round 1; 11% round 2) deficiency were eliminated after round 2. Zinc and magnesium are essential trace elements and normal level of zinc are needed for proper immune system function [52]. However, there is no evidence of reduced performance in zinc deficient endurance athletes [53]. Magnesium supplementation, in turn, may enhance athletic performance in deficient athletes [54], although is not beneficial when magnesium status is normal [55]. Based on current scientific knowledge, antioxidants (level of agreement: 82% round 1; 22% round 2) including Vitamin C and E supplements, may not pro- vide additional benefits for athletes. Finally, the ‘risk of upper respiratory tract infections’ achieved a level of agreement of 67% after round 3 and thus has not been included in the con- sensus report (threshold 70%). Except for injuries, upper respiratory symptoms are the most PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 9 / 15 common medical presentation in endurance athletes [56]. All in all, the factors that did not make it on the consensus report remain subject for discussion and should not be overlooked by athletes and coaches. Research implications This consensus report provides athletes, coaches, and exercise scientists a holistic overview of the key factors contributing to high-level endurance performance. Consequently, these factors need to be recognized and prioritized in the future. The 26 factors included in the consensus report demonstrate where to focus on in the future. For instance, the current results can be integrated into novel research disciplines such as -omics approaches. The aim of such approaches is the implementation of precision medicine in sports [14], which attempt to per- sonalize performance by employing prediction models [16]. The predictive ability of such models tends to increase with the amount and quality of data input [17]. Hence, the identifica- tion of key factors and integration into future models the better the predictions will become [18]. Moreover, the current findings can be integrated into new technology such as wearables for data-informed decision-making enabling endurance athletes and coaches to monitor the athletes’ health, risk of injury, and performance in real time [57]. In near future, through rapid advances in technology, more opportunities will arise to integrate the FENDLE factors. Finally, replication studies are warranted to validate the current findings and update the consensus report. Strengths and limitations We would like to highlight several strengths. i) First, we strictly followed the study protocol and reported the methodological considerations undertaken transparently. ii) In the prepara- tion phase, a literature review was performed. The identified studies served as theoretical framework for establishing a comprehensive list. This list may not have captured all endurance factors, although only one new item was proposed from the expert panel. iii) Before distribu- tion, the survey was pilot tested to assure comprehension of the survey [58]. iv) The steering committee consisted of international experts from the field of sport science who advised on study and survey design, methodology, and content. v) The authors assured geographic disper- sion and anonymity of the panellists. Furthermore, the expert panel did not interact directly with each other so that social pressure was avoided [28]. The panellists had also the option to provide free text comments in round 1 and could reconsider initial ratings. vi) Finally, the response rates in round 2 (66%) and 3 (100%) were high [59]. Engaging 15–20 panellists is suf- ficient as long as the background of panellists is homogenous, and three rounds is appropriate for reaching consensus within a Delphi process [19]. We would like to acknowledge some limitations: i) Purposive sampling might have intro- duced selection bias [60] and personal perceptions of the experts could have influenced the results. ii) We are aware that the final expert selection is a limitation of the current study. We advise to recruit a higher number of experts from more diverse backgrounds/ disciplines (e.g., accredited registered sports dietitians) representing in-depth knowledge in each area. iii) We used different approaches to answer the questions in round 1 and 2. Hence we recommend a follow up study employing coherent methods throughout the Delphi process. iv) Some factors may be classified as “higher-level” (e.g., maximal oxygen consumption) or as “lower-level” fac- tors (e.g., ‘metabolic’, ‘blood’ related factors) with the lower-level factors influencing endur- ance performance indirectly by affecting the higher-level factors. Hence, some potentially meaningful (and likely lower-level) factors may have been omitted (e.g., ‘strength/ power’; level of agreement 33.3%). v) Although experts from various disciplines were recruited, most PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 10 / 15 factors identified pertain to running. Factors related to sport-specific skills, such as ‘technique in swimming’ or ‘power in cycling’ are underrepresented. vi) Factors such as ‘exogenous car- bohydrate/ caffeine/ post exercise protein ingestion’, ‘neuromuscular control’, ‘strategy’ or ‘decision making’ were not examined in the current study, although these factors have exten- sively been reported and linked to endurance performance or recovery [5, 6, 61–64]. vii) Expert opinion remains among the lowest levels of empirical evidence [65] and the findings are only as valid as the opinions of the experts constituting the panel. Nonetheless, the experts of our study had on average 20 years of practical experience, reflecting a trained panel with good knowledge of the topic. viii) Finally, Delphi studies are considered evidence-based approaches with emphasis on value of expert judgement, which is not accessible through clini- cal trials [27]. This technique has the potential to arrive at valid and credible results—if per- formed properly. We therefore strictly followed the quality criteria for conducting and reporting Delphi studies to increase the validity of the results. Conclusions This study provides an expert-derived consensus report identifying the important factors for high-level endurance performance, the FENDLE factors. We offer professional coaches, ath- letes, and scientist insights into 26 key endurance factors and strongly recommend considering these factors when optimizing personalized training strategies and technology in the future. Supporting information S1 Table. CREDES checklist. (PDF) S2 Table. Qualifications and responsibilities of the expert panel. (PDF) S3 Table. List of candidate factors (n = 120). (PDF) S4 Table. Survey outline. (PDF) S5 Table. Results of round 1. (PDF) S6 Table. Results of round 2. (PDF) S7 Table. Results of round 3. (PDF) S8 Table. Consensus decision. (PDF) S9 Table. Moderate level of agreement factors. (PDF) S10 Table. Low level of agreement factors. (PDF) PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 11 / 15 Acknowledgments The research team thanks the 27 panellists. We specially thank Mr. Amatori, Mr. Gonzalez, Mr. Leary, and Mr. Tonessen for participation. We also would like to thank the Steering Com- mittee: Paul A. Solberg, Louis Delhaije, Romain Meeusen, Geert Ruigrok, Gerard Rietjens, and Billy Sperlich. Protocol register name: Open science framework. Registration doi: 10.17605/OSF.IO/PGNWT. Author Contributions Conceptualization: Magdalena J. Konopka, Maurice P. Zeegers, Paul A. Solberg, Louis Del- haije, Romain Meeusen, Geert Ruigrok, Gerard Rietjens, Billy Sperlich. Data curation: Magdalena J. Konopka. Formal analysis: Magdalena J. Konopka. Investigation: Magdalena J. Konopka. Methodology: Magdalena J. Konopka, Maurice P. Zeegers, Paul A. Solberg, Louis Delhaije, Romain Meeusen, Geert Ruigrok, Gerard Rietjens, Billy Sperlich. Project administration: Magdalena J. Konopka. Software: Magdalena J. Konopka. Supervision: Maurice P. Zeegers, Billy Sperlich. Visualization: Magdalena J. Konopka. Writing – original draft: Magdalena J. Konopka, Billy Sperlich. Writing – review & editing: Magdalena J. Konopka, Maurice P. Zeegers, Paul A. Solberg, Louis Delhaije, Romain Meeusen, Geert Ruigrok, Gerard Rietjens, Billy Sperlich. References 1. Midgley AW, McNaughton LR, Jones AM. Training to enhance the physiological determinants of long- distance running performance: can valid recommendations be given to runners and coaches based on current scientific knowledge? Sports Med. 2007; 37(10):857–80. https://doi.org/10.2165/00007256- 200737100-00003 PMID: 17887811 2. Suchomel TJ, Nimphius S, Bellon CR, Hornsby WG, Stone MH. Training for Muscular Strength: Meth- ods for Monitoring and Adjusting Training Intensity. Sports Med. 2021; 51(10):2051–66. https://doi.org/ 10.1007/s40279-021-01488-9 PMID: 34101157 3. Seiler S. What is best practice for training intensity and duration distribution in endurance athletes? Int J Sports Physiol Perform. 2010; 5(3):276–91. https://doi.org/10.1123/ijspp.5.3.276 PMID: 20861519 4. Bytomski JR. Fueling for Performance. Sports Health. 2018; 10(1):47–53. https://doi.org/10.1177/ 1941738117743913 PMID: 29173121 5. Thomas DT, Erdman KA, Burke LM. American College of Sports Medicine Joint Position Statement. Nutrition and Athletic Performance. Med Sci Sports Exerc. 2016; 48(3):543–68. https://doi.org/10.1249/ MSS.0000000000000852 PMID: 26891166 6. Burke LM, Hawley JA, Wong SH, Jeukendrup AE. Carbohydrates for training and competition. J Sports Sci. 2011; 29 Suppl 1:S17–27. https://doi.org/10.1080/02640414.2011.585473 PMID: 21660838 7. Saunders PU, Garvican-Lewis LA, Chapman RF, Pe´riard JD. Special Environments: Altitude and Heat. Int J Sport Nutr Exerc Metab. 2019; 29(2):210–9. https://doi.org/10.1123/ijsnem.2018-0256 PMID: 30676138 8. Gould D, Dieffenbach K, Moffett A. Psychological Characteristics and Their Development in Olympic Champions. J Appl Sport Psychol. 2002; 14(3):172–204. PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 12 / 15 9. Baker J, Horton S, Robertson-Wilson J, Wall M. Nurturing sport expertise: factors influencing the devel- opment of elite athlete. J Sports Sci Med. 2003; 2(1):1–9. PMID: 24616603 10. Muhammad Sayem AS, Hon Teay S, Shahariar H, Fink PL, Albarbar A. Review on Smart Electro-Cloth- ing Systems (SeCSs). Sensors (Basel). 2020; 20(3):587. https://doi.org/10.3390/s20030587 PMID: 31973123 11. Mendes JJ, Vieira ME, Pires MB, Stevan SL. Sensor Fusion and Smart Sensor in Sports and Biomedi- cal Applications. Sensors (Basel). 2016; 16(10):1569. https://doi.org/10.3390/s16101569 PMID: 27669260 12. Rodrigo-Carranza V, Gonza´lez-Mohı´no F, Santos-Concejero J, Gonza´lez-Rave´ JM. The effects of foot- wear midsole longitudinal bending stiffness on running economy and ground contact biomechanics: A systematic review and meta-analysis. Eur J Sport Sci. 2021 Aug 8. https://doi.org/10.1080/17461391. 2021.1955014 PMID: 34369282 13. Sellami M, Elrayess MA, Puce L, Bragazzi NL. Molecular Big Data in Sports Sciences: State-of-Art and Future Prospects of OMICS-Based Sports Sciences. Front Mol Biosci. 2021; 8:815410. https://doi.org/ 10.3389/fmolb.2021.815410 PMID: 35087871 14. Tanisawa K, Wang G, Seto J, Verdouka I, Twycross-Lewis R, Karanikolou A, et al. Sport and exercise genomics: the FIMS 2019 consensus statement update. Br J Sports Med. 2020; 54(16):969–75. https:// doi.org/10.1136/bjsports-2019-101532 PMID: 32201388 15. Pickering C, Kiely J. The Development of a Personalised Training Framework: Implementation of Emerging Technologies for Performance. J Funct Morphol Kinesiol. 2019; 4(2):25. https://doi.org/10. 3390/jfmk4020025 PMID: 33467340 16. Montalvo A, Tse-Dinh Y-C, Liu Y, Swartzon M, Hechtman K, Myer G. Precision Sports Medicine: The Future of Advancing Health and Performance in Youth and Beyond. Strength Cond J. 2017; 39:48–58. 17. Fro¨hlich H, Balling R, Beerenwinkel N, Kohlbacher O, Kumar S, Lengauer T, et al. From hype to reality: data science enabling personalized medicine. BMC Med. 2018; 16(1):150. https://doi.org/10.1186/ s12916-018-1122-7 PMID: 30145981 18. Mattsson CM, Wheeler MT, Waggott D, Caleshu C, Ashley EA. Sports genetics moving forward: les- sons learned from medical research. Physiol Genomics. 2016; 48(3):175–82. https://doi.org/10.1152/ physiolgenomics.00109.2015 PMID: 26757801 19. Hsu C-C, Sandford B. The Delphi Technique: Making Sense Of Consensus. Practical Assessment, Research and Evaluation. 2007; 12:10. 20. Nasa P, Jain R, Juneja D. Delphi methodology in healthcare research: How to decide its appropriate- ness. World J Methodol. 2021; 11(4):116–29. https://doi.org/10.5662/wjm.v11.i4.116 PMID: 34322364 21. Humphrey-Murto S, Varpio L, Wood TJ, Gonsalves C, Ufholz LA, Mascioli K, et al. The Use of the Del- phi and Other Consensus Group Methods in Medical Education Research: A Review. Acad Med. 2017; 92(10):1491–8. https://doi.org/10.1097/ACM.0000000000001812 PMID: 28678098 22. Jorm AF. Using the Delphi expert consensus method in mental health research. Aust N Z J Psychiatry. 2015; 49(10):887–97. https://doi.org/10.1177/0004867415600891 PMID: 26296368 23. Scrivin R, Costa RJ, Pelly F, Lis D, Slater G. Development and validation of a questionnaire investigat- ing endurance athletes practices to manage gastrointestinal symptoms around exercise. Nutr Diet. 2021; 78(3):286–95. https://doi.org/10.1111/1747-0080.12674 PMID: 34047004 24. Tam R, Beck KL, Gifford JA, Flood VM, O’Connor HT. Development of an Electronic Questionnaire to Assess Sports Nutrition Knowledge in Athletes. J Am Coll Nutr. 2020; 39(7):636–44. https://doi.org/10. 1080/07315724.2020.1723451 PMID: 32011971 25. Sinha IP, Smyth RL, Williamson PR. Using the Delphi technique to determine which outcomes to mea- sure in clinical trials: recommendations for the future based on a systematic review of existing studies. PLoS Med. 2011; 8(1):e1000393. https://doi.org/10.1371/journal.pmed.1000393 PMID: 21283604 26. Hasson F, Keeney S, McKenna H. Research guidelines for the Delphi survey technique. J Adv Nurs. 2000; 32(4):1008–15. PMID: 11095242 27. Ju¨nger S, Payne SA, Brine J, Radbruch L, Brearley SG. Guidance on Conducting and REporting DElphi Studies (CREDES) in palliative care: Recommendations based on a methodological systematic review. Palliat Med. 2017; 31(8):684–706. https://doi.org/10.1177/0269216317690685 PMID: 28190381 28. Beiderbeck D, Frevel N, von der Gracht HA, Schmidt SL, Schweitzer VM. Preparing, conducting, and analyzing Delphi surveys: Cross-disciplinary practices, new directions, and advancements. MethodsX. 2021; 8:101401. https://doi.org/10.1016/j.mex.2021.101401 PMID: 34430297 29. Joyner MJ, Coyle EF. Endurance exercise performance: the physiology of champions. J Physiol. 2008; 586(1):35–44. https://doi.org/10.1113/jphysiol.2007.143834 PMID: 17901124 30. Murray B, Rosenbloom C. Fundamentals of glycogen metabolism for coaches and athletes. Nutr Rev. 2018; 76(4):243–59. https://doi.org/10.1093/nutrit/nuy001 PMID: 29444266 PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 13 / 15 31. Hall ECR, Semenova EA, Bondareva EA, Borisov OV, Andryushchenko ON, Andryushchenko LB, et al. Association of muscle fiber composition with health and exercise-related traits in athletes and untrained subjects. Biol Sport. 2021; 38(4):659–66. https://doi.org/10.5114/biolsport.2021.102923 PMID: 34937976 32. Bergh U, Thorstensson A, Sjo¨din B, Hulten B, Piehl K, Karlsson J. Maximal oxygen uptake and muscle fiber types in trained and untrained humans. Med Sci Sports. 1978; 10(3):151–4. PMID: 723502 33. Conley KE. Mitochondria to motion: optimizing oxidative phosphorylation to improve exercise perfor- mance. J Exp Biol. 2016; 219(2):243–9. https://doi.org/10.1242/jeb.126623 PMID: 26792336 34. Groennebaek T, Vissing K. Impact of Resistance Training on Skeletal Muscle Mitochondrial Biogenesis, Content, and Function. Front Physiol. 2017; 8:713. https://doi.org/10.3389/fphys.2017.00713 PMID: 28966596 35. Bishop DJ, Botella J, Genders AJ, Lee MJ, Saner NJ, Kuang J, et al. High-Intensity Exercise and Mito- chondrial Biogenesis: Current Controversies and Future Research Directions. Physiology (Bethesda). 2019; 34(1):56–70. https://doi.org/10.1152/physiol.00038.2018 PMID: 30540234 36. Gaesser GA, Brooks GA. Muscular efficiency during steady-rate exercise: effects of speed and work rate. J Appl Physiol. 1975; 38(6):1132–9. https://doi.org/10.1152/jappl.1975.38.6.1132 PMID: 1141128 37. Otto JM, Montgomery HE, Richards T. Haemoglobin concentration and mass as determinants of exer- cise performance and of surgical outcome. Extrem Physiol Med. 2013; 2(1):33. https://doi.org/10.1186/ 2046-7648-2-33 PMID: 24280034 38. Wood RI, Stanton SJ. Testosterone and sport: current perspectives. Horm Behav. 2012; 61(1):147–55. https://doi.org/10.1016/j.yhbeh.2011.09.010 PMID: 21983229 39. Sasaki R, Masuda S, Nagao M. Erythropoietin: multiple physiological functions and regulation of biosyn- thesis. Biosci Biotechnol Biochem. 2000; 64(9):1775–93. https://doi.org/10.1271/bbb.64.1775 PMID: 11055378 40. Kraemer WJ, Ratamess NA, Hymer WC, Nindl BC, Fragala MS. Growth Hormone(s), Testosterone, Insulin-Like Growth Factors, and Cortisol: Roles and Integration for Cellular Development and Growth With Exercise. Front Endocrinol (Lausanne). 2020 Feb 25. https://doi.org/10.3389/fendo.2020.00033 PMID: 32158429 41. Duclos M, Gouarne C, Bonnemaison D. Acute and chronic effects of exercise on tissue sensitivity to glucocorticoids. J Appl Physiol (1985). 2003; 94(3):869–75. https://doi.org/10.1152/japplphysiol.00108. 2002 PMID: 12433870 42. Crescioli C. Vitamin D Restores Skeletal Muscle Cell Remodeling and Myogenic Program: Potential Impact on Human Health. Int J Mol Sci. 2021; 22(4):1760. https://doi.org/10.3390/ijms22041760 PMID: 33578813 43. Krzysztofik M, Wilk M, Wojdała G, Gołaś A. Maximizing Muscle Hypertrophy: A Systematic Review of Advanced Resistance Training Techniques and Methods. Int J Environ Res Public Health. 2019; 16 (24):4897. https://doi.org/10.3390/ijerph16244897 PMID: 31817252 44. Baker LB. Sweating Rate and Sweat Sodium Concentration in Athletes: A Review of Methodology and Intra/Interindividual Variability. Sports Med. 2017; 47(1):111–28. https://doi.org/10.1007/s40279-017- 0691-5 PMID: 28332116 45. Costa RJS, Knechtle B, Tarnopolsky M, Hoffman MD. Nutrition for Ultramarathon Running: Trail, Track, and Road. Int J Sport Nutr Exerc Metab. 2019; 29(2):130–40. https://doi.org/10.1123/ijsnem.2018-0255 PMID: 30943823 46. Gillet N, Berjot S, Vallerand RJ, Amoura S, Rosnet E. Examining the motivation-performance relation- ship in competitive sport: A cluster-analytic approach. Int J Sport Psychol. 2012; 43(2):79–102. 47. Anshel M, Anderson D. Coping With Acute Stress in Sport: Linking Athletes’ Coping Style, Coping Strat- egies, Affect, and Motor Performance. Anxiety Stress Coping. 2002; 15(2):193–209. 48. Feltz DL. Self-confidence and sports performance. Exerc Sport Sci Rev. 1988; 16:423–57. PMID: 3292264 49. Hays K, Thomas O, Maynard I, Bawden M. The role of confidence in world-class sport performance. J Sports Sci. 2009; 27(11):1185–99. https://doi.org/10.1080/02640410903089798 PMID: 19724964 50. Zając A, Chalimoniuk M, Maszczyk A, Gołaś A, Lngfort J. Central and Peripheral Fatigue During Resis- tance Exercise—A Critical Review. J Hum Kinet. 2015; 49:159–69. https://doi.org/10.1515/hukin-2015- 0118 PMID: 26839616 51. Duteil S, Bourrilhon C, Raynaud JS, Wary C, Richardson RS, Leroy-Willig A, et al. Metabolic and vascu- lar support for the role of myoglobin in humans: a multiparametric NMR study. Am J Physiol Regul Integr Comp Physiol. 2004; 287(6):1441–9. https://doi.org/10.1152/ajpregu.00242.2004 PMID: 15528402 52. Sanna A, Firinu D, Zavattari P, Valera P. Zinc Status and Autoimmunity: A Systematic Review and Meta-Analysis. Nutrients. 2018; 10(1):68. https://doi.org/10.3390/nu10010068 PMID: 29324654 PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 14 / 15 53. Heffernan SM, Horner K, De Vito G, Conway GE. The Role of Mineral and Trace Element Supplementa- tion in Exercise and Athletic Performance: A Systematic Review. Nutrients. 2019; 11(3):696. https://doi. org/10.3390/nu11030696 PMID: 30909645 54. Zhang Y, Xun P, Wang R, Mao L, He K. Can Magnesium Enhance Exercise Performance? Nutrients. 2017; 9(9):946. https://doi.org/10.3390/nu9090946 PMID: 28846654 55. Lukaski HC. Vitamin and mineral status: effects on physical performance. Nutrition. 2004; 20(7–8):632– 44. https://doi.org/10.1016/j.nut.2004.04.001 PMID: 15212745 56. Colbey C, Cox AJ, Pyne DB, Zhang P, Cripps AW, West NP. Upper Respiratory Symptoms, Gut Health and Mucosal Immunity in Athletes. Sports Med. 2018; 48(Suppl 1):65–77. https://doi.org/10.1007/ s40279-017-0846-4 PMID: 29363055 57. Adesida Y, Papi E, McGregor AH. Exploring the Role of Wearable Technology in Sport Kinematics and Kinetics: A Systematic Review. Sensors (Basel). 2019; 19(7):1597. https://doi.org/10.3390/s19071597 PMID: 30987014 58. Schmalz U, Spinler S, Ringbeck J. Lessons Learned from a Two-Round Delphi-based Scenario Study. MethodsX. 2021; 8:101179. https://doi.org/10.1016/j.mex.2020.101179 PMID: 33365260 59. Akins RB, Tolson H, Cole BR. Stability of response characteristics of a Delphi panel: application of boot- strap data expansion. BMC Med Res Methodol. 2005; 5:37. https://doi.org/10.1186/1471-2288-5-37 PMID: 16321161 60. Moser A, Korstjens I. Series: Practical guidance to qualitative research. Part 3: Sampling, data collec- tion and analysis. Eur J Gen Pract. 2018; 24(1):9–18. https://doi.org/10.1080/13814788.2017.1375091 PMID: 29199486 61. Cooke A, Kavussanu M, McIntyre D, Ring C. Effects of competition on endurance performance and the underlying psychological and physiological mechanisms. Biol Psychol. 2011; 86(3):370–8. https://doi. org/10.1016/j.biopsycho.2011.01.009 PMID: 21295108 62. Jeukendrup AE. Carbohydrate and exercise performance: the role of multiple transportable carbohy- drates. Curr Opin Clin Nutr Metab Care. 2010; 13(4):452–7. https://doi.org/10.1097/MCO. 0b013e328339de9f PMID: 20574242 63. Konings MJ, Hettinga FJ. Pacing Decision Making in Sport and the Effects of Interpersonal Competition: A Critical Review. Sports Med. 2018; 48(8):1829–43. https://doi.org/10.1007/s40279-018-0937-x PMID: 29799094 64. Pageaux B. The Psychobiological Model of Endurance Performance: An Effort-Based Decision-Making Theory to Explain Self-Paced Endurance Performance. Sports Med. 2014; 44(9):1319–20. https://doi. org/10.1007/s40279-014-0198-2 PMID: 24809249 65. Evans D. Hierarchy of evidence: a framework for ranking evidence evaluating healthcare interventions. J Clin Nurs. 2003; 12(1):77–84. https://doi.org/10.1046/j.1365-2702.2003.00662.x PMID: 12519253 PLOS ONE Key endurance factors: A Delphi study PLOS ONE | https://doi.org/10.1371/journal.pone.0279492 December 27, 2022 15 / 15

Factors associated with high-level endurance performance: An expert consensus derived via the Delphi technique.

12-27-2022

Konopka, Magdalena J,Zeegers, Maurice P,Solberg, Paul A,Delhaije, Louis,Meeusen, Romain,Ruigrok, Geert,Rietjens, Gerard,Sperlich, Billy

eng

PMC5849329

RESEARCH ARTICLE Advances of the reverse lactate threshold test: Non-invasive proposal based on heart rate and effect of previous cycling experience Leonardo Henrique Dalcheco Messias☯, Emanuel Elias Camolese Polisel☯, Fu´lvia Barros Manchado-Gobatto*☯ School of Applied Sciences, University of Campinas, Limeira, Sao Paulo, Brazil ☯ These authors contributed equally to this work. * [email protected] Abstract Our first aim was to compare the anaerobic threshold (AnT) determined by the incremental protocol with the reverse lactate threshold test (RLT), investigating the previous cycling experience effect. Secondarily, an alternative RLT application based on heart rate was pro- posed. Two groups (12 per group-according to cycling experience) were evaluated on cycle ergometer. The incremental protocol started at 25 W with increments of 25 W at each 3 min- utes, and the AnT was calculated by bissegmentation, onset of blood lactate concentration and maximal deviation methods. The RLT was applied in two phases: a) lactate priming seg- ment; and b) reverse segment; the AnT (AnTRLT) was calculated based on a second order polynomial function. The AnT from the RLT was calculated based on the heart rate (AnTRLT- HR) by the second order polynomial function. In regard of the Study 1, most of statistical pro- cedures converged for similarity between the AnT determined from the bissegmentation method and AnTRLT. For 83% of non-experienced and 75% of experienced subjects the bias was 4% and 2%, respectively. In Study 2, no difference was found between the AnTRLT and AnTRLT-HR. For 83% of non-experienced and 91% of experienced subjects, the bias between AnTRLT and AnTRLT-HR was similar (i.e. 6%). In summary, the AnT determined by the incremental protocol and RLT are consistent. The AnT can be determined during the RLT via heart rate, improving its applicability. However, future studies are required to improve the agreement between variables. Introduction Over the last fifty years, the anaerobic threshold concept (AnT) has been used inside the sports sciences and clinical contexts. This concept was initially proposed for cardiac patients [1], but its application is also considered for control and prescription of individualized exercise inten- sity in trained individuals [2]. Physiological parameters such as respiratory and metabolic are commonly used to identify the AnT [3]. In this sense, the Maximal Lactate Steady State proto- col (MLSS) [4, 5] is considered the gold standard protocol for this determination. The MLSS PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 1 / 20 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Messias LHD, Polisel EEC, Manchado- Gobatto FB (2018) Advances of the reverse lactate threshold test: Non-invasive proposal based on heart rate and effect of previous cycling experience. PLoS ONE 13(3): e0194313. https://doi.org/ 10.1371/journal.pone.0194313 Editor: Pedro Tauler, Universitat de les Illes Balears, SPAIN Received: May 30, 2017 Accepted: February 28, 2018 Published: March 13, 2018 Copyright: © 2018 Messias et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Funding: This work was supported by Conselho Nacional de Desenvolvimento Cientı´fico e Tecnolo´gico (Award Number:302827/2015-3; Recipient: Fu´lvia de Barros Manchado-Gobatto) and Fundac¸ão de Amparo à Pesquisa do Estado de São Paulo (Award Number:2012/06355-2; Recipient: Fu´lvia de Barros Manchado-Gobatto). The funders had no role in study design, data has been characterized as the highest intensity that may be supported over time without blood lactate accumulation [5]. However, the necessity of several evaluation days hinders the MLSS application. Conversely, single-session tests such as the incremental protocol (i.e. graded exercise test) are valid alternatives for the AnT identification. During this application, the elevation of the intensity increases the pyruvate rate production that eventually exceeds the maximal rate of pyruvate oxidation, resulting in the formation of lactate. Thus, the AnT is identified during an incremental protocol based on the non-linear increase or a distinct change in the inclination of blood lactate curve [6, 7]. Although mathematical procedures like bissegmentation of two linear regressions [8, 9], fixed blood lactate concentration [10] or maximal perpendicular dis- tance [11] were proposed and extensively used to overcome disadvantages of the visual identi- fication [12], criticism still persists regarding the AnT determination during the incremental protocol [13, 14]. Based on such criticisms, Dotan [15] proposed a novel-testing coined as Reverse Lactate Threshold test (RLT). This application is conducted in two complementary phases. During the lactate priming segment phase (phase 1) individuals are submitted to a controlled warm-up (i.e. 2–3 stages) followed by two subsequent stages at and above the predicted AnT intensity, respectively. Thereafter, during the reverse segment phase (phase 2) the intensity is stepped down in the subsequent stages. Once intensity declines below the AnT, the blood lactate con- centration falls until the appearance-disappearance balance is attained at the highest point of the reverse plot between intensity vs blood lactate concentration. According to the RLT propo- nent, this point denotes the AnT intensity. Apart from the interesting application, this original study was underpowered by the reduced sample evaluated. Regarding the comparison between the incremental protocol and the RLT, Wahl et al., [16] showed (among other results) high accuracy between these applications to determine the MLSS. However, these authors proposed a modified RLT protocol, and the replication of the Dotan [15] findings is still required, mainly considering a larger sample. Additionally, despite heart rate is a valuable tool to link the laboratory results to field conditions [17, 18], the effect of the heart rate for AnT identification via the RLT was not tested. Due to its non-invasive nature, this analysis represents a valuable tool for athletes and non-athletes. During an incre- mental protocol blood lactate concentration commonly present a curvilinear kinetic [19] while heart rate increases linearly or often presents an exponential decrement [20]. However, during the reverse segment of the RLT, the heart rate presents a smoothed fashion [15] that theoreti- cally enables the identification of AnT by the heart rate. Similar procedure was already con- ducted for healthy [21] and disabled [22] individuals in the lactate minimum test [23], which in turn shares some similarity within the RLT protocol. In order to increase the RLT applica- bility, we investigated the AnT identification during this protocol using the heart rate. Therefore, the first aim of this study was to compare in cycling exercise the AnT deter- mined by distinct mathematical models on traditional incremental protocol with those from the RLT. Secondarily, we aimed to test an alternative and non-invasive proposal for the AnT determination in terms of heart rate during the RLT. Moreover, both aims were tested consid- ering the effect of the previous experience in cycling exercise. Materials and methods Subjects were asked to keep the same individual hydration/food habits and avoid hard physical activity, alcohol and caffeine ingestion at least 96 hours prior the tests. Sample size estimation was performed using the G-Power software [24], considering 1-β = 0.90 and α = 0.05. Accord- ing to it, ten subjects per group were enough for the aims of this study. Therefore, twenty-four Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 2 / 20 collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: The authors have declared that no competing interests exist. active, non-athletes, and non-smokers men were divided in two groups (12 per group). Group 1 consisted of individuals (age = 24±2 years, body mass = 78.1±11.9 kg, height = 178±1 cm, body fat = 8.1±5.3%) that have a minimum of two years’ experience (3 times a week for at least 60 minutes) in cycling exercise, including mountain bike and speed subcategories. On the other hand, in Group 2 were included individuals (age = 23±3 years, body mass = 67.9±7.8 kg, height = 173±1 cm, body fat = 7.9±4.3%) that perform (3 times a week for at least 60 minutes) distinct activities (e.g. resisted training, soccer, basketball, etc.), but did not have any experience in cycling exercise. The level of physical activity was analyzed by the International Physical Activ- ity Questionnaire (IPAQ) (Group 1 = 4985±3196 MET-min/wk; Group 2 = 3218±1803 MET- min/wk). Participants provided written, informed consent authorizing their participation in this study. All experiments were approved by the Faculty of Medical Sciences Ethics Committee (n˚ 41257314.5.0000.5404) and were conducted according to the ethical international standards. Design Twenty four subjects completed three visits to the laboratory, 36–48 hr apart. The first visit was conducted as a familiarization session (i.e. 30 minutes of cycling at fixed 80 rpm) and also for the anthropometric evaluation. During the second session, participants underwent an incremental exercise testing for AnT determination. Such procedure was necessary for the third evaluation session, in which the RLT protocol was carried with basis on the AnT. Regard- ing our first aim, we compared the AnT from a traditional incremental protocol with those from the RLT, considering the previous cycling experience effect. In light of our positive results, we tested the possibility to determine non-invasively the AnT from the RLT protocol in terms of heart rate. The previous experience effect was also considered in such perspective. The approaches from Cohen, Bland and Altman, Pearson and Fisher were used for analysis of magnitude of differences, agreement, relationship and comparison of variances [25–27]. Procedures All procedures were conducted in a controlled environment throughout experiment (temperature = 22˚C±1˚C; relative humidity = 50%±2%; luminosity = ~300lx. The protocols were carried on a Monark cycle ergometer (Monark Ergomedic 894 E, Monark, Sweden) at the same time of day. While experienced cyclists (Group 1) prefer high cadences (>85 rpm), non-experienced (Group 2) have opposite preferences (<75 rpm) [28]. However, evidences conclude the delta efficiency (i.e. indirect measurement of muscular efficiency) seems to be not affected by the cycling experience [28]; therefore, we opted for the 80 rpm, since it is an intermediary cadence between experienced and non-experienced individuals. The power was directly analyzed via a module USB 6008 (National Instruments, TX, US) connected to the cycle ergometer. Throughout testing, signals were collected at 1000 Hz frequency. Thereafter, were processed and analyzed using LabView-Signal-Express 2.0 (National Instruments, TX, US) and Matlab (Mathworks, MA, US), respectively. The power was calculated as: Power = (RF (EDT x PR))/6.12, where RF = resistance on the flywheel (kg); EDT = effective distance travelled (m), which was equal to 6.03 (i.e. circumference of the flywheel resistance track— 1.625 m times the revolutions on the flywheel—3.714 resulted from one complete revolution); PR = pedal revolution; 6.12 is the constant for the conversion of kpm to watts [29]. Aim 1 –Comparison between the incremental protocol and RLT results in experienced and non-experienced subjects The incremental protocol started at initial power output of 25 W, with increments of 25 W at every 3 minutes. The exhaustion criteria were considered as non-maintenance of the predicted Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 3 / 20 cadence, attainment of the predicted maximum heart rate ((210)—age0,65) [30] or volitional exhaustion. Subjects were instructed to maintain the 80 rpm cadence throughout the test. At the end of each stage, capillarized blood samples (25 μl) were taken from the earlobe and deposited into microtubes (Eppendorf 1.5 ml) containing 50 μl of NaF. The blood L-lactate concentration was analyzed by the electrochemical method using a lactimeter YSI- 2300-STAT-Plus (Yellow Springs, OH, USA). Three mathematical models were used to calculate the AnT and blood lactate/heart rate at the AnT. Regarding the bissegmentation method (i.e. two-segment linear fit) [8, 9], blood lac- tate concentration was plotted against power and the lactate breakpoint was visually identified by three independent and experienced researchers; subsequently, the AnT (AnTInc-Biss) was determined by the intersection of the two linear fit. The blood lactate concentration at anaero- bic threshold intensity ([Lac]AnT-Inc-Biss) was calculated by linear interpolation. The mean heart rate of each stage was plotted against power and the heart rate at anaerobic threshold intensity (HRAnT-Inc-Biss) was calculated by linear interpolation. For obtainment of AnT and heart rate at AnT by the onset of blood lactate concentration (OBLA) [10], the fixed blood lactate concentration at 4 (AnTInc-OBLA4 and HRAnT-Inc-OBLA4) and 3 mmol.L-1 (AnTInc-OBLA3 and HRAnT-Inc-OBLA3) were calculated by an exponential inter- polation on the blood lactate concentration and power curve. The blood lactate concentration at these intensities was not calculated in this method since it uses fixed blood lactate concentra- tion. The method defined as the maximal distance between two end points of the blood lactate concentration vs power curve (Dmax) [11] was also conducted. Therefore, the AnT was calcu- lated as the maximal perpendicular distance between a linear regression (considering the two end points) and an exponential (AnTInc-DmaxExp) or 2˚ order polynomial (AnTInc-DmaxPoli) adjustments. The blood lactate concentration at these intensities were calculated by exponen- tial ([Lac]Inc-DmaxExp) or polynomial ([Lac]Inc-DmaxPoli) interpolation. The time to exhaustion (Tex) was considered as the total time of exercise performed during the incremental protocol. Heart rate was monitored using a validated monitor [31] (Polar, RS800, RJ, BR; accuracy ± 1%). As previously described for HRAnT-Inc-Biss analysis, the mean heart rate of each stage was plotted against power, and the HRAnT-Inc-OBLA4, HRAnT-Incremental- OBLA3, HRInc-DmaxExp and HRInc-DmaxPoli were calculated by linear interpolation. In addition, the maximal heart rate was obtained as the highest heart rate registered during the test (HRmax-calc) and by predicted equation (HRmax-predic) (HRmax-predic = 210—age x 0.65) [30]. The percent of maximal heart rate at the thresholds was also calculated considering the HRmax (%HRmax-calc) and by the predicted maximum heart rate (%HRmax-predic). Pulse oxim- etry (SpO2) was also monitored during the test (OXIFAST, Takaoka, SP, BR). The RLT protocol was applied according to the original study [15]. The only difference con- sisted on the stages duration. Dotan [15] opted by the 4-min duration to avoid conflicting con- siderations that shorter on longer durations may result. Since previous reports demonstrated no difference in terms of AnT determination in graded exercise tests [3] within stages of 3-min or 4-min, we opted for the former to avoid lengthen evaluations. The RLT was con- ducted with basis on the AnTInc. Capillarized blood samples were collected at the end of each stage and blood lactate concentration was analyzed as earlier described. Heart rate and pulse oximetry were also monitored. The RLT was applied in 27 minutes of cycling exercise in two phases: a) Lactate-priming segment; and b) Reverse Segment (Fig 1). During the lactate-priming segment three graded stages below the AnTInc were conducted as a controlled warm-up. For instance, whether a subject achieved an AnTInc of 150 W, we considered the intensities of 75, 100, 125 W during the first part of the RLT lactate-priming segment (totalizing 9 minutes of warm-up). Subsequently, individuals performed three Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 4 / 20 minutes in its AnT intensity and right after another stage at 20% above the AnTInc. In sum- mary, the lactate-priming segment consisted in 15 minutes of cycling. The reverse segment was conducted following the lactate-priming segment. In this phase, intensity is retrograde. Decrements of 10 W (8%) were applied at each stage, totaling 12 min- utes of cycling exercise. Considering the blood lactate concentration is elevated at the last part of the lactate-priming segment (i.e 20% above AnTInc), it is expected that this metabolite decays once exercise intensity was below AnTInc (i.e. during the last two stages of the reverse segment). The reverse plot between power and blood lactate concentration was traced using a second order polynomial function. According to the RLT proponent, the AnT from the RLT protocol (AnTRLT) corresponds to the apex of the blood lactate concentration vs power curve. The second order polynomial equation (y = ax2- bx + c) was considered, and the AnTRLT was calculated as y = b / (2a), and the blood lactate concentration at AnTRLT ([Lac]AnT-RLT) as: y = ((a (AnTRLT  AnTRLT))—(b  AnTRLT) + c). The HRmax-calc, %HRmax-calc, %HRmax-predic were also calculated. The HRmax-predic was the same for both RLT and incremental protocol, since considers the subject’s age [30]. Aim 2 –Non-invasive determination of the AnTRLT based on heart rate in experienced and non-experienced subjects During the RLT the heart rate was monitored. The heart rate was plotted against power (Fig 2) and a second order polynomial function was adopted and the apex was considered as the AnTRLT in terms of heart rate (AnTRLT-HR). As described in the latter section, the second order polynomial equation (y = ax2- bx + c) was considered, and the AnTRLT-HR was calculated as y = b / (2a), and the heart rate at AnTRLT (HRAnT-RLT) as: y = ((a (AnTRLT-HR  AnTRLT-HR))—(b  AnTRLT-HR) + c). The success rate of the AnTRLT and AnTRLT–HR was determined considering the goodness of fit (R2) of the polynomial adjust- ment higher than 0.90. Fig 1. Curve of the reverse lactate threshold test considering the blood lactate concentration plotted against power from experienced individual (subject 1 from experienced group). The second order polynomial adjustment was considered for the anaerobic threshold determination (AnTRLT) and blood lactate concentration at anaerobic threshold ([Lac]AnT-RLT). https://doi.org/10.1371/journal.pone.0194313.g001 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 5 / 20 Statistical analyses Statistical procedures were conducted using a statistical software package (STATISTICA 7.0, Statsoft, OK, USA). Mean and standard deviation (SD) were calculated for all studied variables. Levene and Shapiro-Wilk tests confirmed the homogeneity and normality of our data. Com- parison among the results of the incremental protocol (i.e. AnTInc, [Lac]AnT-Inc and HRAnT-Inc, SpO2) and RLT (i.e. AnTRLT and [Lac]AnT-RLT and HRAnT-RLT, SpO2) for both experienced and non-experienced subjects was performed by the two-way ANOVA. Similar procedures were adopted for the comparison among the AnT determined by blood lactate concentration and heart rate in the RLT. In the latter case, the data from the RLT in terms of blood lactate concentration was also used in the second aim of this study. The Scheffe´ post-hoc was consid- ered in all analysis of variance. The comparison of the Tex between experienced and non-expe- rienced individuals was performed by a t-test for independent samples. The same test was adopted for the comparison of the anthropometric characteristics. The agreement between variables was analyzed by the Bland-Altman procedure [25] considering α = 0.05. Pearson product moment (r) coefficient of variation (CV) [27] and effect sizes (ES) [26] were also cal- culated. Cohen’s categories used to evaluate the magnitude of the ES were: small if 0  |d| 0.5; medium if 0.5 < |d|  0.8; and large if |d| > 0.8). In all cases, statistical significance was set at P<0.05. Results Aim 1 –Comparison between the incremental protocol and RLT results in experienced and non-experienced subjects Statistical difference between groups regarding anthropometric characteristics was only visual- ized for body mass (age-P = 0.251; body mass-P = 0.017; height-P = 0.079; body fat-P = 0.982). ANOVA did not point difference regarding the mathematical models used for anaerobic thresh- old intensity determination on the incremental protocol and individual’s cycling experience Fig 2. Curve of the reverse lactate threshold test considering the heart rate plotted against power from experienced individual (subject 1 from experienced group). The second order polynomial adjustment was considered for the anaerobic threshold determination (AnTRLT-HR) and heart rate at anaerobic threshold (HRAnT-RLT). https://doi.org/10.1371/journal.pone.0194313.g002 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 6 / 20 (P = 0.395). Similar results were obtained for blood lactate concentration (P = 0.818) and heart rate (P = 0.469) at the anaerobic threshold intensity. In regard of the comparison between incremental protocol and RLT results (Tables 1–3), no difference was found between the anaerobic threshold intensity determined by the RLT and the incremental protocol; such result was not affected by the previous experience in cycling exercise (Table 1). However, only the comparison between AnTRLT and AnTInc-Biss converge in most of statistical analysis for both non-experienced and experienced individuals. In line with this, the agreement between these intensities were the strongest when compared with the other mathematical methods (Fig 3A, Fig 4A). Apart from the non-difference (for both groups) and significant relationship (only for non- experienced subjects) between [Lac]AnT-RLT and [Lac]AnT-Inc-DmaxExp, distinct results were found between protocols in terms of the blood lactate concentration at the AnT (Table 2). Excepted of the significant relationship between HRAnT-RLT and HRAnT-Inc-Biss, different results were found between protocols for heart rate at the AnT intensity (Table 3). In addition, the Table 4 shows the results of HRmax-calc, HRmax-predic, %HRmax-calc, %HRmax-predic. The Tex Table 1. Comparison between the anaerobic threshold intensity from the incremental protocol analyzed by different mathematical models and the reverse lactate threshold test obtained by subjects without (non-experienced) or with (experienced) experience with cycling exercise. AnTRLT AnTInc-Biss AnTInc-OBLA4 AnTInc-OBLA3 AnTInc-DmaxExp AnTInc-DmaxPoli (W) (W) (W) (W) (W) (W) Non-Experienced (n = 12) Mean 112 108 105 87 117 95 SD 15 17 10 10 13 8 P (Scheffe´) ——— 0.977 0.924 0.180 0.967 0.226 ES ——— 0.24 0.60 1.86 0.34 1.42 % Diff ——— 3.4 6.8 21.9 4.3 15.1 r (P) ——— 0.95 (0.008) 0.50 (0.090) 0.46 (0.127) 0.81 (0.001) 0.67 (0.015) CV ——— 3.1 8.4 10.1 5.6 8.0 Bland-Altman ——— -3.9±5.0 -7.6±13.6 -24.7±14.35 4.8±9.1 -16.9±11.84 Experienced (n = 12) Mean 155 149 167 143 164 133 SD 25 26 41 39 34 23 P (Scheffe´) ——— 0.925 0.722 0.824 0.868 0.062 ES ——— 0.23 0.34 0.37 0.27 0.91 % Diff ——— 3.8 7.2 7.6 5.2 14.3 r (P) ——— 0.98 (0.000) 0.79 (0.002) 0.75 (0.004) 0.95 (0.000) 0.87 (0.000) CV ——— 2.2 12.3 12.5 5.6 6.2 Bland-Altman ——— -6.0±4.8 11.3±26.3 -11.9±26.4 8.1±12.5 -22.3±12.6 Note: The statistical results refers to the comparison between the anaerobic threshold intensity from the reverse lactate threshold and the anaerobic threshold intensity from the mathematical models applied in the incremental protocol. AnTRLT—Anaerobic threshold intensity determined during the reverse lactate threshold test in terms of blood lactate concentration. AnTInc-Biss−Anaerobic threshold intensity determined during the incremental protocol by the bissegmentation method; AnTInc- OBLA4 –Anaerobic threshold intensity determined during the incremental protocol by the fixed blood lactate concentration at 4 mmol.L-1; AnTInc-OBLA3 –Anaerobic threshold intensity determined during the incremental protocol by the fixed blood lactate concentration at 3 mmol.L-1; AnTInc-DmaxExp−Anaerobic threshold intensity determined during the incremental protocol by maximal deviation method using the exponential adjustment; AnTInc-DmaxPoli−Anaerobic threshold intensity determined during the incremental protocol by maximal deviation method using the second order polynomial adjustment; P (Scheffe´)–P value from the Scheffe´ post- hoc analysis; ES–effect size; % Diff–percent difference between means; r (P) Pearson product moment and P value from the correlation; CV–coefficient of variation; Bland-Altman—Bland Altman analysis considering 95% of agreement limits. https://doi.org/10.1371/journal.pone.0194313.t001 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 7 / 20 was different between non-experienced (18min 13s ± 1min 46s) and experienced (22min 54s ± 2min 48s) individuals (P = 0.000). Pulse oximetry was not different during both incremental (non-experienced– 96.4±1.0; experienced– 96.1±1.2%; P = 0.916) and RLT (non-experienced– 96.0±1.0; experienced– 95.4±1.2%; P = 0.695); additionally, ANOVA also did not show interaction regarding the factorials variables (i.e. test and experience) (P = 0.729). Aim 2 –Non-invasive determination of the AnTRLT based on heart rate in experienced and non-experienced subjects The Table 5 shows the AnT determined during the RLT by means of heart rate for both groups. All statistical procedures converged to the similarity between the results, with the exception of the non-significant relationship (for both groups). The agreement among vari- ables are shown on Fig 5. High R2 were obtained on the polynomial adjustment adopted for both groups in terms of AnTRLT (Non-Experienced—range = 0.78–0.99; Experienced— range = 0.90–0.99) and AnTRLT-HR (Non-Experienced—range = 0.90–0.99; Experienced— Table 2. Comparison between the blood lactate concentration at the anaerobic threshold intensity from the incremental protocol analyzed by different mathemati- cal models and the reverse lactate threshold test obtained by subjects without (non-experienced) or with (experienced) experience with cycling exercise. [Lac]AnT-RLT [Lac]AnT-Inc-Biss [Lac]AnT-Inc-OBLA4 [Lac]AnT-Inc-OBLA3 [Lac]AnT-Inc-DmaxExp [Lac]AnT-Inc-DmaxPoli (mmol.L-1) (mmol.L-1) (mmol.L-1) (mmol.L-1) (mmol.L-1) (mmol.L-1) Non-Experienced (n = 12) Mean 6.74 3.26 ——— ——— 5.01 3.20 SD 2.63 1.29 ——— ——— 1.03 0.86 P (Scheffe´) ——— 0.013 ——— ——— 0.417 0.009 ES ——— 1.78 ——— ——— 0.95 2.03 % Diff ——— 51.6 ——— ——— 25.6 52.5 r (P) ——— 0.71 (0.008) ——— ——— 0.70 (0.010) 0.54 (0.065) CV ——— 27.3 ——— ——— 24.6 32.4 Bland-Altman ——— -3.4±1.9 ——— ——— -1.7±2.0 -3.5±2.2 Experienced (n = 12) Mean 6.47 2.27 ——— ——— 4.12 1.88 SD 3.84 1.08 ——— ——— 1.46 0.85 P (Scheffe´) ——— 0.002 ——— ——— 0.166 0.000 ES ——— 1.71 ——— ——— 0.89 1.95 % Diff ——— 64.9 ——— ——— 36.2 70.8 r (P) ——— 0.29 (0.350) ——— ——— 0.38 (0.212) 0.07 (0.816) CV ——— 59.4 ——— ——— 47.3 77.6 Bland-Altman ——— -4.2±3.6 ——— ——— -2.3±3.5 -4.5±3.9 Note: The statistical results refers to the comparison between the anaerobic threshold intensity from the reverse lactate threshold and the anaerobic threshold intensity from the mathematical models applied in the incremental protocol. [Lac]AnT-RLT−Blood lactate concentration at the anaerobic threshold intensity determined during the reverse lactate threshold test; [Lac]AnT-Inc-Biss−Blood lactate concentration at the anaerobic threshold intensity determined during the incremental protocol by the bissegmentation method; [Lac]AnT-Inc-OBLA4 –Blood lactate concentration at the anaerobic threshold intensity determined during the incremental protocol by the fixed blood lactate concentration at 4 mmol.L-1; [Lac]AnT-Inc-OBLA3 –Blood lactate concentration at the anaerobic threshold intensity determined during the incremental protocol by the fixed blood lactate concentration at 3 mmol.L-1; [Lac]AnT-Inc-DmaxExp−Blood lactate concentration at the anaerobic threshold intensity determined during the incremental protocol by maximal deviation method using the exponential adjustment; [Lac]AnT-Inc-DmaxPoli−Blood lactate concentration at the anaerobic threshold intensity determined during the incremental protocol by maximal deviation method using the second order polynomial adjustment; P (Scheffe´)–P value from the Scheffe´ post-hoc analysis; ES–effect size; % Diff–percent difference between means; r (P) Pearson product moment and P value from the correlation; CV–coefficient of variation; Bland-Altman—Bland Altman analysis considering 95% of agreement limits. https://doi.org/10.1371/journal.pone.0194313.t002 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 8 / 20 range = 0.90–0.99). In line with this, high success rate were obtained for the two determina- tions in both groups. Discussion The main findings of the present investigation were the AnT determined on the RLT was con- sistent with those obtained on the incremental protocol calculate by the bissegmentation mathematical model. In addition, our results showed that previous cycling experience did not influence on such consistency. However, due to the different methodologies, additional parameters like blood lactate concentration and heart rate at the anaerobic threshold intensity are not similar. Additionally, for 83% of the evaluated non-experienced individuals the differ- ence between the AnT determined on RLT by blood lactate concentration and heart rate was 6%. Similar difference was also obtained for 91% of the evaluated experienced individuals. Therefore, our results suggest that AnT determination via heart rate on RLT may be consid- ered as a non-invasive alternative. However, future studies are necessary to investigate which factors can improve the agreement between these measures. Table 3. Comparison between the heart rate at the anaerobic threshold intensity from the incremental protocol analyzed by different mathematical models and the reverse lactate threshold test obtained by subjects without (non-experienced) or with (experienced) experience with cycling exercise. HRAnT-RLT HRAnT-Inc-Biss HRAnT-Inc-OBLA4 HRAnT-Inc-OBLA3 HRAnT-Inc-DmaxExp HRAnT-Inc-DmaxPoli (bpm) (bpm) (bpm) (bpm) (bpm) (bpm) Non-Experienced (n = 12) Mean 171 155 152 144 159 148 SD 13 13 15 15 14 13 P (Scheffe´) ——— 0.000 0.001 0.000 0.010 0.000 ES ——— 1.21 1.30 1.89 0.89 1.72 % Diff ——— 9.4 10.9 15.8 7.2 13.5 r (P) ——— 0.64 (0.022) 0.42 (0.165) 0.35 (0.255) 0.45 (0.138) 0.37 (0.226) CV ——— 4.8 3.5 7.4 6.2 6.6 Bland-Altman ——— -16.1±11.1 -18.7±15.4 -27.1±16.4 -12.3±14.5 -23.1±15.0 Experienced (n = 12) Mean 177 159 165 155 163 150 SD 12 10 14 13 11 8 P (Scheffe´) ——— 0.007 0.321 0.007 0.088 0.000 ES ——— 1.62 0.88 1.71 1.17 2.63 % Diff ——— 10.4 6.6 12.3 7.8 15.2 r (P) ——— 0.44 (0.146) 0.09 (0.760) 0.01 (0.959) 0.14 (0.659) 0.03 (0.907) CV ——— 5.1 4.1 7.8 6.4 6.3 Bland-Altman ——— -18.6±12.5 -11.7±19.9 -21.9±18.3 -13.8±15.5 -27.0±14.5 Note: The statistical results refers to the comparison between the anaerobic threshold intensity from the reverse lactate threshold and the anaerobic threshold intensity from the mathematical models applied in the incremental protocol. HRAnT-RLT−Hear rate at the anaerobic threshold intensity determined during the reverse lactate threshold test; HRAnT-Inc-Biss−Heart rate at the anaerobic threshold intensity determined during the incremental protocol by the bissegmentation method; HRAnT-Inc- OBLA4 –Hear rate at the anaerobic threshold intensity determined during the incremental protocol by the fixed blood lactate concentration at 4 mmol.L-1; HRAnT-Inc- OBLA3 –Hear rate at the anaerobic threshold intensity determined during the incremental protocol by the fixed blood lactate concentration at 3 mmol.L-1; HRAnT-Inc- DmaxExp−Hear rate at the anaerobic threshold intensity determined during the incremental protocol by maximal deviation method using the exponential adjustment; HRAnT-Inc-DmaxPoli−Hear rate at the anaerobic threshold intensity determined during the incremental protocol by maximal deviation method using the second order polynomial adjustment; P (Scheffe´)–P value from the Scheffe´ post-hoc analysis; ES–effect size; % Diff–percent difference between means; r (P) Pearson product moment and P value from the correlation; CV–coefficient of variation; Bland-Altman—Bland Altman analysis considering 95% of agreement limits. https://doi.org/10.1371/journal.pone.0194313.t003 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 9 / 20 Fig 3. Bland-Altman analysis performed for anaerobic threshold intensity determined on the reverse lactate threshold test and incremental protocol on non-experienced individuals. a) Agreement between anaerobic threshold intensity determined on incremental protocol by the bissegmentation method (AnTInc-Biss) and reverse lactate threshold test (AnTRLT); b) Agreement between anaerobic threshold intensity determined on incremental protocol by fixed blood lactate concentration at 4 mmol.L-1 (AnTOBLA4) and AnTRLT; c) Agreement between anaerobic threshold intensity determined on incremental protocol by fixed blood lactate concentration at 3 mmol.L-1 (AnTOBLA3) and AnTRLT; d) Agreement between anaerobic threshold intensity determined on incremental protocol by maximal deviation method using the exponential adjustment (AnTDmaxExp) and AnTRLT; e) Agreement between anaerobic threshold intensity determined on incremental protocol by maximal deviation method using the second order polynomial adjustment (AnTDmaxPoli) and AnTRLT. https://doi.org/10.1371/journal.pone.0194313.g003 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 10 / 20 Fig 4. Bland-Altman analysis performed for anaerobic threshold intensity determined on the reverse lactate threshold test and incremental protocol on experienced individuals. a) Agreement between anaerobic threshold intensity determined on incremental protocol by the bissegmentation method (AnTInc-Biss) and reverse lactate threshold test (AnTRLT); b) Agreement between anaerobic threshold intensity determined on incremental Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 11 / 20 Aim 1 –Comparison between the incremental protocol and RLT results in experienced and non-experienced subjects Regarding our first aim, we found no difference between the AnT determined during the RLT and incremental protocol. On the other hand, differences in terms of blood lactate concentra- tion and heart rate at AnT intensity were found. Despite these distinct results, the previous experience in cycling exercise did not affect the results provided by both protocols. Overall, both protocols seem to identify the same phenomenon (i.e. AnT) regardless of previous experi- ence in cycling, but the physiological condition that such identification occurs depends on the characteristic of each protocol. The main criticism suggested by Dotan [15] regarding the incremental exercise is related to the ambiguity of where the highest blood lactate concentration is identified. Some studies con- verge with this approach [13, 14]. Although the RLT seems to deal with such problem, its pro- ponent also highlights other methodological concerns [15]. For instance, the need for “on-the- fly” blood sampling is suggested as a limited factor for sports where this is not possible (e.g. swimming); this limitation can be extended to incremental protocols applied to these exercises. In this sense, long and repetitive exercise breaks can lead to an AnT overestimation, since the glycolytic rate resulted from the previous incremental stage is decreased while oxygen uptake remains elevated, improving the blood lactate concentration clearance [32]. However, such limitation was not present in our study. Moreover, the duration of stages were standardized for the RLT and incremental protocol. We believe that these factors along with the RLT appli- cation based on the incremental protocol were crucial for the similarity in terms of AnT. Indeed, the AnT may be identified during the incremental protocol when an abrupt eleva- tion of blood lactate concentration is visualized. In this sense, the balance of blood lactate con- centration between lactate producing and absorbing compartments is gradually lost. Conversely, the RLT test is conducted under an inverse concept, that is, the organism is already facing the loss of blood lactate concentration balance between compartments (i.e. last part of lactate-priming segment). When intensities are stepped down in the reverse segment, the balance between blood lactate concentration production-removal is gradually recovered. On the other hand, because at least two stages were performed above the predicted AnT (i.e. last stage from the lactate-priming segment and first stage of the reverse segment), the blood lactate concentration removal will be greater than its production only when the stages below the predicted AnT were performed. Such characteristic is responsible by the high blood lactate concentration and heart rate found in our study at the apex of the second order polynomial adjustment, that is, at the AnTRLT (Tables 2–3). According to Dotan [15], further experimentation is required before final validation of the RLT. Thus, we contributed with such perspective by demonstrating that the previous cycling experience did not affect the AnTRLT identification. It is consolidated that muscular efficiency as well as aerobic and anaerobic power/capacity between individuals with or without cycling experience is quite distinct [33, 34]. Despite the encouraging reports, methodological concerns related to fitness level must be investigated. For instance, we considered the upper limit of 20% increment proposed in the original RLT during the last stage of the lactate-priming segment. However, since well-trained individuals present the AnT at high percentage of its maximal protocol by fixed blood lactate concentration at 4 mmol.L-1 (AnTOBLA4) and AnTRLT; c) Agreement between anaerobic threshold intensity determined on incremental protocol by fixed blood lactate concentration at 3 mmol.L-1 (AnTOBLA3) and AnTRLT; d) Agreement between anaerobic threshold intensity determined on incremental protocol by maximal deviation method using the exponential adjustment (AnTDmaxExp) and AnTRLT; e) Agreement between anaerobic threshold intensity determined on incremental protocol by maximal deviation method using the second order polynomial adjustment (AnTDmaxPoli) and AnTRLT. https://doi.org/10.1371/journal.pone.0194313.g004 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 12 / 20 oxygen consumption (VO2max) [35], it is possible that the 20% increment surpass its maximal oxygen consumption intensity, leading to premature exhaustion. Regarding the reverse segment, the individual’s physiological/metabolic condition must be also accounted. The improvement of blood lactate clearance is considered as positive physio- logical adaptation from endurance training [36], which can be extended to well-trained indi- viduals. Therefore, the upper limit of 8% decrement originally proposed for the reverse segment may lead to a faster blood lactate concentration clearance in these individuals, overes- timating the real AnT. Regarding the stage duration, the slight modification used in this study (i.e. 3-min instead of the original 4-min) did not implicate on the AnTRLT determination, since high coefficient of determinations and high success rate were obtained. Other methodo- logical concerns were considered in the study of Wahl et al., [16], in which a modified RLT protocol was proposed to improve the test’s precision, also enabling the VO2max determination during the lactate priming-segment. Within this new approach the RLT was more accurate to determine the MLSS than other threshold concepts, such as OBLA4 [10] or the Dmax [11]. Overall, further studies should consider the methodological concerns highlighted along with the results of Wahl et al., [16], mainly whether the RLT application was conducted to highly trained individuals. Lastly, another RLT limitation is the necessity of the previous AnT determination for inten- sity prescription during the two phases. However, futures studies are encouraged to investigate whether indirect methods or even the reported mean velocity during recent championships/ training sessions to estimate the intensities of the RLT. The marked differences in terms of methodology and application implies in the use of the RLT and incremental protocol inside Table 4. Maximal heart rate predicted (HRmax-predic) and registered (HRmax-calc) during the reverse lactate threshold test and incremental protocol, as well as the percent of maximal heart rate at the anaerobic threshold intensities considering the registered (%HRmax-calc) and predicted (%HRmax-predic) maximal heart rate. Non-Experienced (n = 12) Experienced (n = 12) RLT HRmax-calc (bpm) 177±14 183±13 HRmax-predic (bpm) 197±3 195±2 AnTRLT %HRmax-calc 96±1 96±1 %HRmax-predic 91±7 94±6 Incremental HRmax-calc (bpm) 191±16 195±7 HRmax-predic (bpm) 197±3 195±2 AnTInc-Biss %HRmax-calc 81±6 81±5 %HRmax-predic 80±7 81±5 AnTInc-OBLA4 %HRmax-calc 79±5 85±8 %HRmax-predic 78±8 85±7 AnTInc-OBLA3 %HRmax-calc 76±6 79±7 %HRmax-predic 74±8 80±7 AnTInc-DmaxExp %HRmax-calc 83±5 83±5 %HRmax-predic 81±7 84±6 AnTInc-DmaxPoli %HRmax-calc 77±4 77±5 %HRmax-predic 76±7 77±4 AnTRLT—Anaerobic threshold intensity determined during the reverse lactate threshold test in terms of blood lactate concentration. AnTInc-Biss−Anaerobic threshold intensity determined during the incremental protocol by the bissegmentation method; AnTInc-OBLA4 –Anaerobic threshold intensity determined during the incremental protocol by the fixed blood lactate concentration at 4 mmol.L-1; AnTInc-OBLA3 –Anaerobic threshold intensity determined during the incremental protocol by the fixed blood lactate concentration at 3 mmol.L-1; AnTInc-DmaxExp−Anaerobic threshold intensity determined during the incremental protocol by maximal deviation method using the exponential adjustment; AnTInc-DmaxPoli−Anaerobic threshold intensity determined during the incremental protocol by maximal deviation method using the second order polynomial adjustment https://doi.org/10.1371/journal.pone.0194313.t004 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 13 / 20 the sport and clinical contexts. For instance, evidences have been suggested that blood lactate concentration and heart rate at the AnT are valid indicators for monitoring longitudinal train- ing effects for different populations [17, 37]. In this sense, because during the RLT application high values of [Lac]AnT-RLT and HRAnT-RLT are obtained, it is possible that this protocol is dis- advantaged for utilization of these results for monitoring longitudinal training effects. This detriment is not transposed to the incremental protocol, since the [Lac]AnT-Inc and HRAnT-Inc are consolidated indicators for training control [17, 37]. Moreover, RLT does not requires exhaustion, the HRmax-calc and %HRmax-calc values cannot be used in the same way as for the incremental protocol (Table 4). It is valid to state that the comparison among mathematical models for determining the anaerobic threshold intensity on incremental protocol is beyond the main aim of this study. Such results were provided to strength the comparison between results from the incremental protocol with the reverse lactate threshold test. In line with this, the results on Table 1 showed that only the AnTInc-Biss and the AnTRLT converged in all statistical procedures adopted. In terms of Pearson product moment, both non-experienced and experienced were heteroge- neous regarding AnT (non-experienced 8–15% and experienced 17–27% if considered all mathematical models), which could have improved the correlation coefficients. In regard of the agreement, the bias from Bland-Altman analysis revealed that in some non-experienced subjects the AnTRLT overestimated the AnTInc-Biss in ~9W. This result was Table 5. Comparison between the results from the reverse lactate threshold test analyzed in terms of blood lactate concentration and heart rate obtained by subjects without (non-experienced) or with (experienced) experience with cycling exercise. AnTRLT AnTRLT-HR AnTRLT AnTRLT-HR AnTRLT AnTRLT-HR (W) (W) R2 R2 Success rate (%) Success rate (%) Non-Experienced (n = 12) 83.33 (n = 10) 100 (n = 12) Mean 112 112 0.93 0.96 SD 15 16 0.07 0.03 P (Scheffe´) 0.999 0.331 ES 0.02 0.69 % Diff 0.3 3.9 r (P) 0.95 (0.000) 0.29 (0.354) CV 3.0 5.4 Bland-Altman 0.2±4.8 -0.0±0.0 Experienced (n = 12) 91.66 (n = 11) 100 (n = 12) Mean 155 156 0.96 0.97 SD 25 27 0.04 0.03 P (Scheffe´) 0.999 0.998 ES 0.04 0.10 % Diff 0.6 0.3 r (P) 0.95 (0.000) 0.14 (0.656) CV 3.8 3.8 Bland-Altman 1.0±8.4 0.0±0.0 AnTRLT—Anaerobic threshold intensity determined during the reverse lactate threshold test in terms of blood lactate concentration; AnTRLT-HR—Anaerobic threshold intensity determined during the reverse lactate threshold test in terms of heart rate; AnTRLT R2—goodness of fit of the polynomial adjustment between power and blood lactate concentration (reverse segment); AnTRLT-HR R2—goodness of fit of the polynomial adjustment between power and heart rate (reverse segment); Success rate— goodness of fit of the polynomial adjustment higher than 0.90. P (Scheffe´)–P value from the Scheffe´ post-hoc analysis; ES–effect size; % Diff–percent difference between means; r (P) Pearson product moment and P value from the correlation; CV–coefficient of variation; Bland-Altman—Bland Altman analysis considering 99% of agreement limits. https://doi.org/10.1371/journal.pone.0194313.t005 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 14 / 20 confirmed in subjects 3 and 10, in which the AnTRLT was 14 and 10 W (respectively) higher than the AnTInc-Biss. However, without these subjects the bias is reduced to ~5W. Overall, for 83% of non-experienced subjects this overestimation only represents 4%. This perspective is extended for experienced subjects. In this case the bias is ~8W, which was confirmed for sub- jects 7, 10 and 12 (i.e. AnTRLT overestimated the AnTInc-Biss in 10, 8 and 17 W, respectively). However, without these subjects the bias is reduced to ~4W, which represents an overestima- tion of only 2% for 75% of experienced subjects. Although fixed blood lactate concentration enables the AnT determination without the exhaustion attainment in incremental protocols [10], some studies have demonstrated that depending on the mathematical model adopted, the AnT determination can be influenced [38, 39]. Therefore, despite the excessive blood sampling may be not suitable for practical contexts, the exhaustion attainment is interesting for the robust analysis of the AnT that is being mea- sured. Moreover, the reliable acquisition of additional parameters (e.g. maximal oxygen con- sumption, velocity at maximal oxygen consumption or time to exhaustion) requires that individuals reach exhaustion. However, the exhaustion attainment necessity may be not Fig 5. a) Agreement performed by Bland-Altman analysis for anaerobic threshold intensity determined on the reverse lactate threshold test by blood lactate concentration (AnTRLT) and heart rate (AnTRLT-HR) for non-experienced individuals; b) Agreement performed by Bland-Altman analysis AnTRLT and AnTRLT-HR for experienced individuals; c) Agreement performed by Bland-Altman analysis for coefficient of determination on the reverse lactate threshold test by blood lactate concentration (AnTRLT R2) and heart rate (AnTRLT-HR R2) for non-experienced individuals; d) Agreement performed by Bland-Altman analysis AnTRLT R2 and AnTRLT-HR R2 for experienced individuals. https://doi.org/10.1371/journal.pone.0194313.g005 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 15 / 20 suitable for untrained individuals. In this sense, the non-exhaustive nature of the RLT is an interesting possibility for the AnT identification in peculiar cases, such as untrained or inexpe- rienced subjects. Therefore, although no differences between AnTInc and AnTRLT were found, we suggest that future studies should consider its applications according to the population that will be evaluated. Aim 2 –Non-invasive determination of the AnTRLT based on heart rate in experienced and non-experienced subjects Heart rate has been considered for monitoring training since 1980s [17]. Sooner after the expansion of heart devices, Conconi et al., [40] proposed the AnT estimation based on this physiological parameter. Nowadays the individualized exercise prescription as well as load controlling based on heart rate are commonly used in practice [17, 18, 41, 42], since factors as easy applicability and reduced financial cost improve the heart rate devices utilization. Addi- tionally, in a recent meta-analysis Bellenger et al., [41] concluded the autonomic heart rate reg- ulation may be considered a potential indicator for overreaching status. Since Dotan [15] demonstrated the heart rate presented a smoothed fashion during the reverse segment, it also opened for discussion if the AnT could be determined based on this physiological parameter. In accordance, our results showed the AnT is determined during RLT based on of heart rate. Moreover, this determination was not affected by the previous experience in the cycling exercise. These results improve the scientific knowledge around the RLT application, since the heart rate is a cheap and easy non-invasive tool. The AnT determination via heart rate during the RLT was possible for all subjects, as exposed in Table 5. In fact, the success rate of AnT determination via blood lactate concentra- tion for two non-experienced (subject 8 –R2 = 0.83; subject 9 –R2 = 0.78) and one experienced (subject 11 –R2 = 0.89) was slightly lower. On the other, the success rate for the AnT via heart rate was 100% for both non-experienced and experienced subjects. In fact, while the blood lac- tate concentration tends to decrease significantly once the equilibrium between production- removal is gradually recovery, the cardiovascular system tends to readjust after the stressful stage (i.e. 20% above the AnT) in a smoothed fashion [15](Fig 2). However, such differences did not affect the second order polynomial adjustment, since high R2 were found for all sub- jects in the AnT determination via blood lactate concentration and heart rate. Apart from the non-difference, low effect size and significant correlation, the comparison between AnT determined from both blood lactate concentration and heart rate resulted in a bias of ~8 and ~15W in the Bland-Altman analysis for non-experienced and experienced sub- jects, respectively. As discussed in the previous section, such bias can be related with the results of few subjects. For instance, the difference of AnTRLT and AnTRLT-HR for subjects 6 and 7 (i.e. Group 1) was ~7 and ~8W, respectively. Without this difference, the bias is reduced to ~6W which represents a difference of AnTRLT and AnTRLT-HR of 6% for 83% of evaluated non-expe- rienced subjects. Regarding the experienced individuals, the bias was ~15W. For this group, the heart rate behavior during the RLT reverse segment for one experienced subject was markedly different from the others (Fig 6). It is clear to notice the high decrement of heart rate mainly in the two last stages of the test. Such behavior can be related with a low efficiency in terms of readjust- ment of autonomic nervous system along with the sympathetic-parasympathetic regulation on the cardiovascular system [41]. Within this result, the zero tangent in x-axis of the polynomial adjustment was right shifted, overestimating the AnT via heart rate in 13% (180 W) when compared to the blood lactate concentration determination (159 W). This overestimation is also extended when considered the AnTInc-Biss of this subject (158W). Without this subject, the Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 16 / 20 bias of AnTRLT and AnTRLT-HR for the experienced subjects is reduced for ~10W, which means that for 91% of these individuals the difference may be 6%. Overall, we do not believe that differences invalidate the AnT determination via heart rate during the RLT. Every proposed test to determine the AnT has a bias. The major question in this issue is related within the magnitude of the error. It is consolidated that heart rate is a valu- able maker for monitoring training effects or even determine physiological phenomenon such as the AnT [17, 18, 41]. Our results corroborate with this perspective. However, every mea- surement must be carefully analyzed in order to identify factors that are culminating for dis- crepancies between methods, as demonstrated on Fig 6. Therefore, the results presented in this study strengthen the determination of AnT in the RLT via heart rate, which can a valid and useful tool. Conclusions In summary, the AnT determined from the RLT is consistent with those determined from the incremental protocol by means of the bissegmentation model. However, caution is required regarding the comparison of other physiological results (blood lactate concentration and heart rate at AnT), since protocols are different considering methodology, resulting in distinct results. Additionally, the previous experience in cycling exercise did not affect the AnT from such application, highlighting the robustness of this RLT for AnT determination. Moreover, we demonstrated the AnT determination using heart rate is a valid non-invasive alternative, and it’s also not a function of the previous experience in cycling exercise. However, for 83% and 91% of the evaluated non-experienced and experienced individuals (respectively), the dif- ference between the AnT determined on RLT by blood lactate concentration and heart rate was 6%. Therefore, future studies are necessary to investigate which factors can improve the agreement between these measures. Supporting information S1 File. Datasheet with all data used in the manuscript. Folder “IncrementalTestBiss” con- tains the data regarding the bissegmentation method. Moreover, this folder contains general Fig 6. a) Curve of the reverse lactate threshold test from subject 5 (Experienced group) considering the blood lactate concentration plotted against power. The second order polynomial adjustment was considered for the anaerobic threshold determination (AnTRLT); b) Curve of the reverse lactate threshold test from subject 5 (Experienced group) considering the heart rate plotted against power. The second order polynomial adjustment was considered for the anaerobic threshold determination (AnTRLT-HR). https://doi.org/10.1371/journal.pone.0194313.g006 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 17 / 20 information regarding the incremental protocol, such as the calculated and predicted maximal heart rate (HRmax-Incremental-calculated and HRmax-Incremental-predicted), time to exhaustion (Tex) and pulse oximetry (SpO2). Folder “IncrementalTestOBLA” contains the data regarding the fixed blood lactate concentration method considering 3 and 4 mmol.L-1. Folder “IncrementalTestDMAX” contains the data regarding the maximal perpendicular dis- tance method considering exponential and polynomial second order adjustments. Folder “ReverseLactateThresholdTest”contains the data regarding the Reverse Lactate Threshold Test. (XLSX) Acknowledgments We would like to thank the subjects for the participation on the procedures. Author Contributions Conceptualization: Leonardo Henrique Dalcheco Messias, Emanuel Elias Camolese Polisel, Fu´lvia Barros Manchado-Gobatto. Data curation: Leonardo Henrique Dalcheco Messias, Emanuel Elias Camolese Polisel, Fu´lvia Barros Manchado-Gobatto. Formal analysis: Leonardo Henrique Dalcheco Messias, Emanuel Elias Camolese Polisel, Fu´l- via Barros Manchado-Gobatto. Funding acquisition: Fu´lvia Barros Manchado-Gobatto. Investigation: Leonardo Henrique Dalcheco Messias, Emanuel Elias Camolese Polisel, Fu´lvia Barros Manchado-Gobatto. Methodology: Leonardo Henrique Dalcheco Messias, Emanuel Elias Camolese Polisel, Fu´lvia Barros Manchado-Gobatto. Project administration: Fu´lvia Barros Manchado-Gobatto. Resources: Fu´lvia Barros Manchado-Gobatto. Supervision: Leonardo Henrique Dalcheco Messias, Fu´lvia Barros Manchado-Gobatto. Validation: Leonardo Henrique Dalcheco Messias, Fu´lvia Barros Manchado-Gobatto. Visualization: Leonardo Henrique Dalcheco Messias, Fu´lvia Barros Manchado-Gobatto. Writing – original draft: Leonardo Henrique Dalcheco Messias, Emanuel Elias Camolese Polisel, Fu´lvia Barros Manchado-Gobatto. Writing – review & editing: Leonardo Henrique Dalcheco Messias, Emanuel Elias Camolese Polisel, Fu´lvia Barros Manchado-Gobatto. References 1. Wasserman K, McIlroy MB. Detecting the Threshold of Anaerobic Metabolism in Cardiac Patients dur- ing Exercise. Am J Cardiol. 1964; 14:844–852. PMID: 14232808 2. Heck H, Mader A, Hess G, Mucke S, Muller R, Hollmann W. Justification of the 4-mmol/l lactate thresh- old. Int J Sports Med. 1985; 6:117–130. https://doi.org/10.1055/s-2008-1025824 PMID: 4030186 3. Bentley DJ, Newell J, Bishop D. Incremental exercise test design and analysis: implications for perfor- mance diagnostics in endurance athletes. Sports Med. 2007; 37:575–586. PMID: 17595153 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 18 / 20 4. Bang O. The lactate content of blood during and after muscular exercise in men. Scand Arch Physiol. 1936; 74:51–82. 5. Beneke R. Anaerobic threshold, individual anaerobic threshold, and maximal lactate steady state in rowing. Med Sci Sports Exerc. 1995; 27:863–867. PMID: 7658947 6. Bosquet L, Leger L, Legros P. Methods to determine aerobic endurance. Sports Med. 2002; 32:675– 700. PMID: 12196030 7. Faude O, Kindermann W, Meyer T. Lactate threshold concepts: how valid are they? Sports Med. 2009; 39:469–490. https://doi.org/10.2165/00007256-200939060-00003 PMID: 19453206 8. Hinkley DV. Inference about the intersection in two-phase regression. Biometrika. 1969; 56:495–504. 9. Manchado-Gobatto FB, Vieira NA, Messias LHD, Ferrari HG, Borin JP, Andrade VC, et al. Anaerobic threshold and critical velocity parameters determined by specific tests of canoe slalom:Effects of moni- tored training. Sci Sport. 2014; 29:55–58. 10. Sjo¨din B, Jacobs I. Onset of blood lactate accumulation and marathon running performance. Int J Sports Med. 1981; 3:23–26. 11. Cheng B, Kuipers H, Snyder AC, Keizer HA, Jeukendrup A, Hesselink M. A new approach for the deter- mination of ventilatory and lactate thresholds. Int J Sports Med. 1992; 13:518–522. https://doi.org/10. 1055/s-2007-1021309 PMID: 1459746 12. Thomas V, Costes F, Chatagnon M, Pouilly JP, Busso T. A comparison of lactate indices during ramp exercise using modelling techniques and conventional methods. J Sports Sci. 2008; 26:1387–1395. https://doi.org/10.1080/02640410802104920 PMID: 18923955 13. Svedahl K, MacIntosh BR. Anaerobic threshold: the concept and methods of measurement. Can J Appl Physiol. 2003; 28:299–323. PMID: 12825337 14. Yeh MP, Gardner RM, Adams TD, Yanowitz FG, Crapo RO. "Anaerobic threshold": problems of deter- mination and validation. J Appl Physiol Respir Environ Exerc Physiol. 1983; 55:1178–1186. https://doi. org/10.1152/jappl.1983.55.4.1178 PMID: 6629951 15. Dotan R. Reverse lactate threshold: a novel single-session approach to reliable high-resolution estima- tion of the anaerobic threshold. Int J Sports Physiol Perform. 2012; 7:141–151. PMID: 22180336 16. Wahl P, Manunzio C, Vogt F, Strutt S, Volmary P, Bloch W, et al. Accuracy Of A Modified Lactate Mini- mum Test And Reverse Lactate Threshold Test To Determine Maximal Lactate Steady State. J Strength Cond Res. 2017; 31:3489–3496. https://doi.org/10.1519/JSC.0000000000001770 PMID: 28033123 17. Achten J, Jeukendrup AE. Heart rate monitoring: applications and limitations. Sports Med. 2003; 33:517–538. PMID: 12762827 18. Valenzano A, Moscatelli F, Triggiani AL, Capranica L, De Loannon G, Piacentini MF, et al. Heart-rate changes after an ultraendurance swim from Italy to Albania: a case report. Int J Sports Physiol Perform. 2016; 11:407–409. https://doi.org/10.1123/ijspp.2015-0035 PMID: 26263484 19. Stegmann H, Kindermann W, Schnabel A. Lactate kinetics and individual anaerobic threshold. Int J Sports Med. 1981; 2:160–165. https://doi.org/10.1055/s-2008-1034604 PMID: 7333753 20. Lucia A, Hoyos J, Santalla A, Perez M, Carvajal A, Chicharro JL. Lactic acidosis, potassium, and the heart rate deflection point in professional road cyclists. Br J Sports Med. 2002; 36:113–117. https://doi. org/10.1136/bjsm.36.2.113 PMID: 11916893 21. Strupler M, Mueller G, Perret C. Heart rate-based lactate minimum test: a reproducible method. Br J Sports Med. 2009; 43:432–436. https://doi.org/10.1136/bjsm.2006.032714 PMID: 18308890 22. Perret C, Labruyere R, Mueller G, Strupler M. Correlation of heart rate at lactate minimum and maximal lactate steady state in wheelchair-racing athletes. Spinal Cord. 2012; 50:33–36. https://doi.org/10.1038/ sc.2011.97 PMID: 21894166 23. Tegtbur U, Busse MW, Braumann KM. Estimation of an individual equilibrium between lactate produc- tion and catabolism during exercise. Med Sci Sports Exerc. 1993; 25:620–627. PMID: 8492691 24. Faul F, Erdfelder E, Buchner A, Lang AG. Statistical power analyses using G*Power 3.1: tests for corre- lation and regression analyses. Behav Res Methods. 2009; 41:1149–1160. https://doi.org/10.3758/ BRM.41.4.1149 PMID: 19897823 25. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986; 1:307–310. PMID: 2868172 26. Cohen J. The Earth Is Round. Am Psychol. 1994; 49:997–1003. 27. Hopkins W G. A spreadsheet for deriving a confidence interval, mechanistic inference and clinical infer- ence from a p value. Sports Sci. 2007; 11:16–20. 28. Marsh AP, Martin PE, Foley KO. Effect of cadence, cycling experience, and aerobic power on delta effi- ciency during cycling. Med Sci Sports Exerc. 2000; 32:1630–1634. PMID: 10994916 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 19 / 20 29. Gledhill N, Jamnik R. Determining power outputs for cycle ergometers with different sized flywheels. Med Sci Sports Exerc. 1995; 27:134–135. PMID: 7898329 30. Astrand I. Aerobic work capacity in men and women with special reference to age. Acta Physiol Scand Suppl. 1960; 49:1–92. 31. Essner A, Sjostrom R, Ahlgren E, Lindmark B. Validity and reliability of Polar(R) RS800CX heart rate monitor, measuring heart rate in dogs during standing position and at trot on a treadmill. Physiol Behav. 2013; 114–115:1–5. https://doi.org/10.1016/j.physbeh.2013.03.002 PMID: 23499770 32. Beneke R. Methodological aspects of maximal lactate steady state-implications for performance test- ing. Eur J Appl Physiol. 2003; 89:95–99. https://doi.org/10.1007/s00421-002-0783-1 PMID: 12627312 33. Coppin E, Heath EM, Bressel E, Wagner DR. Wingate anaerobic test reference values for male power athletes. Int J Sports Physiol Perform. 2012; 7:232–236. PMID: 22174179 34. Simon J, Young JL, Blood DK, Segal KR, Case RB, Gutin B. Plasma lactate and ventilation thresholds in trained and untrained cyclists. J Appl Physiol. 1986; 60:777–781. https://doi.org/10.1152/jappl.1986. 60.3.777 PMID: 3957830 35. Allen WK, Seals DR, Hurley BF, Ehsani AA, Hagberg JM. Lactate threshold and distance-running per- formance in young and older endurance athletes. J Appl Physiol. 1985; 58:1281–1284. https://doi.org/ 10.1152/jappl.1985.58.4.1281 PMID: 3988681 36. Donovan CM, Brooks GA. Endurance training affects lactate clearance, not lactate production. Am J Physiol. 1983; 244:E83–92. https://doi.org/10.1152/ajpendo.1983.244.1.E83 PMID: 6401405 37. Yoshida T, Suda Y, Takeuchi N. Endurance training regimen based upon arterial blood lactate: effects on anaerobic threshold. Eur J Appl Physiol Occup Physiol. 1982; 49:223–230. PMID: 6889499 38. Machado FA, Nakamura FY, Moraes SMF. Influence of regression model and incremental test protocol on the relationship between lactate threshold using the maximal-deviation method and performance in female runners. J Sport Sci. 2012; 30:1–8. 39. Alves JCC, Peserico CS, Nogueira GA, Machado FA. The influence of the regression model and final speed criteria on the reliability of lactate threshold determined by the Dmax method in endurance- trained runners. Appl Physiol Nutr Metab. 2016; 41:1039:1044. https://doi.org/10.1139/apnm-2016- 0075 PMID: 27628199 40. Conconi F, Ferrari M, Ziglio PG, Dorghetti P, Codeca L. Determination of the anaerobic threshold by a noninvasive field test in runners. J Appl Physiol. 1982; 52:869–873. https://doi.org/10.1152/jappl.1982. 52.4.869 PMID: 7085420 41. Bellenger CR, Fuller JT, Thomson RL, Davison K, Robertson EY, Buckley JD. Monitoring Athletic Train- ing Status Through Autonomic Heart Rate Regulation: A Systematic Review and Meta-Analysis. Sports Med. 2016; 46:1461–1486. https://doi.org/10.1007/s40279-016-0484-2 PMID: 26888648 42. Borresen J, Lambert MI. Autonomic control of heart rate during and after exercise: measurements and implications for monitoring training status. Sports Med. 2008; 38:633–646. PMID: 18620464 Analysis of the reverse lactate concept PLOS ONE | https://doi.org/10.1371/journal.pone.0194313 March 13, 2018 20 / 20

Advances of the reverse lactate threshold test: Non-invasive proposal based on heart rate and effect of previous cycling experience.

03-13-2018

Messias, Leonardo Henrique Dalcheco,Polisel, Emanuel Elias Camolese,Manchado-Gobatto, Fúlvia Barros

eng

PMC3409063

The Transeurope Footrace Project: longitudinal data acquisition in a cluster randomized mobile MRI observational cohort study on 44 endurance runners at a 64-stage 4,486km transcontinental ultramarathon Schütz et al. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 (19 July 2012) TECHNICAL ADVANCE Open Access The Transeurope Footrace Project: longitudinal data acquisition in a cluster randomized mobile MRI observational cohort study on 44 endurance runners at a 64-stage 4,486km transcontinental ultramarathon Uwe HW Schütz1,2*†, Arno Schmidt-Trucksäss3†, Beat Knechtle4, Jürgen Machann5, Heike Wiedelbach1, Martin Ehrhardt1, Wolfgang Freund1, Stefan Gröninger6, Horst Brunner1, Ingo Schulze7, Hans-Jürgen Brambs1 and Christian Billich1 Abstract Background: The TransEurope FootRace 2009 (TEFR09) was one of the longest transcontinental ultramarathons with an extreme endurance physical load of running nearly 4,500 km in 64 days. The aim of this study was to assess the wide spectrum of adaptive responses in humans regarding the different tissues, organs and functional systems being exposed to such chronic physical endurance load with limited time for regeneration and resulting negative energy balance. A detailed description of the TEFR project and its implemented measuring methods in relation to the hypotheses are presented. Methods: The most important research tool was a 1.5 Tesla magnetic resonance imaging (MRI) scanner mounted on a mobile unit following the ultra runners from stage to stage each day. Forty-four study volunteers (67% of the participants) were cluster randomized into two groups for MRI measurements (22 subjects each) according to the project protocol with its different research modules: musculoskeletal system, brain and pain perception, cardiovascular system, body composition, and oxidative stress and inflammation. Complementary to the diverse daily mobile MR-measurements on different topics (muscle and joint MRI, T2*-mapping of cartilage, MR- spectroscopy of muscles, functional MRI of the brain, cardiac and vascular cine MRI, whole body MRI) other methods were also used: ice-water pain test, psychometric questionnaires, bioelectrical impedance analysis (BIA), skinfold thickness and limb circumference measurements, daily urine samples, periodic blood samples and electrocardiograms (ECG). Results: Thirty volunteers (68%) reached the finish line at North Cape. The mean total race speed was 8.35 km/ hour. Finishers invested 552 hours in total. The completion rate for planned MRI investigations was more than 95%: 741 MR-examinations with 2,637 MRI sequences (more than 200,000 picture data), 5,720 urine samples, 244 blood samples, 205 ECG, 1,018 BIA, 539 anthropological measurements and 150 psychological questionnaires. Conclusions: This study demonstrates the feasibility of conducting a trial based centrally on mobile MR- measurements which were performed during ten weeks while crossing an entire continent. This article is the reference for contemporary result reports on the different scientific topics of the TEFR project, which may reveal * Correspondence: [email protected] † Contributed equally 1Department of Diagnostic and Interventional Radiology, University Hospital of Ulm, Germany Full list of author information is available at the end of the article Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 © 2012 Schütz et al; licensee BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. additional new knowledge on the physiological and pathological processes of the functional systems on the organ, cellular and sub-cellular level at the limits of stress and strain of the human body. Please see related articles: http://www.biomedcentral.com/1741-7015/10/76 and http://www.biomedcentral.com/ 1741-7015/10/77 Background Ultramarathon Various aspects of the physical characteristics of recrea- tional and elite level runners up to marathon distance events have been reported [1-9]. Much less has been written about the anthropometric characteristics of ultra endurance runners [10-14]. The case and field studies of Knechtle et al. developed a growing knowledge about the physical characteristics of multistage ultra endurance runners in the past years [15-22]. The German Ultra- marathon Association (DUV) defines foot-races of 50 km or longer as ultramarathons (UM). Multistage ultramara- thons (MSUM) are races in which each stage has a dis- tance of a UM. Besides a few case reports very little has been reported about the medical aspects of runners doing a transcontinental extended MSUM over several weeks [23]. Until now, there have been no reports pub- lished regarding UM running over more than 1,500 km. However, prolonged MSUM races offer the best opportu- nity to study physical adaptation and the associations of the physiological parameters of athletes in a longitudinal setting day by day. The race Among some very heroic solo runs, the TransEurope Foo- tRace 2009 [24] (TEFR09) was the 11th official transconti- nental competition multistage footrace within living memory (Table 1) [25-33]. This second European trans- continental MSUM took place from 19 April to 21 June 2009 from Bari, South Italy (41° 8’ N, 16° 52’ E) to the North Cape, Norway (71°10’N, 25°47’E) (Figure 1). Sixty- seven ultra endurance runners (mean age 50.7 years, range 26 to 74 years, male 56 (83.6%)) from 12 nations (Ger- many, Japan, Netherlands, France, Switzerland, Norway, Sweden, Finland, Turkey, South Korea, Taiwan, USA) met the challenge and tried to cross six countries (Italy, Aus- tria, Germany, Sweden, Finland, Norway). This comprised running 4,487 km (2,788 miles) in 64 stages without any day of rest. Thus, they expected to complete an average stage distance of 70.1 km, representing 1.7 marathon dis- tances (minimum: 44 km, maximum: 95.1 km) [32]. All participants organized their arrival at Bari on their own. Following breakfast at 5:00 a.m., the daily stage started at 6:00 a.m. The race director, together with his staff, planned the stages with their corresponding distances and ascent or descent and organized the accommodations for the runners in halls as well as the food for each stage. In addition, most of the runners carried individual nutri- tion on their own. Depending on the stage length, five to ten stop points for nutrition were placed on the daily routes. After each stage the runners had time on their own (nutrition, sleeping, regeneration). Depending on the stage length and local situation, dinner was served between 5:00 and 9:00 p.m. The runners slept in camping grounds (mainly in Italy), local sport halls or local commu- nity halls at the stage destinations (9:00 p.m. to 4:00 a.m.). Sometimes the quarters were crowded resulting in difficult sleeping conditions. In total, the runners had about 7 to 13 hours of rest per day for recuperation. The local sani- tary conditions also changed daily and from country to country. The project The TEFR09 project is the first observational cohort study intended to produce unique and comprehensive data from longitudinal measurements of a large sample of ultra endurance runners taking part in one of the most extreme multistage endurance competitions in the world, which takes the participants to a different level, where the race becomes a way of life, and where nutrition, sleep, energy and psychological states have to be carefully managed [34,35]. It is also the first study using a mobile MRI scan- ner for continuous examination of the athletes while per- forming a transcontinental MSUM. The aim was to explain the wide spectrum of adaptive responses in humans being exposed to such a chronic physical endur- ance load with negative energy balancing but without enough time for regeneration and to identify factors asso- ciated with inter-individual variation in these responses. Due to the unique possibility to observe morphological and physiological changes and the reactions of different tissues and functional systems on the systemic, organ and (sub)cellular level with modern MRI techniques, a wide range of multiple questions, hypotheses and unproven assumptions regarding injury, adaptation, regeneration, reparation and overuse processes arose. Therefore, four different project modules of investigations were created, focusing on multiple open questions and unproven hypotheses regarding long distance running: Project module I: musculoskeletal system Pre-race injuries and deformation It is inevitable that pre-race injuries or biomechanical deviations from the norm (for example, mal-alignment) Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 2 of 33 will lead to progressing focal damage of the lower extre- mities when performing a transcontinental MSUM. If any participant who suffers from unhealed pre-race inju- ries or deformation of the lower extremities reaches the finish line at North Cape without deterioration, this hypothesis can be rejected. Furthermore, the TEFR pro- ject tries to detect reasons for not finishing the race in detail. Joints Recent investigations indicate that running a marathon does not increase pathologies of structures of the knee joint [36-39]. Some authors postulate a risk factor of repeated marathon running for osteoarthritis of the knee [40]. However, a protective effect of running for knee joint cartilage is also discussed [41,42]. Nothing is known about the effect of ultra long distance running over weeks as in MSUM on knee structures. For the knee car- tilage, experimental studies on animals using quantitative microspectrophotometry and polarized light microscopy showed a decrease of the glycosaminoglycan content of the superficial femorotibial joint cartilage layer and degradation and reorganization of the superficial collagen network [43-45]. With this study, there are two hypoth- eses to be proven: first, in a MSUM over nine weeks, the well-trained participants show no increase in pathologies of the structures of the knee joint; and second, the carti- lage of the femorotibial joint shows a degradation of the glycosaminoglycan content with quick regeneration after the end of the race. The hypothesis that MSUM running does not indicate a higher risk of osteoarthritis of the knee in well-trained endurance runners has to be proven with the TEFR project. Contrary to the femorotibial joint the following hypothesis for the femoropatellar joint has to be proven: retropatellar cartilage degeneration is not caused by MSUM running. The femoropatellar joint is not a limiting factor for ultra endurance running, although degeneration arthrosis is already present in some MSUM participating athletes, respectively. This hypothesis is based on the fact that retropatellar arthrosis has a high prevalence in older people [46] and long Table 1 History of transcontinental footraces No race: date route total distance days, stages mean stage distance star -tera fin-ishera C.C. Pyle’s International Trans-continental Foot Races (Bunion Derbies 1928, 1929) [25-27] 1. 1928: 4 March - 24 May Los Angeles - New York 5,509 km 3,423 miles 84 65.6 km/d 40.8 miles/d 199 55 28% 2. 1929: 21 March - 8 June New York - Los Angeles 5,509 km 3,423 miles 84 65.6 km/d 40.8 miles/d 80 31 39% Trans America Foot Races 1992-95 [28,29] 3. 1992: 20 June - 22 August Huntington Beach - New York 4,722 km 2,935 miles 64 73.8 km/d 45.9 miles/d 28 13 46% 4. 1993: 19 June - Aug.21 Huntington Beach - New York 4,686 km 2,912 miles 64 73.8 km/d 45.9 miles/d 13 6 46% 5. 1994: June 18 - 20 August Huntington Beach - New York 4,708 km 2,926 miles 64 73.6 km/d 45.7 miles/d 14 5 36% 6. 1995: 17 June 17 - 19 August Huntington Beach - New York 4,676 km 2,906 miles 64 73.1 km/d 45.4 miles/d 14 10 71% Trans Australia Foot Race 2001 7. 2001: 6 January - 11 March Perth - Canberra 4,109 km 2,553 miles 63 67.8 km/d 42.1 miles/d 24 14 58% Run Across America 2002 8. 2002: 15 June - 24 August New York - Huntington Beach 4,961 km 3,084 miles 71 68.9 km/d 42.8 miles/d 11 8 73% 10. 2004: 15 June - 24 August Huntington Beach - New York 4,961 km 3,084 miles 71 68.9 km/d 42.8 miles/d 10 6 60% Trans Europe Foot Races 2003, 2009 [30-33] 9. 2003: 19 April - 21 June Lisbon, Portugal - Moscow 5,020 km 3,119 miles 64 79.5 km/d 49.4 miles/d 44 22 50% 11. 2009: 19 April - 21 June Bari, Italy - North Cape 4,486 km 2,787 miles 64 70.1 km/d 43.6 miles/d 67 46 69% Run Across America 2011: LANY (Los Angeles to New York) 12. 2011: 19 June - 27 August Huntington Beach - New York 5,157 km 3,205 miles 70 73.7 km/d 45.8 miles/d 14 8 57% atotal: 528 starters, 224 finishers. 42.4% (some of these have made more than one crossing) Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 3 of 33 distance running participation rises with age [47]. This TEFR project will also prove the same hypothesis for the joints of the ankle and hindfoot. For the latter no specific studies have been published until now. Soft tissues of the leg The literature on long distance running demonstrates that injuries mostly occur in the active and passive soft tissues of the lower extremities [48-54]. Injuries of the muscles, tendons and fascia of the lower legs are the most obvious limiting factors for performance and the most common reasons for not finishing a transconti- nental footrace. This postulation leads to the following hypothesis which has to be proven by the TEFR project: During a MSUM every participant will suffer from differ- ent injuries and overload of active and passive soft tissues of the lower musculoskeletal system, resulting in long lasting damage due to lack of recreation time. A special problem for endurance runners is a pain syn- drome of the lower leg, so-called ‘shin splint’. Investiga- tions over the past years, especially with MRI, showed that several entities caused by overuse of the lower leg due to long distance running can be differentiated. Some authors indicated that bone or periosteal reactions of the ventral tibia are a typical part of this syndrome [55,56], while others do not consider this as being mandatory [57]. Depending on the involved tissues and the lack of exact knowledge of the pathogenesis, the terminology regarding chronic lower leg pain in runners is broad and has not been differentiated and defined in detail until now: medial tibial stress syndrome, shin splint, anterior muscle syn- drome, (peri-)myositis, periostitis, fasciitis, and so on [58,59]. With our study design we try to prove the hypoth- esis that running-associated chronic lower leg pain includes different entities, such as overuse pathologies of muscles, fascias, tendons and bone tissues of the lower leg. The problem begins in the friction areas of the fascia of the muscles and the tendons (peritendineum) and then extends to other tissues such as the muscles, periosteum and bone if the running burden continues and the pain is ignored by the athlete. As many ultra athletes reported, it seems to be possible to overrun ‘shin splint’ without further damage. Whether medial tibial stress syndrome can end up in a stress fracture or a chronic exertional compartment syndrome when running is continued with- out any further rest is not clear [56,60,61]. Perhaps further understanding and differentiation of soft tissue overuse of the leg is possible due to this observational cohort study with a mobile MRI. Bones Stress fractures of the lower body often occur in people doing extensive walking or running without proper train- ing or adaption to the repeated and persistent mechanical burden their bones have to deal with. As seen in young soldiers doing their first long march with full body equip- ment [62] or amateur runners or beginners in (half-) marathons [63] our hypothesis is that even well trained ultra runners, such as the TEFR participants, can suffer stress fractures, because their skeletal system is not Figure 1 Route of Trans Europe Foot Race 2009 (4,486 km from south to north of Europe). Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 4 of 33 adapted to this tremendous mechanical burden that occurs while crossing a continent by foot with a speed of more than 6 km/hour without any day of rest. The TEFR project with mobile MRI tries to detect early signs of bone reaction (for example, subperiosteal new bone for- mation, adaptive cortical hypertrophy with little true inflammation, local bone edema or bruising) indicating overload as precursors of stress fractures [59]. For asymptomatic healthy marathon runners signifi- cantly higher red bone marrow hyperplasia has been observed compared to healthy volunteers [64]. This is pos- tulated to be a response to ‘sports anemia’, which is com- monly found in highly conditioned trained athletes. The TEFR project wanted to prove the hypothesis that during a MSUM lasting more than nine weeks an increase of red bone marrow occurs in the participants, even though they are adapted and well trained in ultra running. Project module II: brain, mind and pain perception Brain volume Aerobic exercise protects from insular atrophy in the brain of healthy volunteers [65,66]. The normal annual volume loss due to age-related brain atrophy is about 0.11% [66,67]. Increased atrophy is shown for several diseases, such as Alzheimers (2% per year) [68,69] anorexia nervosa [70,71] or malnutrition based on other reasons [72]. Some hypotheses explain this based on the influence of the stress hormone cortisol [70] but further pathophysiological prin- ciples leading to this decrease in brain volume are not understood [73-76]. Marathon-induced changes in endo- crine levels [77,78], in fluid balance [79] and amino acid blood level [80] are known to alter brain metabolism. The hypothesis that ultra endurance running leads to volume reduction of the brain cortex could be investigated by the TEFR project using high resolution cerebral MRI. Com- bined with specific laboratory analyses of different markers (for example S100B [81,82]) in blood and urine, reasons explaining the mechanism may be identified. On the other hand, when we postulate that an ultra long MSUM, such as the TEFR09, modifies the plastic brain (state marker) the hypothesis that the sensomotoric cortex volume - which is responsible for the lower extremities - will increase has also to be proven [83]. Brain lesions Additionally we hypothesize that the well-known exer- cise-induced hyponatremia due to inappropriate arginine vasopressin secretion which can lead to encephalopathy [84-87] is not seen in the highly endurance-trained par- ticipants of the TEFR09. If an MSUM leads to brain lesions, water sensitive cerebral MRI sequences of the TEFR project will show it. Pain perception, mind and mental stress All participants in the TEFR09 had previously finished an ultra marathon. This unique collective of endurance athletes is eminently suitable for examining the hypoth- esis that ultra runners have different mental prerequi- sites (higher auto suggestibility) compared to the normal population (trait marker). Experienced ultra endurance runners often mentioned that finishing an ultra race is more a matter of mind than a matter of the body. The hypothesis that finishers of MSUM differ from non- finishers with regard to pain suppression and willpower can be proven by pain tests combined with functional MRI of the brain and mental stress markers in serum samples. Project module III: cardiovascular system Heart The MSUM TEFR09 results in an extreme prolonged stress for the whole organism. Its effects on the heart can be discussed controversially. Cardiac dysfunction after marathon running is verified with biochemical markers and cardiac ultrasound [88,89]. Investigations with car- diac MRI are inconsistent; cardiac damage such as myo- cardial necrosis is seen in middle-aged marathon runners (57.2 +/- 5.7 years) [90], but not in younger marathon participants (30 to 50 years) [91,92]. Myocardial function disorders are reported with cardiac MRI tagging [93]. Until now, diagnostic analyses of long-lasting running effects on the heart using cardiac biomarkers are difficult to interpret, for example, the increase in brain natriuretic peptide (BNP) after 100 km trials [94,95] could show cytoprotective or growth regulatory effects [89,96] but also myocardial insufficiency. The TEFR project intends to prove the following hypotheses: Even in well trained ultra endurance runners, the running burden of a trans- continental MSUM of more than nine weeks induces a progressive cardiac distress and signs of cardiovascular mal-adaption. Cardiac MRI, stress laboratory tests and ECG might show signs of cardiac structural and func- tional restrictions, dysfunctions, damages and insuffi- ciency. Non-finishers of the TEFR09 will show more restrictive parameters than finishers. Another hypothesis contradicting the Morganroth hypothesis [97,98] is pro- ven: Although an aerobic endurance burden is per- formed, the TEFR09 finisher will show an increase in cardiac output and left ventricular mass index and in ventricular wall thickness, which might be caused by incomplete or critical left ventricular hypertrophy in some cases. Using a MR tagging technique, we will try to prove the hypothesis that the anatomical position of the heart is going to change (steep position) with a prolonged aerobic running burden [99]. Arteries Arteries of the muscular-type, such as the common femoral artery (CFA) adapt structure and function to endurance exercise training and elastic-type arteries, such as the common carotid artery, show functional Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 5 of 33 adaptations to increased endurance exercise training. Schmid-Trucksäss et al. [100] found an increase of dia- meter and compliance and a constant shear rate for CFA using noninvasive vascular ultrasound in highly endurance-trained athletes compared to sedentary con- trols. Other authors postulate that endurance exercise does change arterial wall stiffness and vascular (endothe- lial) function [101-103]. The TEFR project tries to prove the hypotheses that the MSUM burden results in cardi- ovascular adaptations in the form of an increase of arterial wall compliance in the lower extremities asso- ciated with an increase in the vessel lumen diameter of the femoral artery and of the central aortal pulse wave velocity even in well-trained ultra-endurance runners. This will result in a decrease of the peripheral arterial resistance leading to an increase in basic perfusion of the lower extremities. Project module IV: body composition Endurance exercise leads to a reduction of subcutaneous fatty tissue as demonstrated in several field studies [104-107]. It is well known that fat is the main energy-rich substrate for ultra endurance performance [105,107,108]. In contrast, muscle tissue provides lower energy, when being catabolized. A decrease in skeletal muscle mass due to ultra endurance performance has only been demon- strated in case reports [15,23,109] or small series [108]. Different effects for long-lasting or ultra endurance per- formances on body composition are described in the literature and seem to depend on the type of endurance burden. In ultra endurance performance with defined breaks (for example, during the night), body mass may remain stable [110-112] or even increase [105] while body fat is reduced [104,105,113], whereas skeletal muscle mass seems to be spared [111,113,114] or may even increase [104]. Ultra endurance performance over hours, days or weeks without a break, results in a decrease in body mass [23,107,109,115] in which body fat as well as skeletal mus- cle seems to decrease as a few case reports indicate [15,23,109]. With this cohort study we can prove the hypothesis that due to the immense negative energy bal- ance during a transcontinental MSUM, not only fat but also lean tissue is involved in catabolism, even in the leg muscles. With its mobile whole body MRI protocol, the TEFR project will be able to measure the different amount of mass loss in the different functional muscle units of the leg. We also intend to detect the microstructural and intracellular adaption processes of leg muscle tissue with modern MRI methods (MR-spectroscopy, diffusion and perfusion MR imaging). Purpose This report describes the design and conduct of the TEFR project. We report the pattern of chronic endurance running exposure, characteristics of the sub- ject groups and reasons for not finishing. Detailed descriptions of measuring methods in relation to the hypotheses and technical challenges encountered in the realization of the interdisciplinary TEFR project are pre- sented. We discuss the strengths and limitations of the study setting. Methods After commitment to funding by the German Research Society (DFG) the 67 TEFR09 participants were asked to join the TEFR project, which was approved by the local ethics committee of the University Hospital of Ulm (UHU, No.: 270/08-UBB/se), Germany (in accordance with the Declaration of Helsinki) regarding the study design, risk management plan and individual protocols. Verbal and written informed consent was obtained from all concurring subjects. Mobile MRI The most important research tool was a 1.5 Tesla whole- body MR imager (Magnetom Avanto™ mobile MRI 02.05, software version: Syngo™ MR B15, Siemens Ltd., Erlan- gen, Germany) mounted on a mobile unit (MRI-Trailer Model Mob.MRI 02.05, SMIT Mobile Equipment B.V., Division AK Specialty Vehicles, Farnham, UK) pulled by a specially hired truck tractor. The semi-trailer had an inter- nal diesel generator to power the helium cooling circuit for the MRI over the ten-week period. However, it did not generate enough electricity for continuous MRI measure- ments and was therefore supplemented by a more power- ful custom made external diesel generator (150 KVA, Strom Rent™ e.k., Dortmund, Germany) which was pulled by an additional material van. The mobile hardware had a total weight of more than 45 tonnes and was nearly 30 meters long. All of the equipment was installed daily at each stopover and required daily checks and support of all technical systems (Figure 2). Study participants Forty-four (67%) of the race participants (mean age 49.7 years, range 26 to 68 years, male 40 (90.9%), f 4) were recruited for the TEFR project. The inclusion criterion obviously was an official acceptance as a participant at the TEFR09 by the organizers and the race director. The conditions of participation were: minimum age 18 years, the presence of a medical certificate not older than 30 days which indicated physical health and clear proof of appropriate running performance in the field of UM. The specific running history and performance of the indi- vidual subjects can be described by different traits, which were requested before the start of the TEFR09: years of regular endurance running training, finished (ultra-) marathons, personal best times in different defined ultra Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 6 of 33 races and extent of training (volume, duration, intensity) before TEFR09. The investigators additionally performed a resting car- diovascular check using a 12-channel PC-ECG system (Custo cardio 100™, Custo Med Ltd., Ottobrunn, Ger- many) and blood pressure (RR) measurement using a manual sphygmomanometer (BOSO Clinicus, Jungin- genGermany: to the nearest 3 mmHg). Cardiovascular exclusion criteria were resting blood pressure > 200 mmHg systolic and/or > 110 mmHg diastolic, acute sys- temic infection, acute chest pain and new arrhythmias or ECG changes. An orthopedic physical examination was done focusing on contraindications for endurance running such as relevant malalignment and painful joint diseases of the lower extremities. Additional specific exclusion criteria were contraindications against MRI scanning (for example, metallic foreign bodies in dan- gerous locations, specific cochlear or ocular implants, ferromagnetic vascular clips and relevant claustropho- bia). None of the volunteers had to be excluded from study participation due to these criteria. Investigators Four members of the TEFR project comprised the inves- tigator core team that accompanied the TEFR09 for direct data acquisition before and during the race: two physicians, one medical student and one radiological assistant. The latter (HW) was responsible for subject positioning in the scanner and performance of the MR examinations. One of the investigators, the initiator and main organizer of the TEFR project (US), drove the MR-trailer truck, adapted the daily research program to the actual circumstances, controlled and checked the quality of the MR examinations and was responsible for the technical readiness of the whole mobile MRI and its functional circuits and equipment with external and internal diesel generator. Being specialized in radiology and orthopedic surgery, he also did the initial and fol- low-up physical musculoskeletal examinations of the subjects. The second investigator (CB) was responsible for acquisition of daily anthropometric, laboratory and ECG data. The medical student (ME) made the daily anthropological measurements. The two physicians were solely responsible for the study and gave neither training advice nor provided medical help. Study design The study design of the TEFR project is shown in Figure 3. Pre-race Baseline studies were performed within the last four days before the start of the TEFR09 in Bari on every subject. They included group specific MRI examinations and anthropometric and cardiovascular physical mea- surements with urine and venous blood samples. Additionally, body height measuring using a wall- mounted stadiometer (to the nearest 5 mm, standing barefoot) and active range of motion measurement (AROM) of hip and knee joints using a manual double- armed universal goniometer (to the nearest 5°) were done before the start. One experienced orthopedic sur- geon, trained in a standardized procedure for position- ing both the subject and the goniometer, collected these data. Figure 2 Truck trailer with mobile MRI and external generator in working position. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 7 of 33 Adapted from the methods of Paley et al. [116] and Weidelich et al. [117], analysis of lower limb alignment was done on coronal lower body scout views of pre-race MRI with subjects in the supine position and with extended legs. Measured parameters were: leg length (LL), as the straight line from the middle of the femoral head to the midpoint of the upper talus rim; femoroti- bial angle (FTA), as the angle between anatomical femoral and tibial axis; the mechanical axis deviation (MAD) as the distance from the point of intersection between the perpendicular and mechanical axis of the limb (straight line from the middle of the femoral head to the midpoint of the upper talus rim) to the midpoint of the knee (medial tibial eminence) and the femoral to tibial length ratio (F/T) [116-122]. A 240-item, 31-dimensional personality temperament and character inventory (TCI) [123,124] in addition to a 10-item, 4-scaled questionnaire on self expectancy (Gen- eral Self-Efficacy Scale, GSE) [125,126] were also inte- grated into the project before its start. Additionally, 15 of the 44 subjects had an initial sepa- rate pre-race test on pain perception (ice-water test) combined with functional cerebral MRI two weeks before the start of the TEFR09 at UHU on 1 to 3 April 2009, because these examinations could not be imple- mented on the mobile MRI due to technical limitations. Due to the exorbitant physical and mental burden placed on the subjects, there was no opportunity for field experiments, invasive tests or application of psy- chometric instruments during the transcontinental foot- race. Field studies The observant field studies during the TEFR09 were completed between 15 April and 21 June 2009. Every morning from 3:45 to 4:30 a.m. urine samples and anthropometric measurements were taken. The core team broke down their examination units and drove to the next stage destination. Before stage length depen- dent arrival of the first runner they had set their systems ready and had refuelled the generators and vehicles. MRI examination time was between 2:30 p.m. +/- 90 minutes and 9:00 p.m.). At the same time anthropo- metric and cardiovascular physical measurements, blood and urine samples were collected and ECG was done. The daily data acquisition also included measurement Figure 3 Study design of TEFR-project. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 8 of 33 and documentation of daily weather conditions (tem- perature outside and inside, humidity outside) using a calibrated electronic thermometer, the stage length and the individual stage performances of the runners (stage running time). MR measurements For MR measurements two groups (22 subjects each) were cluster randomized according to the different research modules. The MR protocols were created in an interdisciplinary content ensuring multifold specific and diverse but precise analyses and measurements for detailed testing of the mentioned hypotheses concerning long distance running (Table 2). MRI of feet For high resolution investigation of the whole foot a special table fixed boot-like designed 8-channel foot- ankle coil was chosen and a sagittal orientated water sensitive T2w MR sequence (TIRM) configured a wide field of view. If on this sequence any pathology was detected, a transversal oriented focused water sensitive sequence with a more structured T2 sequence (fat satu- rated proton density weighted (PDw)) was added. For investigation of the joint cartilage a specific T2* map- ping MR sequence (syngo™ MapIt FLASH T2*w GRE) in sagittal orientation was used [127-129], allowing quantitative measurement of hydrophilic changes in the cartilage layers of tibiotalar, talocalcaneal, calcaneocu- boid, and calcaneonavicular joints. The specification of these MR sequences (Table 2) was done for detection of typical running associated overuse injuries of the feet [52]: subcutaneous edema, Achilles tendonitis [49,50], extensor digitorum tendonitis [48,49]), plantar fasciitis [50], calcaneal apophysitis, arthritis/arthrosis, stress frac- tures, bone edema, metatarsalgia, Morton’s neuroma, and ankle inversion injuries (Figure 4). MRI of knees With a table-fixed 8-channel knee coil all subjects of group 1 had both knees examined with a sagittal TIRM sequence for water detection in knee-related tissues and evaluation of femorotibial joint. A transversal fat satu- rated PDw sequence was used to assess the femoropatel- lar joint. As for the hindfoot joints, specific T2* mapping MR sequences in sagittal and transversal orientations were done for quantification of cartilage layers of the femoropatellar and femorotibial joints regarding intra- chondral water proportioning [127-129]. The specification of these MR sequences (Table 2) was done to evaluate running-associated overuse injuries in the knees [52]: patella tendonitis (‘runner’s knee’), arthritis/arthrosis [130], stress fracture, bone edema [64], retropatellar pain syndrome [48-50], chondromalacia patellae, meniscal lesions [50] and patellar tendinitis [50] (Figure 5). MRI of hips/pelvis One flexible 6-channel body matrix coil was used to obtain an MR overview of the pelvis with one coronal water sensitive sequence (TIRM: Table 2) to detect injuries in this part of the body: hip arthritis/arthrosis [131], sacroiliac injuries [52], stress fractures of the pel- vic ring [132-134], muscle overuse injuries and so on. Additional case specific sequences were added as necessary (Figure 6). MRI of upper/lower legs With three to four flexible 6-channel body matrix coils total MR examination of upper and lower legs was pos- sible. To get detailed information about soft tissue edema, muscle perfusion and injuries of the legs differ- ent sequences were adapted in transversal orientation (T1w for adipose tissue separation and acute bleeding detection, TIRM for high sensitivity in water detection, fat saturated PDw for structural detailed water sensitive imaging, DWI for perfusion analysis of muscles and separation between intra- and extra-cellular water in the muscles: Table 2). With these sequences all of the typi- cal running-associated syndromes could be detected and differential diagnosis done [49,52]: anterior compart- ment pain/syndrome [48], (medial) tibial stress syn- drome [50,135,136], gastrocnemius injuries, peroneal tendonitis, tibialis posterior injury, calcaneal apophysitis, iliotibial band friction syndrome [50], greater trochan- teric bursitis, gluteus medius - hamstring - adductor - abductor - quadriceps injuries, such as tendonitis, strains or tears. Muscle volumetry of different compart- ments of the upper and lower leg muscles is possible for evaluation of changes in muscle volume: Figure 7. Cerebral MRI, functional MRI As for the muscles in the legs, a MRI guided volumetric analysis of the brain was one focus of the cerebral MRI measurements. Therefore, a T1 weighted high resolution (1 mm) turbo FLASH three-dimensional-sequence was used, making an isovoxel based volumetry (VBM) possi- ble (Figure 8A). For detection of brain lesions and global edema a typical T2-sensitive sequence (FLAIR) in coro- nal orientation was chosen (Figure 8B). With diffusion weighted imaging (DWI), ischemia detection was possi- ble. For all these MR sequences (Table 2) a table inte- grated 12-channel head matrix coil with a head restraint system was used. The same coil was used on the sta- tionary scanner for functional MRI (fMRI) using echo- planar imaging (epi) with blood oxygenation level dependent (BOLD) contrast to analyze pain perception Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 9 of 33 Table 2 MRI protocols of the TEFR project MRI in module I: Musculoskeletal System ankle/foot (PP: FFS, supine): • TIRM 2D sag; PM: FA 140, TE 60; TR 11320, IR 120, ST 2, SBS 2.4, FOV 900, MS 512*512, PS 0.586 iso, PB 130; IAT: 5:37 • PDw TSE fs 2D tra; PM: FA 150, TE 32, TR 5830, ST 4, SBS 4.4, FOV 256, MS 384*384, PS 0.4167 iso, PB: 150; IAT: 3:46 • syngo™ MapIt FLASH 2D sag: T2*w GRE; PM: FA 60, TE 4.5, TR 1010, ST 2.5, SBS 2.75, FOV 182.25, MS 320*320; PS 0.421875 iso, PB: 250, IN: 12; IAT: 4:15 knee (PP: FFS, supine): • TIRM 2D sag; PM: FA 140, TE 50, TR: 4010, TI 150, ST 3, SBS 3.3, FOV 289, MS 256*256, PS 0.664 iso, PB 180; IAT: 3:31 • PDw TSE fs 2D cor; PM: FA 150, TE 31, TR 4400, ST 3, SBS 3.3, FOV 289, MS 512*512, PS 0.332 iso, PB 100; IAT: 4:11 • syngo™ MapIt FLASH 2D tra/sag: T2*w GRE; PM: FA 60, TE 4.18, TR 889/1120, ST 3, SBS 3.3, FOV 289, MS 512*512; PS 0.332 iso, PB: 250, IN: 12; IAT: 4:21/3:58 Hip (PP: FFS): • TIRM 2D cor; PM: FA 150, TE 61, TR 6230, TI 145, ST 3.5, SBS 3.85, FOV 1444, MS 384*384, PS 0.9896 iso, PB 130; IAT: 03:38 upper/lower leg (PP: FFS, supine): • T1w SE 2D tra, PM: FA 90, TE 13, TR 626, ST 5, SBS 5, FOV 1050/722, MS 512*336/256, PS 0.78125/0.7422 iso, PB 115; IAT: 1:54/1:30 • TIRM 2D tra; PM: FA 140, TE 62; TR 12530, TI 130, ST 3, SBS 3.9, FOV 512, MS 384*192, PS 0.833 iso, PB 180; IAT: 2:08 • PDw TSE fs 2D tra; PM: FA 150, TE 39, TR 6730, ST 3, SBS 3.9, FOV 512, MS 320*160, PS 1.0 iso, PB: 150; IAT: 2:14 • DWI (with ADC): SPAIR epi b-value 0-800, 2D tra; PM: FA 90, TE 75, TR 5100, ST 10, SBS 10, FOV 1300/1173.25, MS 128*104, PS 3.125/2.96875 iso, PB 1030; IAT: 2:32/3:01 MRI in module II: Brain and Pain Brain (PP: HFS, supine): • (turbo) FLASH 3D sag: T1 mpr; PM: FA 15, TE 4.75, TR 2100, ST 1, FOV 614.4, MS 240*256, PS 1.0 iso, PB130; IAT: 8:37 • T2w fs FLAIR 2D cor: TIRM; PM: FA 150, TE 120, TR 9000, TI 2500, ST 5, SBS 5.5, FOV, MS 288*384, PS 0.599 iso, PB 150; IAT: 4:43 • DWI (with ADC): SPAIR epi b-value 0-1000, 2D tra; PM: FA 90, TE 98, TR 3700, ST 5, SBS 6, FOV 529, MS 256*256, PS 0, 89844 iso, PB 1000; IAT: 0:49 functional MRI (fMRI) for pain perception: Epi 2d: epi2d_bold; PM: Fa 90, TE 60, TR 2600, ST 5, SBS 6, FOV, MS 384*384 (Start fMRI: 16*16), PS 3.59375, PB 2440; IAT: ~15:20 in total MRI in module III: Cardiovascular System Cardiac cine-MRI (PP: HFS, supine), IAT: ~25:00 in total • Cine SSFP, 2D: GRE cine with retrospective 2d cardiac triggering; PM: FA 80, TE var, TR var, ST 6, SBS 6, FOV 1156/1089, MS 192*192/156, PS 1.771/ 1.71875 iso, PB 930, IN: 30 • Phase contrast acquisition 2D; PM: venc 150, FA 30, TE 2.33var, TR 41.1var, ST 6, FOV, MS 180*192, PS 1.875 iso, PB 555, IN:25 • Cine-tagging SSFP, 2D: GRE cine with retrospective 2d cardiac triggering; PM: FA 20, TE var, TR var, ST 6, SBS 18, FOV 1073.25, MS 212*256, PS 1.40625 iso, PB 500, IN: 21 Vascular cine-MRI (PP: HFS, supine), IAT: ~25:00 var in total • Carotid artery: FLASH 2D tra: T2*w gradient-spoiled GRE cine with prospective 2d cardiac triggering; PM: FA 15, TE ~5.45 var, TR ~34.75 var, ST 6, FOV 289, MS 320 × 320, PS 0.53125 iso, PB 250, IN: 50/RR-cycle; IAT: ~5:30 var • Femoral artery: FLASH 2D tra: see carotid artery; PM: FA 15, TE 5.00 var, TR 26.80 var, ST 6, FOV 768, MS 512 × 384, PS 0.625 iso, PB 250, IN: 50/ RR-cycle; IAT: ~4:20 var • Aortic flow prox./dist.: FLASH 2D tra: PM: FA 20, TE 2.75 var, TR 11.55 var, ST 5, FOV 768, MS 256*192, PS 1.25 iso, PB 590, IN: 100; IAT: ~4:10 MRI in module IV: Morphometry, Body Composition whole body MRI (PP: HFP/FFP, prone): T1w TSE 2D tra; PM: FA 180, TE 12, TR 490, ST 10, SBS 20, FOV 1991, MS 256*196, PS 1.9922 iso, PB 120, IN: 90-120 var; IAT: appr. 20:00 MR spectroscopy (PP: FFS, supine) for evaluation of intramyocellular lipids (IMCL): Single-voxel STEAM, TE 20, TR 2000, voxel of interest 11 × 11 × 20 mm3, 40 acq., IAT: appr. 10:00 Tra, transversal (axial); sag, sagittal; cor, coronal; Seq., sequence; ADC, apparent diffusion coefficient; bold, blood oxygenation level dependent contrast; DWI, diffusion weighted imaging; epi, echo planar imaging; FLAIR, fluid-attenuated inversion recovery; FLASH, fast low angle shot; fs, fat saturated; GRE, gradient echo seq.; IR, inversion recovery; mpr, MPRAGE (magnetization prepared rapid gradient echo); PDw, proton density weighting; SPAIR, spectral adiabatic inversion recovery; SSFP, steady state free precision; STEAM, stimulated-echo acquisition mode; T1w, T1 contrast; T2w, T2 contrast; TIRM, turbo inversion recovery magnitude; (T)SE, (turbo) spin echo; PM, parameters; parameter abbreviations: FA, flip angle (°); TE, time to echo (ms); TR, repetition time (ms); TI, inversion time (ms); ST, slice thickness (mm), SBS, spacing between slices (mm); FOV, field of view (cm2); MS, matrix size (pixel); PS, pixel size (mm); PB, pixel bandwith; IN, image number (n); var, variable dependency; venc, velocity encoded gradient echo imaging (cm/second); IAT, image acquisition time (min:second); PP, patient positioning; HFS, head forward supine; FFS, feet forward supine; HFP, head forward prone; FFP, feet forward prone; cardiac triggering, ECG- gating. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 10 of 33 in 12 participants of the TEFR09 compared to age- related normal volunteers (Figure 8C). Cardiac cine MRI For mobile cardiac cine MRI, a flexible six-channel body matrix coil was used. Cine SSFP gradient echo sequences with retrospective cardiac triggering were generated to obtain plane short axis four-, three- and two-chamber (Figure 9B) views of the heart. The mitral and aortic flow (Figure 9D) was measured using phase contrast sequences with 150 cm/second velocity encoded gradient echo imaging (venc). This protocol ensured measurement or secondary evaluation of parameters, such as ejection fraction (%), end diastolic and systolic volume and, there- fore, stroke volume (ml), cardiac output (L/minute), myocardial mass (g) (Figure 9E), muscle volume of ven- tricles (ml) (Figure 9A), and so on. MR tagging using a Cine SSFP gradient echo sequence with retrospective Figure 4 Mobile MRI of the feet. A: male, 39-years-old, stage 53, 3,669 km (fused colored T2* GRE map sagittal: syngo™ MapIt fusion technique, Siemens Medical solutions, Erlangen, Germany): Colored visualization of focal water concentration in cartilage layers of the ankle joint (1), subtalar joint (2), talonavicular joint (3) and calcneocuboid joint (4). B: male, 61-years-old, stage 23, 2,176 km (T2 TIRM sagittal): Multiple running related signs of tissue alteration: Peritendinous fluid accumulation in tendon sheaths (1). Focal bone edema at insertion of plantar fascia (2) with corresponding subcutaneous edema (3), but without plantar fasciitis. Edema in the Achilles tendon (4). Articular effusion in the ankle joint (5). C: male, 59-years-old, stage 45, 3,082 km (T2 TIRM sagittal): Peritendinous fluid accumulation in tendon sheaths of foot dorsiflexors (1), focal arthrosis of the ankle joint with subchondral bone edema (2), plantar subcutaneous edema (3), articular effusion in subtalar and metatarsophalangeal joint (4). D: male, 54-years-old, stage 32, 2,176 km (T2 TIRM sagittal): Subcutaneous edema (1), arthrosis of midfoot joints (2). E: male, 30-years-old, stage 12, 789 km (PDw TSE fs transversal): Peritendinous fluid accumulation in Achilles tendon sheath (1) with wide local subcutaneous edema (2). F: female, 68-years old, stage 15, 1,003 km (PDw TSE fs transversal): Extensive plantar subcutaneous edema (1). G: male, 47-years-old, stage 52, 3,609 km (PDw TSE fs transversal): Severe arthrosis of midfoot joints (1) with perifocal bone edema (2), hallux valgus (3). Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 11 of 33 Figure 5 Mobile MRI of the knees. Femoral condyle (1), tibial head (2), patella (3), retropetallar cartilage layer (4), ventral femoral cartilage layer (5), dorsal femoral cartilage layer (6), tibial cartilage layer (7), medial meniscus (8), lateral meniscus (9), patellar tendon (10). A: female, 45-years-old, stage 44, 3,021 km (fused colored T2* GRE map: syngo™ MapIt fusion technique, A1: medial femorotibial joint sagittal, A2: lateral femorotibila joint sagittal, A3: femoropatellar joint transversal). B: male, 25-years-old, before start of TEFR09 Cartilage layer segmentation (T2* GRE map: syngo™ MapIt). C: male, 43-years-old, stage 40, 2,738 km (PDw TSE fs, C1: sagittal, C2: coronal, C3: transversal): Severe arthrosis of the patellofemoral joint with retropatellar cartilage defects (4, 5) and wide subchondral bone edema of the patella (3), intrachondral signal alterations of the femoral (6) and tibial (7) cartilage layers. D: male, 26-years-old, stage 40, 2,738 km (PDw TSE fs coronal): Nondescript cartilage layers of femorotibial joint. Figure 6 Mobile MRI of pelvis region and hip joints. A: male, 45-years-old, stage 8, 511 km (PDw TSE fs coronal): Massive edema and (peri-) myositis of right proximal quadriceps muscle (1). B: female, 46-years-old, stage 48, 3,290 km (PDw TSE fs, B1: coronal, B2,3: transversal): Stress fracture of left ventral pelvic ring (1: Ramus superior ossis pubis; 2: Ramus inferior ossis pubis) with perifocal soft tissue edema/inflammation (3). C: male, 41-years old, stage 11, 739 km (T1 TSE coronal): Arthrosis of the left hip (1: acetabular sclerosis, 2: deformation and osteophyte of the femoral head). D: male, 61-years-old, stage 38, 2601 km (PDw TSE fs transversal): Intraosseus edema in the right Ala ossis ileum (1) with massive peri-osseal inflammation of the gluteal muscle origin (2). Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 12 of 33 cardiac triggering in plane short axis four- and two- chamber view (Table 2) made quantification of the myo- cardial motion with its spatial orientation (Figure 9C) possible. Vascular cine MRI For analysis of changes in the arterial aortic stiffness, measurement of the central pulse wave velocity using MRI is the gold standard [137]. With detection and measurement of the proximal and distal aortic flow and diameter using phase contrast acquisition with venc and prospective two-dimensional cardiac triggering on mobile MRI (Figure 10B, C), this and the central hemodynamic changes (peak and mean shear rate differences) and their influence on the vascular (aortic) diameter [100] during the TEFR09 can be calculated. Additionally, T2 weighted cine FLASH gradient echo sequences with prospective two-dimensional cardiac triggering were generated Figure 7 Mobile MRI of upper and lower legs. A: male, 49-years-old (coronal slices): A1: stage 12, 789 km, upper legs (PDw TSE fs): Subfascial intermuscular fluid, superficial (1), deep peri-neural (2). Partial quadriceps tear (M. vastus intermedius: 3). A2: stage 19, 1,260 km, upper legs (PDw TSE fs): Subfascial intermuscular fluid, superficial (1), deep peri-neural (2) and peri-vascular (3). Partial muscle edema of M. vastus intermedius (4). Specific diffusion weight imaging (A3, same slice as A2) is a sensitive method for free water detection. A4: stage 19, 1,260 km, lower legs (T2 TIRM): Subfascial intermuscular (1) and epifascial subcutaneous edema (2) indicating soft tissue inflammation such as perimyositis and panniculitis (shin splints), respectively. B: male, 31-years-old, B1: start, B2: stage 62, 6,358 km (PDw TSE fs transversal): Segmentation of muscle compartments of upper leg for functional muscle volumetry. C: male, 53-years-old: C1: start, C2: stage 46, 3,161 km (T2 TIRM transversal): Segmentation of muscle compartments of lower leg for functional muscle volumetry. Muscle edema in calf muscle (1). Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 13 of 33 Figure 8 MRI of the brain. Male, 52-years-old. A1: stage 57, 3,971 km (T1w turbo FLASH 3D sagittal): High resolution isometric three- dimensional sequence (1 mm) allows isovoxel based volumetry (VBM) of the whole brain. A2: Three-dimensional view shows dominant areas of volume loss (colored) of grey brain matter occurring during the TEFR09. B: stage 36, 2,448 km (T2w FLAIR): Sensitive sequence for detection of brain lesions. In this case, no lesions visible. C: 20 days before the start. Functional MRI (fMRI) using blood oxygenation level dependent (BOLD) contrast for evaluation of pain perception in ultra runners (C1: without pain stimulus, C2: with pain stimulus). C3: Post-processing analysis using statistical parametric mapping (SPM) shows areas of activation. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 14 of 33 (Table 2) to measure compliance changes of the vessel wall of the distal common carotid (Figure 10A) and prox- imal superficial femoral artery (Figure 10D). In total, for vascular MRI three flexible six-channel body matrix coils for aortic and femoral artery measurements, one four- channel phased dual mode neck matrix coil and ECG triggering makes positioning and preparation of the sub- jects very time consuming. Figure 9 Mobile cardiac cine-MRI. A: male, 52-years-old, stage 23, 1,569 km (cine SSFP GRE, 2-chamber view): Focus is the left ventricle (1), myocardium (2). papillary muscles (3), right ventricle (4), lung (5), liver (6), left kidney (7), spleen (8), stomach (9). A2, A3: Specific post-processing makes functional volumetry of left ventricle and myocardium possible (green line: epicardium, red line: endocardium). B: male, 52-years-old, stage 25, 1,706 km (cine SSFP GRE, 3-chamber view): left ventricle (1), myocardium (2), papillary muscles (3), left atrium (4), lung (5), mitral valve (6), aortic valve (7), pulmonary vein (8), aorta (9), thoracic spine (10). C: male, 49-years-old, stage 26, 1,770 km (cine tagging SSFP GRE, four-chamber view): MR tagging of the left ventricle (1) makes quantification of the myocardial (2) motion with its spatial orientation possible. D: female, 45-years-old, stage 38, 2,601 km (phase contrast transversal): Ascending aortic (1) flow is measured by specific velocity-encoded (venc) MR imaging. Descending aorta (2), pulmonary artery (3), liver (4), lung (5). E: Selection of possible cardiac parameters measurable by cardiac cine-MRI. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 15 of 33 Whole body MRI For total body MRI, change of subject positioning from prone head forward to prone feet forward was necessary during a T1 weighted turbo spin echo sequencing using an adapted protocol developed on adipose and diabetic volunteers [138] (Table 2). With topographic tissue seg- mentation and mapping of the athlete’s body using a fuzzy c-means algorithm according to Würslin et al. [139] a simple and time-saving strategy for assessment and standardization of the tissue distribution in the entire body was possible. With additional manual adaption due to the non-fasting condition of the subjects changes in different lean and adipose body compartments could be measured during the TEFR09 (Figure 11). Figure 10 Mobile vascular cine-MRI: male, 52-years-old, stage 27, 1,838 km. A1: MR localizer sagittal, A2,3: FLASH transversal: Automatic functional measurement of common right carotid artery diameter just below carotid bifurcation (1). Left carotid artery (2), right deep jugular vein (3). B1: MR localizer, B2: phase contrast transversal: Ascending aortic (1) diameter and flow is measured by specific velocity-encoded (venc) MR imaging. Descending aorta (2), pulmonary artery (3), lung (4). B3: graphic depiction of aortic pulsatile flow (ml/second). C1: MR localizer coronal, C2: phase contrast transversal: Distal descending aortic (1) diameter and flow is measured by specific velocity-encoded (venc) MR imaging just above the aortic bifurcation. Inferior vena cava (2), liver (3), intestines (4). D1: FLASH transversal: Functional measurement of superficial right femoral artery diameter just below bifurcation (1), left femoral artery (2). D2: Manual diameter measurement (1), D3: Automatic diameter measurement (1). Femoral veins (2). Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 16 of 33 Figure 11 Semiautomatic tissue separation with mobile whole body MRI of a 32-year-old male finisher of the TEFR09. A: Right row (before start): green: total lean tissue, red: somatic adipose soft tissue, yellow: visceral adipose tissue, blue: adipose bone marrow. Left row (after 4,120 km): green: total lean tissue, red: somatic adipose tissue (= somatic adipose soft tissue + adipose bone marrow), yellow: visceral adipose tissue. (selected slices: I: ankles, II: middle of lower legs, III: knees, IV: middle of upper legs, V: hip/pelvis, VI: umbilical level, VII: upper abdomen, VIII: heart/mediastinum, IX: shoulder girth, X: elbows). B: Right row (before start): green: somatic lean tissue, red: somatic adipose tissue, grey: total visceral volume. Left row (after 4,120 km): green: total lean tissue, red: somatic adipose tissue (= somatic adipose soft tissue + adipose bone marrow), yellow: visceral adipose tissue, blue: intraluminal nutrition fat in intestinal tract. (selected slices: V: hip/pelvis, VI: umbilical level, VII: upper abdomen, VIII: heart/mediastinum). C: Loss of total lean and total adipose tissue during the TEFR09. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 17 of 33 MR-spectroscopy Proton MR-spectroscopy with a flexible six-channel body matrix coil for measurement of the intramyocellular lipid (IMCL) content of the tibialis anterior and soleus muscle required the stimulated-echo acquisition mode (STEAM) technique (Table 2) and manual shimming of the magnetic field [140], which makes generation of valuable results on a mobile MRI difficult and unpredictable (Figure 12). Focused supplementary sequences In addition to the mentioned study protocol, additional MR examinations were done on subjects and TEFR par- ticipants, if acute injuries (for example, stress fractures [52]) and pain syndromes (for example, low back pain [49,52]) occurred and a specific diagnostic finding was necessary to prevent further injuries or complications on the endurance runners (Figure 13). Anthropometric and cardiovascular physical measurements Anthropometric and cardiovascular physical measure- ments were done on all subjects (Figure 3) every fourth day. Therefore, the 44 subjects were randomly assigned to one of four groups. Body mass was measured with BIA using a Tanita BC-545™ BIA scale (Arlington Heights, IL, USA: to the nearest 0.1 kg). This balance gave additional results about percentage of body fat and lean body mass based on MR validated calculation pro- cedures [141]. The measurements took place in the morning (between 4 a.m. and 5 a.m.) and after the stage (between 3 p.m. and 9 p.m.) together with measurement of blood pressure and body temperature (T) using an infrared ear thermometer (ThermoScan IRT 4020 ™, Braun, Germany: to the nearest 0.2°C. After the stage between 3 p.m. and 9 p.m., the skinfold (SF) thickness of the same subjects was measured using a skinfold cali- per (GPM ™, Silber and Hegner, Zurich, Switzerland: to the nearest 2 mm) and their segmental body circumfer- ence (CF) was measured using a retractable measuring tape (to the nearest 1 mm). For SF, the mean value was calculated from three consecutive intra-individual mea- surements at eight regions on the right side of the body according to Ball et al. [142]: chest, midaxillary (verti- cal), triceps, subscapular, abdominal (vertical), suprailiac (at anterior axillary), thigh, and calf. For CF, mean value was calculated from three consecutive intra-individual measurements at six regions on the right side of the body according to Lee et al. [143]: upper arm (largest part of the limb), waist, hip, thigh (10 cm/20 cm above upper patella pole), and calf (largest part of the limb). Figure 12 Mobile MRI H1-spectroscopy. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 18 of 33 To avoid inter-observer error all the anthropometric measurements were done by the same, specifically trained investigator. Every 800 km a short term ECG was planned on every subject. Lab samples Midstream urine samples were taken from all subjects twice each day. Before breakfast in the morning between 4:00 a.m. and 5:00 a.m. and after each stage in the eve- ning after dinner between 7:00 p.m. and 9:00 p.m. Blood samples were taken every 1,000 km from the cubital vein after stage. The samples were immediately centri- fuged and frozen (below -20°C) and put on -80°C after the race. Post-race/follow-up On the day they dropped out, non-finishers (NF) had a last complete measurement of all specific MRI protocols and physical examinations (BIA, SF, CF) and provided blood- and urine-samples. Nearly eight months after the TEFR09, 15 of the 44 subjects (all of them finishers of the TEFR09) had a follow up examination at UHU on the same topics involved during the field studies: speci- fic MRI examinations, anthropometric measurements, ECG and blood and urine samples. Statistical analysis For statistical analysis the software ‘SPSS 12.OG for Windows, Version 12.0.1’ was used. Data are presented as mean (SD, range) and median (IQR) as appropriate. The coefficient of variation (CV (%) = 100*SD/mean) was calculated only for measured absolute data on per- formance. The stage severity index (SSI) is an indirect parameter calculated from the mean stage velocity of all runners without a severe handicap v;¯ STAGE* in relation to the total mean velocity of the whole race v;¯ TEFR*. Therefore, the SSI represents the relative burden of each stage, which is dependent on the mentioned multiple external factors that changed daily. It reflects the sum of daily weather and route conditions: SSI = vSTAGE vSTAGE* vTEFR* *: values are only integrated in calculation, if the stage performance of the specific runner is more than 87% of his mean race performance Results Race conditions The mean stage length was 70.1 km (SD 11.8 km, range 44 to 95.1 km) and influenced the SSI positively (Figure 14). Temperature and humidity were also factors influ- encing the SSI and showed a mean (mean of three daily measures at 6:00 a.m., 10:00 a.m. and 2:00 p.m.) of 15.2° C (SD 4.7°C, range 3.7 to 25.1°C) and 55.6% (SD 14.3, range 26.5% to 82.7%), respectively. Altitude differences were not measured. The longest stages occurred in the last third of the race and the coldest, wettest and most humid and, therefore, most severe stages, were at end of the TEFR09 which pushed the runners to their limits (Figure 14). Changes of study plan due to hazards in TEFR For every research topic, distance intervals of measure- ment (MI) throughout the TEFR09 were defined. The discrepancies between these planned and the realized MI can be shown as mean absolute deviations (Figure 15). For MRI, data showed mean deviations between 100 and 300 km. For MR spectroscopy it raised up to 400 km, because this special MR technique was highly dependent on the locations with their local magnetic field disturbances (such as, traffic and so on). However, reasons for the deviations were multifold. Study staff had to deal with many influencing factors, which made daily adaptation of the research plan neces- sary: acute or chronic illness of study staff, bad weather conditions (Figure 14) which sometimes influenced operability of the mobile MRI, accidents and technical problems (Table 3) and local situations at stage destina- tion which sometimes made a nearby commissioning of the mobile MRI difficult. However, the strongest influ- ence forcing the staff to change and adapt the daily research work program was the athlete with more or less daily changes in mental and physical conditions and necessities: pain, fatigue, fears, doubts, illness, nutrition time schedules and specific behavior and rituals regard- ing regeneration from this immense physical and psy- chological stress. Therefore, it was not always possible to ensure the exact time of pre- and post-race measure- ments. Despite these uncertainties, 95% of the measure- ment protocol could be followed. Figure 16 shows all performed examinations and mea- surements done before, during and six months after the TEFR09. The overall work load includes: 741 MRI proto- cols with 2,637 MRI sequences (more than 200,000 pic- ture data), 5,600 urine samples, 1,018 BIA-, 539 SF- and CF-measurements, 250 blood samples and 205 ECGs. Baseline characteristics and performances of subjects The baseline characteristics of the study groups are summarized in Table 3. Age and gender did not differ between the two MR groups. None of the 44 subjects showed different AROM of the hip and knee or any functional or anatomical mal-alignment of the legs com- pared to normal values or between the MR groups. Only active hip extension shows a tendency to be better in endurance runners (Table 4). Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 19 of 33 Regarding the ratio of F and NF there is no relevant difference between subject group and the whole starter group (Table 5, Figure 17), but randomly between the MR groups (Table 4). Reasons for dropping out of the race were multifold (Table 5, Figure 17). The main rea- sons for premature exiting the race were overuse syn- dromes of the soft tissues of the leg, resulting in (peri-) myotendinous inflammations in the lower and upper legs (71.4%). Two subjects suffered a stress fracture in the third part of the race, one high tibia fracture (male, 60 years old) and one ventral pelvic fracture (female, 46 years old). Due to the unspecific pain and the high pain level, they ran with these fractures for about 200 km to 240 km before they gave up. There was one case of a rapidly ascending soft tissue abscess of the upper extre- mity due to an initially minor finger lesion (male, 39 years old) indicating the immense burden of the ultra- endurance performance to the runners and their immu- nological system. Performances Regarding all participants, the mean speed per stage was 8.35 km/hour (SD = 0.32; CV = 3.8%) and the mean total race speed of all finishers was 8.25 km/hour (SD = 1.4, Figure 13 Supplementary mobile MRI examinations during the TEFR09. A: male, 61-years-old, stage 38, 2,601 km (PDw TSE fs, A1: coronal, A2: sagittal): Stress fracture of the proximal tibia (1) with perifocal bone edema (2) and focal subcutaneous edema (3). B: male, 49-years old, stage 52, 3,609 km (PDw TSE fs, B1: sagittal, B2: transversal): Retropatellar chondral ulcer (1) leading to bleeding/hematoma of the patellar bone (2). C: male, 41-years-old, stage 13, 857 km (PDw TSE fs transversal): Massive subcutaneous edema (panniculitis: 1) with focal myositis of deep flectors (2) and tenosynovitis of Achilles tendon (3). Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 20 of 33 CV = 17.1%). Finishers invested 552 hours (SD = 91, CV = 16.5%) for the 4,486 km in total. There was a wide range of performance difference between the best and slowest runner throughout the whole race, independent of the stage severity (Figure 18). The best runner (male, 28 years old) performed the race with a mean speed of 11.9 km/ hour (total running time: 378 hours), nearly twice as fast as the slowest runner (female, 58 years old), with a mean speed of 6.2 km/hour (total running time: 723 hours. In the subject group, mean speed per stage was 8.28 km/hour (SD = 0.33; CV = 3.9%) and the mean total race speed of the subject finishers was 8.25 km/hour (SD = 1.3, CV = 15.3%), ranging from 11.1 km/hour (male, 26 years old, total running time: 407 hours, second rank) to 6.5 km/ hour (male, 63 years old, total running time: 696.4 hours, 45th rank). Subjects mean stage speed was on average 8.32 km/hour (SD = 0.33, CV = 3.9%). Figure 18 shows mean performances in total and per stage. Figure 14 Daily profile of weather and stage conditions during the entire TEFR09. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 21 of 33 Discussion History Looking at transcontinental footrace history, the finisher rate ranges from 28% to 73% (Table 1). The Bunion Derbies of the early twentieth century showed the lowest rate due to lower standards regarding sports equipment, nutrition features, endurance-associated behavioral knowledge and the level of organization. These two Derbies had 150 starters in mean, much more than nowadays, indicating a high rate of rookies with little or no experience in ultra running. In the TEFR09, 31% of the participants (32% of the subjects) did not reach the finish line (Table 5, Figure 17). This is 18% more fin- ishers than at the TEFR03. This could be attributed to the longer distance in 2003 (+540 km) from Lisbon to Moscow, implying a mean difference of 8.3 km per stage between the TEFR09 and the TEFR03. Apart from the TEFR03, running distances of modern transcontinental footraces (1992 to 2009) were approximately equally long and the finishing rate of 68% in the TEFR09 lies in the upper range of the published data (Table 1). Looking at the rate of participation, being more than 200% higher Figure 15 Deviation of measurements from projected intervals [km] during TEFR09. Table 3 Relevant accidents and damage to MRI and vehicles during the TEFR project No. stage location (Figure 1) event (Figure 3) MR down time 1 0 Bari, Southern Italy Defect of MRI table. 24 hours 2 12 Lugo to Alberone, Northern Italy Truck collision on bridge over the river Po. - 3 33 Bad Segeberg, Nothern Germany Roof damage on MRI trailer. - 4 36 Göteborg to Sjövik, Southern Sweden Total system breakdown, damage of one compressor. 16 hours 5 38 Kristinehamn, Central Sweden Truck sunken in football sand-field. 5 hours 6 45 Hackas, Central Sweden Severe ankle fracture of MRI assistant. 16 hours 7 56 Svapparaava, Northern Sweden Total rupture of tractor to trailer cables. 2 hours 8 after race nearby Gällivare, Northern Sweden Reindeer collision on the way back from North Cape: total damage of material van. - Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 22 of 33 than the mean starter rate of all modern transcontinental footraces, the comparably high finishing rate indicates a professional organization and preparation of both the runners and organizers of the TEFR09. Performance Due to the diverse ways and possibilities to exercise long distance running (that is, area, length, altitude, distance, weather, indoor/outdoor, on/off-road, looped course, combinations with other disciplines and so on), it is extremely difficult to compare the performances of ultra athletes in the literature [107,108,145]. Regarding the present literature, an abundant variety of physiological, anthropometrical, pre-race and training variables seem to influence running performance and associated injuries, depending upon the length and duration of the races [146-150]. In MSUM, such as the TEFR09, the daily changing environmental conditions have a direct influ- ence on stage performances. In the last days of the TEFR09 weather conditions became more and more diffi- cult towards the destination, North Cape, leading to a marked decrease of mean running speed (Figure 18). Only their indomitable will not to drop out at the end of race after more than 4,000 km of running, kept many emaciated participants in the race. Drop out and injuries Due to the likely multifactorial nature of running inju- ries, very few firm conclusions can be made based on the existing studies. In general, there are intrinsic factors such as individual biomechanical abnormalities (that is, mal-alignments, muscle imbalance, stiffness, weakness, instability) or extrinsic (mostly avoidable) factors such as poor running technique, improper equipment and improper changes in training extent and mode or dura- tion and frequency of the race burden contributing to overuse injuries [151]. The one year prevalence of run- ning injuries is 55% in male marathon runners; limb overuse injuries are the most common [152]. In UM these entities become much more important. The most common injuries for runners are multiply cited in the literature: anterior knee pain (for example, patella- femoral syndrome), iliotibial band friction syndrome, tibial stress syndrome (shin splint/injuries), plantar Figure 16 Realized clinical and MRI-measurements on all subjects during TEFR09. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 23 of 33 fasciitis, Achilles tendonitis and meniscal injuries of the knee [152-156]. Approximately two thirds of NF dropped out of the race in the first half of the TEFR09 (Figure 17). As our results show, the reasons for premature resignation of subjects were different. Conforming to the literature [20,22], in more than two thirds of the cases, overuse injuries of the lower and upper limb were the most common reasons (Table 5). However, these soft tissue overuse injuries occurred not only in less experienced ultra runners, but also in runners who had already successfully finished transcontinental races such as the TEFR03 or the ‘Run Across America’. There were only a few subjects and run- ners without any overuse problems of the limbs in these 64 days. However, not every soft tissue overuse inflamma- tion leads to the cessation of running. Most runners were able to ‘overrun’ these specific problems. They reduced running speed in adaptation to their problems, used topi- cal application of anti-inflammatory medication and some of them took non-steroidal anti-inflammatory drugs for a Table 4 Baseline characteristics of the TEFR study population all subjects MR group 1 MR group 2 number (%) number (%) number (%) total 44 22 (50.0) 22 (50) men 40 (90.9) 20 (90.9) 20 (90.9) women 4 (9.1) 2 (9.1) 2 (9.1) Finisher (F) 30 (68.2) 19 (86.4) 11 (50.0) Non-finisher (NF) 14 (31.8) 3 (13.6) 11 (50.0) mean/median (SD) mean/median (SD) mean/median (SD) age (years) 49.7 (10.5) 50.3 (9.6) 49.1 (11.5) prerace history: years of regular endurance running 17.9 (7.5) 19.1 (7.5) 17.1 (7.4) finished marathons 91.7 (168.6) 62.0 (93.4) 121.47 (218.8) finished ultra-marathons 85.4 (63.6) 81.1 (59.0) 89.8 (69.0) finished multistage ultra-marathons 5.7 (3.6) 5.1 (4.1) 6.3 (2.9) anthropometry: height (cm) 175 (8) 175 (6) 174 (9) BMI (kg/m2) 23.1 (2.2) 22.8 (1.8) 23.4 (2.6) body fat percentage, BIA (%) 11.2 (4.3) 11.0 (4.1) 11.4 (4.5) body fat percentage, calculateda (%) 16.6 (4.2) 15.5 (3.2) 16.6 (5.0) body fat percentage, MRI (%) - - 22.7 (6.0) muscle percentage, calculatedb (%) 49.8 (5.1) 49.7 (4.7) 50.0 (5.7) somatic lean tissue, MRI (%) - - 65.0 (5.3) active range of hip motion (°) flexion, 121 (26)c 123 (27) 122 (26) 124 (27) extension, 19 (16)c 24 (17) 25 (17) 21 (16) abduction, 42 (22)c 43 (23) 43 (22) 42 (24) internal rotation, 31 (16)c 31 (16) 30 (16) 32 (16) external rotation, 32 (18)c 34 (19) 33 (18) 34 (19) active range of knee motion (°): flexion, 132 (20)c 134 (19) 135 (20) 133 (19) lower limb alignment: - - LL difference [mm], 6 (95th: 11)d 2 (3.3), 95th: 9) FTR, 1.26 (0.05)d 1.17 (0.04) FTA [°], m: 178 (174-182)e 178 (175-182) w: 181 (177-185)e FTA difference 1 (0.8) MAD (mm), 10 (4-16)e 10 (4-17) acalculated by the updated DC (DEXA criterion)-equation according to Ball et al. [142], inputs: age, 7SF (chest, midaxillary, triceps, subscapular, abdomen, suprailiac, thigh); bcalculated estimation of skeletal muscle mass according to Lee et al. [143], inputs: gender, age, race, height, 3 SF and CF (upper arm, thigh, calf); cnormal AROM-values of hip and knee joints (goniometric data) according to the 1st National Health and Nutrition Examination Survey (NHANES I) [144]; dnormal values of side differences in the lower limb [119], generated with computed tomography gold standard method [117]. FTR: femoro-tibial length-ratio. Leg length (LL) differences of > 14 mm (3SD) are seen as pathological [122,119,118]; enormal values of femorotibial angle (FTA) and mean axis distance (MAD) [116,120,121] Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 24 of 33 few days. With adequate behavior many, but not all, ath- letes recovered and were able to continue the race. Pre- sumably some athletes could handle more pain than others [20]. An example is one extreme runner, an experi- enced 49-year-old male subject, who had multiple severe overuse-induced soft tissue inflammations with local mus- cle fiber rupture forcing him to frequently slow down his speed (Figure 7A). He also showed signs of exertional compartment syndrome, but did finish the TEFR09. His ordeal at the TEFR09 is reflected by the red line in Figure 18. Contrary to other reports, Achilles tendonitis or lower limb joint problems were not a reason for subjects to stop the TEFR09. Further results of module II research topics such as specific personality, temperament, character and pain perception will be presented soon. Statement of principal findings The relevance of results in field studies is determined by the appropriateness of the research questions and hypoth- eses, by the practicability of methods and measurements and the consistency of their specific implementation and by the correct interpretation of results. Due to the mani- fold open questions and unproven hypotheses in endur- ance running, the unique opportunity of doing real time observations of changes in the body of athletes while run- ning at the upper limit in a MSUM was demanding. Table 5 Reason for not-finishing the TEFR09 affected region pathology subjects (number = 14, 31.8%) all (number = 21, 31.3%) Soft tissues of legs: 10 (71.4%) 14 (66.7%) lower legs: shin splint: myofasciitis, tenditis 5 (35.7%) 7 (33.3%) Achillodynia - 1 (4.8%) upper legs: myo-tendino-fasciitis, perineuritis 5 (35.7%) 6 (28.6%) Bone/joint of lower body: stress fractures: tibia, pelvis 2 (14.3%) 2 (9.5%) bunion (arthritis) 1 (7.1%) 1 (4.8%) Upper extremities: Phlegmon of the hand 1 (7.1%) 1 (4.8%) Gastrointestinal (GIT): upper GIT-bleeding (NSAID) - 1 (4.8%) GIT infection - 1 (4.8%) Mental problems: Intolerance of crowded small halls at night - 1 (4.8%) GIT, gastrointestinal tract; NSAID, non-steroidal anti-inflammatory drug. Figure 17 Drop-out rate of TEFR09. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 25 of 33 The TEFR project was designed to explore inter-indi- vidual variability in adaption to the tremendous persist- ing physical endurance running load on the different organic and functional systems of the body with regard to the lack of breaks and time for regeneration. All technical equipment was tested by the specific man- ufacturers on reliability and validity under normal clinical conditions and usage. But daily dismantling, transport and setting up of the mobile MRI hardware sets extraor- dinary demands which were initially not totally verifiable and calculable. Despite some technical problems and temporal defects (Table 3), our arrival at the North Cape demonstrated the feasibility of accompanying a large group of endurance runners (67) with a mobile MRI and all its necessary equipment ensuring permanent operabil- ity during the 64 stage ultra marathon. Throughout the whole TEFR09 our time schedules for examinations adapted to the daily changing local circum- stances and the athletes mental state and problems. To avoid additional stress for the subjects, they could not and were not forced to follow the study protocol strictly. How- ever, the efficiency of this strategy was reflected in the high rate of compliance (98%) until the end of the TEFR09. Only one subject who finished the race left the study at stage 36 (km 2,448) due to personal and, expli- citly, not study related problems. Consequently, the com- pletion rate of planned examinations over the whole running distance of 4,486 km was only limited by the drop-out rate of the subjects from the TEFR09 (Table 4, Figure 16). In particular, specific implementation of sta- tionary validated MRI protocols on the mobile MRI on the truck trailer by a team of MRI experts and training of the research staff on the mobile MRI before the start ensures practical experience with the experimental protocols under field conditions and makes modification of them possible where necessary. Strengths of the study The strengths of the TEFR project are the unique chance to do a field study with the large number of 44 subjects, the realization of tests and measurements with the modern technical gold-standard equipment MRI in a daily changing and increasingly harsh and inhospitable environment (Figure 14), the complete baseline control data and the high rate of test completion. Large subject numbers provide the statistical power to discriminate between and, identify associations with, different pat- terns of adaptation as well as to detect differences in response between subgroups. Matched subject race pro- files and baseline measurements before the start of the TEFR09 control for variability of exposure to ascending running distance and, thereby, permit valid inter-indivi- dual comparison of responses to this burden (with sub- jects as their own controls), maximizing the signal (true physiological differences) to noise (variations in expo- sure) ratio. Figure 18 TEFR09 performances. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 26 of 33 The avoidance of invasive or interventional tests on the subjects’ mechanisms during the TEFR09 and the descriptive nature of the data may be considered a weakness of this study. However, the variety of outputs from different measurement techniques (for example, functional and cine MRI, physical anthropometrical measurements and laboratory data including proteomics, plasma and urine biomarkers) allows observation of con- sistent patterns of response that may be strongly sugges- tive of particular mechanisms. In module I, for example, measured data of T2*-map- ping of joint cartilage (Figure 4 A and 4.2A/B) will allow conclusions on the influence of long distance running on the proteoglycans in the cartilage matrix based on the cur- rent experimental experiences [43-45,157]. As in most other mobile MR associated examinations of other mod- ules, additional laboratory analyses using specific para- meters on collected blood and urine samples (for example, cartilage oligomeric matrix protein (COMP) [158-165] for joint cartilage research) will give further information for interpretation and verification of image related results. Another example is the vascular cine MRI studies of module III. In humans, the relationships of blood flow changes to structure, function, and shear rate of conduct- ing arteries have not been thoroughly examined. There- fore, the purpose of the vascular cine MRI study in module III (Figure 10) was to investigate these para- meters of the elastic-type, common carotid artery (CCA) and the muscular-type, common femoral artery (CFA) in long-term running, assuming that the impact of activity- induced blood flow changes on conduit arteries, if any, should be seen in these highly endurance-trained ath- letes. These investigations using the gold standard method, MRI [137], enable further analyses on the cur- rent status of insights on the question of structural and functional vascular adaption and associated exercise- induced blood flow changes on endurance training based on sonographic B-mode measurements [100]. The manifold investigations of Knechtle et al. on ultra endurance athletes [15-22,113,145] focused on the ques- tion which anthropometric parameters of ultra athletes are predictors of ultra endurance performance. These authors postulated some direct connections between spe- cific physical anthropometric markers and ultra endur- ance performance [16,19,166]. Examinations of module I and module IV (morphometry, body composition) of the TEFR project with its possibility of precise and differen- tiated morphometric analysis (for example, segmental and functional muscle volumetry) may be able to verify common experiences and to detect relationships between anthropometry and morphometry of endurance athletes and performance in MSUM. All tissue systems - subcutaneous and visceral adipose tissues, muscles, ligaments, fascia, tendons, bones and cartilage - were studied with special quantitative and quali- tative MR techniques. This should help explain how the different tissues react to the severe stress that continued for days and weeks without any pauses for regeneration or even resting phases as two marathon distances had to be completed every day. Individual performance and ability to deal with injuries and overuse symptoms with regard to decision making for stopping MSUM is a complex psychosomatic process and more or less modulated by character traits. Strong changes of endocrine and metabolic status during mara- thon runs are described [78,79]. Hormonal changes can influence pain sensation and show an influence on speci- fic brain functions [167]. Knowing this, investigations detecting reasons for dropping out of the race (14 sub- jects) can focus not only on MR image analysis, but must also include specific laboratory analysis and psychometric tests as done or planned in the TEFR project. Serotonin, tryptophan and endorphin are described for use as stress markers in UM [3]. The relation of branched-chain to aromatic amino acids as a model (amino acid dysbalance hypothesis) to explain running-associated fatigue is described [80]. The reduction of the pain sensation is known for cortisol [167]. Considering all these particular mechanisms influencing performance and decision mak- ing in ultra athletes, the important dimension of labora- tory analysis possibilities, in addition to MR data analysis, becomes obvious for the different parts of the TEFR project. Overall, the possibility of cross-validations between phy- sical, MR-graphic, -functional and laboratory follow-up data on multiple organic systems during a nearly ten-week ultra run is a unique strength of this study. Weakness of the study The main weakness of this study is the lack of a control group of non endurance experienced subjects. However, this is not a feasible option in field studies under race con- ditions including such an immense amount of physical and mental load. In order to explore the influence of pre- race running experience, we will undertake subgroup ana- lyses investigating the influence of individual pre-race per- formances on our findings. In the pre-race pain study (project module II, MR group 2), we recruited a parallel age-related control group that was tested over the same pain scale and functional MRI protocol as the MSUM exposed subjects. This sub-group is, therefore, not con- founded by self-selection due to prior endurance tolerance. As the first attempt in MR research, we tried to per- form H1-MR-spectroscopy for measurement of IMCL [140,168,169] with a mobile MRI on a truck trailer. MR spectroscopy needs a stable magnetic field and, there- fore, a still and static environment around the scanner. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 27 of 33 Due to the daily changing position of the mobile MRI, the possibility and feasibility of manual shimming was not predictable. This is the only measurement with uncertain validity due to changing environmental condi- tions in the TEFR project. Environmental factors, such as ambient weather condi- tions (Figure 14), subject de- or hyperhydration and con- current illnesses may also have confounded results. However, indoor temperature (18.7°C, SD 3.0°C, range 11.7 to 28.5°C) and temperature in the MR trailer (20.5°C, SD 0.8°C, range 18.5 to 21.8°C) was much less variable than outdoor temperature (15.2°C, SD 4.7°C, range 3.7 to 25.1°C). All subjects were encouraged to maintain adequate hydration (guided by the production of good quanti- ties of pale urine). There was only one Japanese subject identified with a severe illness during the race, suffering from a severe cough which persisted from stage 12 till stage 32 (day of drop out). Another weakness of the study was that there was only rough documentation of nutrition. Nutrition depended on food availability at the TEFR stages and was provided by the TEFR organization. The use of doping substances was forbidden by the terms of participation but not controlled. Runners did not agree to close measurement and docu- mentation of food and caloric intake, because this would have meant too much disturbance of their daily running routine and compromised compliance due to additional stress conducted by the research work. Despite initial con- cerns, mobile MRI examinations did not result in addi- tional stress for the athletes. On the contrary, most of them enjoyed relaxing in the MR scanner, having no other noises and people around them while listening to their favorite music via headphones. Strengths and weakness in relation to other studies In comparison to previous field, laboratory and radiologi- cal, especially stationary MR studies focusing on long dis- tance running and its effects on the human body, our study is unique in several aspects: ultra-long distance run- ning without any day of rest, cohort size of subjects and use of a mobile MRI throughout the whole race. This is the first MR-based follow-up ultra marathon field study that ensures unique data based on repeated measurements on ascending distance burden. We explored the possibility of conducting this study with a stationary MRI in a fixed local setting. However, this is not realizable with a large cohort size, because not many ultra endurance athletes took the challenge to run ultra long distances in circles in local regions or stadiums day by day. For example, at the Sri Chinmoy Self-Transcendence 3,100 Mile Race over 5,649 laps of one extended city block in Jamaica, Queens, New York (http://www.3100.srichin- moyraces.org) only 10 to 14 participants started regularly. If a study like this is planned, it has to be adapted to the race circumstances and not the race conditions to the study. Only exceptional runners would be willing to take such a burden under laboratory conditions. It is the experi- ence of the distance and the environment that motivates these athletes to run thousands of kilometers. In addition, such an approach might have incurred significant addi- tional costs; our subjects were entirely self-funded, whereas volunteers in chamber studies often expect remuneration. Unanswered questions and future research Further research arising from this study will follow two themes. First, studies in patients to explore the validity of our model by applying the findings of this study to pathophysiological problems in a clinical setting. Sec- ond, collecting additional healthy volunteer data from subjects exposed to an ultra endurance burden (ultra marathon, ultra triathlon, ultra cycling and) in further field studies and chamber studies. Whether it is possible to initiate future projects using this model of a mobile MRI field study is critical. First, this was a unique cohort size in transcontinental ultra running and it would be difficult to find a size like this again: the latest Run Across America (Table 1) had only 14 participants. Second, in addition to sufficient funding a bit of luck is necessary to finish a field study success- fully when using a sensitive and high-maintenance tech- nical piece of equipment such as a mobile MRI. Future studies might answer additional questions by using alternative or additional measurement techniques or undertaking novel intervention trials. Conclusions The TEFR project was both a challenge and risk together. It demonstrates the feasibility and safety of conducting a large ultra endurance cohort study with a mobile MRI under ‘natural’ conditions over 64 stages and daily chan- ging environment on the way across all of Europe. Thanks to the possibility offered by a modern mobile MR-imager diverse research topics from different fields of medicine could be implemented in the measurement protocol to study human adaption to an ultra endurance burden. Systematic measurements of a large set of vari- ables were achieved with high-fidelity in 44 subjects and up to 4,500 kilometers distance running. The resulting dataset is a unique resource for the study of regeneration and adaption in relation to a high impact ultra endurance running burden which may improve specific or general scientific understanding of responses to critical illness at the limits of stress and strain of the human body. Acknowledgements Contributions to the study Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 28 of 33 The TransEurope FootRace 2009 Project is a research project coordinated by the Department of Diagnostic and Interventional Radiology, University Hospital of Ulm, Germany. The aim of the project is to conduct research into ultra endurance running in order to improve understanding of regeneration and adaption processes in different kind of tissues, organs and functional body systems. The TEFR project research groups members all contributed to the collection and analysis of data described in this paper. Study funding sources This project was mainly supported in part by the German Research Association (DFG: ‘Deutsche Forschungsgemeinschaft’), under Grants SCHU 2514/1-1 and SCHU 2514/1-2. Other non-public funds were received from Siemens Healthcare Ltd. and the Medical Faculty of the University of Ulm. All funding was unrestricted. None of the funding bodies had any role in the study design, data collection, data analysis, data interpretation, manuscript preparation or decision to publish. We cordially thank all endurance athletes who took part at this project. Considering their immense physical and mental stress they showed an extraordinary compliance every day of the TransEurope FootRace 2009. Author details 1Department of Diagnostic and Interventional Radiology, University Hospital of Ulm, Germany. 2Outpatient Rehabilitation Centre at University Hospital of Ulm, Germany. 3Institute of Exercise and Health Sciences, Sports Medicine, University of Basel, Switzerland. 4Health Center St. Gallen and Department of General Practice, University Hospital of Zürich, Switzerland. 5Section on Experimental Radiology, Department of Diagnostic and Interventional Radiology, University Hospital of Tübingen, Germany. 6Siemens Healthcare, Magnetic Resonance, Stuttgart, Germany. 7Main organizer and race director TransEurope FootRace 2009, Horb, Germany. Authors’ contributions US contributed to the conception and design of the study, to the funding, to the acquisition of data, the analysis of data, the interpretation of data and drafted the manuscript. AST contributed to the conception and design of the study, to the acquisition of data, and the analysis and interpretation of data. BK contributed to the design of the study and the interpretation of data. JM contributed to the design of the study, to specific MR sequence protocols and to the analysis and interpretation of data. HW contributed to the acquisition of data. ME contributed to the acquisition and analysis of data. WF contributed to the conception of the study, to the acquisition of data, the analysis of data and the interpretation of data. IS contributed to the conception and to the acquisition of data. SG contributed to the creation of specific MR sequence protocols. HB contributed to specific MR sequence protocols. IS contributed to the conception of the study. HJB contributed to the acquisition of data. CB contributed to the conception and design of the study, to the acquisition of data and the analysis and interpretation of data. All authors read and approved the final draft. Cooperators and coworkers in data post processing and analysis are permanently rising. At time of manuscript writing they are as follows: F. Birklein, M. Breimhorst, DC. Cheng, J. Ellermann, S. Faust, S. Göd, L. Heisterkamp, E. Kitzenmaier, K. König, S. König, TC. Mamisch, A. Reiner, D. Schoss, C. Tassler, C. Trattnig, F. Weber, S. Wuchenauer, A. Wunderlich and C. Würslin. Authors’ information Dr Uwe Schütz may also be contacted using his alternative email address: [email protected] Competing interests The authors declare that they have no competing interests. Received: 16 May 2012 Accepted: 19 July 2012 Published: 19 July 2012 References 1. Costill DL, Bowers R, Kammer WF: Skinfold estimates of body fat among marathon runners. Med Sci Sports 1970, 2:93-95. 2. Davies CTM, Thompson MW: Aerobic performance of female marathon and male ultramarathon athletes. Eur J Appl Physiol 1979, 41:233-245. 3. Tittel K, Wutscherk H: Anatomical and anthropometric fundamentals of endurance. In Endurance in Sport. Edited by: Shephard RJ, Ästrand PO. Oxford: Blackwell Scientific Publications; 1992:35-45. 4. Svedenhag J, Sjödin B: Body-mass-modified running economy and step length in elite male middle- and long-distance runners. Int J Sports Med 1994, 15:305-310. 5. Speechly DP, Taylor SR, Rogers GG: Differences in ultra-endurance exercise in performance-matched male and female runners. Med Sci Sports Exerc 1996, 28:59-365. 6. Hetland ML, Haarbo J, Christiansen C: Regional body composition determined by dual-energy X-ray absorptiometry. Relation to training, sex hormones, and serum lipids in male long-distance runners. Scand J Med Sci Sports 1998, 8:102-108. 7. Abe T, Kumagai K, Brechue WF: Fascicle length of leg muscles is greater in sprinters than distance runners. Med Sci Sports Exerc 2000, 32:1125-1129. 8. Hoffman MD: Anthropometric characteristics of ultramarathoners. Int J Sports Med 2008, 29(19):808-811. 9. Freund W, Billich C, Brambs HJ, Weber F, Schütz UHW: MRI changes of Achilles tendon in experienced runners and beginners during training and after (half-) marathon competition. Z Orthop Unfall 2011, 149:407-417. 10. Davies CTM, Thompson MW: Physiological responses to prolonged exercise in ultramarathon athletes. J Appl Physiol 1986, 61:611-617. 11. Sharwood K, Collins M, Goedecke J, Wilson G, Noakes T: Weight changes, sodium levels, and performance in the South African Ironman Triathlon. Clin J Sport Med 2002, 12:391-399. 12. Tomaszewski M, Charchar FJ, Przybycin M, Crawford L, Wallace AM, Gosek K, Lowe GD, Zukowska-Szczechowska E, Grzeszczak W, Sattar N, Dominiczak AF: Strikingly low circulating CRP concentrations in ultramarathon runners independent of markers of adiposity: how low can you go? Arterioscler Thromb Vasc Biol 2003, 23:1640-1644. 13. Knechtle B, Wirth A, Knechtle P, Zimmermann K, Kohler G: Personal best marathon performance is associated with performance in a 24-h run and not anthropometry or training volume. Br J Sports Med 2009, 43(11):836-839. 14. Knechtle B, Knechtle P, Rosemann T, Senn O: What is associated with race performance in male 100-km ultra-marathoners - anthropometry, training or marathon best time? J Sports Sci 2011, 23:1-7. 15. Knechtle B, Bircher S: Changes in body composition during an extreme endurance run. Praxis 2005, 94:371-377. 16. Knechtle B, Knechtle P, Schulze I, Kohler G: Upper arm circumference is associated with race performance in ultra-endurance runners. Br J Sports Med 2008, 42:295-299. 17. Knechtle B, Kohler G: Running 338 km within 5 days has no effect on body mass and body fat but reduces skeletal muscle mass - the Isarrun 2006. J Sports Sci Med 2007, 6:401-407. 18. Knechtle B, Duff B, Schulze I, Kohler G: The effects of running 1,200 km within 17 days on body composition in a female ultrarunner- Deutschlandlauf 2007. Res Sports Med 2008, 16:167-188. 19. Knechtle B, Duff B, Welzel U, Kohler G: Body mass and circumference of upper arm are associated with race performance in ultraendurance runners in a multistage race–the Isarrun 2006. Res Q Exerc Sport 2009, 80:262-268. 20. Knechtle B, Duff B, Schulze I, Rosemann T, Senn O: Anthropometry and pre-race experience of finishers and nonfinishers in a multistage ultra- endurance run–Deutschlandlauf 2007. Percept Mot Skills 2009, 109:105-118. 21. Knechtle B, Wirth A, Knechtle P, Rosemann T: Training volume and personal best time in marathon, not anthropometric parameters, are associated with performance in male 100-km ultrarunners. J Strength Cond Res 2010, 24:604-609. 22. Knechtle B, Knechtle P, Rosemann T: Race performance in male mountain ultra-marathoners: anthropometry or training? Percept Mot Skills 2010, 110:721-735. 23. Knechtle B, Enggist A, Jehle T: Energy turnover at the Race across America (RAAM): a case report. Int J Sports Med 2005, 26:499-503. 24. Transeurope Footrace. [http://www.transeurope-footrace.org]. 25. Kastner CB: Bunion Derby: The 1928 Footrace Across America. 1 edition. University of New Mexico Press; 2007. 26. Williams G: C.C. Pyle’s Amazing Foot Race: The True Story of the 1928 Coast- to-Coast Run Across America. Illustrated edition. Rodale Books; 2007. 27. Griffis ML: Great American Bunion Derby. Eakin Press;, 1 2003. 28. Lewis B: Running the Trans America Footrace: Trials and Triumphs of Life on the Road. 1 edition. Stackpole Books; 1994. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 29 of 33 29. Shapiro JE: Meditations from the Breakdown Lane: Running Across America Houghton Mifflin Press; 1983. 30. Schulze I: Transeuropalauf 2003. Lissabon - Moskau 5036 km in 64 Tagesetappen. 1 edition. Engelsdorfer Verlagsgesellschaft; 2004. 31. Biermanski B: Der härteste Lauf der Welt: In 64 Tagen von Lissabon nach Moskau (Trans Europa Lauf). 1 edition. Books on Demand GmbH; 2005. 32. Schulze I: TransEurope-FootRace 2009: Bari - Nordkap - 4.487,7 km in 64 Tagesetappen. Engelsdorfer Verlag , 1 2010. 33. Secker R: Running across countries Create Space; 2009. 34. Knechtle B, Knechtle P, Schulze I, Kohler G: Vitamins, minerals and race performance in ultra-endurance runners-Deutschlandlauf 2006. Asia Pac J Clin Nutr 2008, 17:194-198. 35. Knechtle B, Schulze I: Nutritional behaviours in ultra-endurance runners- Deutschlandlauf 2006. Praxis (Bern 1994) 2008, 97:243-251. 36. Shellock FG, Deutsch AL, Mink JH, Kerr R: Do asymptomatic marathon runners have an increased prevalence of meniscal abnormalities? An MR study of the knee in 23 volunteers. AJR Am J Roentgenol 1991, 157:1239-1241. 37. Shellock FG, Mink JH: Knees of trained long-distance runners: MR imaging before and after competition. Radiology 1991, 179:63. 38. Hohmann E, Wörtler K, Imhoff A: MR imaging of the hip and knee before and after marathon running. Am J Sports Med 2004, 32:55-9. 39. Schueller-Weidekamm C, Schueller G, Uffmann M, Bader TR: Does marathon running cause acute lesions of the knee? Evaluation with magnetic resonance imaging. Eur Radiol 2006, 16:2179-2185. 40. Krampla W, Mayrhofer R, Malcher J, Kristen KH, Urban M, Hruby W: MR imaging of the knee in marathon runners before and after competition. Skeletal Radiol 2001, 30:72-76. 41. Cymet TC, Sinkov V: Does long-distance running cause osteoarthritis? J Am Osteopath Assoc 2006, 106:342-345. 42. Krampla WW, Newrkla SP, Kroener AH, Hruby WF: Changes on magnetic resonance tomography in the knee joints of marathon runners: a 10- year longitudinal study. Skeletal Radiol 2008, 37:619-626. 43. Arokoski J, Kiviranta I, Jurvelin J, Tammi M, Helminen HJ: Long-distance running causes site-dependent decrease of cartilage glycosaminoglycan content in the knee joints of beagle dogs. Arthritis Rheum 1993, 36:1451-1459. 44. Arokoski J, Jurvelin J, Kiviranta I, Tammi M, Helminen HJ: Softening of the lateral condyle articular cartilage in the canine knee joint after long distance (up to 40 km/day) running training lasting one year. Int J Sports Med 1994, 15:254-260. 45. Arokoski JP, Hyttinen MM, Lapveteläinen T, Takács P, Kosztáczky B, Módis L, Kovanen V, Helminen H: Decreased birefringence of the superficial zone collagen network in the canine knee (stifle) articular cartilage after long distance running training, detected by quantitative polarised light microscopy. Ann Rheum Dis 1996, 55:253-264. 46. Lacey RJ, Thomas E, Duncan RC, Peat G: Gender difference in symptomatic radiographic knee osteoarthritis in the Knee Clinical Assessment–CAS(K): a prospective study in the general population. BMC Musculoskelet Disord 2008, 9:82. 47. Leyk D, Erley O, Gorges W, Ridder D, Rüther T, Wunderlich M, Sievert A, Essfeld D, Piekarski C, Erren T: Performance, training and lifestyle parameters of marathon runners aged 20-80 years: results of the PACE- study. Int J Sports Med 2009, 30:360-365. 48. Bishop GW, Fallon KE: Musculoskeletal injuries in a six-day track race: ultramarathoner’s ankle. Clin J Sport Med 1999, 9:216-220. 49. Fallon KE: Musculoskeletal injuries in the ultramarathon: the 1990 Westfield Sydney to Melbourne run. Br J Sports Med 1996, 30:319-323. 50. Fredericson M, Misra AK: Epidemiology and aetiology of marathon running injuries. Sports Med 2007, 37:437-439. 51. Schueller-Weidekamm C: Long-term success and risk for marathon runners. Radiologe 2010, 50:444-452. 52. Taunton JE, Ryan MB, Clement DB, McKenzie DC, Lloyd-Smith DR, Zumbo BD: A retrospective case-control analysis of 2002 running injuries. Br J Sports Med 2002, 36:95-101. 53. Taunton JE, Ryan MB, Clement DB, McKenzie DC, Lloyd-Smith DR, Zumbo BD: A prospective study of running injuries: the Vancouver Sun Run “In Training” clinics. Br J Sports Med 2003, 37:239-244. 54. Van Middelkoop M, Kolkman J, van Ochten J, Bierma-Zeinstra SM, Koes BW: Course and predicting factors of lower-extremity injuries after running a marathon. Clin J Sport Med 2007, 17:25-30. 55. Johnell O, Rausing A, Wendeberg B, Westlin N: Morphological bone changes in shin splints. Clin Orthop Relat Res 1982, 167:180-184. 56. Aoki Y, Yasuda K, Tohyama H, Ito H, Minami A: Magnetic resonance imaging in stress fractures and shin splints. Clin Orthop Relat Res 2004, 421:260-267. 57. Wilder RP, Sethi S: Overuse injuries: tendinopathies, stress fractures, compartment syndrome, and shin splints. Clin Sports Med 2004, 23:55-81. 58. Detmer DE: Chronic shin splints: classification and management of medial tibial stress syndrome. Sports Med 1986, 3:436-446. 59. Batt ME: Shin splints–a review of terminology. Clin J Sport Med 1995, 5:53-57. 60. Verleisdonk EJ, van Gils A, van der Werken C: The diagnostic value of MRI scans for the diagnosis of chronic exertional compartment syndrome of the lower leg. Skeletal Radiol 2001, 30:321-325. 61. Bong MR, Polatsch DB, Jazrawi LM, Rokito AS: Chronic exertional compartment syndrome: diagnosis and management. Bull Hosp Jt Dis 2005, 62:77-84. 62. Lee D, Armed Forces Health Surveillance Center (AFHSC): Stress fractures, active component, U.S. Armed Forces, 2004-2010. MSMR 2011, 18:8-11. 63. Harrast MA, Colonno D: Stress fractures in runners. Clin Sports Med 2010, 29:399-416. 64. Shellock FG, Morris E, Deutsch AL, Mink JH, Kerr R, Boden SD: Hematopoietic bone marrow hyperplasia: high prevalence on MR images of the knee in asymptomatic marathon runners. AJR Am J Roentgenol 1992, 158:335-338. 65. Gondoh Y, Sensui H, Kinomura S, Fukuda H, Fujimoto T, Masud M, Nagamatsu T, Tamaki H, Takekura H: Effects of aerobic exercise training on brain structure and psychological well-being in young adults. J Sports Med Phys Fitness 2009, 49:129-135. 66. Peters J, Dauvermann M, Mette C, Platen P, Franke J, Hinrichs T, Daum I: Voxel-based morphometry reveals an association between aerobic capacity and grey matter density in the right anterior insula. Neuroscience 2009, 163:1102-1108. 67. Good CD, Johnsrude IS, Ashburner J, Henson RN, Friston KJ, Frackowiak RS: A voxel-based morphometric study of ageing in 465 normal adult human brains. Neuroimage 2001, 14:21-36. 68. O’Brien JT, Paling S, Barber R, Williams ED, Ballard C, McKeith IG, Gholkar A, Crum WR, Rossor MN, Fox NC: Progressive brain atrophy on serial MRI in dementia with Lewy bodies, AD, and vascular dementia. Neurology 2001, 56:1386-1388. 69. Fox NC, Crum WR, Scahill RI, Stevens JM, Janssen JC, Rossor MN: Imaging of onset and progression of Alzheimer’s disease with voxel-compression mapping of serial magnetic resonance images. Lancet 2001, 358:201-205. 70. Katzman DK, Lambe EK, Mikulis DJ, Ridgley JN, Goldbloom DS, Zipursky RB: Cerebral gray matter and white matter volume deficits in adolescent girls with anorexia nervosa. J Pediatr 1996, 129:794-803. 71. Castro-Fornieles J, Bargalló N, Lázaro L, Andrés S, Falcon C, Plana MT, Junqué C: A cross-sectional and follow-up voxel-based morphometric MRI study in adolescent anorexia nervosa. J Psychiatr Res 2009, 43:331-340. 72. Geibprasert S, Gallucci M, Krings T: Alcohol-induced changes in the brain as assessed by MRI and CT. Eur Radiol 2010, 20:1492-1501. 73. Gunston GD, Burkimsher D, Malan H, Sive AA: Reversible cerebral shrinkage in kwashiorkor: an MRI study. Arch Dis Child 1992, 67:1030-1032. 74. Golden NH, Ashtari M, Kohn MR, Patel M, Jacobson MS, Fletcher A, Shenker IR: Reversibility of cerebral ventricular enlargement in anorexia nervosa, demonstrated by quantitative magnetic resonance imaging. J Pediatr 1996, 128:296-301. 75. Knudsen L, Drummond PD: Cold-induced limb pain decreases sensitivity to pressure-pain sensations in the ipsilateral forehead. Eur J Pain 2009, 13:1023-1029. 76. Van den Eynde F, Treasure J: Neuroimaging in eating disorders and obesity: implications for research. Child Adolesc Psychiatr Clin N Am 2009, 18:95-115. 77. Kraemer WJ, Fragala MS, Watson G, Volek JS, Rubin MR, French DN, Maresh CM, Vingren JL, Hatfield DL, Spiering BA, Yu-Ho J, Hughes SL, Case HS, Stuempfle KJ, Lehmann DR, Bailey S, Evans DS: Hormonal responses to a 160-km race across frozen Alaska. Br J Sports Med 2008, 42:116-120. 78. Agawa H, Yamada N, Enomoto Y, Suzuki H, Hosono A, Arakawa K, Ghadimi R, Miyata M, Maeda K, Shibata K, Tokudome M, Goto C, Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 30 of 33 Tokudome Y, Hoshino H, Imaeda N, Marumoto M, Suzuki S, Kobayashi M, Tokudome S: Changes of mental stress biomarkers in ultramarathon. Int J Sports Med 2008, 29:867-871. 79. Hew-Butler T, Noakes TD, Soldin SJ, Verbalis JG: Acute changes in endocrine and fluid balance markers during high-intensity, steady-state, and prolonged endurance running: unexpected increases in oxytocin and brain natriuretic peptide during exercise. Eur J Endocrinol 2008, 159:729-737. 80. Lehmann M, Huonker M, Dimeo F, Heinz N, Gastmann U, Treis N, Steinacker JM, Keul J, Kajewski R, Häussinger D: Serum amino acid concentrations in nine athletes before and after the 1993 Colmar ultra triathlon. Int J Sports Med 1995, 16:155-159. 81. Pickering A, Carter J, Hanning I, Townend W: Emergency department measurement of urinary S100B in children following head injury: can extracranial injury confound findings? Emerg Med J 2008, 25:88-89. 82. Gazzolo D, Frigiola A, Bashir M, Iskander I, Mufeed H, Aboulgar H, Venturini P, Marras M, Serra G, Frulio R, Michetti F, Petraglia F, Abella R, Florio P: Diagnostic accuracy of S100B urinary testing at birth in full-term asphyxiated newborns to predict neonatal death. PLoS One 2009, 4: e4298. 83. Jäncke L, Koeneke S, Hoppe A, Rominger C, Hänggi J: The architecture of the golfer’s brain. PLoS One 2009, 4:e4785. 84. Stuempfle KJ, Lehmann DR, Case HS, Bailey S, Hughes SL, McKenzie J, Evans D: Hyponatremia in a cold weather ultraendurance race. Alaska Med 2002, 44:51-55. 85. Hew TD, Chorley JN, Cianca JC, Divine JG: The incidence, risk factors, and clinical manifestations of hyponatremia in marathon runners. Clin J Sport Med 2003, 13:41-47. 86. Almond CS, Shin AY, Fortescue EB, Mannix RC, Wypij D, Binstadt BA, Duncan CN, Olson DP, Salerno AE, Newburger JW, Greenes DS: Hyponatremia among runners in the Boston Marathon. N Engl J Med 2005, 352:1550-1556. 87. Noakes TD, Sharwood K, Speedy D, Hew T, Reid S, Dugas J, Almond C, Wharam P, Weschler L: Three independent biological mechanisms cause exercise-associated hyponatremia: evidence from 2,135 weighed competitive athletic performances. Proc Natl Acad Sci USA 2005, 102:18550-18555. 88. Urhausen A, Scharhag J, Herrmann M, Kindermann W: Clinical significance of increased cardiac troponins T and I in participants of ultra-endurance events. Am J Cardiol 2004, 94:696-698. 89. Neilan TG, Januzzi JL, Lee-Lewandrowski E, Ton-Nu TT, Yoerger DM, Jassal DS, Lewandrowski KB, Siegel AJ, Marshall JE, Douglas PS, Lawlor D, Picard MH, Wood MJ: Myocardial injury and ventricular dysfunction related to training levels among nonelite participants in the Boston marathon. Circulation 2006, 114:2325-2333. 90. O’Connor DT, Frigon RP: Chromogranin A, the major catecholamine storage vesicle soluble protein. Multiple size forms, subcellular storage, and regional distribution in chromaffin and nervous tissue elucidated by radioimmunoassay. J Biol Chem 1984, 259:3237-3247. 91. Mousavi N, Czarnecki A, Kumar K, Fallah-Rad N, Lytwyn M, Han SY, Francis A, Walker JR, Kirkpatrick ID, Neilan TG, Sharma S, Jassal DS: Relation of biomarkers and cardiac magnetic resonance imaging after marathon running. Am J Cardiol 2009, 103:1467-1472. 92. Trivax JE, Franklin BA, Goldstein JA, Chinnaiyan KM, Gallagher MJ, deJong AT, Colar JM, Haines DE, McCullough PA: Acute cardiac effects of marathon running. J Appl Physiol 2010, 108:1148-1153. 93. Hanssen H, Keithahn A, Hertel G, Drexel V, Stern H, Schuster T, Lorang D, Beer AJ, Schmidt-Trucksäss A, Nickel T, Weis M, Botnar R, Schwaiger M, Halle M: Magnetic resonance imaging of myocardial injury and ventricular torsion after marathon running. Clin Sci (Lond) 2011, 120:143-152. 94. Scharhag J, Herrmann M, Urhausen A, Haschke M, Herrmann W, Kindermann W: Independent elevations of N-terminal pro-brain natriuretic peptide and cardiac troponins in endurance athletes after prolonged strenuous exercise. Am Heart J 2005, 150:1128-1134. 95. Scharhag J, George K, Shave R, Urhausen A, Kindermann W: Exercise- associated increases in cardiac biomarkers. Med Sci Sports Exerc 2008, 40:1408-1415. 96. Montgomery HE, Clarkson P, Dollery CM, Prasad K, Losi MA, Hemingway H, Statters D, Jubb M, Girvain M, Varnava A, World M, Deanfield J, Talmud P, McEwan JR, McKenna WJ, Humphries S: Association of angiotensin- converting enzyme gene I/D polymorphism with change in left ventricular mass in response to physical training. Circulation 1997, 96:741-747. 97. Morganroth J, Henry WL, Maron RJ, Clark CE, Epstein SE: Idiopathic left ventricular hypertrophy. N Engl J Med 1974, 290:1047-1050. 98. Morganroth J, Maron BJ, Henry WL, Epstein SE: Comparative left ventricular dimensions in trained athletes. Ann Intern Med 1975, 82:521-524. 99. Naylor LH, George K, O’Driscoll G, Green DJ: The athlete’s heart: a contemporary appraisal of the ‘Morganroth hypothesis’. Sports Med 2008, 38:69-90. 100. Schmidt-Trucksäss A, Schmid A, Brunner C, Scherer N, Zäch G, Keul J, Huonker M: Arterial properties of the carotid and femoral artery in endurance-trained and paraplegic subjects. J Appl Physiol 2000, 89:1956-1963. 101. Kasikcioglu E, Oflaz H, Kasikcioglu HA, Kayserilioglu A, Umman S, Meric M: Endothelial flow-mediated dilatation and exercise capacity in highly trained endurance athletes. Tohoku J Exp Med 2005, 205:45-51. 102. Petersen SE, Wiesmann F, Hudsmith LE, Robson MD, Francis JM, Selvanayagam JB, Neubauer S, Channon KM: Functional and structural vascular remodeling in elite rowers assessed by cardiovascular magnetic resonance. J Am Coll Cardiol 2006, 48:790-797. 103. Naylor LH, O’Driscoll G, Fitzsimons M, Arnolda LF, Green DJ: Effects of training resumption on conduit arterial diameter in elite rowers. Med Sci Sports Exerc 2006, 38:86-92. 104. Raschka C, Plath M, Cerull R, Bernhard W, Jung K, Leitzmann C: The body muscle compartment and its relationship to food absorption and blood chemistry during an extreme endurance performance. Z Ernährungswiss 1991, 30:276-288. 105. Raschka C, Plath M: Body fat compartment and its relationship to food intake and clinical chemical parameters during extreme endurance performance. Schweiz Z Sportmed 1992, 40:13-25. 106. Höchli D, Schneiter T, Ferretti G, Howald H, Claassen H, Moia C, Atchou G, Belleri M, Veicsteinas A, Hoppeler H: Loss of muscle oxidative capacity after an extreme endurance run: the Paris-Dakar Foot-Race. Int J Sports Med 1995, 16:343-346. 107. Helge JW, Lundby C, Christensen DL, Langfort J, Messonnier L, Zacho M, Andersen JL, Saltin B: Skiing across the Greenland icecap: divergent effects on limb muscle adaptations and substrate oxidation. J Exp Biol 2003, 206:1075-1083. 108. Frykman PN, Harman EA, Opstad PK, Hoyt RW, DeLany JP, Friedl KE: Effects of a 3-month endurance event on physical performance and body composition: the G2 trans-Greenland expedition. Wilderness Environ Med 2003, 14:240-248. 109. Bircher S, Enggist A, Jehle T: Effects of an extreme endurance race on energy balance and body composition: a case study. J Sports Sci Med 2006, 5:154-162. 110. Nagel D, Seiler D, Franz H, Leitzmann C, Jung K: Effects of an ultra-long- distance (1000 km) race on lipid metabolism. Eur J Appl Physiol 1989, 59:16-20. 111. Dressendorfer RH, Wade CE: Effects of a 15-d race on plasma steroid levels and leg muscle fitness in runners. Med Sci Sports Exerc 1991, 23:954-958. 112. Väänänen II, Vihko V: Physiological and psychological responses to 100 km crosscountry skiing during 2 days. J Sports Med Phys Fitness 2005, 45:301-305. 113. Knechtle B, Salas OF, Andonie JL, Kohler G: Effect of a multistage ultra- endurance triathlon on body composition: World Challenge Deca Iron Triathlon 2006. Br J Sports Med 2008, 42:121-125. 114. Reynolds RD, Lickteig JA, Deuster PA, Howard MP, Conway JM, Pietersma A, deStoppelaar J, Deurenberg P: Energy metabolism increases and regional body fat decreases while regional muscle mass is spared in humans climbing Mt. Everest. J Nutr 1999, 129:1307-1314. 115. Lehmann M, Huonker M, Dimeo F: Serum amino acid concentrations in nine athletes before and after the 1993 Colmar Ultra Triathlon. Int J Sports Med 1995, 16:155-159. 116. Paley D, Herzenberg JE, Tetsworth K, McKie J, Bhave A: Deformity planning for frontal and sagittal plane corrective osteotomies. Orthop Clin North Am 1994, 25:425-465. 117. Waidelich HA, Strecker W, Schneider E: Computed tomographic torsion- angle and length measurement of the lower extremity. The methods, normal values and radiation load. Rofo 1992, 157:245-251. Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 31 of 33 118. Blustein SM, D’Amico JC: Limb length discrepancy. Identification, clinical significance, and management. J Am Podiatr Med Assoc 1985, 75:200-206. 119. Strecker W, Keppler P, Gebhard F, Kinzl L: Length and torsion of the lower limb. J Bone Joint Surg Br 1997, 79:1019-1023. 120. Keppler P, Strecker W, Kinzl L: Analysis of leg geometry–standard techniques and normal values. Chirurg 1998, 69:1141-1152. 121. Keppler P: Normwerte des Beines. 8. Kurs zur Analyse und Korrektur von Beindeformitäten. Schloss Reisensburg, Günzburg, Germany 2005, 03:16-19. 122. Keppler P, Kinzl L: Diagnostik der posttraumatischen Beindeformitäten. Trauma Berufskrankh 2006, 8(Suppl 3):303-316. 123. Nagoshi CT, Waler D, Muntaner C, Haertzen CA: Validation of the Tridimensional Personality Questionnaire in a sample of male drug users. Personality and Individual Differences 1992, 13:401-409. 124. Hansenne M, Delhez M, Cloninger CR: Psychometric properties of the temperament and character inventory-revised (TCI-R) in a Belgian sample. J Pers Assess 2005, 85:40-49. 125. Schwarzer R, Jerusalem M: Generalized Self-Efficacy scale. In Measures in Health Psychology: A User’s Portfolio. Causal and Control Beliefs. Edited by: Weinman J, Wright S, Johnston M. Windsor; 1995:35-37. 126. Luszczynska A, Scholz U, Schwarzer R: The general self-efficacy scale: multicultural validation studies. J Psychol 2005, 139:439-457. 127. Mamisch TC, Trattnig S, Quirbach S, Marlovits S, White LM, Welsch GH: Quantitative T2 mapping of knee cartilage: differentiation of healthy control cartilage and cartilage repair tissue in the knee with unloading– initial results. Radiology 2010, 254:818-826. 128. Welsch GH, Trattnig S, Hughes T, Quirbach S, Olk A, Blanke M, Marlovits S, Mamisch TC: T2 and T2* mapping in patients after matrix-associated autologous chondrocyte transplantation: initial results on clinical use with 3.0-Tesla MRI. Eur Radiol 2010, 20:1515-1523. 129. Apprich S, Welsch GH, Mamisch TC, Szomolanyi P, Mayerhoefer M, Pinker K, Trattnig S: Detection of degenerative cartilage disease: comparison of high-resolution morphological MR and quantitative T2 mapping at 3.0 Tesla. Osteoarthritis Cartilage 2010, 18:1211-1217. 130. Schueller-Weidekamm C, Schueller G, Uffmann M, Bader T: Incidence of chronic knee lesions in long-distance runners based on training level: findings at MRI. Eur J Radiol 2006, 58:286-293. 131. Schmitt H, Rohs C, Schneider S, Clarius M: Is competitive running associated with osteoarthritis of the hip or the knee? Orthopade 2006, 35:1087-1092. 132. Major NM, Helms CA: Sacral stress fractures in long-distance runners. AJR Am J Roentgenol 2000, 174:727-729. 133. Shah MK, Stewart GW: Sacral stress fractures: an unusual cause of low back pain in an athlete. Spine 2002, 27:E104-108. 134. Alsobrook J, Simons SM: Sacral stress fracture in a marathon runner. Curr Sports Med Rep 2007, 6:39-42. 135. Fredenicson M, Bergman AG, Hoffman KL, Dillingham MS: Tibial stress reaction in runners: correlation of clinical symptoms and scintigraphy with a new magnetic resonance imaging grading system. Am Sports Med 1995, 23:472-481. 136. Bergman AG, Fredericson M: MR imaging of stress reactions, muscle injuries, and other overuse injuries in runners. Magn Reson Imaging Clin N Am 1999, 7:151-174, ix. 137. Bolster BD Jr, Atalar E, Hardy CJ, McVeigh ER: Accuracy of arterial pulse- wave velocity measurement using MR. J Magn Reson Imaging 1998, 8:878-888. 138. Machann J, Thamer C, Schnoedt B, Haap M, Haring HU, Claussen CD, Stumvoll M, Fritsche A, Schick F: Standardized assessment of whole body adipose tissue topography by MRI. J Magn Reson Imaging 2005, 21:455-462. 139. Würslin C, Machann J, Rempp H, Claussen C, Yang B, Schick F: Topography mapping of whole body adipose tissue using a fully automated and standardized procedure. J Magn Reson Imaging 2010, 31:430-439. 140. Brechtel K, Niess AM, Machann J, Rett K, Schick F, Claussen CD, Dickhuth HH, Haering HU, Jacob S: Utilisation of intramyocellular lipids (IMCLs) during exercise as assessed by proton magnetic resonance spectroscopy (1H-MRS). Horm Metab Res 2001, 33:63-66. 141. Jebb SA, Cole TJ, Doman D, Murgatroyd PR, Prentice AM: Evaluation of the novel Tanita body-fat analyser to measure body composition by comparison with a four-compartment model. Br J Nutr 2000, 83:115-122. 142. Ball SD, Altena TS, Swan PD: Comparison of anthropometry to DXA: a new prediction equation for men. Eur J Clin Nutr 2004, 58:1525-1531. 143. Lee RC, Wang Z, Heo M, Ross R, Janssen I, Heymsfield SB: Total-body skeletal muscle mass: development and cross-validation of anthropometric prediction models. Am J Clin Nutr 2000, 72:796-803. 144. Roach KE, Miles TP: Normal hip and knee active range of motion: the relationship to age. Phys Ther 1991, 71:656-665. 145. Knechtle B, Knechtle P, Andonie JL, Kohler G: Influence of anthropometry on race performance in extreme endurance triathletes: World Challenge Deca Iron Triathlon 2006. Br J Sports Med 2007, 41:644-648. 146. Marti B, Abelin T, Minder CE: Relationship of training and life-style to 16- km running time of 4000 joggers. The ‘84 Berne “Grand-Prix” Study. Int J Sports Med 1988, 9:85-91. 147. Marti B, Vader JP, Minder CE, Abelin T: On the epidemiology of running injuries. The 1984 Bern Grand-Prix study. Am J Sports Med 1988, 16:285-294. 148. Macera CA, Pate RR, Woods J, Davis DR, Jackson KL: Postrace morbidity among runners. Am J Prev Med 1991, 7:194-198. 149. Pate RR, Macera CA, Bailey SP, Bartoli WP, Powell KE: Physiological, anthropometric, and training correlates of running economy. Med Sci Sports Exerc 1992, 24:1128-1133. 150. Saunders PU, Pyne DB, Telford RD, Hawley JA: Factors affecting running economy in trained distance runners. Sports Med 2004, 34:465-485. 151. Wilder RP, Sethi S: Overuse injuries: tendinopathies, stress fractures, compartment syndrome, and shin splints. Clin Sports Med 2004, 23:55-81. 152. Van Middelkoop M, Kolkman J, Van Ochten J, Bierma-Zeinstra SM, Koes BW: Risk factors for lower extremity injuries among male marathon runners. Scand J Med Sci Sports 2008, 18:691-697. 153. Wen DY, Puffer JC, Schmalzried TP: Injuries in runners: a prospective study of alignment. Clin J Sport Med 1998, 8:187-194. 154. Satterthwaite P, Norton R, Larmer P, Robinson E: Risk factors for injuries and other health problems sustained in a marathon. Br J Sports Med 1999, 33:22-26. 155. Taunton JE, Ryan MB, Clement DB, McKenzie DC, Lloyd-Smith DR, Zumbo BD: A retrospective case-control analysis of 2002 running injuries. Br J Sports Med 2002, 36:95-101. 156. Fredericson M, Misra AK: Epidemiology and aetiology of marathon running injuries. Sports Med 2007, 37:437-439. 157. Mosher TJ, Liu Y, Torok CM: Functional cartilage MRI T2 mapping: evaluating the effect of age and training on knee cartilage response to running. Osteoarthritis Cartilage 2010, 18:358-364. 158. Neidhart M, Müller-Ladner U, Frey W, Bosserhoff AK, Colombani PC, Frey- Rindova P, Hummel KM, Gay RE, Häuselmann H, Gay S: Increased serum levels of non-collagenous matrix proteins (cartilage oligomeric matrix protein and melanoma inhibitory activity) in marathon runners. Osteoarthritis Cartilage 2000, 8:222-229. 159. Kim HJ, Lee YH, Kim CK: Biomarkers of muscle and cartilage damage and inflammation during a 200 km run. Eur J Appl Physiol 2007, 99:443-447. 160. Kim HJ, Lee YH, Kim CK: Changes in serum cartilage oligomeric matrix protein (COMP), plasma CPK and plasma hs-CRP in relation to running distance in a marathon (42.195 km) and an ultra-marathon (200 km) race. Eur J Appl Physiol 2009, 105:765-770. 161. Addison S, Coleman RE, Feng S, McDaniel G, Kraus VB: Whole-body bone scintigraphy provides a measure of the total-body burden of osteoarthritis for the purpose of systemic biomarker validation. Arthritis Rheum 2009, 60:3366-3373. 162. Liphardt AM, Mündermann A, Koo S, Bäcker N, Andriacchi TP, Zange J, Mester J, Heer M: Vibration training intervention to maintain cartilage thickness and serum concentrations of cartilage oligometric matrix protein (COMP) during immobilization. Osteoarthritis Cartilage 2009, 17:1598-1603. 163. Berry PA, Maciewicz RA, Wluka AE, Downey-Jones MD, Forbes A, Hellawell CJ, Cicuttini FM: Relationship of serum markers of cartilage metabolism to imaging and clinical outcome measures of knee joint structure. Ann Rheum Dis 2010, 69:1816-1822. 164. Niehoff A, Kersting UG, Helling S, Dargel J, Maurer J, Thevis M, Brüggemann GP: Different mechanical loading protocols influence serum cartilage oligomeric matrix protein levels in young healthy humans. Eur J Appl Physiol 2010, 110:651-657. 165. Pelletier JP, Raynauld JP, Caron J, Mineau F, Abram F, Dorais M, Haraoui B, Choquette D, Martel-Pelletier J: Decrease in serum level of matrix metalloproteinases is predictive of the disease-modifying effect of Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 32 of 33 osteoarthritis drugs assessed by quantitative MRI in patients with knee osteoarthritis. Ann Rheum Dis 2010, 69:2118-2124. 166. Knechtle B, Knechtle P, Rosemann T: Upper body skinfold thickness is related to race performance in male Ironman triathletes. Int J Sports Med 201, 32:20-27. 167. al’Absi M, Petersen KL, Wittmers LE: Adrenocortical and hemodynamic predictors of pain perception in men and women. Pain 2002, 96:197-204. 168. Thamer C, Machann J, Bachmann O, Haap M, Dahl D, Wietek B, Tschritter O, Niess A, Brechtel K, Fritsche A, Claussen C, Jacob S, Schick F, Häring HU, Stumvoll M: Intramyocellular lipids: anthropometric determinants and relationships with maximal aerobic capacity and insulin sensitivity. J Clin Endocrinol Metab 2003, 88:1785-1791. 169. Machann J, Etzel M, Thamer C, Haring HU, Claussen CD, Fritsche A, Schick F: Morning to evening changes of intramyocellular lipid content in dependence on nutrition and physical activity during one single day: a volume selective 1H-MRS study. MAGMA 2011, 24:29-33. Pre-publication history The pre-publication history for this paper can be accessed here: http://www.biomedcentral.com/1741-7015/10/78/prepub doi:10.1186/1741-7015-10-78 Cite this article as: Schütz et al.: The Transeurope Footrace Project: longitudinal data acquisition in a cluster randomized mobile MRI observational cohort study on 44 endurance runners at a 64-stage 4,486km transcontinental ultramarathon. BMC Medicine 2012 10:78. Submit your next manuscript to BioMed Central and take full advantage of: • Convenient online submission • Thorough peer review • No space constraints or color figure charges • Immediate publication on acceptance • Inclusion in PubMed, CAS, Scopus and Google Scholar • Research which is freely available for redistribution Submit your manuscript at www.biomedcentral.com/submit Schütz et al. BMC Medicine 2012, 10:78 http://www.biomedcentral.com/1741-7015/10/78 Page 33 of 33

The TransEurope FootRace Project: longitudinal data acquisition in a cluster randomized mobile MRI observational cohort study on 44 endurance runners at a 64-stage 4,486 km transcontinental ultramarathon.

07-19-2012

Schütz, Uwe H W,Schmidt-Trucksäss, Arno,Knechtle, Beat,Machann, Jürgen,Wiedelbach, Heike,Ehrhardt, Martin,Freund, Wolfgang,Gröninger, Stefan,Brunner, Horst,Schulze, Ingo,Brambs, Hans-Jürgen,Billich, Christian

eng

PMC9918280

Citation: Battistini, J.I.; Mastrorilli, V.; Nicolis di Robilant, V.; Saraulli, D.; Marinelli, S.; Farioli Vecchioli, S. Role of Running-Activated Neural Stem Cells in the Anatomical and Functional Recovery after Traumatic Brain Injury in p21 Knock-Out Mice. Int. J. Mol. Sci. 2023, 24, 2911. https://doi.org/10.3390/ijms24032911 Academic Editors: Daniele Bottai and Ola Hermanson Received: 8 December 2022 Revised: 28 January 2023 Accepted: 30 January 2023 Published: 2 February 2023 Copyright: © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/). International Journal of Molecular Sciences Article Role of Running-Activated Neural Stem Cells in the Anatomical and Functional Recovery after Traumatic Brain Injury in p21 Knock-Out Mice Jonathan Isacco Battistini 1 , Valentina Mastrorilli 2, Vittoria Nicolis di Robilant 3, Daniele Saraulli 4, Sara Marinelli 1 and Stefano Farioli Vecchioli 1,* 1 Institute of Biochemistry and Cell Biology, Institute of Biochemistry and Cell Biology, National Research Council (IBBC/CNR), Monterotondo, 00015 Rome, Italy 2 Plaisant S.R.L., 00128 Rome, Italy 3 Experimental Translational Oncology Department at Menarini Ricerche, Pomezia, 00071 Rome, Italy 4 Department of Law, Economics, Politics and Modern Languages, LUMSA University, 00193 Rome, Italy * Correspondence: [email protected] Abstract: Traumatic brain injury (TBI) represents one of the most common worldwide causes of death and disability. Clinical and animal model studies have evidenced that TBI is characterized by the loss of both gray and white matter, resulting in brain atrophy and in a decrease in neurological function. Nowadays, no effective treatments to counteract TBI-induced neurological damage are available. Due to its complex and multifactorial pathophysiology (neuro-inflammation, cytotoxicity and astroglial scar formation), cell regeneration and survival in injured brain areas are strongly hampered. Recently, it has been proposed that adult neurogenesis may represent a new approach to counteract the post- traumatic neurodegeneration. In our laboratory, we have recently shown that physical exercise induces the long-lasting enhancement of subventricular (SVZ) adult neurogenesis in a p21 (negative regulator of neural progenitor proliferation)-null mice model, with a concomitant improvement of olfactory behavioral paradigms that are strictly dependent on SVZ neurogenesis. On the basis of this evidence, we have investigated the effect of running on SVZ neurogenesis and neurorepair processes in p21 knock-out mice that were subject to TBI at the end of a 12-day session of running. Our data indicate that runner p21 ko mice show an improvement in numerous post-trauma neuro-regenerative processes, including the following: (i) an increase in neuroblasts in the SVZ; (ii) an increase in the migration stream of new neurons from the SVZ to the damaged cortical region; (iii) an enhancement of new differentiating neurons in the peri-lesioned area; (iv) an improvement in functional recovery at various times following TBI. All together, these results suggest that a running-dependent increase in subventricular neural stem cells could represent a promising tool to improve the endogenous neuro-regenerative responses following brain trauma. Keywords: subventricular zone; adult neurogenesis; neural stem cells; traumatic brain injury; p21 1. Introduction Traumatic brain injury (TBI) is one of the most common causes of death and disability in young people [1]. TBI severity is assessed on the basis of the Glasgow Coma Scale (GCS) score [2], with further modifications and can be graded as mild, moderate or severe [3,4]. The symptoms of mild patients are short-term memory and concentration difficulties and they usually show complete neurological recovery [5]. Moderate patients are lethargic and stuporous, whilst severe subjects are comatose, unable to open their eyes or follow commands [6]. Severe patients also have a higher risk of hypotension, hypoxemia and brain swelling, all effects that, if not prevented, can cause complications and lead to death [7,8]. Moreover, TBI strongly increases a patient’s susceptibility to neurodegenerative diseases such as Alzheimer’s and Parkinson’s disease [9,10]. Int. J. Mol. Sci. 2023, 24, 2911. https://doi.org/10.3390/ijms24032911 https://www.mdpi.com/journal/ijms Int. J. Mol. Sci. 2023, 24, 2911 2 of 24 The adult mammalian brain is able to respond to damage with structural and func- tional modifications and regeneration mechanisms. In this context, the role of the endoge- nous neurogenic post-traumatic response is the subject of a wide range of studies [11,12]. Under physiological conditions, the neural stem cells (NSCs, named type B cells) that reside in the SVZ represent a population of relatively quiescent cells [13,14], which give rise, in the course of neurogenic differentiation, to the following two classes of cells: type C cells that have a high proliferative rate [15], which in turn give rise to neuroblasts called type A cells [16]. These cells exit the cell cycle and migrate along the rostral migratory stream to reach the olfactory bulb, where they mature into inhibitory GABAergic neurons [17,18]. Af- ter brain injury, endogenous neural stem cells can be activated to modulate the timing and rate of proliferation/differentiation, with the aim of facilitating brain repair [19,20]. This process is also finely regulated by a series of changes in the environment of the neurogenic niche, such as an increase in vasculature permeability that favors the migration of NSCs and neuroblasts to the injured cortex [21]. Furthermore, transcriptomic studies based on single-cell RNA sequencing (RNA-seq) have shown a strong increase in the post-injured SVZ of the proportion of “primed” quiescent and active NSCs, which highly express genes related to protein synthesis and cell cycle regulation [22]. These findings highlighted the existence of reactive SVZ NSCs capable of conferring protection following TBI, suggesting an important role of the activation of SVZ NSCs in providing beneficial outcomes for post-TBI brain repair. The p21Waf1/Cip1 gene represents one of the main regulators of the cell cycle and plays a primary role in modulating the transition between quiescence and activation of NSCs in adult neurogenic niches [23]. This gene is part of the Cip/Kip family of cyclin-dependent kinase inhibitors (CKIs), which also include the p27 and p57 genes with the function of neg- atively regulating cell cycle progression [24,25]. Within neurogenic niches, p21 expression correlates with the maintenance of NSCs in a quiescent state and with the restraining of progenitor proliferation [26–28]. Constitutive deletion of the p21 gene in mouse models has led to the rapid and powerful activation in the cell cycle of quiescent NSCs at the post-natal stage, with a consequent reduction in the self-renewal capacity of NSCs, the onset of replicative stress in the hyper-proliferating NSCs and progenitors and a reduction in neurogenesis in adult mice [28–31]. Previous work by our group demonstrated that in the absence of the p21 gene, SVZ neurogenesis and olfactory behavior are significantly enhanced by 12 days of voluntary running. These results strongly indicate that p21-null NSCs retain their high neurogenic potential, which is specifically triggered by physical activity [28]. Based on these findings, in this project, we investigated the possibility that the enhancement of SVZ neurogenesis that occurs in a p21-null mouse model that undergoes a 12 day-session of voluntary running could be an effective mechanism that contributes to the neuroanatomical and functional recovery processes following pathological conditions such as TBI. Our results demonstrate that at different time-points following TBI, the combination of running and of p21 knockdown induces an increase in the migration of SVZ NSCs and progenitors toward the cortical lesion, an enhancement of new differentiating neurons in the peri-lesioned area and partial functional recovery in the injured mice. These data suggest a potential model for the strengthening of post-traumatic endogenous neurogenesis that is capable of effectively counteracting the anatomical and functional dysfunctions induced by cortical damage and accelerating neurorepair processes. 2. Results The aim of this work is to analyze the post-traumatic neurogenic effects of a running session of 12 days in a mouse model with the deletion of p21. The 12-day running paradigm was chosen in accordance with previous data that demonstrate that this running protocol provokes the peak of running-dependent increments of neurogenesis in the SVZ of p21 ko mice [28]. The experimental procedure is detailed in the Material and Method section and is shown in Figure 1. Int. J. Mol. Sci. 2023, 24, 2911 3 of 24 paradigm was chosen in accordance with previous data that demonstrate that this running protocol provokes the peak of running-dependent increments of neurogenesis in the SVZ of p21 ko mice [28]. The experimental procedure is detailed in the Material and Method section and is shown in Figure 1. Figure 1. Graphical representation of experimental procedures. (A) Experimental timeline. The animals have been subjected to a running session for 12 days. From day 7 of the running session until the day of the lesion, BrdU is administered in the drinking water of the animals. On the 12th day, the mice undergo the CCI surgical procedure. The functional outcome has been evaluated by the Ladder Rung Walking Task (LWT) 24 h before CCI (pre-TBI) to determine the baseline number of errors, and two, seven, fourteen and thirty-three days after the surgery (P2, P7, P14 and P30). At P7, P14 and P30, the animals have been sacrificed to perform immunofluorescence assays (IF). (B) Figure 1. Graphical representation of experimental procedures. (A) Experimental timeline. The animals have been subjected to a running session for 12 days. From day 7 of the running session until the day of the lesion, BrdU is administered in the drinking water of the animals. On the 12th day, the mice undergo the CCI surgical procedure. The functional outcome has been evaluated by the Ladder Rung Walking Task (LWT) 24 h before CCI (pre-TBI) to determine the baseline number of errors, and two, seven, fourteen and thirty-three days after the surgery (P2, P7, P14 and P30). At P7, P14 and P30, the animals have been sacrificed to perform immunofluorescence assays (IF). (B) Coronal section of a lesioned brain. To evaluate the cellular processes that occur in the brain after the trauma, we considered the following: the ipsi- and contralateral SVZ, the migratory route and the low, medial and lateral sides of the lesioned cortex. Scale bar = 400 µm. Int. J. Mol. Sci. 2023, 24, 2911 4 of 24 2.1. TBI Induces the Significant Activation of Type B NSCs in the SVZ of Injured Mice To understand the impact of TBI on type B Glia-like NSC recruitment and proliferation, we analyze, at different time points from the trauma (7, 14 and 30 days), the ipsi- and contralateral SVZ of mice subjected to brain injury or in the SHAM condition. The NSC sub-population was identified through the co-localization of the GFAP marker and the Nestin GFP transgene. The data show a strong increase in the SVZ of both hemispheres of type B cell recruit- ment from quiescence in the TBI groups, in comparison with their SHAM counterpart at 7 days post TBI (recruitment: ratio of Ki67+NestinGFP+GFAP+ cells/NestinGFP+GFAP+ total cells, ipsilateral: p < 0.001, Figure 2A–C, Supplementary Figure S1A–H; contralateral: p < 0.001, Supplementary Figure S2A) and their proliferation (Ki67+NestinGFP+GFAP+ cells, ipsilateral: p < 0.001, Figure 2A,B,D; contralateral: p < 0.001, Supplementary Figure S2B), as well as 14 days after injury (recruitment: ipsilateral, p < 0.001, Figure 2E, Supplementary Figure S3A–H; contralateral: p < 0.001, Supplementary Figure S2C; proliferation: ipsilateral: p < 0.001, Figure 2F, Supplementary Figure S3A–H; contralateral: p < 0.001, Supplementary Figure S2D). After 30 days from TBI, we observe in the SVZ of both hemispheres a main lesion effect of the type B cells’ pool of TBI groups compared to the SHAM animals, due to the strong increase observed in the KO TBI and KO RUN TBI mice, compared to the respective SHAM mice (NestinGFP+GFAP+ cells, ipsilateral: p < 0.001, Figure 2G, Supplementary Figure S4A–H; contra-lateral: p = 0.002, Supplementary Figure S2E). These data let us hypothesize that TBI might induce the powerful activation of type B cells, in terms of recruitment from quiescence and proliferation, leading in the long term to an expansion of the pool compared to the SHAM groups. 2.2. Influence of p21 Deletion and Running Session on TBI-Induced NSC Activation Moreover, we evaluated the different responses to trauma within the injured groups, with the aim of assessing whether the deletion of p21 and/or the physical activity before the TBI might promote the neurogenic response. The comparative analysis within the four TBI groups (WT TBI, WT RUN TBI, KO TBI and KO RUN TBI) indicates that at 7 days post trauma, no changes were detectable in the recruitment or expansion of type B cells. Instead, after 14 days, we observe, in the contralateral SVZ, a TBI-dependent expansion of type B cells in the KO RUN TBI, with respect to the other groups in terms of the enhancement of type B proliferation (KO RUN TBI vs. WT TBI and KO TBI, p < 0.001, Supplementary Figure S2D). In addition, 30 days after the TBI, in the ipsilateral SVZ, no significant variations were detected, while in the contralateral SVZ, a strong expansion of type B pool size in the KO RUN TBI mice was detected (NestinGFP+GFAP+ cells: KO RUN TBI vs. WT TBI, WT RUN TBI and KO TBI, p < 0.001, Supplementary Figure S2E). From this first analysis, it emerges that the deletion of p21 and/or physical activity is ineffective in triggering the post-traumatic NSC response in the ipsilateral hemisphere, while a strong effect of running and p21 deletion on NSC activation is observed in the contralateral hemisphere. 2.3. Time-Course of Neural Stem Progenitor Cell (NSPC) Modulation after TBI The count of NestinGFP+ cells and their proliferating fraction (Ki67+ NestinGFP+ cells) allowed us to evaluate the variations in the stem cells and the transit amplifying progenitors (small fraction of type B and C cells), hereinafter collectively referred to as neural stem/progenitor cells (NSPCs). Int. J. Mol. Sci. 2023, 24, 2911 5 of 24 Int. J. Mol. Sci. 2023, 24, x FOR PEER REVIEW 5 of 25 Figure 2. Modulation of type B cells after TBI. (A,B) Representative images in coronal sections of the neurogenic processes that occur in the ipsilateral SVZ of p21 ko mice 7 days after TBI (KO TBI, (A)) and after SHAM (KO SHAM, (B)). The images show an increment in the pool of proliferating type B cells (Ki67+/GFAP+/NestinGFP+ cells) in the KO TBI animals, with respect to the KO SHAM mice at 7 days after the surgery (N = 5 mice/group). (C) Graph shows the increase in type B recruitment at 7 days post TBI in the ipsilateral SVZ of the mice subjected to injury in comparison to their SHAM groups (ratio of Ki67+ NestinGFP+ GFAP+ cells/NestinGFP+ GFAP+ total cells, ipsilateral: lesion effect: F(1,60) = 91.45, p < 0.001). (D) Graph shows the enhancement of type B proliferation in the ipsi-lateral SVZ of TBI groups (Ki67+ NestinGFP+ GFAP+ cells, ipsilateral: lesion effect: F(1,60) = 45.8 p < 0.001). (E) Histograms illustrate the increase in type B recruitment in the ipsilateral SVZ in mice after 14 days from TBI (lesion effect: F(1,56) = 188, p < 0.001, $). (F) Graphs indicate an increase in type B proliferation in the ipsi-lateral SVZ of mice subjected to TBI 14 days after the trauma (lesion effect: F(1,56) = 85.9, p < 0.001). (G) After 30 days from TBI, we observed a significant increase in the type B cell population in the TBI mice with respect to their SHAM groups (NestinGFP+ GFAP+ cells, ipsilateral: lesion effect: F(1,44) = 21.76, p < 0.001, $). Statistical significance of main lesion effect between SHAM and TBI Figure 2. Modulation of type B cells after TBI. (A,B) Representative images in coronal sections of the neurogenic processes that occur in the ipsilateral SVZ of p21 ko mice 7 days after TBI (KO TBI, (A)) and after SHAM (KO SHAM, (B)). The images show an increment in the pool of proliferating type B cells (Ki67+/GFAP+/NestinGFP+ cells) in the KO TBI animals, with respect to the KO SHAM mice at 7 days after the surgery (N = 5 mice/group). (C) Graph shows the increase in type B recruitment at 7 days post TBI in the ipsilateral SVZ of the mice subjected to injury in comparison to their SHAM groups (ratio of Ki67+ NestinGFP+ GFAP+ cells/NestinGFP+ GFAP+ total cells, ipsilateral: lesion effect: F(1,60) = 91.45, p < 0.001). (D) Graph shows the enhancement of type B proliferation in the ipsi-lateral SVZ of TBI groups (Ki67+ NestinGFP+ GFAP+ cells, ipsilateral: lesion effect: F(1,60) = 45.8 p < 0.001). (E) Histograms illustrate the increase in type B recruitment in the ipsilateral SVZ in mice after 14 days from TBI (lesion effect: F(1,56) = 188, p < 0.001, $). (F) Graphs indicate an increase in type B proliferation in the ipsi-lateral SVZ of mice subjected to TBI 14 days after the trauma (lesion effect: F(1,56) = 85.9, p < 0.001). (G) After 30 days from TBI, we observed a significant increase in the type B Int. J. Mol. Sci. 2023, 24, 2911 6 of 24 cell population in the TBI mice with respect to their SHAM groups (NestinGFP+ GFAP+ cells, ipsilateral: lesion effect: F(1,44) = 21.76, p < 0.001, $). Statistical significance of main lesion effect between SHAM and TBI groups: $ p < 0.001. Multifactorial analysis with the following three independent variables: genotype, treatment and running, followed by Fisher’s LSD post hoc tests. Magnification = 20×. Scale bar = 100 µm. SVZ = subventricular zone. LV = lateral ventricle. At 7 days post TBI, we observe in both ipsi- and contralateral regions a strong increase in NSPC proliferation in the WT and KO groups subjected to TBI, with respect to their SHAM groups (Ki67+NestinGFP+ cells, ipsilateral: WT TBI vs. WT SHAM and KO TBI vs. KO SHAM p = 0.015, Figure 3A,C, Supplementary Figure S5A,B,E,F; contralateral: p = 0.013, Supplementary Figure S6A), which induces a significant increase in total proliferation (Ki67+ cells, ipsilateral WT TBI vs. WT SHAM, p = 0.03, KO TBI vs. KO SHAM p < 0.001, Figure 3A,D; contralateral: p = 0.0013, Supplementary Figure S6B). On the other hand, in the KO RUN TBI group, we detect in the ipsi- and contralateral SVZ a net decrease, compared to the KO RUN SHAM group, of proliferating NSPCs (Ki67+NestinGFP+ cells, ipsilateral: KO RUN TBI vs. KO RUN SHAM, p = 0.009, Figure 3A,C, Supplementary Figure S5G,H; contralateral: KO RUN TBI vs. KO RUN SHAM p < 0.001, Supplementary Figure S6A) as well as of total proliferation (Ki67+ cells, ipsilateral: KO RUN TBI vs. KO RUN SHAM, p < 0.001, Figure 3A,D, Supplementary Figure S5G,H; contralateral: KO RUN TBI vs. KO RUN SHAM p < 0.001, Supplementary Figure S6B). The analysis carried out 14 days post TBI demonstrates a significant decrease in the SVZ of both hemispheres of the NestinGFP+ cell pool in all the TBI groups compared to their SHAM counterparts (ipsilateral: p < 0.001, Figure 3B,E; contralateral: p = 0.0011, Supplementary Figure S6C). Moreover, we observe in the ipsilateral hemisphere of the WT TBI and KO RUN TBI groups a sharp decrease in total proliferation, compared to the respective WT SHAM and KO RUN SHAM groups (Ki67+ cells: WT TBI vs. WT SHAM, p = 0.029; KO RUN TBI vs. KO RUN SHAM p < 0.001, Figure 3B,F), as well as in the number of Ki67+ NestinGFP+ cells (WT TBI vs. WT SHAM, p = 0.06; KO RUN TBI vs. KO RUN SHAM p = 0.04). After 30 days from the TBI, we observed the main lesion effect on the decrease in NestinGFP+ cells in the ipsilateral SVZ of the TBI groups compared to the SHAM mice (p = 0.003, Figure 3G). These data demonstrate the initial expansion of NestinGFP+ cells in the WT TBI and KO TBI mice, while a decrease in NSPCs is observed in the KO RUN TBI group with respect to the KO RUN SHAM mice, which display a powerful increase in SVZ neurogenesis, as previously shown [28]. Later, the decline in NestinGFP+ cells also becomes evident in the other TBI groups compared to their SHAM counterparts. 2.4. Influence of p21 Deletion and Running Session on NSPC Regulation after TBI The analysis within the injured groups did not reveal at 7 days post TBI any significant differences in the NestinGFP+ sub-populations. In addition, 14 days after TBI, we observe in the ipsilateral SVZ an increase in the proliferating Nestin GFP+ and in the total proliferation in the WT RUN TBI, KO TBI and KO RUN TBI group compared to the WT TBI group (WT TBI vs. WT RUN TBI, p = 0.01, vs. KO TBI and KO RUN TBI p < 0.001; Ki67+ cells: WT TBI vs. WT RUN TBI, p = 0.01, vs. KO TBI, and KO RUN TBI, p < 0.001, Figure 3F). In the contralateral SVZ, we observe a strong proliferative response of NestinGFP+ cells in the KO RUN TBI group, compared to the WT TBI mice (Ki67+ NestinGFP+: KO RUN TBI vs. WT TBI, p = 0.024, Supplementary Figure S6D), leading to an increase in NestinGFP+ cell pool size (KO RUN TBI vs. KO TBI, p = 0.002, Supplementary Figure S6C), and consequently in total proliferation (KO RUN TBI vs. WT TBI, p = 0.0048, vs. KO TBI, p = 0.006, Supplementary Figure S6E). At 30 days post TBI, we did not find any difference within the injured groups in either of the hemispheres. Int. J. Mol. Sci. 2023, 24, 2911 7 of 24 These data demonstrate the initial expansion of NestinGFP+ cells in the WT TBI and KO TBI mice, while a decrease in NSPCs is observed in the KO RUN TBI group with respect to the KO RUN SHAM mice, which display a powerful increase in SVZ neurogenesis, as previously shown [28]. Later, the decline in NestinGFP+ cells also becomes evident in the other TBI groups compared to their SHAM counterparts. Figure 3. Time course of NSPCs and type A neuroblast following TBI. (A) Micrographs show the decreased proliferation of neural stem/progenitor cells (NSPCs, Ki67+/NestinGFP+ cells) in the KO TBI RUN group with respect to the KO TBI mice at 7 days post TBI. (B) Representative pictures indicate the significantly decreased number of Ki67+/NestinGFP+ cells in the KO TBI RUN mice in comparison to their respective SHAM littermate 14 days after TBI. (C) Histograms show a significant increase in the proliferating NSPCs of WT TBI and KO TBI mice in comparison with their respective WT SHAM and KO groups (Ki67+ NestinGFP+ cells, ipsilateral: genotype x run x lesion interaction: F(1,112) = 11.71, p < 0.001, followed by LSD post-test, WT TBI vs. WT SHAM and KO TBI vs. KO SHAM p = 0.015). (D) Graph indicates a significant increase in the total proliferation of WT TBI and KO TBI mice with respect to the WT and KO SHAM groups (Ki67+ cells, genotype x run x lesion interaction, ipsilateral: F(1,108) = 14.38, p < 0.001, followed by LSD post-test, WT TBI vs. WT SHAM, p = 0.03, KO TBI vs. KO SHAM p < 0.001). (C,D) It is also possible to observe the decrease in the NSPCs proliferation of KO RUN TBI mice with respect to the KO RUN SHAM group (KO RUN TBI vs. KO RUN SHAM, p < 0.001, (C)) and total proliferation (KO RUN TBI vs. KO RUN SHAM, p < 0.001, (D)). Figure 3. Time course of NSPCs and type A neuroblast following TBI. (A) Micrographs show the decreased proliferation of neural stem/progenitor cells (NSPCs, Ki67+/NestinGFP+ cells) in the KO TBI RUN group with respect to the KO TBI mice at 7 days post TBI. (B) Representative pictures indicate the significantly decreased number of Ki67+/NestinGFP+ cells in the KO TBI RUN mice in comparison to their respective SHAM littermate 14 days after TBI. (C) Histograms show a significant increase in the proliferating NSPCs of WT TBI and KO TBI mice in comparison with their respective WT SHAM and KO groups (Ki67+ NestinGFP+ cells, ipsilateral: genotype x run x lesion interaction: F(1,112) = 11.71, p < 0.001, followed by LSD post-test, WT TBI vs. WT SHAM and KO TBI vs. KO SHAM p = 0.015). (D) Graph indicates a significant increase in the total proliferation of WT TBI and KO TBI mice with respect to the WT and KO SHAM groups (Ki67+ cells, genotype x run x lesion interaction, ipsilateral: F(1,108) = 14.38, p < 0.001, followed by LSD post-test, WT TBI vs. WT SHAM, p = 0.03, KO TBI vs. KO SHAM p < 0.001). (C,D) It is also possible to observe the decrease in the NSPCs proliferation of KO RUN TBI mice with respect to the KO RUN SHAM group (KO RUN TBI vs. KO RUN SHAM, p < 0.001, (C)) and total proliferation (KO RUN TBI vs. KO RUN SHAM, p < 0.001, (D)). (E) Graph displays the reduced proliferation of NSPCs in mice analyzed 14 days after TBI, when compared to their respective SHAM counterparts (NestinGFP+ cells, ipsilateral: lesion effect: F(1,105) = 17.21, p < 0.001, $). (F) Histogram shows the decreased total proliferation after 14 days Int. J. Mol. Sci. 2023, 24, 2911 8 of 24 after TBI in the WT TBI and KO RUN TBI mice with respect to their SHAM littermates (Ki67+ cells: genotype x run x lesion interaction: F(1,99) = 17.04, p < 0.001, followed by LSD post-test, WT TBI vs. WT SHAM, p = 0.029; KO RUN TBI vs. KO RUN SHAM p < 0.001). In the TBI groups, we detected a significant increase in proliferation in the WT RUN TBI mice (F(1,108) = 5.84, p = 0.017, followed by LSD post-test, WT TBI vs. WT RUN TBI, p = 0.01), as well as an increase in the KO TBI and KO RUN TBI mice (p < 0.01), in comparison with the WT TBI group. (G) Graph shows the decreased NSPC population in the ipsilateral SVZ of TBI mice with respect to their SHAM genotypes, 30 days after the TBI (NestinGFP+ cells: lesion effect: F(1,89) = 9.2, p = 0.003, #). (H) Histogram indicates at 30 days post TBI the increased number of neuroblasts (DCX+ cells) in the TBI group in comparison with their SHAM littermates (DCX+ cells, ipsilateral, lesion effect: F(1,53) = 20.7, p < 0.001, $). The comparison within the TBI groups (black histograms) show a significant increase in neuroblasts in the KO RUN TBI mice in comparison with WT TBI, WT RUN and KO TBI mice (genotype x run interaction: F(1,53) = 7.77, p = 0.007, followed by LSD post-test, KO RUN TBI vs. WT TBI = 0.015, vs. WT RUN TBI < 0.001, vs. KO TBI = 0.001, a, b, c, respectively). N = 5 mice/group. Statistical significance of LSD post hoc analysis: * p < 0.05, ** p < 0.01 and *** p < 0.001. Statistical significance of main lesion effect between SHAM and TBI groups: $ p < 0.001 and # p < 0.01. Multifactorial analysis with the following three independent variables: genotype, treatment and running, followed by Fisher’s LSD post hoc tests. Magnification = 20×. Scale bar = 100 µm. SVZ = subventricular zone. LV = lateral ventricle. These data suggest a transitory increase 14 days after the TBI in contralateral sub- ventricular cell proliferation, which is dependent on physical activity and the lack of the p21 gene. 2.5. TBI Triggers a Late Increase in Type A Neuroblasts The analysis of the TBI-induced modulation of type A neuroblasts was carried out by using the specific marker DCX. At 7 and 14 days post TBI in the ipsi-lateral region, we did not detect any significant variation in DCX+ neuroblasts in the injured groups compared to the corresponding SHAM groups, while in the contralateral SVZ, our data showed at 7 days post TBI a lesion effect on the increase in neuroblasts of the injured groups, compared to their SHAM counterparts (DCX+ cells: p = 0.004, Supplementary Figure S7A); after 14 days post TBI, we observed in the contralateral SVZ a lesion effect on the decrease in DCX+ cells of the groups subjected to TBI (p < 0.029, Supplementary Figure S7B). At 30 days post TBI, we observe a strong increase in both hemispheres in the number of neuroblasts in the TBI groups, compared to the corresponding SHAM groups (DCX+ cells, ipsilateral: p < 0.001, Figures 3H and 4A,B, Supplementary Figure S8A–H; contra-lateral: p < 0.001, Supplementary Figure S7C). Collectively, these data highlight how the neuroblast population tends to respond late to TBI, through an increase of the cell number in the ipsilateral SVZ at 30-days following the trauma, compared to the SHAM groups. 2.6. Influence of p21 Deletion and Running Session on TBI-Induced Neuroblast Modulation The analysis between the injured groups at 7 and 14 days post TBI does not show any TBI-dependent modulation of DCX+ cells in either of the ipsi- and the contralateral region. At 30 days after the trauma, on the other hand, we find an increase in neuroblasts in the ipsilateral SVZ of the KO RUN TBI group, compared to the other groups (KO RUN TBI vs. WT TBI = 0.015, vs. WT RUN TBI and KO TBI < 0.001, a, b, c, respectively, Figures 3H and 4B,C, Supplementary Figure S8B,D,F,H). Int. J. Mol. Sci. 2023, 24, 2911 9 of 24 2.6. Influence of p21 Deletion and Running Session on TBI Induced Neuroblast Modulation The analysis between the injured groups at 7 and 14 days post TBI does not show any TBI-dependent modulation of DCX+ cells in either of the ipsi- and the contralateral region. At 30 days after the trauma, on the other hand, we find an increase in neuroblasts in the ipsilateral SVZ of the KO RUN TBI group, compared to the other groups (KO RUN TBI vs. WT TBI = 0.015, vs. WT RUN TBI and KO TBI < 0.001, a, b, c, respectively, Figures 3H and 4B,C, Supplementary Figure S8B,D,F,H). Figure 4. (A–C). Increase in DCX+ cells in the KO RUN TBI group. Confocal representative micro- graphs show the increased number of DCX+ cells in the KO RUN TBI mice (C), in comparison with the KO SHAM (A) and KO TBI (B) mice, 30 days after TBI. Magnification = 20×. Scale bar = 100 µm. SVZ = subventricular zone. LV = lateral ventricle. 2.7. TBI Triggers a Boost in Migration toward the Peri-Lesion Cortical Region of KO RUN Mice The next step was to study the migratory flow of new cells that originated in the SVZ and were redirected to the damage site in an attempt to contribute to the tissue repair Int. J. Mol. Sci. 2023, 24, 2911 10 of 24 process. For this study, in the four groups subjected to trauma, both the NSPCs (Nestin GFP+) and neuroblasts (DCX+) in their path from the SVZ to the injured cortex were analyzed. We also analyzed the proliferative fraction of NSPCs and neuroblasts. The data obtained show that 7 days after TBI in the KO RUN TBI group, there is a significant increase in both NestinGFP+ and Ki67+NestinGFP+ cells within the migration region towards the injured cortex, compared to the other groups (NestinGFP+: KO RUN TBI vs. WT TBI, p = 0.017, vs. WT RUN TBI, p = 0.031 and vs. KO TBI p = 0.027, Figure 5A,I,J; Ki67+ NestinGFP+ cells: KO RUN TBI vs. WT TBI, p = 0.027, vs. WT RUN TBI, p = 0.038 and vs. KO TBI p = 0.049, Figure 5B,I,J). Int. J. Mol. Sci. 2023, 24, x FOR PEER REVIEW 11 of 25 Figure 5. Cell migration following TBI. (A) Graph shows the increase in the migratory stream (MS) of the KO RUN TBI mice in comparison to the other groups 7 days after the lesion of NestinGFP+ cells (genotype x running: F(1.18) = 5.98, p = 0.025, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.017, vs. WT RUN TBI, p = 0.031 and vs. KO TBI p = 0.027). (B) Graph shows the increase in proliferating NSPCs in the migratory stream (MS) of the KO RUN TBI mice 7 days after the lesion (Ki67+ Nestin GFP+ cells: genotype x running: F(1,18) = 5.51, p = 0.03, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.027, vs. WT RUN TBI, p = 0.038 and vs. KO TBI p = 0.049). (C) Histograms illustrate the enhancement in the MS of KO RUN TBI mice with respect to the other experimental condition 14 days after the TBI of NestinGFP+ cells (genotype x run interaction: F(1,18) = 6.55, p = 0.019, followed by LSD post-test, KO RUN TBI vs. WT TBI, p < 0.001, vs. WT RUN TBI, p = 0.01 and vs. KO TBI, p = 0.023). (D) Graph shows the increment, with respect to the other experimental conditions of DCX+ cells, in the MS of KO RUN TBI 14 days after TBI (genotype x run interaction: F(1,18) = 7.07, p = 0.016, followed by LSD post-test, KO RUN TBI vs. WT RUN TBI, p = 0.024 and vs. KO TBI, p = 0.02). (E) Graph shows the increase in the MS of KO RUN TBI mice with respect to the other groups 14 days after the TBI of proliferating NestinGFP+ (Ki67+/NestinGFP+: genotype x run interaction: F(1,18) = 5.07, p = 0.037, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.007, WT RUN TBI, p = 0.038 and vs. KO TBI, p = 0.0073). (F) Graph shows the increments, with respect to the other experimental conditions of proliferating DCX+ cells, in the MS of KO RUN TBI 14 days after TBI (Ki67+ DCX+ cells: genotype x run interaction: F(1,18) = 7.07, p = 0.016, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.004, WT RUN TBI, p = 0.008 and vs. KO TBI, p = 0.0018). (G) Graph indicates an increase in migrating NestinGFP+ cells in the KO TBI and KO TBI RUN mice with respect to the WT TBI mice 30 days after TBI (genotype effect: F(1 25) = 3 34 p = 0 011 ^) (H) Figure 5. Cell migration following TBI. (A) Graph shows the increase in the migratory stream (MS) of the KO RUN TBI mice in comparison to the other groups 7 days after the lesion of NestinGFP+ cells (genotype x running: F(1.18) = 5.98, p = 0.025, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.017, vs. WT RUN TBI, p = 0.031 and vs. KO TBI p = 0.027). (B) Graph shows the increase in proliferating NSPCs in the migratory stream (MS) of the KO RUN TBI mice 7 days after the lesion (Ki67+ Nestin GFP+ cells: genotype x running: F(1,18) = 5.51, p = 0.03, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.027, vs. WT RUN TBI, p = 0.038 and vs. KO TBI p = 0.049). (C) Histograms illustrate the enhancement in the MS of KO RUN TBI mice with respect to the other experimental condition 14 days after the TBI of NestinGFP+ cells (genotype x run interaction: F(1,18) = 6.55, p = 0.019, followed by LSD post-test, KO RUN TBI vs. WT TBI, p < 0.001, vs. WT RUN TBI, p = 0.01 and vs. KO TBI, p = 0.023). (D) Graph shows the increment, with respect to the other experimental conditions of DCX+ cells, in the MS of KO RUN TBI 14 days after TBI (genotype x run interaction: F(1,18) = 7.07, p = 0.016, followed by LSD post-test, KO RUN TBI vs. WT RUN TBI, p = 0.024 and vs. KO TBI, p = 0.02). (E) Graph shows the increase in the MS of KO RUN TBI mice Int. J. Mol. Sci. 2023, 24, 2911 11 of 24 with respect to the other groups 14 days after the TBI of proliferating NestinGFP+ (Ki67+/NestinGFP+: genotype x run interaction: F(1,18) = 5.07, p = 0.037, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.007, WT RUN TBI, p = 0.038 and vs. KO TBI, p = 0.0073). (F) Graph shows the increments, with respect to the other experimental conditions of proliferating DCX+ cells, in the MS of KO RUN TBI 14 days after TBI (Ki67+ DCX+ cells: genotype x run interaction: F(1,18) = 7.07, p = 0.016, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.004, WT RUN TBI, p = 0.008 and vs. KO TBI, p = 0.0018). (G) Graph indicates an increase in migrating NestinGFP+ cells in the KO TBI and KO TBI RUN mice with respect to the WT TBI mice 30 days after TBI (genotype effect: F(1,25) = 3.34, p = 0.011, ˆ). (H) Histograms show an enhancement of migrating DCX+ cells in the KO TBI and KO TBI RUN group with respect to the WT TBI mice 30 days after TBI (genotype effect: F(1.25) = 12.64, p = 0.0015, #). (I,J) Confocal micrographs show the increased density of migrating NestinGFP+ cells observed in the KO RUN TBI mice with respect to the WT TBI mice, 7 days from TBI. (K,L) Confocal micrographs illustrate the enhancement of NSPCs (NestinGFP+) and neuroblasts (DCX+) along the MS of the KO RUN TBI mice compared to the WT TBI group, 14 days post TBI. Arrow indicates the presence of proliferating NestinGFP+ cells and arrowheads indicate proliferating DCX+ cells. N = 5 mice/group. Statistical significance of LSD post hoc analysis: * p < 0.05, ** p < 0.01 and *** p < 0.001. Statistical significance of main genotype effect between WT and KO groups: # p < 0.01, ˆ p < 0.05. Two-way ANOVA analysis followed by Fisher’s LSD post hoc tests. Magnification = 20×. Scale bar = 100 µm. SVZ = subventricular zone. MS = migratory stream. In addition, 14 days after the TBI, our data show in the migratory stream (MS) of the KO RUN TBI group a significant increase, compared to the other experimental groups, in NestinGFP+ cells (KO RUN TBI vs. WT TBI, p < 0.001, vs. WT RUN TBI, p = 0.01 and vs. KO TBI, p = 0.023, Figure 5C,K,L) and DCX+ cells (KO RUN TBI vs. WT RUN TBI, p = 0.024 and vs. KO TBI, p = 0.02, Figure 5D,K,L). We also observed a marked enhancement in proliferating migrating cells in the KO RUN TBI group, both in terms of total proliferation (KO RUN TBI vs. WT TBI, p = 0.046, vs. WT RUN TBI, p = 0.04 and vs. KO TBI, p = 0.0028), and regarding the proliferative rate of Nestin GFP+ cells (Ki67+ NestinGFP+: KO RUN TBI vs. WT TBI, p = 0.007, vs. WT RUN TBI, p = 0.038 and vs. KO TBI, p = 0.0073, Figure 5E,K,L), and neuroblasts (Ki67+ DCX+ cells: KO RUN TBI vs. WT TBI, p = 0.004, vs. WT RUN TBI, p = 0.008 and vs. KO TBI, p = 0.0018, Figure 5F,K,L). At 30 days post TBI, we observe a significant genotype effect with an increase in the p21 KO groups in density within the MS of Nestin GFP+ and DCX+ cells, compared to the WT animals (NestinGFP+: p = 0.011, Figure 5G; DCX+: p = 0.0015, Figure 5H). Altogether, these data suggest that voluntary physical activity promotes widespread TBI-dependent migration from SVZ toward the injured cortex of NSCs and neuroblasts in mice lacking the p21 gene, likely enhancing the neuroprotective and regenerative processes that occur after the lesion. 2.8. p21 Deletion and Running Session Strongly Influence the TBI-Induced New-Born Cell Localization in the Peri-Lesion Cortex Moreover, we characterized the presence of newborn cells, likely derived from the SVZ, in the damage site 7, 14 and 30 days after TBI. To this aim, we considered three different areas of the damage, which were as follows: (i) one area lateral to the lesion (lateral); (ii) one on the medial side (medial); (iii) the last on the central lower edge of the lesion (low). In these regions, we counted the total number of NestinGFP+ and DCX+, as well as the number of NestinGFP+ and DCX+ cells co-expressing BrdU, assuming that most of these cells that express BrdU might originate in and derive from the SVZ through the migratory redirection previously analyzed. At 7 days post TBI, we observe a substantial increase in BrdU+ NestinGFP+ and BrdU+ DCX+ cells in the KO RUN TBI group of animals, compared to the other experimental conditions both in the lateral region (BrdU+ Nestin GFP+: KO RUN TBI vs. WT TBI, p = 0.02, vs. WT RUN TBI, p = 0.041 and vs. KO TBI p = 0.016, Figure 6A,K,L; BrdU+ DCX+: KO RUN TBI vs. WT TBI, p = 0.026, vs. WT RUN TBI, p = 0.015 and vs. KO TBI p = 0.011, Figure 6B,K,L), as well as in the medial peri-lesion Int. J. Mol. Sci. 2023, 24, 2911 12 of 24 site (BrdU+ NestinGFP+: KO RUN TBI vs. WT TBI, p = 0.04, vs. WT RUN TBI, p = 0.027 and vs. KO TBI p = 0.049, Figure 6C; BrdU+ DCX+: KO RUN TBI vs. WT TBI, p = 0.026, vs. WT RUN TBI, p = 0.015 and vs. KO TBI p = 0.011, Figure 6D). Int. J. Mol. Sci. 2023, 24, x FOR PEER REVIEW 13 of 25 Figure 6. Distribution of NSPCs and neuroblasts in the peri-lesion cortex. Seven days post TBI (A,C). Graphs show the increased density in the lateral (A) and medial (C) regions that border the cortical lesion of the KO RUN TBI mice of Brdu+NestinGFP+ cells (lateral, genotype x running: F(1,24) = 4.64, p = 0.041, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.02, vs. WT RUN TBI, p = 0.041 and vs. KO TBI p = 0.016, (A); medial, genotype X running: F(1,21) = 7.26, p = 0.013, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.04, vs. WT RUN TBI, p = 0.027 and vs. KO TBI p = 0.049, (C)). (B,D) Histograms illustrate in the lateral (B) and medial (D) peri-lesioned cortical area of the KO RUN TBI mice the enhanced BrdU+DCX+ cell density (lateral, genotype X running: F(1,21) = 7.32, p = 0.012, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.026, vs. WT RUN TBI, p = 0.015 and vs. KO TBI p = 0.011, (B); medial, BrdU+ DCX+: genotype X running: F(1,21) = 7.32, p = 0.012, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.026, vs. WT RUN TBI, p = 0.015 and vs. KO TBI p = 0.011, (D)). Fourteen days post TBI. (E,F,G) Graphs show, in the medial (E) and low (F) peri-injured cortical region of the KO RUN TBI, a significant increase in DCX+ cells (medial, Figure 6. Distribution of NSPCs and neuroblasts in the peri-lesion cortex. Seven days post TBI (A,C). Graphs show the increased density in the lateral (A) and medial (C) regions that border the cortical lesion of the KO RUN TBI mice of Brdu+NestinGFP+ cells (lateral, genotype x running: F(1,24) = 4.64, p = 0.041, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.02, vs. WT RUN TBI, p = 0.041 and vs. KO TBI p = 0.016, (A); medial, genotype X running: F(1,21) = 7.26, p = 0.013, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.04, vs. WT RUN TBI, p = 0.027 and vs. KO TBI p = 0.049, (C)). (B,D) Histograms illustrate in the lateral (B) and medial (D) peri-lesioned cortical area of the KO RUN TBI mice the enhanced BrdU+DCX+ cell density (lateral, genotype X running: F(1,21) = 7.32, Int. J. Mol. Sci. 2023, 24, 2911 13 of 24 p = 0.012, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.026, vs. WT RUN TBI, p = 0.015 and vs. KO TBI p = 0.011, (B); medial, BrdU+ DCX+: genotype X running: F(1,21) = 7.32, p = 0.012, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.026, vs. WT RUN TBI, p = 0.015 and vs. KO TBI p = 0.011, (D)). Fourteen days post TBI. (E,F,G) Graphs show, in the medial (E) and low (F) peri-injured cortical region of the KO RUN TBI, a significant increase in DCX+ cells (medial, genotype x run interaction: F(1,20) = 14.14, p = 0.0012, followed by LSD post-test, KO RUN TBI vs. WT RUN TBI, p = 0.007 and vs. KO TBI, p = 0.03, (E) and low (genotype x run interaction: F(1,20) = 7.52, p = 0.012, followed by LSD post-test, KO RUN TBI vs. WT TBI and WT RUN TBI, p < 0.001, vs. KO TBI, p = 0.03, (F)), as well as an increase in BrdU+DCX+ cells in the low side ((G), genotype x run interaction: F(1,20) = 7.52, p = 0.012, followed by LSD post-test, KO RUN TBI vs. WT TBI and WT RUN TBI, p < 0.001, vs. KO TBI, p = 0.03). Thirty days post TBI. (H,I) Histograms illustrate the significant increment in the KO RUN TBI group of NestinGFP+ cells in the medial (H) and low (I) regions (medial: genotype x run interaction: F(1,15) = 8.26, p = 0.011, followed by LSD post-test, KO RUN TBI vs. WT TBI, p < 0.001, vs. WT RUN TBI, p = 0.008, vs. KO TBI, p = 0.01, (H); low: genotype x run interaction: F(1,15) = 6.71, p = 0.02, followed by LSD post-test, KO RUN TBI vs. WT TBI, p < 0.001, vs. WT RUN TBI and KO TBI p = 0.01, (I)). (J) Graph shows the increase in DCX+ cells in the lateral cortical regions of KO RUN TBI mice, 30 days after TBI (genotype x run interaction: F(1,15) = 6.51, p = 0.022, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.015, vs. WT RUN TBI, and KO TBI, p = 0.047). (K,L) Confocal representative pictures show the specific localization of NestinGFP+ (arrows) and DCX+ (arrowheads) cells in the lateral cortical region lining the lesion of the KO RUN TBI mice, at 7 days post TBI (K). At the same time-point in the WT TBI group, we observe only a limited number of NestinGFP+ (arrow) cells in the peri-lesioned lateral side. (M,N) Confocal representative micrographs show that after 14 days from TBI, it is possible to observe in the medial side of the lesion of KO RUN TBI mice (M) a high density of NestinGFP+ (green), and DCX+ (blue) cells co-localizing with BrdU (red). In the same area of the WT TBI (N) mice, we detect a much lower density of cells. (O,P) The confocal pictures show the accumulation of DCX+ and BrdU+DCX+ cells in the low cortical peri-lesion area of the KO RUN TBI mice (P), which is not detectable in the WT TBI group (O). N = 5 mice/group. Statistical significance: * p < 0.05, ** p < 0.01 and *** p < 0.001. Two-way ANOVA analysis and Fisher’s LSD post hoc tests. Magnification = 20×. Scale bar = 100 µm. We also observed a genotype effect on the number of total NestinGFP+ and DCX+ cells in the lateral region (NestinGFP+ cells: p < 0.001, Supplementary Figure S9A; DCX+ cells: p = 0.02, Supplementary Figure S9B) and in the medial region (NestinGFP+ cells: p < 0.001, Supplementary Figure S9C; DCX+ cells: p < 0.001, Supplementary Figure S9D), suggesting that deletions of the p21 gene may play an important role in increasing the number of new neurons in the injured cortical regions. At 14 days post TBI, a significant effect in the lateral peri-lesion region of p21 dele- tion and running on the increase in NestinGFP+ cells (genotype effect: p = 0.0013; run effect: p = 0.0038, Supplementary Figure S9E) and BrdU+ NestinGFP+ cells (genotype effect: F(1,20) = 9.65, p = 0.0056; run effect: F(1.20) = 4.64, p = 0.043, Supplementary Figure S9F) occurs. In the medial and low region of the lesion, we observed a significant increase in the DCX+ populations in the KO RUN TBI group (medial: KO RUN TBI vs. WT RUN TBI, p = 0.007 and vs. KO TBI, p = 0.03, Figure 6E,M,N; low: KO RUN TBI vs. WT TBI, WT RUN TBI and KO TBI, p < 0.001, Figure 6F) and in BrdU+ DCX+ limited to the low zone (KO RUN TBI vs. WT TBI and WT RUN TBI, p < 0.001, vs. KO TBI, p = 0.03, Figure 6G). At 30 days post TBI, a significant effect of genotype on the increase in NestinGFP+ cells can be observed in the lateral peri-lesion region (lateral: genotype effect: p < 0.001). Moreover, we detect a significant increase in NestinGFP+ cells in the KO RUN TBI group when compared to the other conditions in the medial and low cerebral region (medial: KO RUN TBI vs. WT TBI, p < 0.001, vs. WT RUN TBI, p = 0.008, vs. KO TBI, p = 0.01, Figure 6H; low: KO RUN TBI vs. WT TBI, p < 0.001, vs. WT RUN TBI and KO TBI p = 0.01, Figure 6I,O,P). Finally, in the lateral region, we observed a strong increase in DCX+ cells in the KO RUN TBI animals, compared to the other three injured groups (KO RUN TBI vs. WT TBI, p = 0.015, vs. WT RUN TBI, and KO TBI, p = 0.047, Figure 6J,O,P), as well as in Int. J. Mol. Sci. 2023, 24, 2911 14 of 24 BrdU+ neuroblasts (DCX+ BrdU+ cells: KO RUN TBI vs. WT TBI, p = 0.013, vs. WT RUN TBI, p = 0.047 and vs. KO TBI p = 0.05). Taken together, these data demonstrate how TBI produces a significant migratory flow of stem cells and neuroblasts towards the injured region in the KO RUN TBI group, with the consequent accumulation of these populations in the cortical area closely adjacent to the lesion. 2.9. Volume of the Lesion Is Not Affected by p21 Deletion or Running Subsequently, we wanted to evaluate whether the cellular dynamics of proliferation, differentiation and especially migration that occur after TBI could modulate the macro- scopic neuroanatomical recovery in our experimental groups. To this end, we measured the volume of the lesions in the cerebral cortex directly affected by the lesion of the four groups at 14 and 30 days post lesion, by the application of Cavalieri’s estimator of morpho- metric volume. Our analysis does not evidence any significant difference within the four experimental conditions (Figure 7A,B). Moreover, the data related to the lesion’s variation over time evidence the significant effect of time in contributing to lesion volume reduction (time effect: p < 0.05, Figure 7C), confirming that after TBI, the brain activates mechanisms that promote the processes of regeneration and anatomical recovery. 2.10. Running and p21 Deletion Induce Partial Post-TBI Functional Recovery Since the controlled cortical impact (CCI) procedure was conducted in the cortical area corresponding to the primary motor cortex that controls the right forelimb, the next step was the evaluation of the putative relevance of injury-induced SVZ neurogenesis in ameliorating the functional recovery after TBI. To this aim, we used the Ladder Rung Walking test as a behavioral task to assess the skilled walking and right forelimb stepping, placing and coordination of the animals by measuring the number of mistakes in foot placement of the right forelimb during a 50 cm walk. We carried out this analysis at different time points, including 1 day before TBI (pre-TBI) to determine the functional baseline of the mice and 2, 7, 14 and 30 days after TBI. In the statistical analysis (ANOVA analysis for repeated measures), we considered the following four independent variables and their interaction: (1) the genotype of the animals (p21 WT or KO), (2) the surgical procedure (TBI or SHAM), (3) the physical activity (running or sedentary) and (4) the effect of the period of training (time). First of all, the data reveal the functional recovery over time, as demonstrated by the decrease in mistakes in the different TBI groups through the time points and by the statistical analysis of the time variable (time effect: p < 0.05, Figure 7D). Moreover, even if the statistical analysis of the independent variables genotype and running was not significant, the study of the interaction among genotype, physical activity and treatment (TBI or SHAM) was significant in influencing the functional outcome (genotype x surgery x running interaction, p < 0.05, Figure 7D). We did not find any significant differences among the SHAM groups that maintained a similar number of errors at every time point considered. Moreover, at P2, we noticed an increased number of errors in the four TBI groups if compared to SHAM animals, an effect that seems to be more prominent in WT RUN animals. This fact confirms that the surgical procedure by itself does not provoke any impairment in the functional performance of the animals and that contusion is a key factor for the higher number of mistakes committed by animals (TBI vs. SHAM, p < 0.001, Figure 7D). The graph even shows that at P7, the KO RUN TBI mice display better functional performances after the trauma than the other TBI groups (P7: KO RUN TBI vs. WT TBI, WT RUN TBI and KO TBI, p < 0.001). Moreover, we observe that only in KO RUN TBI mice, the number of errors is not significantly different compared to the SHAM groups at 7, 14 and 30 days after TBI (P7, P14, P33: p > 0.05 KO RUN TBI vs. all SHAM groups, Figure 7D). The area under the curve analysis confirms the TBI-dependent functional deterioration (TBI vs. SHAM, p < 0.001), while within the TBI groups, we observe a significant decrease in errors in the KO RUN TBI group compared to the other experimental groups (KO RUN TBI vs. WT TBI, p < 0.05, vs. WT RUN TBI and KO TBI, Int. J. Mol. Sci. 2023, 24, 2911 15 of 24 p < 0.001, Figure 7E). As a whole, our data suggest that the addictive effect of running and p21 deletion could be effective in promoting a partially higher functional outcome after TBI. Int. J. Mol. Sci. 2023, 24, x FOR PEER REVIEW 15 of 25 Figure 7. Anatomical and functional recovery following TBI. (A,B) Graphs show the average volumes of the lesions 14 (A) and 30 (B) days after the TBI procedure in the four experimental groups. (C) Histograms indicate the decreased average volumes of the lesions from 14 to 30 days post TBI in the four different experimental conditions (time effect: F(4,28) = 22.08; p < 0.05). (D) The graph indicates the mean number of errors per group at each time point. The same color shows an Figure 7. Anatomical and functional recovery following TBI. (A,B) Graphs show the average volumes of the lesions 14 (A) and 30 (B) days after the TBI procedure in the four experimental groups. (C) Histograms Int. J. Mol. Sci. 2023, 24, 2911 16 of 24 indicate the decreased average volumes of the lesions from 14 to 30 days post TBI in the four different experimental conditions (time effect: F(4,28) = 22.08; p < 0.05). (D) The graph indicates the mean number of errors per group at each time point. The same color shows an experimental condition and its relative SHAM control. The statistical analysis evidenced significant effects of the time variable (time effect F = 9.949 p < 0.05, Figure 7D). Within the TBI groups, we found that at 7 days post TBI, the KO RUN TBI animals demonstrated better performances with respect to the other conditions and, notably, a number of mistakes comparable to their SHAM control time effect (P7: genotype x run interaction: F(1,64) = 7.65 p = 0.032, followed by LSD post-test, KO RUN TBI vs. WT TBI, WT RUN TBI and KO TBI, p < 0.001). For the statistical analyses of the volumes of the lesions, we performed a multifactorial analysis with the following three independent variables: genotype; running and time. (E) The histograms represent the area under the curve analysis and show the increase in errors in the TBI groups (lesion effect F(1,48) = 153; p < 0.001, $); moreover, in the TBI groups, the statistical analysis indicates a decrease in errors in the KO RUN TBI mice in comparison with the other TBI groups (genotype x run x lesion interaction, F(1,48) = 12.92, p < 0.001, followed by LSD post-test, KO RUN TBI vs. WT TBI, p = 0.011, vs. WT RUN TBI, and KO TBI, p < 0.001). The behavioral data of the Ladder Rung Walking task have been analyzed by a multifactorial analysis with the following four independent variables: genotype, running, time and the treatment. The area under the curve statistics were evaluated by multifactorial analysis with the following three independent variables: genotype, treatment and running. (N = 8 mice/group). The post hoc analyses have been conducted via by Fisher’s LSD post hoc tests. Statistical significance: * p < 0.05 and *** p < 0.001. 3. Discussion Several studies have suggested that NSCs may retain or even potentiate their self- renew ability following injury, in order to produce additional neural progenitors and neuroblasts, which migrate toward the damaged tissue and contribute to the post-traumatic neuro-regenerative processes [32,33]. If this hypothesis is confirmed, new strategies to enhance neurogenesis could be very useful to increase the number of new neurons that can benefit the cortical region that is directly involved in the damage. In this study, we have demonstrated that the concomitant deletion of p21 and physical activity play a powerful role in enhancing the subventricular neurogenic post-traumatic response and improving functional recovery. From the comparison between the TBI and SHAM groups, we can observe a clear dynamic in the post-traumatic SVZ neurogenic response, which results in the early activa- tion of type B cells, followed over time by a net decrease in the Nestin GFP+ population and a concomitant increase in type A neuroblasts, leading to a significant increase in this population at 30 days post TBI. From such evidence, we can hypothesize that in our ex- perimental model, TBI triggers a highly specific pro-neurogenic process within the SVZ, characterized by temporally different neurogenic responses within the sub-populations considered. These data are in apparent contradiction with a study that demonstrates that transit-amplifying cells (type C cells) are the main cell type responsible for the injury- induced increase in cell proliferation, with no contribution from either GFAP+ or DCX+ cells [34]. However, in a recent work, a gradual increase over time in proliferation (from day 1 to day 7 post TBI) associated with an increase in NestinGFP+/GFAP+ NSCs and DCX progenitors within the SVZ of rats has been observed [21]. In another single-cell transcriptomics study, it has been demonstrated that brain injury is able to transform dormant NSCs into primed quiescent and active NSCs, with the concomitant activation of protein synthesis and cell cycle genes [22]. From these and other studies clearly emerges a strong discrepancy in the post-traumatic SVZ neurogenic response, derived from the high heterogeneity of the experimental protocols used, differing significantly from each other in the type of damage applied, in the proliferative/differentiative markers used and in the length of post-traumatic time-points, as well as by the utilization of different mice strains [35]. The comparison of the post-traumatic neurogenic response within TBI groups clearly demonstrates that the p21 deletion and pre-traumatic voluntary physical activity over a Int. J. Mol. Sci. 2023, 24, 2911 17 of 24 12-day period exert a potent proneurogenic effect, which belatedly results in a significant increase in the population of DCX+ neuroblasts at 30 days post TBI. This stimulating effect of post-TBI neurogenesis observed in the KO RUN TBI mice differs from what was observed in the KO RUN SHAM group, in which we detected a considerable increase in subventricular proliferation, which was significantly higher than in the KO RUN TBI group 7 and 14 days after injury. In this regard, we might speculate that the strong proliferative increase in the KO RUN mice at 7 and 14 days post SHAM could be the result of the hyper-proliferation of running-activated NSCs as previously observed [28], an event that was not observed in the KO RUN TBI group, which displayed a late pro-neurogenic response. These differences show how the modulation of the sub-populations of newborn cells within the SVZ is very diversified, depending on the external stimuli, and furthermore lead us to hypothesize that in the p21 ko mice, the temporally consequential effects of physical activity and trauma establish a well-defined series of cellular and microenvironment modifications that, on the one hand, require well-defined timing to optimize their proneurogenic effect whereas on the other hand, the differ profoundly from the effects observed if these external stimuli are provided separately. Alternatively, we can hypothesize that a third pro-neurogenic stimulation, the TBI, to the SVZ neurogenic niche of the KO RUN SHAM mice could lead to an excessive overload in the proliferative rate, with the consequent depletion of the NSC/progenitor pool, as previously observed following a prolonged running period (21 days) in p21 ko mice [28]. Furthermore, our study examined the effects of p21 deletion and running on the migration of newborn neurons after cortical brain injury. In the non- injured SVZ, post-mitotic neuroblasts move from SVZ to the olfactory bulb along the rostral migratory stream (RMS) [36]. When cortical regions are injured, neural progenitors start to migrate outside the RMS into the adjacent tissues, including the corpus callosum (CC) and the peri-lesioned cortex [35,37]. In this context, interesting evidence that arose from our study is represented by the observation of the widespread migration of NestinGFP+ and DCX+ cells from the SVZ towards the damaged peri-cortical regions in the KO RUN TBI group. A very intriguing aspect to take into consideration is the high fraction of proliferation observed in the migrating cells, testifying the presence of cytogenesis within the migratory flow, far from the SVZ neurogenic niche. The cytogenesis observed within the migratory flow both in the NestinGFP+, and to a lesser extent in the DCX+ cells, could represent an additional neurogenic response that is able to further increase the pool of NSCs and progenitor cells involved in the unknown tissue repair processes and post-traumatic functional improvement. Both processes are observed in all the groups subjected to TBI, although the deletion of p21 associated with physical activity greatly increases its extent. This evidence confirms what was observed in a previous study, indicating a dramatic increase in the migration of post-mitotic neuroblasts along the RMS in p21 ko mice after a 5- and 12-day running session [28]. Moreover, the widespread migration of NestinGFP+ cells from the SVZ towards the injured cortex could account for the transient decrease in the NestinGFP+ population observed within the SVZ at 14 days post TBI; in this case, a process of cellular evasion could be outlined in the phases immediately following the trauma, which leads to the gradual impoverishment of the NestinGFP+ population within the SVZ. An explanation for the increased cell migration in the KO RUN TBI mice is still required and further studies will evaluate the impact of p21 deletion and running on the main pathways that regulate neuroblast migration. In this regard, several groups have demonstrated the involvement of trophic factors and their receptors in the microenvironment that promotes neuroblast migration in both the naïve and post-lesioned brain, including stromal cell- derived factor 1 (SDF)/C-X-C motif chemokine 4 (CXCR4), brain-derived neurotrophic factor (BDNF)/tropomyiosin receptor kinase B (TrkB) and vascular endothelial growth factors (VEGF)/VEGF receptor [37–39]. In particular, it was found that BDNF caused SVZ cells to emigrate toward cerebral regions [40,41] in a concentrated manner [42]; and also that either pre- or post-traumatic exercise increased cerebral BDNF protein expression as compared to non-exercised animals [43–46], suggesting a putative role of BDNF in the increase in migration in pre-exercised animal trauma. Int. J. Mol. Sci. 2023, 24, 2911 18 of 24 In our study, the large migratory flow of NSCs/neuroblasts observed in the KO RUN TBI group translates into a significant accumulation of newly generated neurons in the tissue bordering the cortical lesion. Using a long-term BrdU assay, we have also demonstrated that a large fraction of these neuroblasts are newly born and are most likely of subventricular origin. We do not know exactly the differentiative fate of these cells nor their functional role within the lesioned microenvironment, because migrating NSPCs take 1 to 3 months to fully maturate into neurons [47]; furthermore, in our study, we detected only a minimal fraction of cells that co-expressed the markers BrdU and NeuN (marker of mature neurons) at 30 days post injury, in full agreement with previous works [48,49], demonstrating the nearby absence of newly born mature neurons in peri- lesion regions. The formation of newly generated mature neurons in the cortex following injury is controversial. Recent works suggest that the cortex microenvironment may favor glial differentiation, resulting in the downregulation of DCX expression and the concomitant cortical up-regulation of Olig2 [37,50] and of Shh [51], which trigger glial differentiation [35]. Pous et al. identified the fibrinogen released in the microenvironment SVZ by the leaky vasculature after injury, a key factor driving the differentiation of NSPCs into astrocytes via the activation of the BMP signaling pathway [52]. Whatever the fate of the NSPCs of p21 runner mice after TBI, we hypothesize that their considerable increase in the peri-lesioned cortical region could be beneficial for the neuro-reparative response, as evidenced by the improvement of the Ladder Walking Test observed in the KO RUN TBI mice. In agreement with our hypothesis, Dixon et al. demonstrated that selective NSPC ablation induces a reduction in the number of neuroblasts migrating toward the injury with the consequent decrease in residential neuron and glial cells in the peri-lesion cortex and reduced locomotor recovery [53]. We believe that a strong increase in the number of NSPCs in the peri-injury cortical region of KO RUN TBI mice could greatly enhance the tissue stabilization processes of the injury milieu, allowing neuroprotection through the increased influx of neuro-protective factors, such as BDNF and VEGF, which can in turn promote neuronal survival, local glial proliferation, reduced gliosis and functional recovery. Our study is descriptive and one of the main limitations of this study is the lack of analysis of molecular mechanisms that can partially explain the results obtained. In this regard, however, some of our preliminary evaluations have highlighted the specific role of p21 deletion and physical activity in the subventricular neurogenic niche capable of downregulating the expression of anti-neurogenic genes in the BMP2 pathway. This event could trigger NSCs to exit from the quiescent state, allowing them to differentiate into the neuronal lineage [54]. 4. Material and Methods 4.1. Animals Male wild-type and p21-null mice [55] of the same genetic background (129Sv/c57BL6; 50:50; https://www.jax.org/strain/003263, accessed on 12 February 2015) were housed under a continuous 12 h light/12 h dark cycle at a constant temperature of 21 ◦C, with complete availability of water and food. Nestin green fluorescent protein mice (C57BL/6 background; kindly provided by Dr. G. Enikolopov) express GFP driven by the Nestin promoter [56]. Nestin-GFP mice were crossed with WT and knockout mice to obtain WT and KO/NestinGFP+ mice, which were interbred at least four times before further analysis, generating the different genotypes under study. All experiments were performed blind for the different experimental conditions. We analyzed 5 mice per group in the immunohistochemistry study and 8 mice group in the functional analysis. 4.2. Running Paradigm and BrdU Administration Mice subjected to the 12-day running session were housed in a running cage (2 mice per cage) and their running activity was measured with a speedometer. Moreover, the mice had been treated with BrdU administered in their drinking water (B5002, Sig-ma; 0.5 g/L) Int. J. Mol. Sci. 2023, 24, 2911 19 of 24 from day 7 until the end of the running session, in order to label proliferating NPCs and neuroblasts and to follow the fate of their progeny. Depending on their genotype (WT or p21 KO) and on their surgical procedure (SHAM or TBI), the mice were subdivided according to the following sedentary or running protocols: WT sedentary (WT SHAM), WT running (WT RUN SHAM), p21 ko sedentary (KO SHAM) and ko running (KO RUN SHAM), and WT sedentary (WT TBI), WT running (WT RUN TBI), p21 ko sedentary (KO TBI) and p21 ko running (KO RUN TBI). 4.3. Controlled Cortical Impact (CCI) Injury Mice were anesthetized with isoflurane and positioned within a mouse stereotaxic frame. Following a longitudinal skin incision, a 3 mm diameter craniotomy was per- formed at the following stereotaxic coordinates: antero-posterior (AP): +0.5 mm; lateral −0.5 mm [57]. Traumatic brain injury was performed at the cortical level with a flat, 3 mm diameter metal tip attached to the CCI device (PinPoint Precision Impactor, Stoelting, Wood Dale, IL, USA), at an impact speed of 3 m/sec, time of impact of 150 ms and a depth of 2 mm below the dura, corresponding to the cerebral region of the primary cortex, which controls the fine movements of the right forelimb (TBI groups). After the impact, the animals were sutured with absorbable suture thread, housed in their home cage and put on a heated plate for 3/4 h in order to control their body temperature during their recovery from anesthesia. Animals were treated following the Italian Ministry of Health and directive 2010/63/EU guideline nr 785/19 PR. Animals subjected to the surgical procedures described above without cortical impact represented the SHAM groups. 4.4. Experimental Procedures First, 12–14-week-old male p21 wt and knockout mice ran for 12 days in free running wheels; from 7 to 12 days of running, each group of animals received BrdU dissolved in drinking water, to mark the cells that underwent DNA replication and to follow the fate of their progeny. On the twelfth day of running, the mice were subjected to the controlled cortical impact (CCI) surgical procedure as described before. The mice were sacrificed at the following three different time points: seven (P7), fourteen (P14) and thirty (P30) days after the CCI procedure (Figure 1A). By immunohis- tochemical assays, we analyzed the following three different regions of the injured brain: (i) the ipsi- and contralateral SVZ to the lesion to evaluate the post-traumatic neurogenic response; (ii) the migratory route of new-born cells originating in the SVZ that were re- directed toward the injured cortical regions; (iii) 3 different cortical sites (medial, low and lateral) of the lesion to detect the new-born progenitors that reached the injured area and that are supposed to contribute to ameliorating the outcome after TBI (Figure 1B). To evaluate the possible functional recovery of the mice subjected to TBI, we performed the Ladder Rung Walking Test at 2, 7, 14 and 30 days post TBI (Figure 1A). 4.5. Immunohistochemistry At 7, 14 and 30 days post TBI, the animals were sacrificed by trans-cardiac perfusion with 4% paraformaldehyde (PFA) in phosphate-buffered saline (PBS); the brains were collected and kept overnight at −4 ◦C in PFA. They were subsequently equilibrated in sucrose diluted at 30% and finally cryopreserved at −80 ◦C. Slicing was carried out by embedding the brain in Tissue-Tek OCT (Sakura, Torrence, CA, USA) and then cut using a cryostat at −25 ◦C throughout the whole rostro-caudal extent. The coronal sections were processed in a one-in-six series protocol at a 40 µm thickness. Sections were then stained for multiple labelling using different fluorescence techniques. Sections were initially washed with 0.1 M glycine for 10 min, followed by permeabilization using 0.3% Triton X-100 in PBS for another 10 min. The sections were then incubated for 30 min in a blocking solution that contained 3% normal donkey serum (NDS) in 0.3% Triton X-100 in PBS to saturate the specific sites, followed by incubation with the same blocking solution that contained primary antibodies for 16–18 h at 4 ◦C. The primary antibodies used were goat polyclonal Int. J. Mol. Sci. 2023, 24, 2911 20 of 24 antibodies, which were used against DCX (Santa Cruz Biotechnology, Dallas, TX, USA; Cat# Sc-8066; 1:300); a rabbit monoclonal antibody was used against Ki67 (Lab Vision, South San Francisco, CA, USA, Cat# RM-9106-S; 1:150), whereas a mouse monoclonal antibody was used against GFAP (Sigma, St. Louis, MO, USA, Cat# G6171; 1:500). The detection of BrdU-positive cells consisted of denaturing DNA with 2N HCl for 45 min at 37 ◦C to facilitate antibody access. The sections were then incubated with 0.1 M sodium borate buffer at pH 8.5, followed by overnight incubation at 4 ◦C with a rat anti-BrdU primary antibody (Abcam, Cambridge, UK, Cat# ab6326; 1:300) diluted in TBS that contained 0.1% Triton, 0.1% Tween, and 3% normal donkey serum (blocking solution). To observe primary antibody binding, donkey secondary antibodies against rat (BrdU) and rabbit (Ki67) conjugated to Cy3 (Jackson ImmunoResearch, West Grove, PA, USA; 1:200 in PBS), and against goat (DCX) and mouse (GFAP) antibodies conjugated to Alexa-647 (Invitrogen, San Diego, CA, USA; 1:300 in PBS) were used. Nuclei were observed by incubating sections with Hoechst (1:500). 4.6. Cell Counting Cell numbers in the SVZ were obtained with stereological analysis, by counting the cells that expressed the indicated markers and were visualized with confocal microscopy throughout the whole rostro-caudal extent of the SVZ in a one-in-ten series of 40 µm free-floating serial coronal sections (240 µm apart). The cell numbers obtained for each SVZ section were divided for the corresponding area of the section to obtain the average number of SVZ cells per 100 mm2. The areas were obtained by tracing the outline of the whole SVZ bulb, identified by the presence of cell nuclei stained by Hoechst on a digital picture captured and measured using ImageJ software (Version 1.52t released 30 January 2020) [58]. A CellSens Standard system (OLYMPUS) was used to record z-stack images, and thus confirm the colocalization of multiple labeled cells in the SVZ. To assess the neural stem/progenitor cell and neuroblast cell numbers in the migratory stream and perilesional region, non-biased cell number estimations were performed on the 5 most central rostro- caudal sections around the injury epicenter (as determined using cresyl violet-stained sections), which were 30 µm thick and 180 µm apart. The count of the labelled cells in the regions bordering the lesion was carried out within a frame of 300,000 mm2. 4.7. Estimation of Lesion Volume On a Polysine microscope slide (Thermo Scientific, Waltham, MA, USA), a number of brain slices that permitted us to comprehend the entire damage extension were placed. Then, images of the sections were acquired by fluorescent microscopy at a magnification of 4×. The areas of the damage (mm2) on each section were estimated via the “Poly- gon Selection” tool of ImageJ and following this the total volume of brain damage (mm) was calculated by Cavalieri’s estimator of morphometric volume, which is as follows: VC = d (Σ yi) − (t) Ymax, where d is the distance between the sections contained in a well (d = 240 µm), yi is the area of a single section, t is the section thickness (t = 40 µm) and yMAX is the maximum value of y. The factor (t) yMAX is subtracted from the basic equation as a correction for overprojection. 4.8. Ladder Rung Walking Task The apparatus is made by two see-through walls of Plexiglass of 1 m of length and 20 cm of height with removable metal rungs (3 mm of diameter) at a minimum distance of 1 cm. The ladder is placed at a minimum of 50 cm above the ground with a neutral cage on one side, from which the mice start the task, and the home cage with the littermates at the other. The width of the apparatus can be adjusted in order to prevent the animals from turning around. Animals were tested with a regularly 2 cm interspaced rung pattern one day before the CCI procedure (pre TBI) to determine the baseline scores and 2 (P2), 7 (P7), 14 (P14) and 30 (P30) days after the surgery to evaluate the functional recovery. Every mouse underwent 4 trials with a ladder length of 50 cm during which they were recorded Int. J. Mol. Sci. 2023, 24, 2911 21 of 24 with a camera placed at one side of the apparatus in order to obtain a clear view of the considered limb (right forelimb). The scoring was carried out through the observation of the recordings in slow-motion and counting the number of errors for each trial (foot placement accuracy analysis). The following errors were considered: total miss (the limb completely missed a rung), deep slip (the limb initially reached the rung but then slipped off, causing a fall when weight-bearing) and slight slip (the limb slipped off when weight-bearing but not causing a fall or an interruption of the gait). The mean number of errors per trial was then standardized for a ladder length of 1 m. The areas under the curve were calculated with Prism 5. 4.9. Statistical Analysis The data on the cellular responses after TBI in the SVZ have been analyzed through a three-way ANOVA, with genotype, running and treatment as the independent variables. Cell migration and number of cells in the peri-lesioned area have been analyzed through a two-way ANOVA, with genotype and running as the independent variables. Lesion volumes at different time-points have been analyzed through a mixed ANOVA, with genotype, running and time as the independent variables. Finally, the behavioral data from the Ladder Rung Walking task have been analyzed through a mixed ANOVA, with genotype, running, time and treatment as the independent variables. Fisher’s LSD post hoc tests have been conducted whenever necessary. Analyses have been performed with GraphPad Prism 5 software for the two-way ANOVA and with Statistica software 14.0 (Dell Software) for the three-way and mixed ANOVA studies. 5. Conclusions The data obtained in this study reveal how the interaction between physical activity and p21 gene deficiency plays an important role in neurogenic and migratory mechanisms in response to traumatic injury. Our data do not reveal whether the increase in neuroblasts within the SVZ and their subsequent migration towards the lesioned cortex is a process capable of increasing the rate of functionally active newly mature neurons. However, the data obtained in the Ladder Rung Walking Test represented an indirect clue relating the correlation within the KO RUN TBI group between the cellular processes that take place in the post-traumatic sequelae and an improvement in the functional response underlying that particular task. We believe that this study can offer interesting perspectives for future pre-clinical strategies aimed at investigating the role of physical activity and NSCs in the post-traumatic neuro-regenerative response. Supplementary Materials: The supporting information can be downloaded at: https://www.mdpi. com/article/10.3390/ijms24032911/s1. Author Contributions: Conceptualization, S.F.V.; methodology, J.I.B., V.M., V.N.d.R., S.M., S.F.V.; software, V.M., D.S.; validation, S.F.V.; formal analysis, J.I.B., V.M., V.N.d.R., S.M., S.F.V.; investigation, J.I.B., V.M., V.N.d.R., D.S., S.M., S.F.V.; resources, S.F.V.; data curation, J.I.B., S.F.V.; writing—original draft preparation, J.I.B., S.F.V.; writing—review and editing, S.F.V., J.I.B., V.M.; visualization S.F.V.; supervision, S.F.V.; funding acquisition, S.F.V. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by Filas Regione Lazio: GAEPFARIOLI; IBCN/CNR Starting Grant Fund: GAEPFARIOLI. Institutional Review Board Statement: The animal study protocol was approved by the Ethics Committee of Ministry of Health (785/PR, 29/11/2019). Informed Consent Statement: Not applicable. Data Availability Statement: Not applicable. Conflicts of Interest: The authors declare no conflict of interest. Int. J. Mol. Sci. 2023, 24, 2911 22 of 24 References 1. Summers, C.R.; Ivins, B.; Schwab, K.A. Traumatic brain injury in the United States: An epidemiologic overview. Mt. Sinai J. Med. 2009, 76, 105–110. [CrossRef] [PubMed] 2. Hawryluk, G.W.; Manley, G.T. Classification of traumatic brain injury: Past, present, and future. Handb. Clin. Neurol. 2015, 127, 15–21. [PubMed] 3. Bae, I.S.; Chun, H.J.; Yi, H.J.; Bak, K.H.; Choi, K.S.; Kim, D.W. Modified Glasgow Coma Scale Using Serum Factors as a Prognostic Model in Traumatic Brain Injury. World Neurosurg. 2019, 126, e959–e964. [CrossRef] [PubMed] 4. Bodien, Y.G.; Barra, A.; Temkin, N.R.; Barber, J.; Foreman, B.; Vassar, M.; Robertson, C.; Taylor, S.R.; Markowitz, A.J.; Manley, G.T.; et al. TRACK-TBI Investigators. Diagnosing Level of Consciousness: The Limits of the Glasgow Coma Scale Total Score. J. Neurotrauma. 2021, 38, 3295–3305. [CrossRef] [PubMed] 5. Pavlovic, D.; Pekic, S.; Stojanovic, M.; Popovic, V. Traumatic brain injury: Neuropathological, neurocognitive and neurobehavioral sequelae. Pituitary 2019, 22, 270–282. [CrossRef] 6. Lundin, A.; de Boussard, C.; Edman, G.; Borg, J. Symptoms and disability until 3 months after mild TBI. Brain Inj. 2006, 20, 799–806. [CrossRef] 7. Wang, K.K.; Yang, Z.; Zhu, T.; Shi, Y.; Rubenstein, R.; Tyndall, J.A.; Manley, G.T. An update on diagnostic and prognostic biomarkers for traumatic brain injury. Expert Rev. Mol. Diagn. 2018, 18, 165–180. [CrossRef] 8. Ghajar, J. Traumatic brain injury. Lancet 2000, 356, 923–929. [CrossRef] 9. Sivanandam, T.M.; Thakur, M.K. Traumatic brain injury: A risk factor for Alzheimer’s disease. Neurosci. Biobehav. Rev. 2012, 36, 1376–1381. [CrossRef] 10. Marras, C.; Hincapié, C.A.; Kristman, V.L.; Cancelliere, C.; Soklaridis, S.; Li, A.; Borg, J.; af Geijerstam, J.L.; Cassidy, J.D. Systematic review of the risk of Parkinson’s disease after mild traumatic brain injury: Results of the International Collaboration on Mild Traumatic Brain Injury Prognosis. Arch. Phys. Med. Rehabil. 2014, 95 (Suppl. 3), S238–S244. [CrossRef] 11. Wang, X.; Gao, X.; Michalski, S.; Zhao, S.; Chen, J. Traumatic Brain Injury Severity Affects Neurogenesis in Adult Mouse Hippocampus. J. Neurotrauma. 2016, 33, 721–733. [CrossRef] 12. Marzano, L.A.S.; de Castro, F.L.M.; Machado, C.A.; de Barros, J.L.V.M.; Macedo E Cordeiro, T.; Simões E Silva, A.C.; Teixeira, A.L.; Silva de Miranda, A. Potential Role of Adult Hippocampal Neurogenesis in Traumatic Brain Injury. Curr. Med. Chem. 2022, 29, 3392–3419. [CrossRef] [PubMed] 13. Doetsch, F.; Caillé, I.; Lim, D.A.; García-Verdugo, J.M.; Alvarez-Buylla, A. Subventricular zone astrocytes are neural stem cells in the adult mammalian brain. Cell 1999, 97, 703–716. [CrossRef] 14. Garcia, A.D.; Doan, N.B.; Imura, T.; Bush, T.G.; Sofroniew, M.V. GFAP-expressing progenitors are the principal source of constitutive neurogenesis in adult mouse forebrain. Nat. Neurosci. 2004, 11, 1233–1241. [CrossRef] [PubMed] 15. Cesetti, T.; Obernier, K.; Bengtson, C.P.; Fila, T.; Mandl, C.; Hölzl-Wenig, G.; Wörner, K.; Eckstein, V.; Ciccolini, F. Analysis of stem cell lineage progression in the neonatal subventricular zone identifies EGFR+/NG2- cells as transit-amplifying precursors. Stem Cells 2009, 27, 1443–1454. [CrossRef] [PubMed] 16. Urbán, N.; Guillemot, F. Neurogenesis in the embryonic and adult brain: Same regulators, different roles. Front. Cell Neurosci. 2014, 8, 96. [CrossRef] 17. Saghatelyan, A.; Carleton, A.; Lagier, S.; de Chevigny, A.; Lledo, P.M. Local neurons play key roles in the mammalian olfactory bulb. J. Physiol. Paris 2003, 97, 517–528. [CrossRef] 18. Kohwi, M.; Osumi, N.; Rubenstein, J.L.; Alvarez-Buylla, A. Pax6 is required for making specific subpopulations of granule and periglomerular neurons in the olfactory bulb. J. Neurosci. 2005, 25, 6997–7003. [CrossRef] 19. Grande, A.; Sumiyoshi, K.; López-Juárez, A.; Howard, J.; Sakthivel, B.; Aronow, B.; Campbell, K.; Nakafuku, M. Environmental impact on direct neuronal reprogramming in vivo in the adult brain. Nat. Commun. 2013, 4, 2373. [CrossRef] 20. Benner, E.J.; Luciano, D.; Jo, R.; Abdi, K.; Paez-Gonzalez, P.; Sheng, H.; Warner, D.S.; Liu, C.; Eroglu, C.; Kuo, C.T. Protective astrogenesis from the SVZ niche after injury is controlled by Notch modulator Thbs4. Nature 2013, 497, 369–373. [CrossRef] 21. Kang, E.M.; Jia, Y.B.; Wang, J.Y.; Wang, G.Y.; Chen, H.J.; Chen, X.Y.; Ye, Y.Q.; Zhang, X.; Su, X.H.; Wang, J.Y.; et al. Downregulation of microRNA-124-3p promotes subventricular zone neural stem cell activation by enhancing the function of BDNF downstream pathways after traumatic brain injury in adult rats. CNS Neurosci. Ther. 2022, 28, 1081–1092. [CrossRef] [PubMed] 22. Llorens-Bobadilla, E.; Zhao, S.; Baser, A.; Saiz-Castro, G.; Zwadlo, K.; Martin-Villalba, A. Single-Cell Transcriptomics Reveals a Population of Dormant Neural Stem Cells that Become Activated upon Brain Injury. Cell Stem Cell 2015, 7, 329–340. [CrossRef] [PubMed] 23. Farioli-Vecchioli, S.; Tirone, F. Control of the Cell Cycle in Adult Neurogenesis and its Relation with Physical Exercise. Brain Plast. 2015, 1, 41–54. [CrossRef] [PubMed] 24. Sherr, C.J.; Roberts, J.M. CDK inhibitors: Positive and negative regulators of G1-phase progression. Genes Dev. 1999, 13, 1501–1512. [CrossRef] 25. Coqueret, O. New roles for p21 and p27 cell-cycle inhibitors: A function for each cell compartment? Trends Cell Biol. 2003, 13, 65–70. [CrossRef] 26. Pechnick, R.N.; Chesnokova, V. Adult neurogenesis, cell cycle and drug discovery in psychiatry. Neuropsychopharmacology 2009, 34, 244. [CrossRef] Int. J. Mol. Sci. 2023, 24, 2911 23 of 24 27. Zonis, S.; Ljubimov, V.A.; Mahgerefteh, M.; Pechnick, R.N.; Wawrowsky, K.; Chesnokova, V. p21Cip restrains hippocampal neuro- genesis and protects neuronal progenitors from apoptosis during acute systemic inflammation. Hippocampus 2013, 23, 1383–1394. [CrossRef] 28. Nicolis di Robilant, V.; Scardigli, R.; Strimpakos, G.; Tirone, F.; Middei, S.; Scopa, C.; De Bardi, M.; Battistini, L.; Saraulli, D.; Farioli Vecchioli, S. Running-Activated Neural Stem Cells Enhance Subventricular Neurogenesis and Improve Olfactory Behavior in p21 Knockout Mice. Mol. Neurobiol. 2019, 56, 7534–7556. [CrossRef] 29. Qiu, J.; Takagi, Y.; Harada, J.; Rodrigues, N.; Moskowitz, M.A.; Scadden, D.T.; Cheng, T. Regenerative response in ischemic brain restricted by p21cip1/waf1. J. Exp. Med. 2004, 199, 937–945. [CrossRef] 30. Porlan, E.; Morante-Redolat, J.M.; Marqués-Torrejón, M.Á.; Andreu-Agulló, C.; Carneiro, C.; Gómez-Ibarlucea, E.; Soto, A.; Vidal, A.; Ferrón, S.R.; Fariñas, I. Transcriptional repression of Bmp2 by p21(Waf1/Cip1) links quiescence to neural stem cell maintenance. Nat. Neurosci. 2013, 16, 1567–1575. [CrossRef] 31. Marqués-Torrejón, M.Á.; Porlan, E.; Banito, A.; Gómez-Ibarlucea, E.; Lopez-Contreras, A.J.; Fernández-Capetillo, O.; Vidal, A.; Gil, J.; Torres, J.; Fariñas, I. Cyclin-dependent kinase inhibitor p21 controls adult neural stem cell expansion by regulating Sox2 gene expression. Cell Stem Cell 2013, 12, 88–100. [CrossRef] [PubMed] 32. Finkel, Z.; Esteban, F.; Rodriguez, B.; Fu, T.; Ai, X.; Cai, L. Diversity of Adult Neural Stem and Progenitor Cells in Physiology and Disease. Cells 2021, 8, 2045. [CrossRef] [PubMed] 33. Weston, N.M.; Sun, D. The Potential of Stem Cells in Treatment of Traumatic Brain Injury. Curr. Neurol. Neurosci. Rep. 2018, 25, 1. [CrossRef] [PubMed] 34. Thomsen, G.M.; Le Belle, J.E.; Harnisch, J.A.; Mc Donald, W.S.; Hovda, D.A.; Sofroniew, M.V.; Kornblum, H.I.; Harris, N.G. Traumatic brain injury reveals novel cell lineage relationships within the subventricular zone. Stem Cell Res. 2014, 13, 48–60. [CrossRef] [PubMed] 35. Chang, E.H.; Adorjan, I.; Mundim, M.V.; Sun, B.; Dizon, M.L.; Szele, F.G. Traumatic Brain Injury Activation of the Adult Subventricular Zone Neurogenic Niche. Front. Neurosci. 2016, 10, 332. [CrossRef] 36. Lim, D.A.; Alvarez-Buylla, A. The Adult Ventricular-Subventricular Zone (V-SVZ) and Olfactory Bulb (OB) Neurogenesis. Cold Spring Harb. Perspect. Biol. 2016, 8, a018820. [CrossRef] 37. Saha, B.; Peron, S.; Murray, K.; Jaber, M.; Gaillard, A. Cortical lesion stimulates adult subventricular zone neural progenitor cell proliferation and migration to the site of injury. Stem Cell Res. 2013, 3, 965–977. [CrossRef] 38. Snapyan, M.; Lemasson, M.; Brill, M.S.; Blais, M.; Massouh, M.; Ninkovic, J.; Gravel, C.; Berthod, F.; Götz, M.; Barker, P.A.; et al. Vasculature guides migrating neuronal precursors in the adult mammalian forebrain via brain-derived neurotrophic factor signaling. J. Neurosci. 2009, 29, 4172–4188. [CrossRef] 39. Kokovay, E.; Goderie, S.; Wang, Y.; Lotz, S.; Lin, G.; Sun, Y.; Roysam, B.; Shen, Q.; Temple, S. Adult SVZ lineage cells home to and leave the vascular niche via differential responses to SDF1/CXCR4 signaling. Cell Stem Cell 2010, 7, 163–173. [CrossRef] 40. Pencea, V.; Bingaman, K.D.; Freedman, L.J.; Luskin, M.B. Neurogenesis in the subventricular zone and rostral migratory stream of the neonatal and adult primate forebrain. Exp. Neurol. 2001, 172, 1–16. [CrossRef] 41. Fon, D.; Zhou, K.; Ercole, F.; Fehr, F.; Marchesan, S.; Minter, M.R.; Crack, P.J.; Finkelstein, D.I.; Forsythe, J.S. Nanofibrous scaffolds releasing a small molecule BDNF-mimetic for the re-direction of endogenous neuroblast migration in the brain. Biomaterials 2014, 35, 2692–2712. [CrossRef] [PubMed] 42. Petridis, A.K.; El Maarouf, A. Brain-derived neurotrophic factor levels influence the balance of migration and differentiation of subventricular zone cells, but not guidance to the olfactory bulb. J. Clin. Neurosci. 2011, 18, 265–270. [CrossRef] [PubMed] 43. Griesbach, G.S.; Hovda, D.A.; Molteni, R.; Wu, A.; Gomez-Pinilla, F. Voluntary exercise following traumatic brain injury: Brain-derived neurotrophic factor upregulation and recovery of function. Neuroscience 2004, 125, 129–139. [CrossRef] [PubMed] 44. Griesbach, G.S.; Gómez-Pinilla, F.; Hovda, D.A. Time window for voluntary exercise-induced increases in hippocampal neuro- plasticity molecules after traumatic brain injury is severity dependent. J. Neurotrauma 2007, 24, 1161–1171. [CrossRef] 45. Griesbach, G.S.; Tio, D.L.; Vincelli, J.; McArthur, D.L.; Taylor, A.N. Differential effects of voluntary and forced exercise on stress responses after traumatic brain injury. J. Neurotrauma 2012, 29, 1426–1433. [CrossRef] 46. Chytrova, G.; Ying, Z.; Gomez-Pinilla, F. Exercise normalizes levels of MAG and Nogo-A growth inhibitors after brain trauma. Eur. J. Neurosci. 2008, 27, 1–11. [CrossRef] 47. Arvidsson, A.; Collin, T.; Kirik, D.; Kokaia, Z.; Lindvall, O. Neuronal replacement from endogenous precursors in the adult brain after stroke. Nat. Med. 2002, 8, 963–970. [CrossRef] 48. Costine, B.A.; Missios, S.; Taylor, S.R.; McGuone, D.; Smith, C.M.; Dodge, C.P.; Harris, B.T.; Duhaime, A.C. The subventricular zone in the immature piglet brain: Anatomy and exodus of neuroblasts into white matter after traumatic brain injury. Dev. Neurosci. 2015, 37, 115–130. [CrossRef] 49. Goodus, M.T.; Guzman, A.M.; Calderon, F.; Jiang, Y.; Levison, S.W. Neural stem cells in the immature, but not the mature, subventricular zone respond robustly to traumatic brain injury. Dev. Neurosci. 2015, 37, 29–42. [CrossRef] 50. Buffo, A.; Vosko, M.R.; Ertürk, D.; Hamann, G.F.; Jucker, M.; Rowitch, D.; Götz, M. Expression pattern of the transcription factor Olig2 in response to brain injuries: Implications for neuronal repair. Proc. Natl. Acad. Sci. USA 2005, 102, 18183–18188. [CrossRef] 51. Amankulor, N.M.; Hambardzumyan, D.; Pyonteck, S.M.; Becher, O.J.; Joyce, J.A.; Holland, E.C. Sonic hedgehog pathway activation is induced by acute brain injury and regulated by injury-related inflammation. J. Neurosci. 2009, 29, 10299–10308. [CrossRef] [PubMed] Int. J. Mol. Sci. 2023, 24, 2911 24 of 24 52. Pous, L.; Deshpande, S.S.; Nath, S.; Mezey, S.; Malik, S.C.; Schildge, S.; Bohrer, C.; Topp, K.; Pfeifer, D.; Fernández-Klett, F.; et al. Fibrinogen induces neural stem cell differentiation into astrocytes in the subventricular zone via BMP signaling. Nat. Commun. 2020, 11, 630. [CrossRef] [PubMed] 53. Dixon, K.J.; Theus, M.H.; Nelersa, C.M.; Mier, J.; Travieso, L.G.; Yu, T.S.; Kernie, S.G.; Liebl, D.J. Endogenous neural stem/progenitor cells stabilize the cortical microenvironment after traumatic brain injury. J. Neurotrauma 2015, 32, 753–764. [CrossRef] [PubMed] 54. Nicolis di Robilant, V. Physical Exercise Induces an Increase of Adult Subventricular Neurogenesis and an Improvement of Olfactory Capacities in a Mouse Model Lacking the P21 Gene, Mediated by Modulation of the BMP Pathway. Bachelor Thesis, University La Sapienza, Rome, Italy, 2017. 55. Brugarolas, J.; Chandrasekaran, C.; Gordon, J.I.; Beach, D.; Jacks, T.; Hannon, G.J. Radiation-induced cell cycle arrest compromised by p21 deficiency. Nature 1995, 377, 552–557. [CrossRef] 56. Mignone, J.L.; Kukekov, V.; Chiang, A.S.; Steindler, D.; Enikolopov, G. Neural stem and progenitor cells in nestin-GFP transgenic mice. J. Comp. Neurol. 2004, 469, 311–324. [CrossRef] 57. Paxinos, G.; Franklin, K. The Mouse Brain in Stereotaxic Coordinates, 5th ed.; Academic Press: Cambridge, MA, USA, 2019; ISBN 9780128161579. 58. Mastrorilli, V.; Scopa, C.; Saraulli, D.; Costanzi, M.; Scardigli, R.; Rouault, J.P.; Farioli-Vecchioli, S.; Tirone, F. Physical exercise rescues defective neural stem cells and neurogenesis in the adult subventricular zone of Btg1 knockout mice. Brain Struct. Funct. 2017, 6, 2855–2876. [CrossRef] [PubMed] Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Role of Running-Activated Neural Stem Cells in the Anatomical and Functional Recovery after Traumatic Brain Injury in p21 Knock-Out Mice.

02-02-2023

Battistini, Jonathan Isacco,Mastrorilli, Valentina,Nicolis di Robilant, Vittoria,Saraulli, Daniele,Marinelli, Sara,Farioli Vecchioli, Stefano

eng

PMC9794057

1 S9 Table. Moderate level of agreement factors. Factors that achieved a level of agreement of 40-69% after all three rounds (n=20). Factor Level of agreement (%) Metabolism Angiogenesis 55,6 Body Muscle fibre transformation capacity 55,6 Tendon stiffness 55,6 Total fat mass 50,0 Weight / BMI 44,4 Lean mass 44,4 Hormones Growth hormone level 66,7 Insulin-like growth factor-1 (IGF-1) level 55,6 Nutrition Vitamin B complex vitamins (B1-12) deficiency 55,6 Immune system Blood pressure regulation 50,0 Healing function of soft tissue 50,0 Injuries Risk of joint injuries 66,7 Risk of upper respiratory tract infections 66,7 Psychological Emotion regulation 66,7 Pain sensitivity 50,0 Self-control 50,0 Resilience capacity 50,0 Concentration capacity 44,4 Environment Altitude training sensitivity 55,6 Heat resistance capacity 50,0

Factors associated with high-level endurance performance: An expert consensus derived via the Delphi technique.

12-27-2022

Konopka, Magdalena J,Zeegers, Maurice P,Solberg, Paul A,Delhaije, Louis,Meeusen, Romain,Ruigrok, Geert,Rietjens, Gerard,Sperlich, Billy

eng

PMC3448081

Nutrients 2012, 4, 949-966; doi:10.3390/nu4080949 nutrients ISSN 2072-6643 www.mdpi.com/journal/nutrients Article Pre-Exercise Hyperhydration-Induced Bodyweight Gain Does Not Alter Prolonged Treadmill Running Time-Trial Performance in Warm Ambient Conditions Pierre-Yves Gigou 1,2, Tommy Dion 1,2, Audrey Asselin 1,2, Felix Berrigan 2 and Eric D. B. Goulet 1,2,* 1 Research Centre on Aging, University of Sherbrooke, Sherbrooke, PQ J1H 4C4, Canada; E-Mails: [email protected] (P.-Y.G.); [email protected] (T.D.); [email protected] (A.A.) 2 Faculty of Physical Education and Sports, University of Sherbrooke, Sherbrooke, PQ J1K 2R1, Canada; E-Mail: [email protected] * Author to whom correspondence should be addressed; E-Mail: [email protected]; Tel.: +1-819-780-2220 (ext. 45226); Fax: +1-819-829-7141. Received: 9 May 2012; in revised form: 17 July 2012 / Accepted: 7 August 2012 / Published: 13 August 2012 Abstract: This study compared the effect of pre-exercise hyperhydration (PEH) and pre-exercise euhydration (PEE) upon treadmill running time-trial (TT) performance in the heat. Six highly trained runners or triathletes underwent two 18 km TT runs (~28 °C, 25%–30% RH) on a motorized treadmill, in a randomized, crossover fashion, while being euhydrated or after hyperhydration with 26 mL/kg bodyweight (BW) of a 130 mmol/L sodium solution. Subjects then ran four successive 4.5 km blocks alternating between 2.5 km at 1% and 2 km at 6% gradient, while drinking a total of 7 mL/kg BW of a 6% sports drink solution (Gatorade, USA). PEH increased BW by 1.00 ± 0.34 kg (P < 0.01) and, compared with PEE, reduced BW loss from 3.1% ± 0.3% (EUH) to 1.4% ± 0.4% (HYP) (P < 0.01) during exercise. Running TT time did not differ between groups (PEH: 85.6 ± 11.6 min; PEE: 85.3 ± 9.6 min, P = 0.82). Heart rate (5 ± 1 beats/min) and rectal (0.3 ± 0.1 °C) and body (0.2 ± 0.1 °C) temperatures of PEE were higher than those of PEH (P < 0.05). There was no significant difference in abdominal discomfort and perceived exertion or heat stress between groups. Our results suggest that pre-exercise sodium-induced hyperhydration of a magnitude of 1 L does not alter 80–90 min running TT performance under warm conditions in highly-trained runners drinking ~500 mL sports drink during exercise. OPEN ACCESS Nutrients 2012, 4 950 Keywords: hyperhydration; hydration; exercise; running; endurance performance; running economy 1. Introduction It has been believed that exercise-induced bodyweight (BW) loss (EIBWL) of ≥2% impairs endurance performance (EP) during exercises conducted under temperate, warm and hot environmental temperatures [1,2]. However, under conditions of temperate and hot ambient temperatures Goulet [3] recently demonstrated through a meta-analysis that EIBWL ≤4% does not impair EP during laboratory-based cycling time-trials (TT) emulating real-world exercise conditions. Of particular importance is that none of the studies included in Goulet [3] showed a statistically significant impairment in cycling TT performance with EIBWL [4–8]. As surprising as it may be, only one laboratory-based study to date has looked at how EIBWL influences running performance. Fallowfield et al. [9] demonstrated that EIBWL of 2% decreased power output by 2.2% during a running test to exhaustion conducted at 70% maximal oxygen consumption (VO2max), 20 °C and 55% relative humidity (RH). This finding is hardly relevant for competitive runners whose racing goal is to run a fixed distance as fast as possible. Moreover, studies have shown that exercise intensity during racing conditions never remains constant but rather constantly varies throughout either on a macro- or micro-scale [10,11]. Finding an answer to how EIBWL impacts running TT performance is imperative in order to widen the evidence base and improve fluid intake guidelines. Several recent field studies have observed a significant relationship between EIBWL and running EP, with the fastest athletes showing the greatest loss in BW [12–14]. Such findings are not easy to explain, but a suggested possibility is that EIBWL could improve running economy, such that athletes who lose the greatest BW are those that best optimize their running speed and, hence, EP [15]. Such assertion makes sense given that there is a strong association between running economy and distance running performance and that this variable is a better predictor of EP than VO2max in elite runners with similar VO2max [16]. Despite the proposed benefits of maintaining EIBWL loss <2% during exercise, real-life competitions are associated with high relative and absolute speeds preventing competitive distance runners from drinking a sufficient volume of fluid to maintain adequate hydration [17]. For example, highly-trained runners have been demonstrated to consume ~150–300 mL/h of fluid and lose more than 2% BW during 15 to 21 km competitive runs [18,19]. Instead of attempting to increase fluid intake during exercise, it could be wiser for runners to hyperhydrate before exercise. In fact, in addition to potentially delaying or preventing EIBWL ≥2%, this technique has been demonstrated to improve cardiovascular and thermoregulatory functions [20] and increase cycling endurance capacity [21], compared with starting an exercise euhydrated. One potential drawback of PEH, however, is that the extra fluid load to be carried could impair running economy and hinder EP, although Beis et al. [22] recently showed that PEH does not alter running economy during a 30 min run conducted at a low intensity. Nutrients 2012, 4 951 This study compared the effect of PEH and PEE in highly-trained runners during an 18 km running TT performed in warm temperature and comprising 8 km at 6% gradient (480 m of vertical climbing). If indeed PEH-associated gain in BW reduces running economy and consequently EP, it is believed that such a protocol would capture it, at least during the inclined parts of the run. We hypothesized that in highly-trained runners the hyperhydration-associated gain in BW would not be sufficient to significantly impact running speed, would improve cardiovascular and thermoregulatory functions and, under an exercise situation where athletes can adjust their speed according to body cues and knowledge of the distance completed, would not provide a hydration-related performance advantage, compared with PEE. 2. Methods 2.1. Subjects Six non-heat-acclimatized, highly-trained competitive male athletes agreed to participate in this study. Among them, three were marathon runners, two were long-distance triathletes (Half-Ironman™ and Ironman™ distance) and one was a short-distance triathlete (Olympic distance). Their mean (±SD) age, height, BW, % fat mass (FM), % fat-free mass (FFM), maximal heart rate and peak oxygen consumption (VO2peak) were 31 ± 7 years, 179 ± 7 cm, 78 ± 10 kg, 11% ± 4%, 89% ± 4%, 192 ± 9 beats/min and 69 ± 3 mL/kg/min, respectively. Subjects were tested over the winter and early spring months of 2011 and were in the preparation phase of their training. The procedures and risks of the study were explained to the six volunteers and informed written consent was obtained. Since the subjects could not be blinded to the treatments they received, neither the specific goals of the study nor the hypotheses tested were explained. All procedures were approved by the University of Sherbrooke Institutional Review Board. 2.2. Overview of the Study After a preliminary visit and a familiarization phase, subjects underwent two experimental trials, started in either a hyperhydrated or euhydrated state, which were conducted in a randomized, crossover fashion, 7–10 days apart, at the same time of the day. After their arrival at the laboratory, participants either passively waited (PEE) or hyperhydrated (PEH) during a 110 min period, after which they underwent an 18 km running TT (comprising a total of 480 m of vertical climbing) on a motorized treadmill at an ambient temperature of ~28 °C and 25%–30% RH. A distance of 18 km was chosen since it was estimated that the course would be completed in ~80 to 90 min, which is the typical time well-trained runners require to complete a half-marathon. The TT was performed under an ambient temperature of 28 °C since no study had yet evaluated the effect of EIBWL upon EP at this temperature. A schematic description of the research protocol is presented in Figure 1. Nutrients 2012, 4 952 Figure 1. Schematic representation of the research protocol. (D or ND) Drink or no drink; (A) Measurement of urine volume, urine specific gravity, bodyweight, heart rate, perceived thirst, abdominal bloating and pain, nausea and dizziness; (B) Measurement of urine volume, urine specific gravity and bodyweight; (C) Measurement of rectal temperature, skin temperature and heart rate; (D) Measurement of perceived exertion, perceived thirst, perceived heat stress and abdominal discomfort; (E) Consumption of 1 mL/kg bodyweight of sports drink; (F) Measurement of perceived exertion and perceived thirst; PEH: Pre-exercise hyperhydration; PEE: Pre-exercise euhydration; * this 4.5 km block was repeated four times; & after running 9 km, subjects stopped for measurement of urine volume, urine specific gravity and bodyweight and to consume one sports gel. 2.3. Preliminary Testing Four to seven days before the familiarization trial, subjects underwent a measurement of height, BW, body composition, VO2peak and maximal heart rate. Height was determined to the nearest 0.5 cm with a wall stadiometer and with subjects wearing only socks. Bodyweight was measured in the nude and post-void to the nearest 100 g with a digital scale (Seca 707, Seca, Germany). Fat mass and FFM were measured using dual-energy X-ray absorptiometry technology (Lunar Prodigy, GE Healthcare, USA). Peak VO2 was measured on a motorized treadmill using an Oxycon Pro (Jaeger, Germany) expired gas analysis system that had been automatically calibrated with gases of known concentration. After subjects had warmed-up for 10–15 min at a self-selected pace and 1% gradient, treadmill speed was adjusted to 10 km/h with a speed increment of 1 km/h/min until 15 km/h, followed by a 2% gradient increment/min until volitional exhaustion of subjects. 2.4. Pre-Experimental Protocol Over the study period (21–27 days), subjects were allowed to continue their training routine but refrained from any physical activity and diuretic substances such as alcohol and caffeine 24 h prior to the three running trials (familiarization run and two experimental runs). Lower leg strength training and dietary supplement intake were forbidden for 48 h prior to the trials. For the last 24 h prior to the familiarization trial, subjects kept and filled a fluid and diet log, which were replicated over the last 24 h prior to the experimental runs. Subjects went to sleep at the same time of the night prior to the running trials. Prior to bedtime and 90 min before their arrival at the laboratory before each trial, subjects consumed 500 mL water. In order to ensure a similar nutritional and hormonal state prior to Nutrients 2012, 4 953 the trials, subject drank a 240 kcal, 237 mL nutritional drink (Boost®, Nestlé, Switzerland) 120 min before reporting to the laboratory. After the drink had been consumed, subjects remained fasted (except for water intake) until the start of the running trials. 2.5. Familiarization Trial Seven to ten days prior to the first experiment a familiarization trial was conducted to minimize any learning effect, familiarize subjects with the measurement techniques and optimize subjects’ pacing strategy for the upcoming two experimental trials. Subjects were required to run as fast as possible during an 18 km TT conducted under the same ambient temperature, RH and wind speed, while wearing the same experimental equipments, clothes and running shoes, drinking the same sports drink, eating the same energy gel, running the same course, listening to the same music and following the same experimental procedures as during the two forthcoming experimental runs. 2.6. Pre-Exercise Hyperhydration and Euhydration Periods Upon arrival at the laboratory, subjects provided a midstream urine sample for urine specific gravity assessment (PAL-10S, Atago, USA), voided their bladder completely (graduated urinal), were weighed in the nude with a high precision scale (Bx-300+, Atron Systems, USA) and instrumented with a T-31 Polar electrode (Polar USA, USA). Following the measurement of heart rate after a 2–3 min seated rest period, subjects rated on a scale of 1 (none) to 5 (extreme) different subjective parameters (perceived thirst, abdominal bloating and pain, nausea and dizziness). Then the 110 min long PEE or PEH period began. No fluid was given to subjects during PEE. During PEH, subjects drank a total of 26 mL fluid/kg BW of a 130 mmol/L (7.5 g NaCl), 4 °C aspartame-flavored (5 g/L) (Crystal Light, Kraftfoods, USA) sodium solution, provided at a rate of 6.5 mL/kg BW every 20 min for the first 60 min. The design of the PEH protocol (total fluid volume, length, rate of ingestion, fluid temperature) was inspired by that previously used by Goulet et al. [21], which was associated with no untoward side-effects. Subjects were required to drink each volume of fluid within 5 min to standardize the time between each urine collection and weighting period. Heart rate, subjective perceptions, urine volume (graduated urinal), urine specific gravity and equipment-corrected BW were sequentially measured in the 18th, 38th, 58th, 78th and 108th min. Subjects were instrumented with the skin probes between the 40th and 80th min, whereas the rectal probe was installed in the 80th min. The changes in BW from before to after the PEE and PEH periods were taken as a reflection of the changes in body water status. Insensible water loss was not measured and was assumed to be similar between trials. A sodium solution was used to induce hyperhydration since the use of glycerol had been banned by the WADA in January 2010 [23]. Results of a pilot study conducted in our laboratory in two highly-trained subjects showed that a 130 mmol/L sodium solution was well-tolerated and produced levels of hyperhydration equivalent to those of glycerol-induced hyperhydration [24]. 2.7. Eighteen Kilometer Time-Trial The 18 km TT run, starting from the 110th minute, consisted of four successive, 4.5 km blocks alternating between 2.5 km at 1% [25] and 2 km at 6% gradient performed on a calibrated motorized Nutrients 2012, 4 954 treadmill (TMX 22, Trackmaster, USA) at ~28 °C (DVTH, Supco, USA) and 25%–30% RH (psychometric chart). To simulate radiant heat stress, two, 500-watts halogen lights (Workshop, Globe electric Company, QC, Canada) were placed ~50 cm above and ~20 cm behind the subject’s head. Before the start of exercise and at the end of each running block, measures of rectal temperature, skin temperature, heart rate, perceived exertion (Borg scale, 20-point scale, 6: very, very light; 20: very, very hard), perceived thirst (11-point scale, 1: none; 11: extreme), perceived heat stress (7-point scale, 1: none; 7: extreme) and abdominal discomfort (5-point scale, 1: none; 5: extreme) were taken. Moreover, for each block, measures of rectal and skin temperatures and heart rate were taken at 2.5 km and 3.5 km, perceived exertion and thirst at 2 km and 3.5 km and abdominal discomfort and perceived heat stress at 2 km. During the runs, subjects received continuous fan-cooling (wind speed of 200 m/min), consumed 1 mL/kg BW of a 6% sports drinks solution (Gatorade, Pepsico, USA) at 2 km and 4.5 km of each block, for a total of 7 mL/kg BW, were encouraged throughout, and were made aware of the distance completed but not of their running speed. To help understand the relationship between BW loss and running speed, subjects were required to stop running for 8 min at 9 km, during which time they were removed from the treadmill, voided their bladder, toweled dry, ate one 110 (27 g carbohydrates (CHO)) kcal energy gel (CarbBoom, Carb Boom Sports Nutrition, Canada) and were weighed while holding the disconnected cables tight to their chest. In addition to these procedures, the 8 min long resting period was necessary for the subjects to reach a respiratory rate that allows a valid measurement of BW and to disconnect and reconnect the skin probes from and to the switch box. At the end of the runs, subjects quickly stepped off the treadmill, voided their bladders, toweled dry and their BW was again measured. Subjects then rapidly removed the rectal and skin probes, running shoes, clothes and Polar electrode worn during the runs and a final nude BW was taken. Finally, a measurement of room temperature and RH was taken. 2.8. Bodyweight Upon arrival at the laboratory and at the end of the running TTs, the running shoes, clothes and equipment worn by subjects during the PEE and PEH periods as well as during the 18 km TT were weighed using a digital compact scale (Symmetry, Cole Parmer, USA). The weight of the tape used to hold the probe cables and that of the energy gel were also measured. Hence, when necessary, measurement of BW was carefully corrected to take into account any excess weight. At 9 km, BW was corrected for half the sweat trapped in the subjects’ clothes, running shoes and tape measured at the end of the running TTs. 2.9. Heart Rate and Rectal, Skin and Body Temperatures Heart rate was measured continuously using a Vantage NV Polar heart rate monitor (Polar USA, USA). Rectal temperature was measured with a YSI 401 rectal probe (Yellow Springs Instrument, USA) inserted 10-cm beyond the anal sphincter and securely held in place with the aid of a lightweight harness developed in our laboratory. Skin temperature was measured with YSI 409 B probes (Yellow Springs Instrument, USA) placed on the left side of the body at the leg, chest and arm level and held in place with Transpore tape (3M, USA). Mean skin and body temperatures were determined as Nutrients 2012, 4 955 suggested by Grucza et al. [26]. The rectal and skin probes were connected to a switch box linked to a high precision digital thermometer (Traceable 4005, Control Company, USA). 2.10. Sweat Loss Sweat loss was computed using the change in BW from the post PEE or PEH period to the post TT period and was corrected for the weight of the energy gel, fluid intake and urine loss during exercise. No correction was made for insensible water loss and the loss of mass associated with the respiratory exchange of O2 and CO2, and all were assumed to be similar between TTs. 2.11. Percent Bodyweight Loss Percent BW loss was computed using the following formula: Percent BW loss = time 0 PEE (or PEH) BW - post TT BW (post void) 100% time 0 pre PEE (or PEH) BW × 2.12. Statistical Analysis The key outcome variable in this study was the difference in TT performance between interventions. On the basis of an estimated CV of 1.5% for the 18 km TT [27,28], a power analysis (α = 0.05, β = 0.2) revealed that six subjects would be sufficient to detect a 2.5% change in TT performance. Data were tested for normality of distribution using the Shapiro-Wilk test and analyzed using either paired sample t-tests and one- or two-way (treatment × time) repeated measures analyses of variance (ANOVA). Sphericity was verified and, if violated, a Greenhouse-Geisser correction was applied. Significance was defined as P < 0.05. Data reported in the text are expressed as means ± SD, and for sake of clarity, those in figures as means ± SEM. Analyses were performed with Microsoft Office Excel 2003 (version 11.8341.8341) and SPSS (version 12.0.0) softwares. 3. Results 3.1. Laboratory Temperature and Relative Humidity For both ambient temperature and RH, a time effect was observed between groups over the course of the study period. Specifically, ambient temperature increased from 27.6 ± 0.1 °C at the onset of the PEE and PEH periods to 27.7 ± 0.2 °C (P = 0.02) at the end of the TTs, whereas the RH changed from 25% ± 2% to 30% ± 2% (P < 0.01) over this time period. No trial or interaction effect was observed. 3.2. Pre-Exercise Hyperhydration and Euhydration Periods 3.2.1. State of Hydration of Subjects at the Arrival at the Laboratory Subjects were adequately and similarly hydrated when they arrived at the laboratory for both trials. This is supported by non-significant differences observed between PEE and PEH with respect to urine specific gravity (1.014 ± 0.010 vs. 1.014 ± 0.004 g/mL, P = 0.97), BW (78.4 ± 9.3 vs. 78.3 ± 9.4 kg, N P 5 3 5 n in B v in u − P w im Nutrients 20 P = 0.49), 56 ± 10 beat 3.2.2. Fluid During P 5.9 ± 0.9 g not ingest m ngest more BW from be volume and nteraction e urine volum −2.52 ± 1.2 PEH. Urine well-hydrate mmediately Figure retenti * Sign 012, 4 urine produ ts/min, P = Balance PEH, subje of sodium. more than 2 than 24 mL efore to afte fluid retent effect in th me (−198 ± 6 mL/kg B specific gr ed state) th y before star e 2. Chang ion (C) dur nificant trial uction (189 0.47). ects ingeste Due to the 23 mL/kg B L/kg BW of er the PEE a tion through he BW cha ± 101 vs. 8 W vs. 1156 ravity was han PEE (1 rting the TT ge in bodyw ring the pr l effect. Res 9 ± 60 vs. ed 1976 ± e time const BW of the f the sodium and PEH pe hout the PE anges (−0.3 819 ± 273 6 ± 309 mL significantly 1.020 ± 0. Ts. weight (A) re-exercise sults are me 130 ± 98 ± 294 mL traints impo sodium-con m-containing eriods, wher EE and PEH 34 ± 0.12 v mL, P < 0 L, 14.81 ± 3 y lower du 006 g/mL, and accum euhydratio an ± SEM. mL, P = 0 of asparta osed at each ntaining flu g fluid. Figu reas Figure H periods, re vs. 1.00 ± 0.01), and 3.32 mL/kg ring PEH ( indicative mulated urin n (■) and 0.53) and h ame-flavore h drinking p uid, and an ure 2A dem 2B,C show espectively. 0.34 kg, P fluid reten g BW, P < (1.005 ± 0.0 of euhydr ne producti hyperhydra heart rate ( ed fluid, to point, one s nother subje monstrates th ws the accum There was P < 0.01), ntion (−198 0.01) betwe 003 g/mL, rated state) ion (B) an ation (●) p 95 (60 ± 11 v ogether wit subject coul ect could no he changes i mulated urin a significan accumulate ± 101 mL een PEE an indicative o (P < 0.05 d fluid periods. 56 vs. th ld ot in ne nt ed L, nd of 5) Nutrients 2012, 4 957 Figure 2. Cont. 3.2.3. Heart Rate and Perceptual Responses Heart rate decreased over time (P = 0.04), but no trial or interaction (P = 0.75) effect between groups was observed. An interaction (P < 0.01) effect was observed for the change in perceived thirst between trials, with thirst increasing over time with PEE but decreasing with PEH. Subjective feelings of dizziness, nausea and abdominal bloating and pain were not significantly different between groups (P > 0.05). 3.3. Time-Trial Periods 3.3.1. Performance There was no trial order effect (P = 0.82), indicating that there was no learning effect from Trial 1 to Trial 2. Moreover, no order effect was observed when the familiarization trial was taken into account (P = 0.28). The TT performance time CV from the familiarization phase to Trial 1 and from Trial 2 to Trial 3 was 2.0% and 2.7%, respectively. There was no difference in the time that the subjects took to complete the 18 km TT (PEE: 85.6 ± 11.6 min; PEH: 85.3 ± 9.6 min, P = 0.82). The first 9 km were completed in 41.5 ± 5.1 min for PEE and 41.8 ± 4.4 min for PEH, respectively (P = 0.64), compared with 44.1 ± 6.8 min for PEE and 43.5 ± 5.3 min for PEH (P = 0.49) for the last 9 km. Figure 3 reports the individual TT performance times observed during the familiarization period and with PEE and PEH. No difference in performance was observed between the familiarization trial (87.20 ± 11.48 min) and the PEE and PEH trials (P = 0.18). As depicted in Figure 4, there was a time effect (P < 0.01) but no trial effect (P = 0.91) or interaction effect (P = 0.45) between groups in running speed throughout the TT. No trial, time or interaction effects between groups were detected regarding the running speed maintained by subjects during the flat portion of the run. However, during the inclined portion of the run, a time effect (P = 0.02) was observed between groups but no trial effect or interaction effect was observed. Nutrients 2012, 4 958 Figure 3. Individual time-trial performance time during the familiarization period and with pre-exercise euhydration (PEE) and hyperhydration (PEH). The dashed line represents the mean (±SEM) time-trial performance time observed in each of the three running trials. Figure 4. Change in running speed throughout the run with pre-exercise euhydration (■) and hyperhydration (●). Results are mean ± SEM. 3.3.2. Fluid Balance Subjects consumed a total volume of 547 ± 69 mL of sports drink for an hourly drinking rate of 403 ± 70 vs. 401 ± 61 mL/h with PEE and PEH, respectively. Total CHO consumption was 60 ± 4 g, with a total of 44 ± 6 and 44 ± 5 g/h for PEE and PEH, respectively. Total sweat loss and hourly sweat rate (PEE: 2642 ± 426 mL, 1878 ± 372 mL/h; PEH: 2652 ± 433 mL, 1876 ± 318 mL/h) did not differ significantly between trials. No time effect, trial effect (P = 0.11) or interaction effect between groups was observed in the change in urine production during exercise (PEE: 30 ± 12 mL; PEH: 73 ± 51 mL). As shown in Figure 5, there were a time effect (P < 0.01) and trial effect (P < 0.01) but no interaction effect between groups regarding the cumulated change in BW from the start of the PEE and PEH periods to the end of the TT. With PEH, a relative loss of BW at the end of the TT amounted to 1.4% ± 0.4%, whereas with PEE this loss reached 3.1% ± 0.3% (P < 0.01). After the first 9 km, subjects had lost 1.7% ± 0.2% of their BW with PEE, compared to 0.00% ± 0.4% with PEH (P < 0.01). Nutrients 2012, 4 959 At the 9 km mark, the difference in BW between groups was 1.22 ± 0.35 kg (P < 0.01), whereas at the end of the TT the difference was 1.19 ± 0.31 kg (P < 0.01). When the loss of BW associated with the waiting period before exercise start was discarded, TT-induced BW loss reached 2.6% ± 0.2% with PEE, which is still significantly more important than PEH (P < 0.01). During exercise, urine specific gravity increased more with PEE than with PEH (P = 0.02), with final values of 1.025 ± 0.003 vs. 1.020 ± 0.005 g/mL, respectively. Figure 5. Accumulated change in bodyweight from the start to the end of the time-trial with pre-exercise euhydration (■) and hyperhydration (●). * Significant trial effect. Results are mean ± SEM. 3.3.3. Heart Rate As depicted in Figure 6, heart rate was significantly lower (−5 ± 1 beats/min) with PEH than PEE throughout the TT, with a time effect (P < 0.01) and group effect (P = 0.03) observed during both the flat and inclined portions of the run. Subjects exercised at a lower percent of maximal heart rate during PEH than PEE (88% ± 5% vs. 90% ± 6%, P = 0.01). Figure 6. Change in heart rate throughout the run with pre-exercise euhydration (■) and hyperhydration (●). * Significant trial effect. Results are mean ± SEM. N 3 a th s 0 s (P (P Nutrients 20 3.3.4. Rectal As shown and mean sk he change stress more t 0.3 ± 0.1 and skin temper P < 0.01) a P = 0.42) an Figure run w Result 012, 4 l and Mean n in Figure kin (B) and in rectal (P than PEE. O d 0.2 ± 0.1 rature. A si across time nd body tem e 7. Change ith pre-exer ts are mean Body and S 7, there wa body (C) te P < 0.01) an On average, °C, respecti ignificant in e between t mperature (P e in rectal (A rcise euhyd ± SEM. Skin Tempe as a signific emperatures nd mean bo , PEH decre ively. No tr nteraction e trials, but n P = 0.06). A) and mea dration (■) erature cant time ef s during the ody (P = 0 eased rectal rial effect (P effect was o no such effe n skin (B) a and hyperh ffect (all P TT. A sign .04) temper l and mean b P = 0.39) w observed in ects were o and body (C hydration ( < 0.01) in t nificant trial rature, with body tempe as observed n the chang observed wi C) temperatu ●). * Signi the change l effect was h PEH redu erature durin d for the cha ge in rectal ith regard t ure through ificant trial 96 in rectal (A observed fo ucing therm ng the TT b ange in mea temperatur to mean ski hout the effect. 60 A) or al by an re in N 3 o s (P d 2 Nutrients 20 3.3.5. Percep Figure 8 of the TT w significantly P = 0.89) a discomfort i 2.2 ± 0.3 un Figure run w Result 012, 4 ptual Respo depicts the with PEE and y lower with and heat str increased s its (P = 0.0 e 8. Change ith pre-exer ts are mean onses changes in d PEH. In a h PEH than ress (P = 0. ignificantly 1), but no tr e in perceiv rcise euhyd ± SEM. perceived e all cases, a t with PEE ( .57) betwee y over time rial (P = 0.1 ved exertion dration (■) exertion (A) time effect (P < 0.01). en groups. F e with a me 17) or intera n (A), thirst and hyperh ), thirst (B) (P < 0.01) There was For both PE ean starting action effect t (B) and h hydration ( and heat st was observ no differen EE and PEH g and final t between g heat stress (C ●). * Signi tress (C) ov ved. Perceiv ce in percei H, feeling o value of 1 groups was o C) through ificant trial 96 ver the cours ved thirst wa ived exertio of abdomin .3 ± 0.2 an observed. out the effect. 61 se as on al nd Nutrients 2012, 4 962 4. Discussion To the best of our knowledge, this is the first study to examine the impact of PEH upon running TT performance. Perhaps even more importantly, this is the first time that a study examines how EIBWL per se influences prolonged running TT performance. The goal of PEH is to maximize EP by preventing a ≥2% EIBWL. Since running economy is closely associated with running EP but PEH can potentially impair this variable, the findings of this study are therefore relevant for runners. Several novel and valuable findings were obtained in the current study. For highly-trained runners who drink ~500 mL of fluid during an 80–90 min intense run, the results suggest that (1) PEH does not contribute to improved running EP although it prevents >2% EIBWL; and (2) a PEH-induced BW gain of 1 kg does not impact running speed. Finally, our results suggest that a 130 mmol/L sodium solution could prove as efficient as glycerol in preventing diuresis-induced fluid loss. It has generally been accepted that ≥2% EIBWL systematically hinders EP regardless of whether the exercise is conducted under temperate, warm or hot temperatures [1]. In the only study conducted using running as a means to test the impact of EIBWL on performance, Fallowfield et al. [9] demonstrated that subjects with a <2% EIBWL took longer to reach exhaustion than those with 2% EIBWL. Goulet et al. [21] demonstrated that, compared with PEE, glycerol-induced hyperhydration that was sufficient to prevent >2% EIBWL significantly increased endurance capacity during an ~12 min long incremental cycling test to exhaustion following 2 h of steady-state cycling. Ebert et al. [29] showed that, despite reducing the metabolic cost of exercise, EIBWL of ~2 kg decreased hill-climb time-to-exhaustion following 2 h of fixed-intensity cycling exercise, compared with a well-hydrated state. Results of the current study clearly contradict those of the aforementioned studies. These discrepant findings between studies can be reconciled on the basis that the three studies showing a positive effect of hydration used fixed-exercise intensity to exhaustion exercise protocols, whereas the current study used a TT performance protocol. In fact, from a statistical point of view, no study to this day has been able to demonstrate that an EIBWL of ≤4% BW impairs EP, compared with a well-hydrated state [4–8]. In a recent meta-analysis, Goulet [3] even demonstrated that EIBWL improves (+0.06%), albeit not significantly, rather than decreases cycling EP. The present study adds a new finding to the literature in showing that attempting to prevent ≥2% EIBWL through the use of PEH is unlikely to confer an increase in running TT performance. Some runners are reluctant to use PEH as they worry about losing speed due to the extra amount of BW needing to be carried. One way through which PEH could potentially reduce running speed is by increasing the O2 cost of running for a given speed. In fact, the relationship between running economy and performance has been well documented, with many studies demonstrating a strong relationship between running economy and distance running performance [30,31]. However, Armstrong et al. [32] demonstrated that the maintenance of euhydration did not increase O2 cost during 10 min of running at 70% and 85% VO2max in highly trained runners, compared with a 5.5% BW loss (2.6 kg) induced by water deprivation. Recently, and of more relevance to the current study, Beis et al. [22] showed that a 0.9 kg hyperhydration BW gain induced by creatine and glycerol did not alter running economy in trained runners completing a 30 min run at 60% VO2max under ambient temperatures of 10 °C and 35 °C. An important limitation of those studies is that no measure of EP was taken. Our findings therefore extend those of the actual literature and suggest that carrying an additional water load of 1 kg Nutrients 2012, 4 963 is unlikely to interfere with running speed and presumably not decrease running economy in trained athletes, both under flat and hilly conditions. Up until 2010, the year when WADA banned glycerol due to its possible masking effect [23], glycerol-induced hyperhydration was the technique of choice used by those athletes wanting to start exercise with an extra fluid reservoir. In fact, in comparison with water-induced hyperhydration, the addition of glycerol (1–1.2 g/kg BW) to a large fluid load (20–26 mL/kg BW) had been demonstrated to increase fluid retention by ~1000 mL, or 13 mL/kg BW, during ~135 min long hyperhydration protocols [24]. Results of the present study suggest that a 130 mmol/L sodium solution ingested at a volume of 26 mL/kg BW provides a fluid retention capability that compares favorably well with the now banned glycerol-induced hyperhydration technique. This effect was likely due to the hypertonic nature of the hyperhydration solution (~330 mOsmol/kg H2O) which, in the face of an accrued body water, likely allowed a slight increase in or at least prevented the excessive decrease of the production of ADH. Hence this enabled a significant reduction of the subsequent rate of water excretion at the kidney level, compared to a situation where only tap water had been ingested. How long could the present PEH protocol sustain a marked increase in body water at rest is not clear. Interestingly, Sharon et al. [33] have demonstrated that the ingestion of 26 mL/kg BW of a 80 mEq/L NaCl solution over 2 h maintains hyperhydration for 4 h, compared with water- and glycerol-induced hyperhydration which sustains hyperhydration for 3 h and 5 h, respectively. Despite these promising findings, further studies directly comparing glycerol- and sodium-induced hyperhydration are needed before any recommendations can confidently be made to athletes. Pre-exercise hyperhydration reduced both heart rate and rectal temperature during exercise, which is in agreement with results of most studies that have compared the effect of PEE and PEH upon physiological functions during exercises ≥45 min [20]. Increased cardiovascular and thermoregulatory strain has been believed to be key causal mechanisms explaining the EIBWL-associated reduction in EP [1]. Clearly, the greater cardiovascular and thermoregulatory challenges encountered by athletes during PEE did not perturb their running ability. Moreover, perceived exertion was not more elevated with PEE, suggesting that these physiological perturbations were not sensed by the brain as adding significant strain to an already severely stressed body. Since no study has demonstrated a deleterious effect of EIBWL upon EP, it must be that the aforementioned relationship between thermoregulatory and cardiovascular strain and EP was established from results of studies that used fixed-intensity exercise protocols. To this effect, Atkinson et al. [34] have argued that when exercise work rate is self-selected, there is evidence to suggest that the pacing strategy, or the “selection” of work rate by athletes, is regulated specifically to ensure that factors that are classically implicated as causing fatigue are instead regulated so that they do not adversely affect physiological variables before the known endpoint of exercise is reached. During fixed-intensity testing protocols where the end-point of exercise is unknown, thereby depriving the brain of a key anchor point, and where the body cannot appropriately deal with physiological insults, commands from the brain could be sent to working muscles to prematurely stop exercising in order to prevent a catastrophic failure from happening [35]. Our findings suggest that despite EIBWL alters the internal milieu in an “unfavorable” manner, the implication and physiological relevance of such changes during running TT performance are likely insignificant, at least in our studied population and within ≤3% EIBWL. Nutrients 2012, 4 964 A possible limitation of the present study relates to its small sample size. However, based on a predicted TT CV of 1%, 1.5% and 2%, a conventional power analysis reveals that 67, 147 and 259 participants, respectively, would have been needed to detect the difference in EP time (0.35%) observed between trials in the current study. Obviously, it would have been impossible to recruit that many highly-trained runners to participate in the study and, from a financial point of view, to test that number of subjects. In order to precisely track changes in BW during the TT, subjects were required to stop running at 9 km. We do not believe that this confounded our findings since running speed remained the same between groups from 7 km to 13.5 km and the pattern of changes in heart rate and rectal temperature between groups was also not altered after 9 km. 5. Conclusions In conclusion, our findings suggest that, although pre-exercise sodium-induced hyperhydration of a magnitude of 1 L and sufficient to prevent >2% EIBWL decreases cardiovascular and thermoregulatory stress, it does not alter 80–90 min running TT performance in highly-trained runners. Further studies using a larger sample size are needed to confirm the present findings. Acknowledgments The authors wish to thank all subjects who participated in this study. This study was made possible through a research grant provided by the Université de Sherbrooke. Conflict of Interest The authors declare no financial or commercial conflict of interest. References 1. Cheuvront, S.N.; Carter, R., III; Sawka, M.N. Fluid balance and endurance exercise performance. Curr. Sports Med. Rep. 2003, 2, 202–208. 2. Sawka, M.N.; Burke, L.M.; Eichner, E.R.; Maughan, R.J.; Montain, S.J.; Stachenfeld, N.S. American college of sports medicine position stand. Exercise and fluid replacement. Med. Sci. Sports Exerc. 2007, 39, 377–390. 3. Goulet, E.D. Effect of exercise-induced dehydration on time-trial exercise performance: A meta-analysis. Br. J. Sports Med. 2011, 45, 1149–1156. 4. Kay, D.; Marino, E.F. Failure of fluid ingestion to improve self-paced exercise performance in moderate-to-warm humid environments. J. Therm. Biol. 2003, 28, 29–34. 5. Robinson, T.A.; Hawley, J.A.; Palmer, G.S.; Wilson, G.R.; Gray, D.A.; Noakes, T.D.; Dennis, S.C. Water ingestion does not improve 1-h cycling performance in moderate ambient temperatures. Eur. J. Appl. Physiol. Occup. Physiol. 1995, 71, 153–160. 6. Bachle, L.; Eckerson, J.; Albertson, L.; Ebersole, K.; Goodwin, J.; Petzel, D. The effect of fluid replacement on endurance performance. J. Strength Cond. Res. 2001, 15, 217–224. 7. Backx, K.; van Someren, K.A.; Palmer, G.S. One hour cycling performance is not affected by ingested fluid volume. Int. J. Sport Nutr. Exerc. Metab. 2003, 13, 333–342. Nutrients 2012, 4 965 8. Dugas, J.P.; Oosthuizen, U.; Tucker, R.; Noakes, T.D. Rates of fluid ingestion alter pacing but not thermoregulatory responses during prolonged exercise in hot and humid conditions with appropriate convective cooling. Eur. J. Appl. Physiol. 2009, 105, 69–80. 9. Fallowfield, J.L.; Williams, C.; Booth, J.; Choo, B.H.; Growns, S. Effect of water ingestion on endurance capacity during prolonged running. J. Sports Sci. 1996, 14, 497–502. 10. Angus, S.D.; Waterhouse, B.J. Pacing strategy from high frequency field data: More evidence for neural regulation? Med. Sci. Sports Exerc. 2011, 43, 2405–2411. 11. Lambert, M.I.; Dugas, J.P.; Kirkman, M.C.; Mokone, G.G.; Waldeck, M.R. Changes in running speeds in a 100 km ultra-marathon race. J. Sports Sci. Med. 2004, 3, 167–173. 12. Kao, W.F.; Shyu, C.L.; Yang, X.W.; Hsu, T.F.; Chen, J.J.; Kao, W.C.; Polun-Chang; Huang, Y.J.; Kuo, F.C.; Huang, C.I.; Lee, C.H. Athletic performance and serial weight changes during 12-and 24-hour ultra-marathons. Clin. J. Sport Med. 2008, 18, 155–158. 13. Zouhal, H.; Groussard, C.; Minter, G.; Vincent, S.; Cretual, A.; Gratas-Delamarche, A.; Delamarche, P.; Noakes, T.D. Inverse relationship between percentage body weight change and finishing time in 643 forty-two-kilometre marathon runners. Br. J. Sports Med. 2011, 45, 1101–1105. 14. Beis, L.Y.; Wright-Whyte, M.; Fudge, B.; Noakes, T.; Pitsiladis, Y.P. Drinking behaviors of elite male runners during marathon competition. Clin. J. Sport Med. 2012, 22, 254–261. 15. Coyle, E.F. Fluid and fuel intake during exercise. J. Sports Sci. 2004, 22, 39–55. 16. Saunders, P.U.; Pyne, D.B.; Telford, R.D.; Hawley, J.A. Factors affecting running economy in trained distance runners. Sports Med. 2004, 34, 465–485. 17. Noakes, T.D. Fluid replacement during exercise. Exerc. Sport Sci. Rev. 1993, 21, 297–330. 18. Burke, L.M.; Wood, C.; Pyne, D.B.; Telford, D.R.; Saunders, P.U. Effect of carbohydrate intake on half-marathon performance of well-trained runners. Int. J. Sport Nutr. Exerc. Metab. 2005, 15, 573–589. 19. Millard-Stafford, M.; Rosskopf, L.B.; Snow, T.K.; Hinson, B.T. Water versus carbohydrate-electrolyte ingestion before and during a 15-km run in the heat. Int. J. Sport Nutr. 1997, 7, 26–38. 20. Goulet, E.D.B. Pre-Exercise hyperhydration: Comments on the 2007 ACSM position stand on exercise and fluid replacement. J. Exerc. Physiol. Online 2008, 11, 64–74. 21. Goulet, E.D.; Rousseau, S.F.; Lamboley, C.R.; Plante, G.E.; Dionne, I.J. Pre-Exercise hyperhydration delays dehydration and improves endurance capacity during 2 h of cycling in a temperate climate. J. Physiol. Anthropol. 2008, 27, 263–271. 22. Beis, L.Y.; Polyviou, T.; Malkova, D.; Pitsiladis, Y.P. The effects of creatine and glycerol hyperhydration on running economy in well trained endurance runners. J. Int. Soc. Sports Nutr. 2011, 16, 24. 23. Koehler, K.; Braun, H.; de Marees, M.; Geyer, H.; Thevis, M.; Mester, J.; Schaenzer, W. Urinary excretion of exogenous glycerol administration at rest. Drug Test. Anal. 2011, 3, 877–882. 24. Goulet, E.D.; Aubertin-Leheudre, M.; Plante, G.E.; Dionne, I.J. A meta-analysis of the effects of glycerol-induced hyperhydration on fluid retention and endurance performance. Int. J. Sport Nutr. Exerc. Metab. 2007, 17, 391–410. 25. Jones, A.M.; Doust, J.H. A 1% treadmill grade most accurately reflects the energetic cost of outdoor running. J. Sports Sci. 1996, 14, 321–327. Nutrients 2012, 4 966 26. Grucza, R.; Szczypaczewska, M.; Kozłowski, S. Thermoregulation in hyperhydrated men during physical exercise. Eur. J. Appl. Physiol. Occup. Physiol. 1987, 56, 603–607. 27. Rollo, I.; Williams, C.; Nevill, A. Repeatability of scores on a novel test of endurance running performance. J. Sports Sci. 2008, 26, 1379–1386. 28. Russell, R.D.; Redmann, S.M.; Ravussin, E.; Hunter, G.R.; Larson-Meyer, D.E. Reproducibility of endurance performance on a treadmill using a preloaded time trial. Med. Sci. Sports Exerc. 2004, 36, 717–724. 29. Ebert, T.R.; Martin, D.T.; Bullock, N.; Mujika, I.; Quod, M.J.; Farthing, L.A.; Burke, L.M.; Withers, R.T. Influence of hydration status on thermoregulation and cycling hill climbing. Med. Sci. Sports Exerc. 2007, 39, 323–329. 30. Di Prampero, P.E.; Capelli, C.; Pagliaro, P.; Antonutto, G.; Girardis, M.; Zamparo, P.; Soule, R.G. Energetics of best performances in middle-distance running. J. Appl. Physiol. 1993, 74, 2318–2324. 31. Conley, D.L.; Krahenbuhl, G.S. Running economy and distance running performance of highly trained athletes. Med. Sci. Sports Exerc. 1980, 12, 357–360. 32. Armstrong, L.E.; Whittlesey, M.J.; Casa, D.J.; Elliott, T.A.; Kavouras, S.A.; Keith, N.R.; Maresh, C.M. No effect of 5% hypohydration on running economy of competitive runners at 23 degrees C. Med. Sci. Sports Exerc. 2006, 38, 1762–1769. 33. Griffin, S.E.; Ghiasvand, F.; Gibson, A.; Orri, J.; Burns, S.; Robergs, R.A. Comparing hyperhydration resulting from the ingestion of glycerol, carbohydrate and saline solutions. Presented at American Society of Exercise Physiologists 2nd Annual Meeting, Albuquerque, NM, USA, October 1999; Abstract #2. Available online: http://faculty.css.edu/tboone2/asep/sharon.pdf (accessed on 10 March 2012). 34. Atkinson, G.; Peacock, O.; St Clair Gibson, A.; Tucker, R. Distribution of power output during cycling: Impact and mechanisms. Sports Med. 2007, 37, 647–667. 35. Noakes, T.D.; St Clair Gibson, A.; Lambert, E.V. From catastrophe to complexity: A novel model of integrative central neural regulation of effort and fatigue during exercise in humans. Br. J. Sports Med. 2004, 38, 511–514. © 2012 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution license (http://creativecommons.org/licenses/by/3.0/).

Pre-exercise hyperhydration-induced bodyweight gain does not alter prolonged treadmill running time-trial performance in warm ambient conditions.

08-13-2012

Gigou, Pierre-Yves,Dion, Tommy,Asselin, Audrey,Berrigan, Felix,Goulet, Eric D B

eng

PMC10496601

Sao Paulo Med J. 2016; 134(3):193-8 193 ORIGINAL ARTICLE DOI: 10.1590/1516-3180.2014.8921512 Post-analysis methods for lactate threshold depend on training intensity and aerobic capacity in runners. An experimental laboratory study Métodos de pós-análise do limiar do lactato dependem da intensidade de treinamento e da capacidade aeróbica dos corredores. Um estudo laboratorial experimental Tiago Lazzaretti FernandesI, Rômulo dos Santos Sobreira NunesII, Cesar Cavinato Cal AbadIII, Andrea Clemente Baptista SilvaIV, Larissa Silva SouzaIV, Paulo Roberto Santos SilvaV, Cyro AlbuquerqueVI, Maria Cláudia IrigoyenVII, Arnaldo José HernandezVIII Laboratório do Estudo do Movimento (LEM), Instituto de Ortopedia e Traumatologia (IOT), Hospital das Clínicas (HC), Faculdade de Medicina da Universidade de São Paulo (FMUSP), and Sports Medicine Group of FMUSP, FIFA Medical Centre of Excellence, IOT HC-FMSUP, São Paulo, Brazil ABSTRACT CONTEXT AND OBJECTIVE: This study aimed to evaluate different mathematical post-analysis methods of determining lactate threshold in highly and lowly trained endurance runners. DESIGN AND SETTING: Experimental laboratory study, in a tertiary-level public university hospital. METHOD: Twenty-seven male endurance runners were divided into two training load groups: lowly trained (frequency < 4 times per week, < 6 consecutive months, training velocity ≥ 5.0 min/km) and highly trained (frequency ≥ 4 times per week, ≥ 6 consecutive months, training velocity < 5.0 min/km). The subjects performed an incremental treadmill protocol, with 1 km/h increases at each subsequent 4-minute stage. Fingerprint blood-lactate analysis was performed at the end of each stage. The lactate threshold (i.e. the running velocity at which blood lactate levels began to exponentially increase) was measured using three different methods: increase in blood lactate of 1 mmol/l at stages (DT1), absolute 4 mmol/l blood lactate concentration (4 mmol), and the semi-log method (semi-log). ANOVA was used to compare different lactate threshold methods and training groups. RESULTS: Highly trained athletes showed significantly greater lactate thresholds than lowly trained run- ners, regardless of the calculation method used. When all the subject data were combined, DT1 and semi- log were not different, while 4 mmol was significantly lower than the other two methods. These same trends were observed when comparing lactate threshold methods in the lowly trained group. However, 4 mmol was only significantly lower than DT1 in the highly trained group. CONCLUSION: The 4 mmol protocol did not show lactate threshold measurements comparable with DT1 and semi-log protocols among lowly trained athletes. RESUMO CONTEXTO E OBJETIVO: O objetivo do presente estudo é avaliar modelos matemáticos de pós-análise do limiar de lactato em grupos de corredores de longa distância muito ou pouco treinados. TIPO DE ESTUDO E LOCAL: Estudo laboratorial experimental. Hospital Público Universitário Terciário. MÉTODO: Vinte e sete corredores homens foram divididos em: pouco treinados (frequência < 4 vezes por semana, < 6 meses, velocidade ≥ 5,0 minutos/km) e muito treinados (frequência ≥ 4 vezes por sema- na, ≥ 6  meses, velocidade < 5,0 minutos/km). Os participantes foram submetidos a protocolo de esteira escalonado (1% inclinação) = 1 km/h por fase (4 minutos). Ao fim de cada estágio, análise da “impressão digital” metabolômica foi realizada. O limiar do lactato (i.e. velocidade em que o lactato sanguíneo aumen- ta exponencialmente) foi medido utilizando-se três métodos: aumento de 1 mmol/l da concentração, concentração absoluta de 4 mmol e método semi-log. ANOVA foi utilizada para comparar os diferentes limiares de lactato e grupos. RESULTADO: Atletas muito treinados apresentaram limiares de lactato maiores que os corredores pouco treinados, independentemente do método de cálculo utilizado. Comparando todos os corredores juntos, as análises de aumento de 1 mmol/l e semi-log não foram diferentes, enquanto a concentração absoluta de 4 mmol/l foi significativamente mais baixa que as dos dois outros métodos. Essas mesmas tendências foram observadas ao se compararem os métodos de limiar de lactato no grupo menos treinado. Entretan- to, a análise absoluta de 4 mmol/l foi menor do que a do aumento de 1 mmol/l no grupo muito treinado. CONCLUSÃO: O método concentração absoluta de 4 mmol não mostrou mensurações comparáveis de limiar do lactato quando comparado com os protocolos aumento de 1 mmol/l e semi-log nos atletas pouco treinados. IMD, MSc. Doctoral Student and Attending Physician, Sports Medicine Group, FIFA Medical Centre of Excellence, Faculdade de Medicina da Universidade de São Paulo (FMUSP), and Instituto de Ortopedia e Traumatologia (IOT), Hospital das Clínicas (HC), São Paulo, Brazil. IIUndergraduate Student, Faculdade de Medicina da Universidade de São Paulo (FMUSP), São Paulo, Brazil. IIIMSc, PhD. Heart Institute, Faculdade de Medicina da Universidade de São Paulo (FMUSP), and Instituto do Coração (InCor), São Paulo, Brazil. IVMD. Sports Medicine Group, Faculdade de Medicina da Universidade de São Paulo (FMUSP), and Instituto de Ortopedia e Traumatologia (IOT), Hospital das Clínicas (HC), São Paulo, Brazil. VPhD. Sports Medicine Group, FIFA Medical Centre of Excellence, Faculdade de Medicina da Universidade de São Paulo (FMUSP), and Instituto de Ortopedia e Traumatologia (IOT), Hospital das Clínicas (HC), São Paulo, Brazil. VIMSC, PhD. Assistant Professor, Department of Mechanical Engineering, Centro Universitário da FEI, São Bernando do Campo, Brazil. VIIMD, PhD. Professor, Faculdade de Medicina da Universidade de São Paulo (FMUSP), and Instituto do Coração (InCor), São Paulo, Brazil. VIIIPhD. Assistant Professor, Director of Sports Medicine Group, FIFA Medical Centre of Excellence, Faculdade de Medicina da Universidade de São Paulo (FMUSP), and Instituto de Ortopedia e Traumatologia (IOT), Hospital das Clínicas (HC), São Paulo, Brazil. KEY WORDS: Lactic acid. Physical endurance. Anaerobic threshold. Oxygen consumption. Exercise test. Sports medicine. PALAVRAS-CHAVE: Ácido láctico. Resistência física. Limiar anaeróbio. Consumo de oxigênio. Teste de esforço. Medicina esportiva. ORIGINAL ARTICLE | Fernandes TL, Nunes RSS, Abad CCC, Silva ACB, Souza LS, Silva PRS, Albuquerque C, Irigoyen MC, Hernandez AJ 194 Sao Paulo Med J. 2016; 134(3):193-8 INTRODUCTION Blood lactate evaluation commonly complements endurance training regimens.1,2 It has been recommended as an efficient method for evaluating training intensity and recovery, and for improving the performance of endurance athletes.3-6 During incremental exercise, the lactate threshold (LT) is defined as the abrupt transition from slow increases to rapid exponential increases in blood lactate levels.7 The evaluation of lactate threshold in athletes has evolved, from the 4 mmol universal lactate threshold, to the more indi- vidualized Onset of Blood Lactate Accumulation, and to the cur- rent Maximal Lactate Steady State standard. This progression has been due to better understanding of the physiological processes of lactate production and clearance, and the role of lactate during prolonged and submaximal exercise.5,8-12 However, most published studies on lactate threshold have compared homogeneous groups of athletes with similar aero- bic capacity, or have made regression analyses on these data.13-16 Comparisons between different methods on lactate thresh- old acquisition also remain controversial in the literature.17-19 To  our knowledge, there is no comparative study evaluat- ing lactate threshold methods in both lowly and highly trained endurance athletes. OBJECTIVE The purpose of this study was to evaluate different lactate thresh- old methods, and determine which methods are most reliable for athletes with different physical conditioning and training programs. METHODS This was an experimental laboratory study performed within the Sports Medicine Group of Faculdade de Medicina da Universidade de São Paulo. Twenty-seven male endurance run- ners were recruited for this study from university campus run- ning clubs. For the primary outcome (post-analysis method for the lactate threshold in the same group), the sample size was cal- culated after a five-athlete pilot study, taking P < 0.05 and power = 80%. The sample size was estimated as 10 individuals per group. We added a minimum of 20% more subjects to account for potential data loss. The subjects were divided into two distinct groups based on the responses to a questionnaire: 15 highly trained runners (min- imum of 4 training runs per week for 6 consecutive months, and a long-distance training pace less than or equal to 5.0 min/km) and 12 lowly trained runners (long-distance training pace greater than 5.0 min/km, with a maximum of 3 runs per week and a maximum of 6 consecutive training months). The exclu- sion criteria were previous cardiorespiratory disease and musculoskeletal running-related injuries. No athlete was cur- rently taking any medications. Oxygen consumption (VO2) was measured continuously and monitored by means of a breath-by-breath gas analyzer on a treadmill (h/p/cosmos, Pulsar, Germany) using a metabolic ana- lyzer (CPX/D Med Graphics, St. Paul, USA) The mean physiological characteristics of the highly trained group were: age 33.7 ± 10.3 years; training velocity: 4.0 ± 0.6 min/km; resting heart rate 68.7 ± 14.7 bpm; and VO2max: 52.4 ± 5.3 ml/kg/min. Characteristics of the lowly trained group were: age 37.2 ± 9.3 years; training velocity: 5.3 ± 0.9 min/km; rest heart rate 79.3 ± 15.2 bpm; and VO2max: 43.4 ± 5.7ml/kg/min. The Institutional Review Board approved this research and informed consent was obtained from each subject prior to partici- pation. This research followed the Helsinki Declaration principles.20 Lactate protocol A washout period of 24 hours with no physical activity was requested for all participants prior to the experiment. The sub- jects then performed an incremental treadmill test to directly measure their lactate threshold. All subjects did the test at the same location, with the same equipment, and under similar ther- mal conditions (temperature 21-26°C, humidity 33-66%, baro- metric pressure 688 mmHg). Throughout the protocol, treadmill elevation was kept constant at a 1% grade to duplicate the energy cost of over-ground running.21 The subjects first performed a 3-minute warm-up run at 30% of their long-distance training velocity. At the beginning of the incremental test, the treadmill velocity was set at 70% of the esti- mated long distance training velocity, depending on the running ability of each participant (it is known that performance in com- petition is an appropriate criterion for valid laboratory tests).3,22 Heart rate and Borg scale were recorded each minute. Stage length was set at 4 minutes,21 with running velocity increases of 1 km/h per stage until volitional exhaustion was reached (as mea- sured from the Borg scale). Fingerprint whole-blood samples were taken between the points of 3.5 and 4 minutes in each stage and were immediately analyzed in an automated blood-lactate ana- lyzer (Accutrend Lactate, Typ3012522) without treadmill proto- col interruption. Blood samples were collected for two additional stages following exponential inflection of the lactate point. Calculating lactate threshold The basis for determining the lactate threshold is that there is an inflection point at a given workload (i.e. running velocity) where blood lactate exponentially increases with a corresponding increase in workload.17,18,21,23,24 It is used to define the highest work rate or O2 uptake (oxygen consumption) at which athletes can maintain their efforts over a specified time frame.25 An individual blood-lactate Post-analysis methods for lactate threshold depend on training intensity and aerobic capacity in runners. An experimental laboratory study | ORIGINAL ARTICLE Sao Paulo Med J. 2016; 134(3):193-8 195 profile was created for each subject by plotting running velocity (km/h) at each stage of the test (x-axis) versus blood-lactate con- centration attained at each stage (y-axis).10,21,24,26,27 Three methods commonly cited in the literature were used to define the inflection point (Figure 1): 1. Increase of 1 mmol/l blood lactate (DT1): the work rate that just precedes a rise in blood lactate concentration of > 1 mmol/l between two stages estimates the lactate threshold.10,21,26,27 2. Absolute value of 4 mmol/l blood lactate (4 mmol): work- load when the concentration of lactate in the blood reaches 4 mmol/l.10,21,26,27 3. Semi-log method (semi-log): based on a logarithmic scale (blood lactate) in which the exponential blood lactate curve is divided into two linear segments that cross each other; the point of intersection is the lactate threshold.17,18 Statistical analysis The normality curve was addressed by means of histograms and it was decided to use parametric tests. Baseline characteristics were analyzed first to demonstrate homogeneity. The threshold values of each method were compared with repeated-measurement analyses of variance (ANOVA). When a significant difference was attained, Tukey’s post-hoc test was performed. Statistical significance was denoted as P < 0.05 (STATA-9 for Windows). RESULTS Before analyzing the relationship between lactate threshold and velocity, the associations between lactate threshold and base- line characteristics such as age, heart rate, VO2max and train- ing regularity were assessed. This analysis showed that neither demographic nor baseline characteristics could explain associa- tions with lactate threshold, except, logically, for the dependent variables of training regularity between groups and VO2max. As expected, the lactate thresholds of the highly trained group were obtained at higher velocity stages than those of the lowly trained group in all tested methods. When considering all subjects (both the highly trained and the lowly trained groups), comparison of lactate threshold methods showed significant dif- ferences between the DT1 and 4 mmol methods, and between the semi-log and 4 mmol methods. There was no statistical difference between DT1 and semi-log (Figure 2). 10 1 mmol/l increase method (DT1) (a) 1 mmol/l increase threshold method 9 8 7 6 5 4 3 2 11 12 13 14 15 16 Velocity (km/h) Lactate (mmol/l) Log lactate (mmol/l) 17 18 19 20 21 1 (b) semi-log method (intersection) 0.9 0.8 0.7 0.6 0.5 11 12 13 14 15 16 Velocity (km/h) 17 18 19 20 21 Figure 1. Example of the three post-analysis methods for lactate threshold applied to one subject: (a) the velocity before the increase of 1 mmol/l blood lactate (DT1); the velocity at which the blood lactate exceeds the value of 4 mmol/l (4 mmol); and (b) the velocity at the intersection of two interpolated lines on the semi-logarithmic scale (semi-log). 18 *# 16 14 12 10 8 6 4 DT1 semi-log Velocity (km/h) 4 mmol Figure 2. Box plot of the velocity of the lactate threshold of all subjects obtained using the DT1, semi-log and 4 mmol post-analysis methods. *Significant difference between DT1 and 4 mmol methods (P < 0.05), and #significant difference between semi-log and 4 mmol (P < 0.05), ANOVA (analysis of variance), Bonferroni post-hoc test. ORIGINAL ARTICLE | Fernandes TL, Nunes RSS, Abad CCC, Silva ACB, Souza LS, Silva PRS, Albuquerque C, Irigoyen MC, Hernandez AJ 196 Sao Paulo Med J. 2016; 134(3):193-8 When the groups were compared separately (highly trained and lowly trained), the 4 mmol measurement was found to be sig- nificantly lower than the DT1 and semi-log measurements in the lowly trained group. The DT1 and semi-log measurements were not statistically different in this group (Figure 3). In the highly trained group, a significant difference was only found between the DT1 and the 4 mmol methods. DISCUSSION The most important finding of this study was the differences in lactate threshold measurement methods between highly and lowly trained endurance runners. The method with fixed blood lactate of 4 mmol/l underestimated the lactate threshold in the lowly trained group. Sargent et  al.28 identified differences in lactate threshold between different groups of subjects, such as men versus women. On the other hand, Smekal et  al.29 reported that blood lactate concentration at the maximal lactate steady state was indepen- dent of both endurance capacity and sex. Other authors have showed comparisons between trained and untrained individuals through using cardiorespiratory tests.30,31 One notable character- istic of our study is that we only used male subjects and made comparisons between controlled training levels (high and low) instead of between trained and sedentary subjects. One explanation for the different values of measurement methods is that error is introduced when the curves do not fol- low the mathematical physiological functions.19 Subjects with different performance levels often have differ- ent mechanical running responses and consequently different metabolic demands.32 Our study also agreed with the literature regarding higher lactate threshold values in trained individuals. Kumagai et al.33 showed that aerobic training increased the lac- tate threshold, with a concomitant improvement in both endur- ance and middle-distance performance. Individuals with greater endurance capacity have faster oxy- gen kinetics.34 The higher values of lactate thresholds in the highly trained subjects may reflect more efficient peripheral and central exchange during exercise.34 During low-intensity exercise, blood lactate formation and removal depends on the intracellular/tissue balance among the glycolytic (cytosol) and oxidative (mitochon- dria) processes.35 It seems that in trained individuals, these vari- ables are more predictable and have controlled behavior. Joyner et al.36 suggested that running performance could be explained by VO2max, running economy and fractional utiliza- tion of VO2max. Moreover, they suggested that the lactate thresh- old integrates all three of these variables and is the best physio- logical predictor of distance running performance1, given that it is detectable in both trained and untrained individuals.10 It is known that the average value for the lactate threshold in normal subjects is 3.7 mmol and that serum blood lactate at the lactate threshold is not equal for all individuals (range: 1.5 to 7.5 mmol) and also changes in a single individual.37 Although the 4 mmol lactate protocol is an easy method for estimating lactate threshold, the fixed value of 4.0 mmol does not take these physi- ological conditions into consideration and may underestimate lac- tate threshold, as shown in this study.38 The clinical relevance of this study relates to the populations tested. Most people are not competitive endurance athletes, yet still need predictions of aerobic threshold and exercise prescrip- tions for health issues. Our results suggest that lowly trained sub- jects would benefit from semi-log or DT1 lactate threshold meth- ods in clinical practice. The main limitation of this study relates to the treadmill pro- tocol, such as the stage duration and initial running velocity. Due to the large variation of treadmill protocols in the lactate threshold literature, direct comparisons of our results with pre- vious studies may not be appropriate. Despite training group characteristics that were very specific (frequency, intensity and duration of training), we believe that they represent objective inclusion criteria and, because of that, the results may be repro- ducible. Future studies should examine lactate threshold meth- ods and cardiorespiratory performance in both highly trained and lowly trained groups. We suggest that these groups should be stratified according to training frequency, intensity and duration. CONCLUSION The 4 mmol protocol did not show lactate threshold measure- ments comparable with with DT1 and semi-log protocols among lowly trained athletes. 18 *# * 16 14 12 10 8 6 4 DT1 semi-log highly trained lowly trained 4 mmol DT1 semi-log 4 mmol Velocity (km/h) Figure 3. Box plot of the velocities at the lactate threshold obtained using the DT1, semi-log and 4 mmol post-analysis methods in each group (highly trained and lowly trained). *Significant difference between DT1 and 4 mmol methods (P < 0.05), and #significant difference between semi-log and 4 mmol (P < 0.05), using ANOVA (analysis of variance) and the Bonferroni post-hoc test. Post-analysis methods for lactate threshold depend on training intensity and aerobic capacity in runners. An experimental laboratory study | ORIGINAL ARTICLE Sao Paulo Med J. 2016; 134(3):193-8 197 REFERENCES 1. Bassett DR Jr, Howley ET. Limiting factors for maximum oxygen uptake and determinants of endurance performance. Med Sci Sports Exerc. 2000;32(1):70-84. 2. Hollmann W. 42 years ago--development of the concepts of ventilatory and lactate threshold. Sports Med. 2001;31(5):315-20. 3. Baldari C, Videira M, Madeira F, Sergio J, Guidetti L. Blood lactate removal during recovery at various intensities below the individual anaerobic threshold in triathletes. J Sports Med Phys Fitness. 2005;45(4):460-6. 4. Beneke R. Methodological aspects of maximal lactate steady state-implications for performance testing. Eur J Appl Physiol. 2003;89(1):95-9. 5. Demarle AP, Heugas AM, Slawinski JJ, et al. Whichever the initial training status, any increase in velocity at lactate threshold appears as a major factor in improved time to exhaustion at the same severe velocity after training. Arch Physiol Biochem. 2003;111(2):167-76. 6. Sjödin B, Jacobs I, Svedenhag J. Changes in onset of blood lactate accumulation (OBLA) and muscle enzymes after training at OBLA. Eur J Appl Physiol Occup Physiol. 1982;49(1):45-57. 7. Svedahl K, MacIntosh BR. Anaerobic threshold: the concept and methods of measurement. Can J Appl Physiol. 2003;28(2):299-323. 8. Farrell PA, Wilmore JH, Coyle EF, Billing JE, Costill DL. Plasma lactate accumulation and distance running performance. Med Sci Sports. 1979;11(4):338-44. 9. Faude O, Kindermann W, Meyer T. Lactate threshold concepts: how valid are they? Sports Med. 2009;39(6):469-90. 10. Nicholson RM, Sleivert GG. Indices of lactate threshold and their relationship with 10-km running velocity. Med Sci Sports Exerc. 2001;33(2):339-42. 11. Tanaka K, Matsuura Y, Matsuzaka A, et al. A longitudinal assessment of anaerobic threshold and distance-running performance. Med Sci Sports Exerc. 1984;16(3):278-82. 12. Weltman A. The blood lactate response to exercise. Champaign: Human Kinetics Publisher; 1995. 13. Carlsson M, Carlsson T, Hammarström D, et  al. Validation of physiological tests in relation to competitive performances in elite male distance cross-country skiing. J Strength Cond Res. 2012;26(6):1496-504. 14. de Sousa NM, Magosso RF, Pereira GB, et  al. The measurement of lactate threshold in resistance exercise: a comparison of methods. Clin Physiol Funct Imaging. 2011;31(5):376-81. 15. Sperlich B, Krueger M, Zinner C, et  al. Oxygen uptake, velocity at lactate threshold, and running economy in elite special forces. Mil Med. 2011;176(2):218-21. 16. Weltman A, Snead D, Seip R, et  al. Prediction of lactate threshold and fixed blood lactate concentrations from 3200-m running performance in male runners. Int J Sports Med. 1987;8(6):401-6. 17. Bishop D, Jenkins DG, Mackinnon LT. The relationship between plasma lactate parameters, Wpeak and 1-h cycling performance in women. Med Sci Sports Exerc. 1998;30(8):1270-5. 18. Davis JA, Rozenek R, DeCicco DM, Carizzi MT, Pham PH. Comparison of three methods for detection of the lactate threshold. Clin Physiol Funct Imaging. 2007;27(6):381-4. 19. Tokmakidis SP, Léger LA, Pilianidis TC. Failure to obtain a unique threshold on the blood lactate concentration curve during exercise. Eur J Appl Physiol Occup Physiol. 1998;77(4):333-42. 20. WMA Declaration of Helsinki - Ethical Principles for Medical Research Involving Human Subjects. Available from: http://www.wma.net/ en/30publications/10policies/b3/. Accessed in 2015 (Jan 7). 21. McGehee JC, Tanner CJ, Houmard JA. A comparison of methods for estimating the lactate threshold. J Strength Cond Res. 2005;19(3):553-8. 22. Ament W, Verkerke GJ. Exercise and fatigue. Sports Med. 2009;39(5):389-422. 23. Robergs RA, Roberts SO. Exercise physiology: exercise, performance and clinical applications. St Louis: Mosby; 1996. 24. Yoshida T, Chida M, Ichioka M, Suda Y. Blood lactate parameters related to aerobic capacity and endurance performance. Eur J Appl Physiol Occup Physiol. 1987;56(1):7-11. 25. Lorenzo S, Minson CT, Babb TG, Halliwill JR. Lactate threshold predicting time-trial performance: impact of heat and acclimation. J Appl Physiol (1985). 2011;111(1):221-7. 26. Machado FA, de Moraes SM, Peserico CS, Mezzaroba PV, Higino WP. The Dmax is highly related to performance in middle-aged females. Int J Sports Med. 2011;32(9):672-6. 27. Machado FA, Nakamura FY, Moraes SM. Influence of regression model and incremental test protocol on the relationship between lactate threshold using the maximal-deviation method and performance in female runners. J Sports Sci. 2012;30(12):1267-74. 28. Sargent C, Scroop GC. Plasma lactate accumulation is reduced during incremental exercise in untrained women compared with untrained men. Eur J Appl Physiol. 2007;101(1):91-6. 29. Smekal G, von Duvillard SP, Pokan R, et al. Blood lactate concentration at the maximal lactate steady state is not dependent on endurance capacity in healthy recreationally trained individuals. Eur J Appl Physiol. 2012;112(8):3079-86. 30. Del Coso J, Hamouti N, Aguado-Jimenez R, Mora-Rodriguez R. Respiratory compensation and blood pH regulation during variable intensity exercise in trained versus untrained subjects. Eur J Appl Physiol. 2009;107(1):83-93. 31. Takeshima N, Kobayashi F, Watanabe T, et al. Cardiorespiratory responses to cycling exercise in trained and untrained healthy elderly: with special reference to the lactate threshold. Appl Human Sci. 1996;15(6):267-73. 32. Ramsbottom R, Williams C, Fleming N, Nute ML. Training induced physiological and metabolic changes associated with improvements in running performance. Br J Sports Med. 1989;23(3):171-6. ORIGINAL ARTICLE | Fernandes TL, Nunes RSS, Abad CCC, Silva ACB, Souza LS, Silva PRS, Albuquerque C, Irigoyen MC, Hernandez AJ 198 Sao Paulo Med J. 2016; 134(3):193-8 33. Kumagai S, Tanaka K, Matsuura Y, et al. Relationships of the anaerobic threshold with the 5 km, 10 km, and 10 mile races. Eur J Appl Physiol Occup Physiol. 1982;49(1):13-23. 34. Whipp BJ, Davis JA, Torres F, Wasserman K. A test to determine parameters of aerobic function during exercise. J Appl Physiol Respir Environ Exerc Physiol. 1981;50(1):217-21. 35. Brooks GA. Cell-cell and intracellular lactate shuttles. J Physiol. 2009;587(Pt 23):5591-600. 36. Joyner MJ, Coyle EF. Endurance exercise performance: the physiology of champions. J Physiol. 2008;586(1):35-44. 37. Pedersen PK, Sj G, Juel C. Plasma acid-base status and hyperventilation during cycling at MAXLASS in low and high lactate responders. Medicine & Science in Sports & Exercise. 2011;33(5):S314. 38. Coyle EF. Integration of the physiological factors determining endurance performance ability. Exerc Sport Sci Rev. 1995;23:25-63. Acknowledgements: The authors would like to thank Sean J. Driscoll for proofreading this manuscript Sources of funding: Fundação de Amparo à Pesquisa do Estado de São Paulo (Fapesp) – Protocolo 2010/19631-2 Conflict of interest: None Date of first submission: May 11, 2014 Last received: November 19, 2014 Accepted: December 15, 2014 Address for correspondence: Tiago Lazzaretti Fernandes Laboratório do Estudo do Movimento Instituto de Ortopedia e Traumatologia, Hospital das Clínicas Faculdade de Medicina da Universidade de São Paulo Dr. Ovídio Pires de Campos, 333 — 2o andar São Paulo (SP) — Brasil CEP 05403-010 Tel. (+55 11) 2661-6486 E-mail: [email protected]

Post-analysis methods for lactate threshold depend on training intensity and aerobic capacity in runners. An experimental laboratory study.

[]

Fernandes, Tiago Lazzaretti,Nunes, Rômulo Dos Santos Sobreira,Abad, Cesar Cavinato Cal,Silva, Andrea Clemente Baptista,Souza, Larissa Silva,Silva, Paulo Roberto Santos,Albuquerque, Cyro,Irigoyen, Maria Cláudia,Hernandez, Arnaldo José

eng

PMC9295982

RESEARCH ARTICLE A 7-min halftime jog mitigated the reduction in sprint performance for the initial 15-min of the second half in a simulated football match Sooil Bang1, Jihong ParkID2* 1 Athletic Training Laboratory, Graduate School of Physical Education, Kyung Hee University, Yongin, Korea, 2 Department of Sports Medicine, Athletic Training Laboratory, Kyung Hee University, Yongin, Korea * [email protected] Abstract This study compared the effects of a 7-min shuttle jog during halftime to a control condition (seated rest) on subsequent athletic performance and lower-leg temperature in the second half. Eighteen male football players (22 years, 179 cm, 70 kg, 10 years of athletic career) ran- domly performed a 20-m shuttle jog (at an intensity of 70% of heart rate maximum) and a seated rest (sitting on a bench) during halftime in two separate sessions. A 5-min football sim- ulation protocol consisting of football-specific activities (jumping, sprinting, kicking, passing, and dribbling at various intensities and distances) was repeated nine times to mimic the first and second half of a football match. Athletic performance (maximal vertical jump height, 20- m sprint time, and the Arrowhead agility test time) recorded during a 15-min period were aver- aged to represent each time point (first half: T1 to T3; second half: T4 to T6). Lower-leg skin and muscle (using the insulation disk technique) temperature was recorded before and after the first and second half. There was no condition effect over time in maximal vertical jump: F5,187 = 0.53, p = 0.75, Arrowhead agility test time: F5,187 = 1.25, p = 0.29, and lower-leg tem- perature (skin: F3,119 = 1.40, p = 0.25; muscle: F3,119 = 1.08, p = 0.36). The 20-m sprint time between conditions during the initial 15-min of the second half was different (condition × time: F5,187 = 2.42, p = 0.04) that subjects who performed the shuttle jog ran 0.09 sec faster (3.08 sec, p = 0.002, ES = 0.68), as compared with those who did the seated rest (3.17 sec). The results of our study confirmed that a decremental effect of the static rest on sprinting perfor- mance during the initial period of the second halftime can be attenuated by a halftime warm-up. Introduction A deterioration in athletic performance during the second half, as compared with the first half in football matches is common. For example, professional footballers have shown a decline in the total distance covered [1] and the distance covered by high-intensity running [2]. Within the second half, the first 15-min is the time that players’ physical performance is the lowest [3– 5]. Since players’ performance generally improves afterwards, this performance decrement during the initial phase of the second half is thought to be from the lack of recovery from the first half and/or preparation for high-intensity activities in the second half. Therefore, there is PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 1 / 15 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Bang S, Park J (2022) A 7-min halftime jog mitigated the reduction in sprint performance for the initial 15-min of the second half in a simulated football match. PLoS ONE 17(7): e0270898. https://doi.org/10.1371/journal. pone.0270898 Editor: Shigehiko Ogoh, Tokyo Joshi Ika Daigaku Toyo Igaku Kenkyujo Clinic, JAPAN Received: December 14, 2021 Accepted: June 19, 2022 Published: July 19, 2022 Copyright: © 2022 Bang, Park. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting information files. Funding: The author(s) received no specific funding for this work. Competing interests: The authors have declared that no competing interests exist. a need for halftime conditioning strategies, which help not only to enhance recovery from fatigue but also to be ready for the second half [6, 7]. Traditional static halftime rest (e.g., sitting on a bench) could change the body’s regulatory systems (e.g., autonomic nervous, thermoregulatory, and cardiovascular), resulting in a reduc- tion of the core body and working muscle temperature [8]. For example, a 10-min passive rest after a warm-up activity can result in a 0.4˚C decrease in body temperature, which could fur- ther lead to a 13% reduction in maximal vertical jump capacity [9]. Therefore, taking a passive rest during halftime is not considered as a recommended halftime strategy to prepare for the subsequent performance in the second half in terms of maintaining the “warmed-up body”. To preserve exercise-induced heat, halftime conditioning strategies such as cycling [10], plyo- metric and agility exercise [11], football-specific activity [12], and jogging [8, 13] have previ- ously been tested. Common endpoints to determine the effectiveness of such halftime activities include body temperature (core [12] and muscle [14]) and athletic performance (sprint time [10, 11] and maximal vertical jump height [13]). While the aforementioned studies reported an advantage in favour of warming-up during halftime, as compared with a passive rest, a couple of limitations still need to be addressed. First, the previous halftime warm-up strategies [8, 13] were examined in actual football matches. Although the results of these filed studies are advantageous for ecological validity, the sources of variation among subjects (e.g., movements or distance covered) are potential for data heterogeneity; thus, the results might be confounded. Although there have been studies [10, 12, 14, 15] using simulation protocols across interventions, the protocols were not con- sisted of football-related running and kicking activities [10, 15] or the effectiveness was exam- ined in relatively small samples (n = 7 [12]; n = 10 [15]; n = 10 [14]; n = 13 [10]). Recently, a 90-min long simulated football match (a 5-min of football simulation protocol: FSP ×9 to mimic the first and second halves) with a fixed distance and intensity has been introduced and validated [3]. The recorded average heart rate (163 bpm) and energy expenditure (1,227 kcal) during the simulated football match fell within the typical ranges of real-football matches (heart rate: 150 and 175 bpm [16]; calorie expenditure (1,200 to 1,500 kcal) [17]). Distance cov- ered by running without ball possession in this simulation protocol (maximal: 720-m; submax- imal 1,514-m) was also similar to an actual football match (maximal: 542 ± 214-m; submaximal 1,590 ± 488-m) [18]. Therefore, the purpose of this study was to quantify the effect of a halftime shuttle jog and seated rest on the second half performance in collegiate footballers. For the objective compari- sons (without possible confounders), the previously validated FSP [3] was employed using a within-subject design. The endpoints were athletic performance (maximal vertical jump height, 20-m sprint time, and Arrowhead agility test time performed during the football simu- lation), lower-leg temperature (skin and muscle), and heart rate during a 90-min simulated football matches. A previous study [8] observed a decrement in core and muscle (quadriceps) temperature with a static rest during halftime, which further impaired sprinting performance. Another study [13] reported that a halftime shuttle jog prevented sprinting and jumping per- formance deterioration in the initial period of the second half. According to the previous data [8, 13], we hypothesised that players who performed a shuttle jog would show better athletic performance and higher lower-leg temperature than those who experienced a seated rest. Materials and methods Study design and experimental approach A two-way (condition × time) crossover field study with repeated measures on time was used. Subjects randomly performed one of the two different conditioning strategies (shuttle jog or PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 2 / 15 seated rest) during halftime each session. The random order was generated by a spreadsheet software. Dependent measurements were athletic performance (maximal vertical jump height, 20-m sprint time, and Arrowhead agility test time), lower-leg temperature (skin and muscle), and heart rate. A 5-min FSP [3] (Fig 1) consisting of football-related activities (a full videoclip is available at https://www.youtube.com/watch?v=9FgWbWMWAa0) were repeated nine times to make up each half of a football match. The athletic performance and heart rate data recorded during the three consecutive FSPs were combined (a total time of 15-min); thus six-time points (first half: T1 through T3; second half: T4 through T6) were analysed and compared. The lower-leg temperatures were measured before and after the first and second half (four time points). Subjects visited the football pitch (regular-sized natural grass) same time of day on Two dif- ferent sessions, each 72-h apart. Subjects were asked to maintain their normal diet and wear the same socks and football stud throughout the experimental period and allowed to consume 1.5 L of water during each session (500 mL during each of the first and second halves and half- time). Data were not collected on rainy days. Air temperature and relative humidity were recorded with a digital thermometer (Kestrel Drop, Nielsen-Kellerman Co., Boothwyn, PA, USA) each session and were not different throughout the data collection period (air tempera- ture: t = 0.15, p = 0.87; relative humidity: t = –0 .52, p = 0.60). The average values of air temper- ature and relative humidity during the data collection period were 17.5 ± 4.6˚C and 40.2 ± 15.9%, respectively. Participants Eighteen male elite collegiate footballers (age: 22 ± 1 years; height: 179 ± 4 cm; mass: 70 ± 5 kg) volunteered for this study. All subjects had to be registered in the Korea Football Association (years of training: 10 ± 1 years) and football trained at least six years. Subjects were excluded if they had any musculoskeletal lower-extremity injury in the past six months, history of back or Fig 1. Football simulation protocol. This protocol was repeated three times during each time point. The protocol was to follow the numerical order as follows (1: 10-m run, 2: 10-m short pass ×2, 3: 10-m jog, 4: 40-m Arrowhead agility test, 5: 10-m run, 6: maximal vertical jump ×2, 7: 10-m side step, 8: 20-m walk, 9: 10-m dribble, 10: 10-m walk, 11: 30-m long kick ×2, 12: 10-m run, 13: 10-m back step, 14: 10-m side step, 15: 10-m jog, 16: 10-m short pass ×2, 17: 10-m jog, 18: 20-m sprint, 19: 10-m dribble, 20: walk, 21: 30-m long kick ×2, 22: 10-mjog, 23: 40-m Arrowhead agility test, 24: 10-m back step, 25: maximal vertical jump ×2, 26: 10-m side step, 27: jog, 28: 10-m walk, 29: 20-m sprint). https://doi.org/10.1371/journal.pone.0270898.g001 PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 3 / 15 lower-extremity surgery, or neuromuscular disorders. All subjects read the study procedures, approved by the university’s institutional review board (protocol #: KHSIRB-18-012), and gave written informed consent prior to taking part in the study. Testing procedures Upon arrival to the football pitch, testing procedures were instructed and written informed consent were obtained. Each session began with a self-directed 10- to 15-min warm-up (light jog and dynamic stretch). Subjects were equipped with the thermistor probes and heart rate monitor, the baseline lower-leg temperature and heart rate were recorded 1-min before the first half (Fig 2). After the thermistor probes were detached, subjects performed nine repeti- tions of the 5-min FSP (Fig 1) to represent the first and second halves. Once halftime begun, subjects walked to the bench (took 1-min), sat on the shaded bench (located in the sideline Fig 2. Testing procedures. Subjects were randomly experienced the condition of shuttle jog or seated rest each session. Heart rate was recorded throughout the experiment. The arrows indicate time points for lower-leg temperature measurements. FSP: football simulation protocol. https://doi.org/10.1371/journal.pone.0270898.g002 PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 4 / 15 area), and rested for 6-min while consuming water (500 mL). During this time, temperature probes were attached again to the lower-leg. After the temperature data were taken, subjects randomly performed either condition (shuttle jog or seated rest) for the next 7-min at each ses- sion. Subjects performed a 20-m shuttle jog at a constant intensity (average 70% of heart rate maximum recorded during the first half) on a football pitch [8]. To maintain the intensity of the shuttle jog, we continuously monitored their heart rate and gave continuous verbal feed- back (e.g., “slow down or a little faster”) to players. For the seated rest, subjects maintained the position (sitting on the bench) for another 7-min. During the last minute of halftime, lower- leg temperatures were taken again for both conditions. Subjects then completed another nine repetitions of the FSP for the second half (same activity as the first half). The temperature probes were attached again to the lower-leg immediately after the second half. Heart rate was recorded throughout the whole experiment. For the second session, subjects came back for another condition and went through the same experimental procedures. Endpoints For the maximal vertical jump, subjects were instructed to place their feet on the pitch within a triangle area (Fig 1) and to vertically jump as high as they could using both legs. Take-off move- ments were self-selected and performed pre-stretch using their lower-extremity joints and both arms’ swinging. Two cameras were set up as 240 fps with 1/1000 shutter speed: One camera (C1 in Fig 1: 1-m away from the center of the jumping area and located 1-m high from the grass) videotaped subjects’ feet to determine flight time while the other camera (C2: 1-m away from the centre of the jumping area and located 20 cm high from the grass) videotaped subjects’ whole-body. Video clips from the two cameras were exported into a motion analysis software (Kinovea 0.8.15, Kinovea Org., France) and synchronised [19]. Flight time (t: ms) as counting the number of frames from take-off to landing was first calculated. Subjects’ last foot to take-off and the first foot to touch the ground were considered as take-off and landing, respectively. Flight time was then inserted into the previously established formula [h = (1000 × t2) × 1.22625 / 10000] [20] to calculate the maximal vertical jump height (h: cm). This countermovement ver- tical jump was performed twice in the FSP; thus, a total of six jumps were averaged to represent each time point. For the 20-m sprint, subjects were asked to run straight as fast as they could with a standing start position (50 cm away from the start line). Two pairs of infrared timing sen- sors (Brower Timing System, Salt Lake City, USA) were set up at the start and finish lines, located 20-m apart. This activity was performed twice in the FSP; thus, a total of six 20-m sprint times were averaged to represent each time point. Subjects started 50 cm away from the start line for the Arrowhead agility test. Arrowhead agility test started towards cone-A (placed 5-m away from the start line), moved to cone-B and cone-C, and sprinted to the finish line [21] (Fig 1). Timing systems were used as the start line and finish line to record the time taken to sprint the agility course. This activity was performed twice in the FSP; thus, a total of six Arrowhead agility test times were averaged to represent each time point. To sample skin and muscle temperatures of the lower-leg, two thermistor probes, attached as two separate channels, to the digital thermometer logger (sampling rate: 60 Hz; NT logger, NKTC, Tokyo, Japan) were used. Each thermistor probe was attached to the gastrocnemius medialis. Two thermistor probes (connected to each channel) were attached to the middle of the gastrocnemius medialis muscle belly [22] in the dominant leg (the foot to kick a penalty shootout). Channel 1 and 2 of the probes were attached into the distal 1/3 (skin) and the proxi- mal 1/3 (muscle). The thermistor probe (channel 1) for skin temperature measurement was secured with film dressing (Tegaderm Film, 3M, St, Paul, USA). The thermistor probe (chan- nel 2) for the muscle temperature was covered by neoprene fabric and secured with the film PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 5 / 15 dressing (Fig 3). This device has been validated [23] and shown high measurement reliability with the intraclass correlation coefficient of 0.93 [24]. These data were collected four times (before and after the first and second half: Fig 2). To record heart rate, a strap monitor (Polar H7, Polar Electro Oy, Kempele, Finland) was worn on the subjects’ chest. Once subjects’ demographic information (sex, height, and mass) was entered, the strap was wirelessly con- nected to a cellphone application (WearLink, sampling rate: 60 Hz). This device has been vali- dated and shown high measurement reliability with the intraclass correlation coefficient of 0.80 to 0.86 [25]. Heart rate data in every 15-min interval (900 data points) were averaged to represent each time point. Statistical analysis An a priori power analysis was performed based on the previous data about athletic perfor- mance. An expected mean difference in maximal vertical jump of 4.3 cm with a standard devi- ation of 6.3 cm [26] and in 20-m sprint time of 0.13 sec with a standard deviation of 0.18 sec [27] estimated that a minimum of 17 and 14 subjects were necessary, respectively (an alpha of 0.05 and a beta of 0.2). Therefore, a sample size in this study was determined as 17. To test the condition effects over time in athletic performance, lower-leg temperature, and heart rate, we performed mixed model analysis of variance (random variable: subject; fixed variable: condition and time). The least significant difference was used for post hoc pairwise comparisons (SAS Institute Inc, Cary, USA, p<0.05 for all tests). If statistical differences appeared, we also calculated between-time effect sizes (ES) using the formula ½ES ¼ ½X1 Fig 3. Lower-leg temperature measurements. The thermistor probes were attached to the lower-leg (Channel 1: skin temperature, Channel 2: muscle temperature). https://doi.org/10.1371/journal.pone.0270898.g003 PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 7 / 15 Fig 4. Changes in maximal vertical jump height (A), 20-m sprint time (B), and Arrowhead agility test time (C) over time. Values are means and the upper and lower bounds of 95% confidence intervals.  A difference between conditions during T4 (p = 0.002, 3.08 vs. 3.17 sec, 3%, ES = 0.68). https://doi.org/10.1371/journal.pone.0270898.g004 PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 8 / 15 Fig 5A) and muscle (condition × time: F3,119 = 1.08, p = 0.36, η2 = 0.013; condition: F1,119 = 0.72, p = 0.40, η2 = 0.003; time: F3,119 = 0.20, p = 0.90, η2 = 0.002, Fig 5B) temperature of the lower-leg. Heart rate The average heart rate during the whole match was recorded as 160 ± 11 bpm. We did not observe condition effects throughout any time point on heart rate (condition × time: F5,187 = 0.31, p = 0.91, η2 = 0.003, Fig 5C). Regardless of time (condition: F1,187 = 7.09, p = 0.008, η2 = 0.013), subjects with the shuttle jog showed 2% less heart rate than those with the seated rest (159 vs. 161 bpm, ES = 0.22). Regardless of condition (time: F5,187 = 10.05, p<0.0001, η2 = 0.086), subjects’ heart rate was lower during the initial 15-min of the second half (T4: 154 bpm), as compared with other time points (T1: 159 bpm, p = 0.003, 3%, ES = 0.48; T2: 164 bpm, p<0.0001, 6%, ES = 0.91; T3: 163 bpm, p<0.0001, 6%, ES = 0.79; T5: 160 bpm, p = 0.0003, 4%, ES = 0.55; T6: 161 bpm, p<0.0001, 4%, ES = 0.62). We were interested in studying the effect of a 7-min halftime shuttle jog on subsequent lower-leg temperature and performance change in the second half as compared with a seated rest. Our hypotheses were partly accepted as subjects who performed the halftime shuttle jog maintained the sprint performance during the initial 15-min period of the second half when compared with those who did the seated rest. Our data confirmed that a decremental effect of the static rest on sprinting performance can be attenuated by a halftime warm-up. The limita- tions from previous data such as inter-individual variability [8, 13], non-football-specific activ- ity [10, 15], and small sample sizes [10, 12, 14, 15] have been overcome by the results of our study. In connection with the lower-leg muscle temperature, a 3% difference in sprint perfor- mance (3.08 vs. 3.17 sec) during T4 is reinforced by a previous suggestion that a 1˚C difference in muscle temperature may lead to a variation of performance capacity ranged between 2% and 5% [28, 29]. Our data on sprint performance change in both conditions during T4 has practical implica- tions. Specifically, the difference in the speed of each performance based on the records was calculated as 0.18 m/s (6.49 vs. 6.31 m/s), which would result in a difference in the distance covered in a given amount of time. For example, a player with the halftime warm-up would be approximately 0.4-m ahead of a player with the seated rest, when running 2 sec (13.0-m vs. 12.6-m). Assuming all other personal and environmental conditions are similar, this could be the main scene of the match when considering the importance of sprinting in football. Since > 90% of sprints were finished within 5 sec [30] and receiving a pass (travelled > 10-m) was the most common attacking option for goal scoring [31] in actual football matches, our estimation in sprint distance was practically meaningful. The difference of lower-leg muscle temperature between conditions during the initial 15-min showed a 1.0˚C, which was not statistically different. Previously a higher quadriceps muscle temperature (1.2˚C [8] and 1.0˚C [14]) after a halftime shuttle jog [8] or agility exercise [14] was reported, as compared with the seated rest. These temperature differences led to a dif- ference in athletic performance in which subjects with the halftime jog [8] or agility exercise [14] ran 0.20 sec [8] or 0.10 sec faster in 30-m or 10-m sprints [14], respectively. The linear relationship between rate of torque development and muscle temperature was also reported in football players [32]. Based on the previous data [8, 14, 33], a muscle temperature of 1.0˚C seems to be a sufficient amount for performance change. The measurement using the insula- tion disk technique [34] allowed to estimate the tissue temperature at an approximate depth of 2.2 cm from the thermistor probe (Fig 3). Additionally, the lower-leg has a smaller muscle mass and distally located than the upper leg, the whole leg temperature was probably 1.0˚C or PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 9 / 15 Fig 5. Changes in lower-leg skin (A) and muscle (B) temperature, and heart rate (C) over time. Values are means and the upper and lower bounds of 95% confidence intervals. https://doi.org/10.1371/journal.pone.0270898.g005 PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 10 / 15 higher in subjects who performed the halftime jog. Therefore, a 0.09 sec faster sprint time dur- ing T4 could have been, at least partially, involved with the prevention effect of halftime jog on the temperature reduction associated with the seated rest. While the lower-leg muscle temper- ature for the halftime jog during T4 was recorded as 31.7˚C, we still do not know if this is an optimal temperature for sprint performance. Either a lower or elevated muscle temperature could impair muscle function [35], future studies should attempt to determine the ideal muscle temperature in relation to the given environmental conditions. Due to the intermittent nature of football, various movements are performed at different intensities. The FSP [3] used in our study provided a similar situation of the intermittent char- acteristics of a real football match. For example, our subjects’ sprint performance did not decline during the last 30-min of the second half (T4 and T5) although fatigue and dehydration could have been occurred towards to the end of match [36]. A previous report on the effects of fatigue on football kick-related performance also showed that the ground reaction force and joint kinematics during instep kicks were not altered by fatigue during a 90-min intermittent exercises [37]. Additionally, a previous study reported that 70% of movements in a common adult football match was low-intensity (e.g., jogging or walking) [38]). The distance covered by high-intensity activities (e.g., 20-m sprint and Arrowhead agility test) in the 5-min FSP was 65-m, which is about 28% of the total distance covered (235-m) in our study. While the data under the control condition (e.g., seated rest) in our study mimics a typical football match, those under the halftime jog condition could, therefore, be applied to real football. The halftime jog was not effective in preventing performance hinderance in jump height and agility time associated with the seated rest. Contrarily, a couple of previous studies [13, 14] reported a halftime warm-up as an effective strategy to prevent the decrements in jump perfor- mance. The discrepancy between our results and those in previous studies [13, 14] could be explained by the several different testing procedures, including simulation protocol (non-foot- ball-specific [14]), measurement time point (single measurements at 46-min [13, 14] and 60-min [14]), jump height assessment techniques (force plate [14] and motion capture system [13]), and environmental conditions (indoor laboratory [14] and 5–6˚C with 84–87% humid- ity [13]). Additionally, the reported difference in jump heights between conditions (warm-up vs. seated rest during halftime) was ranged between 1 [13] and 2 cm [14], which was practically meaningless. In combination with the previous data [13, 14], our results suggest that the mag- nitude of deterioration in athletic performance due to muscle temperature reduction may be related to the frequency of the sport specific activities. Since running (17%) was a more fre- quent activity than jumping (4%) in football [39], decrements in sprint performance could have been larger than jump performance. This is indirectly supported by our data of athletic performance during T4 that the associated 95% confidence intervals of mean difference in ver- tical jump (Fig 6A) crosses zero while running activities (Fig 6B & 6C) does not. Heart rate was secondarily analysed to back up any change in dependent measurements. We observed a condition effect that subjects with the halftime jog (159 bpm) showed 2 bpm less heart rate than subjects with the seated rest (161 bpm, ES = 0.22) in the second half. These results in heart rate values were very similar to those in actual football matches (halftime jog: 157 bpm; seated rest: 161 bpm) [13], indicating the applicability of the FSP used in our study. Change in the cardiopulmonary system due to training or detraining took at least 6 weeks [40], supposing the amount of stroke volume and cardiac output in subjects in our study was not different between the sessions. Therefore, our observation in less heart rate in subjects with the halftime jog could be explained by circulatory efficiency [41]. Under the autonomic regulation, the cardiovascular and hemodynamic adjustments occur to adequately deliver oxy- gen supply to the working muscles (e.g., lower-extremity) when exercising [42]. We believe that subjects who performed the halftime jog (average heart rate: 128 bpm) led to more active PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 11 / 15 Fig 6. The 95% confidence intervals for difference in mean (maximal vertical jump height: A; 20-m sprint time: B; Arrowhead agility test time: C). https://doi.org/10.1371/journal.pone.0270898.g006 PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 12 / 15 hemodynamics (e.g., exchanging nutrients and wastes) in the lower-extremity when compared with players who did the seated rest. Since the same workloads (the 5-min FSP ×9) to both conditions were given, the lower-extremity with a worse circulatory efficiency (e.g., seated rest condition) required a continuous blood supply to accommodate for the energy requirements. As a result, subjects with the halftime jog showed less heart rate. Our study has several limitations. First, the data were collected under the environmental conditions (air temperature: 17.5˚C, relative humidity 40%, and pitch condition: natural grass) on a specific population (collegiate male footballers with an athletic career of 10 years), which should be considered when generalising the results. In connection with the weather condition, a difference in core temperature between conditions should have certainly had an impact on athletic performance [43]. It should also be noted that the jog intensity was guided by the research assistant, not self-determined by the player. In case of under- or over-pacing, we rec- ommend coaches and players practice getting used to the optimal intensity of the halftime jog. Conclusions Our data demonstrated potential benefits of 7-min halftime jog on sprint performance during the initial 15-min of the second half. Our results are meaningful that the limitations of the cur- rent existing data [8, 10, 12–15] have been overcome by obtaining data from a 90-min simu- lated football match performed by a larger number of footballers. We recommend performing a moderate-intensity aerobic exercise (70% of the maximum heart rate) during halftime to avoid the negative effects associated with the sedentary halftime period. Supporting information S1 Data. (XLSX) Author Contributions Conceptualization: Jihong Park. Data curation: Sooil Bang. Formal analysis: Jihong Park. Investigation: Jihong Park. Methodology: Sooil Bang, Jihong Park. Supervision: Jihong Park. Visualization: Sooil Bang, Jihong Park. Writing – original draft: Sooil Bang, Jihong Park. Writing – review & editing: Sooil Bang, Jihong Park. References 1. Dalen T, Jørgen I, Gertjan E, Havard HG, Ulrik W. Player load, acceleration, and deceleration during forty-five competitive matches of elite soccer. J Strength Cond Res. 2016; 30: 351–359. https://doi.org/ 10.1519/JSC.0000000000001063 PMID: 26057190 2. Russell M, Sparkes W, Northeast J, Cook CJ, Love TD, Bracken RM, et al. Changes in acceleration and deceleration capacity throughout professional soccer match-play. J Strength Cond Res. 2016; 30: 2839–2844. https://doi.org/10.1519/JSC.0000000000000805 PMID: 25474342 PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 13 / 15 3. Park J, Kim D, Park J. Lower-leg cooling during halftime does not affect second-half performance in a football match. Int J Appl Sports Sci. 2019; 31: 32–42. 4. Lovell R, Barrett S, Portas M, Weston M. Re-examination of the post half-time reduction in soccer work- rate. J Sci Med Sport. 2013; 16: 250–254. https://doi.org/10.1016/j.jsams.2012.06.004 PMID: 22824313 5. Weston M, Batterham AM, Castagna C, Portas MD, Barnes C, Harley J, et al. Reduction in physical match performance at the start of the second half in elite soccer. Int J Sports Physiol Perfom. 2011; 6: 174–182. https://doi.org/10.1123/ijspp.6.2.174 PMID: 21725103 6. Goncalves B, Coutinho D, Travassos B, Folgado H, Caixinha P, Sampaio J. Speed synchronization, physical workload and match-to-match performance variation of elite football players. PLoS One 2018; 13: e0200019. https://doi.org/10.1371/journal.pone.0200019 PMID: 30040849 7. De Ste Croix MBA, Priestley AM, Lloyd RS, Oliver JL. ACL injury risk in elite female youth soccer: Changes in neuromuscular control of the knee following soccer-specific fatigue. Scand J Med Sci Sports. 2015; 25: e531–e538. https://doi.org/10.1111/sms.12355 PMID: 25556396 8. Mohr M, Krustrup P, Nybo L, Nielsen J, Jung, Bangsbo J. Muscle temperature and sprint performance during soccer matches–beneficial effect of re-warm-up at half-time. Scand J Med Sci Sports. 2004; 14: 156–162. https://doi.org/10.1111/j.1600-0838.2004.00349.x PMID: 15144355 9. Galazoulas C, Tzimou A, Karamousalidis G, Mougios V. Gradual decline in performance and changes in biochemical parameters of basketball players while resting after warm-up. Eur J Appl Physiol. 2012; 112: 3327–3334 https://doi.org/10.1007/s00421-012-2320-1 PMID: 22262012 10. Yanaoka T, Kashiwabara K, Masuda Y, Yamagami J, Kurata K, Takagi S, et al. The effect of half-time re-warm up duration on intermittent sprint performance. J Sports Sci Med. 2018; 17: 269–278. PMID: 29769828 11. Abade E, Sampaio J, Gonc¸alves B, Baptista J, Alves A, Viana J. Effects of different re-warm up activi- ties in football players’ performance. PLoS One 2017; 2: e0180152. https://doi.org/10.1371/journal. pone.0180152 PMID: 28662123 12. Lovell R, J, Kirke I, Siegler J, McNaughton LR, Greig MP. Soccer half-time strategy influences thermo- regulation and endurance performance. J sports Med Phys Fitness. 2007; 47: 263–269. PMID: 17641591 13. Edholm P, Krustrup P, Randers M, Bredsgaard. Half-time re-warm up increases performance capacity in male elite soccer players. Scand J Med Sci Sports. 2015; 25: e40–e49. 14. Lovell R, Midgley A, Barrett S, Carter D, Small K. Effects of different half-time strategies on second half soccer-specific speed, power and dynamic strength. Scand J Med Sci Sports. 2013; 23: 105–113. https://doi.org/10.1111/j.1600-0838.2011.01353.x PMID: 21812822 15. Zois J, Bishop D, Fairweather I, Ball K, Aughey R, J. High-intensity re-warm-ups enhance soccer perfor- mance. Int J Sports Med. 2013; 34: 800–805. https://doi.org/10.1055/s-0032-1331197 PMID: 23444096 16. Bangsbo J. The physiology of soccer—with special reference to intense intermittent exercise. Acta Phy- siol Scand Suppl. 1994; 619: 1–155. PMID: 8059610 17. Stølen T, Chamari K, Castagna C, Wisløff U. Physiology of soccer. Sports Med. 2005; 35: 501–36. https://doi.org/10.2165/00007256-200535060-00004 PMID: 15974635 18. Bradley PS, Di Mascio M, Peart D, Olsen P, Sheldon B. High-intensity activity profiles of elite soccer players at different performance levels. J Strength Cond Res. 2010; 24: 2343–2351. https://doi.org/10. 1519/JSC.0b013e3181aeb1b3 PMID: 19918194 19. Balsalobre-Ferna´ndez C, Tejero-Gonza´lez CM, del Campo-Vecino J, Bavaresco N. The concurrent validity and reliability of a low-cost, high-speed camera-based method for measuring the flight time of vertical jumps. J Strength Cond Res. 2014; 28: 528–533. 20. Glatthorn JF, Gouge S, Nussbaumer S, Stauffacher S, Impellizzeri FM, Maffiuletti NA. Validity and reli- ability of Optojump photoelectric cells for estimating vertical jump height. J Strength Cond Res. 2011; 25: 556–560. https://doi.org/10.1519/JSC.0b013e3181ccb18d PMID: 20647944 21. Rago V, Brito J, Figueiredo P, Ermidis G, Barreira D, Rebelo A. The Arrowhead agility test: reliability, minimum detectable change, and practical applications in soccer players. J Strength Cond Res. 2020; 34: 483–494. https://doi.org/10.1519/JSC.0000000000002987 PMID: 30676390 22. Quesada JIP, Carpes FP, Bini RR, Palmer RS, Pe´rez-Soriano P, de Anda RMCO. Relationship between skin temperature and muscle activation during incremental cycle exercise. J Therm Bio. 2015; 48: 28–35. 23. Flouris AD, Webb P, Kenny GP. Noninvasive assessment of muscle temperature during rest, exercise, and postexercise recovery in different environments. J Appl Physiol. 2015; 118: 1310–1320. https://doi. org/10.1152/japplphysiol.00932.2014 PMID: 25814638 PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 14 / 15 24. Kikuchi K, Asano M, Tagami H, Kato M, Aiba S. Comparison of the measuring efficacy of transepider- mal water loss of a reasonably priced, portable closed-chamber system device H4500 with that of rather expensive, conventional devices such as Tewameter® and Vapometer®. Skin Res Tech. 2017; 23: 597–601. 25. Nunan D, Donovan G, Jakovljevic DG, Hodges LD, Sandercock GR, Brodie DA. Validity and reliability of short-term heart-rate variability from the Polar S810. Med Sci Sports Exerc. 2009; 41: 243–250. https://doi.org/10.1249/MSS.0b013e318184a4b1 PMID: 19092682 26. Lee M, Kim S, Choi H-M, Park J. Ankle or knee joint cooling alters countermovement but not squat jump height in healthy collegiate athletes. Isokinet Exerc Sci. 2017; 25: 1–8. 27. Kim H, Lee D, Choi H-M, Park J. Joint cooling does not hinder athletic performance during high-intensity intermittent exercise. Int J Sports Med. 2016; 37: 641–646. https://doi.org/10.1055/s-0035-1559687 PMID: 27119166 28. Racinais S, Oksa J. Temperature and neuromuscular function. Scand J Med Sci Sports. 2010; 3: 1–18. https://doi.org/10.1111/j.1600-0838.2010.01204.x PMID: 21029186 29. Bergh U, Ekblom B. Influence of muscle temperature on maximal muscle strength and power output in human skeletal muscles. Acta Physiol Scand. 1979; 107: 33–7. https://doi.org/10.1111/j.1748-1716. 1979.tb06439.x PMID: 525366 30. Andrzejewski M, Chmura J, Pluta B, Strzelczyk R, Kasprzak A. Analysis of sprinting activities of profes- sional soccer players. J Strength Cond Res. 2013; 27: 2134–2140. https://doi.org/10.1519/JSC. 0b013e318279423e PMID: 23168374 31. Michailidis Y, Michailidis C, Primpa E. Analysis of goals scored in European Championship 2012. 2013, rua.ua.es/dspace/handle/10045/279734. 32. Racinais S, Cocking S, Pe´riarda d, Julien D. Sports and environmental temperature: From warming-up to heating-up. Temperature. 2017; 4: 227–257. https://doi.org/10.1080/23328940.2017.1356427 PMID: 28944269 33. Sargeant AJ. Effect of muscle temperature on leg extension force and short-term power output in humans. Eur J Appl Physiol Occup Physiol. 1987; 56: 693–698. https://doi.org/10.1007/BF00424812 PMID: 3678224 34. Brajkovic D, Ducharme M, B, Webb P, Reardon F, D, Kenny G, P. Insulation disks on the skin to esti- mate muscle temperature. Eur J Appl Physiol. 2006; 97: 761–765. https://doi.org/10.1007/s00421-005- 0113-5 PMID: 16721613 35. Locke M, Celotti C. The effect of heat stress on skeletal muscle contractile properties. Cell Stress Chap- erones. 2014; 19: 519–527. https://doi.org/10.1007/s12192-013-0478-z PMID: 24264930 36. Mohr M, Krustrup P, Bangsbo J. Match performance of high-standard soccer players with special refer- ence to development of fatigue. J Sports Sci. 2003; 21: 519–528. https://doi.org/10.1080/ 0264041031000071182 PMID: 12848386 37. Kellis E, Katis A, Vrabas I. Effects of an intermittent exercise fatigue protocol on biomechanics of soccer kick performance. Scand J Med Sci Sports. 2006; 16: 334–344. https://doi.org/10.1111/j.1600-0838. 2005.00496.x PMID: 16978253 38. Bangsbo J, Mohr M, Krustrup P. Physical and metabolic demands of training and match-play in the elite football player. J Sports Sci. 2006; 24: 665–674. https://doi.org/10.1080/02640410500482529 PMID: 16766496 39. Tessitore A, Meeusen R, Tiberi M, Cortis C, Pagano R, Capranica L. Aerobic and anaerobic profiles, heart rate and match analysis in older soccer players. Ergonomics. 2005; 48: 1365–1377. https://doi. org/10.1080/00140130500101569 PMID: 16338706 40. Macpherson RE, Hazell TJ, Olver TD, Paterson DH, Lemon PW. Run sprint interval training improves aerobic performance but not maximal cardiac output. Med Sci Sports Exerc. 2011; 43: 115–122. https://doi.org/10.1249/MSS.0b013e3181e5eacd PMID: 20473222 41. Fernandes JF, Goubergrits L, Bru¨ning J, Hellmeier F, Nordmeyer S, da Silva TF, et al. Beyond pressure gradients: the effects of intervention on heart power in aortic coarctation. PLoS One. 2017; 12: e0168487. https://doi.org/10.1371/journal.pone.0168487 PMID: 28081162 42. Fadel PJ. Reflex control of the circulation during exercise. Scand J Med Sci Sports. 2015; 4: 74–82. https://doi.org/10.1111/sms.12600 PMID: 26589120 43. Faulkner SH, Richard AF, Gerrett N, Hupperets M, Hodder SG Havenith G. Reducing muscle tempera- ture drop after warm-up improves sprint cycling performance. Med Sci Sports Exerc. 2013; 45: 359–65. https://doi.org/10.1249/MSS.0b013e31826fba7f PMID: 22935735 PLOS ONE A halftime jog on athletic performance and lower-leg temperature in a simulated football match PLOS ONE | https://doi.org/10.1371/journal.pone.0270898 July 19, 2022 15 / 15

A 7-min halftime jog mitigated the reduction in sprint performance for the initial 15-min of the second half in a simulated football match.

07-19-2022

Bang, Sooil,Park, Jihong

eng

PMC10086059

Exogenous Lactate Augments Exercise-Induced Improvement in Memory but not in Hippocampal Neurogenesis Deunsol Hwang1,2,†, Jisu Kim1,2,†, Sunghwan Kyun1,2, Inkwon Jang1,2, Taeho Kim1,2, Hun-Young Park1,2, and Kiwon Lim1,2,3,* 1Laboratory of Exercise and Nutrition, Department of Sports Medicine and Science in Graduate School, Konkuk University, Seoul, the Republic of Korea 2Physical Activity and Performance Institute (PAPI), Konkuk University, Seoul, the Republic of Korea 3Department of Physical Education, Konkuk University, Seoul, the Republic of Korea *[email protected] †These authors have contributed equally to this work  Supplementary Table 1. Information of used antibodies for immunoblotting in the experiment. Primary antibody Secondary antibody Target Molecular Weight (kDa) Concentration Cat.No (manufacturer) Target Concentration Cat.No (manufacturer) FNDC5 24, 48 1:1000 ab174833 (Abcam) anti- rabbit 1:2000 sc-2357 (Santa Cruz Biotechnology) BDNF 15 1:1000 ab108319 (Abcam) 1:4000 PGC1α 92 - 105 1:1000 ab54481 (Abcam) 1:2000 MCT2 43 1:200 sc-166925 (Santa Cruz Biotechnology) anti- mouse 1:1000 sc-516102 (Santa Cruz Biotechnology) MCT1 54 1:500 ab93048 (Abcam) anti- rabbit 1:1000 sc-2357 (Santa Cruz Biotechnology) VEGFA 23, 45 1:1000 ab46154 (Abcam) 1:10000 HCAR1 40 1:2000 NLS2095 (NONUS Biologicals) 1:4000 beta- actin 43 1:2000 sc-47778 (Santa Cruz Biotechnology) anti- mouse 1:4000 sc-516102 (Santa Cruz Biotechnology) 130 93 70 53 41 30 22 18 14 VEH LAC EXE+VEH EXE+LAC a plantaris muscle sample a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC No. 1 membrane beta-actin beta-actin a plantaris muscle sample 14 18 22 30 41 53 70 93 130 a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC VEH LAC EXE+VEH EXE+LAC FNDC5 FNDC5 14 18 22 30 41 53 70 130 93 a plantaris muscle sample a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC VEH LAC EXE+VEH EXE+LAC BDNF BDNF Supplementary Figure 1. The original blots of hippocampal protein expression of FNDC5 and BDNF that presented in Fig. 5D. The result of (A) beta-actin, (B) FNDC5, and (C) BDNF is obtained from No. 1 membrane. The result of (D) beta-actin, (E) FNDC5, and (F) BDNF is obtained from No. 2 membrane. To check analysis conditions of plantaris muscle for another study, one plantaris muscle sample was running together. VEH, sedentary without lactate; LAC, sedentary with lactate; EXE+VEH, exercise without lactate; EXE+LAC, exercise with lactate; FNDC5, fibronectin type Ⅲ domain-containing protein 5; BDNF, brain derived neurotrophic factor. (A) (D) (B) (E) (C) (F) No. 1 membrane No. 1 membrane No. 2 membrane No. 2 membrane No. 2 membrane kDa kDa kDa Supplementary Figure 2. The original blots of hippocampal protein expression of PGC1α that presented in Fig. 5D. The result of (A) beta-actin and (B) PGC1α is obtained from No. 3 membrane. The result of (C) beta-actin and (D) PGC1α is obtained from No. 4 membrane. To check analysis conditions of plantaris muscle for another study, one plantaris muscle sample was running together. VEH, sedentary without lactate; LAC, sedentary with lactate; EXE+VEH, exercise without lactate; EXE+LAC, exercise with lactate; PGC1α, peroxisome proliferator- activated receptor gamma coactivator 1-alpha. 130 130 93 70 53 41 30 22 18 93 70 53 41 30 22 18 a plantaris muscle sample a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC VEH LAC EXE+VEH EXE+LAC beta-actin beta-actin a plantaris muscle sample a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC VEH LAC EXE+VEH EXE+LAC PGC1α PGC1α No. 3 membrane No. 4 membrane (A) (B) (C) (D) No. 3 membrane No. 4 membrane kDa kDa Supplementary Figure 3. The original blots of hippocampal protein expression of MCT2/1 that presented in Fig. 6C. The result of (A) beta-actin, (B) MCT2, and (C) MCT1 is obtained from No. 5 membrane. The result of (D) beta-actin, (E) MCT2, and (F) MCT1 is obtained from No. 6 membrane. To check analysis conditions of plantaris muscle for another study, one plantaris muscle sample was running together. VEH, sedentary without lactate; LAC, sedentary with lactate; EXE+VEH, exercise without lactate; EXE+LAC, exercise with lactate; MCT1/2, monocarboxylate transporter 1/2. 130 93 70 53 41 30 22 18 14 VEH LAC EXE+VEH EXE+LAC a plantaris muscle sample a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC No. 5 membrane No. 6 membrane beta-actin beta-actin a plantaris muscle sample a plantaris muscle sample 22 18 14 41 30 130 93 70 53 VEH LAC EXE+VEH EXE+LAC VEH LAC EXE+VEH EXE+LAC MCT2 MCT2 22 18 14 41 30 130 93 70 53 a plantaris muscle sample a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC MCT1 MCT1 VEH LAC EXE+VEH EXE+LAC (A) (D) (B) (E) (C) (F) No. 5 membrane No. 6 membrane No. 5 membrane No. 6 membrane kDa kDa kDa Supplementary Figure 4. The original blots of hippocampal protein expression of VEGFA that presented in Fig. 7C. The result of (A) beta-actin and (B) VEGFA is obtained from No. 7 membrane. The result of (C) beta-actin and (D) VEGFA is obtained from No. 8 membrane. To check analysis conditions of plantaris muscle for another study, one plantaris muscle sample was running together. VEH, sedentary without lactate; LAC, sedentary with lactate; EXE+VEH, exercise without lactate; EXE+LAC, exercise with lactate; VEFGA, vascular endothelial growth factor A. 18 30 22 41 70 93 130 53 a plantaris muscle sample No. 7 membrane No. 8 membrane a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC VEH LAC EXE+VEH EXE+LAC beta-actin beta-actin 18 22 30 41 70 93 130 53 a plantaris muscle sample a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC VEH LAC EXE+VEH EXE+LAC VEGFA VEGFA (A) (B) (C) (D) No. 7 membrane No. 8 membrane kDa kDa Supplementary Figure 5. The original blots of hippocampal protein expression of HCAR1 that presented in Fig. 7C. The result of (A) beta-actin and (B) HCAR1 is obtained from No.9 membrane. The result of (C) beta-actin and (D) HCAR1 is obtained from No. 10 membrane. To check analysis conditions of plantaris muscle for another study, one plantaris muscle sample was running together. VEH, sedentary without lactate; LAC, sedentary with lactate; EXE+VEH, exercise without lactate; EXE+LAC, exercise with lactate; HCAR1, hydroxycarboxylic acid receptor 1. 18 22 30 41 70 93 130 53 14 a plantaris muscle sample a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC VEH LAC EXE+VEH EXE+LAC beta-actin beta-actin No. 9 membrane 9 No. 10 membrane 18 22 30 41 70 93 130 53 14 a plantaris muscle sample a plantaris muscle sample VEH LAC EXE+VEH EXE+LAC VEH LAC EXE+VEH EXE+LAC HCAR1 HCAR1 (A) (B) (C) (D) No. 9 membrane 9 No. 10 membrane kDa kDa

Exogenous lactate augments exercise-induced improvement in memory but not in hippocampal neurogenesis.

04-10-2023

Hwang, Deunsol,Kim, Jisu,Kyun, Sunghwan,Jang, Inkwon,Kim, Taeho,Park, Hun-Young,Lim, Kiwon

eng

PMC10703220

RESEARCH ARTICLE Exploring running styles in the field through cadence and duty factor modulation Anouk NijsID*, Melvyn Roerdink, Peter Jan BeekID Department of Human Movement Sciences, Amsterdam Movement Sciences, Vrije Universiteit Amsterdam, Amsterdam, Netherlands * [email protected] Abstract According to the dual-axis model, running styles can be defined by cadence and duty factor, variables that have been associated with running performance, economy and injury risk. To guide runners in exploring different running styles, effective instructions to modulate cadence and duty factor are needed. Such instructions have been established for treadmill running, but not for overground running, during which speed can be varied. In this study, five participants completed eight field training sessions over a 4-week training period with acous- tic instructions to modulate cadence, duty factor, and, in combination, running style. Instruc- tions were provided via audio files. Running data were collected with sports watches. Participants’ experiences with guided-exploration training were evaluated with the user experience questionnaire. Data analysis revealed acoustic pacing and verbal instructions to be effective in respectively modulating cadence and duty factor, albeit with co-varying effects on speed and the non-targeted variable (i.e. duty factor or cadence). Combining acoustic pacing and verbal instructions mitigated these co-varying effects considerably, allowing for running-style modulations in intended directions (particularly towards the styles with increased cadence and increased duty factor). User experience of this form of guided- exploration training was overall positive, but could be improved in terms of autonomy (dependability). In conclusion, combining acoustic pacing and verbal instructions for run- ning-style modulation is effective in overground running. Introduction Running is a popular sport practiced by many people worldwide [1]. Plausible reasons for this are that running is an easily accessible type of physical activity in terms of preparation, location and cost [1], which is associated with significant health benefits [2]. Many runners use a sports watch or mobile application to monitor their running, but they seldom adapt their running style [3]. Runners can benefit from running-style modifications, especially if they are prone to injury or recently began running [4–6]. This suggests that incorporating effective instructions in existing mobile applications could help runners improve their running style in terms of injury prevention, running performance, or running economy, depending on the running var- iables being targeted. PLOS ONE PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 1 / 13 a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Nijs A, Roerdink M, Beek PJ (2023) Exploring running styles in the field through cadence and duty factor modulation. PLoS ONE 18(12): e0295423. https://doi.org/10.1371/journal. pone.0295423 Editor: Laurent Mourot, University of Bourgogne France Comte´, FRANCE Received: June 21, 2023 Accepted: November 20, 2023 Published: December 7, 2023 Peer Review History: PLOS recognizes the benefits of transparency in the peer review process; therefore, we enable the publication of all of the content of peer review and author responses alongside final, published articles. The editorial history of this article is available here: https://doi.org/10.1371/journal.pone.0295423 Copyright: © 2023 Nijs et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the manuscript and its Supporting Information files. Funding: This work was supported by the Dutch Research Council (NWO; https://www.nwo.nl/) Cadence has been related to injury risk [7, 8] and running economy [6, 9, 10]. A common instruction method for modifying cadence is acoustic pacing. A metronome beat or music specifying the desired cadence is played, and the runner synchronizes their steps to the corre- sponding rhythm [11–13]. Acoustic pacing for cadence modulation was proven to be effective for running on a treadmill in the laboratory [11, 12]. An important difference between tread- mill running and overground running is that changes in speed are restricted in treadmill run- ning but not in overground running. In overground running, cadence is related to speed, with a higher speed corresponding to a higher cadence [14], although this is not a consistent finding [22]. Because of this, acoustic pacing to increase cadence could lead to a concomitant increase in speed during overground running. It is therefore important to examine the effectiveness of acoustic pacing in changing the cadence in overground running, while simultaneously consid- ering speed changes. Some studies have used acoustic pacing in overground running on a track [13] or outdoors [15, 16] and found it to be effective. Participants in these studies were instructed to keep running speed constant, but possible changes in speed as a result of the cadence manipulation were often not analyzed specifically. Counterintuitively, te Brake and colleagues [15] reported an increase in cadence combined with a reduction in speed after a four-week music-based intervention aimed at increasing cadence. Besides cadence, the duty factor (i.e. the ratio of stance time relative to step time) has been associated with injury risk [17], running economy, and performance [18]. Verbal instructions to change stance time and flight time were found to be effective in changing the duty factor when running at a constant speed on a treadmill [19–21]. To the best of our knowledge, instructions to change the duty factor have not been investigated in overground running to date. The duty factor is also associated with speed in that a higher speed corresponds to a lower duty factor (i.e. a shorter stance time relative to the step time [14]). Hence, similar to cadence, instructions to change the duty factor could elicit concomitant variations in speed. Furthermore, as cadence and duty factor are both associated with step time, instructions to change either variable could affect the other variable as well. A change in cadence without a change in duty factor requires a change in both stance time and flight time without changing the ratio between them. When studying the effects of modulating a specific running variable, it is therefore important not only to quantify the effects for that specific variable but also to quan- tify potential co-varying effects on other running variables, especially in studies on running style modulation. Such co-varying effects have not been reported in the literature for modu- lated overground running, while for constant-speed treadmill running instructions aimed at changing the duty factor did not affect cadence [19]. According to the dual-axis model [22], running styles can be categorized at a certain speed through the combination of cadence and duty factor (Fig 1). The basic idea behind the model is that these two (‘distal’) variables respectively reflect the (‘proximal’) horizontal and vertical displacement of the center of mass, resulting from the interplay of many kinematic and kinetic factors [22]. The dual-axis model distinguishes five different running styles. The ‘Sit’, located at the center of the model, represents the average runner, with an average cadence and duty factor. The ‘Hop’, located on the left of the model, represents a high cadence, and thus a short step, leading to a small horizontal displacement per step. The ‘Push’ on the contrary, located on the right side, reflects a low cadence, and thus a long step and large horizontal displacement per step. The ‘Bounce’, located at the top of the model, reflects a low duty factor, and thus a rel- atively long flight phase, corresponding to a larger vertical displacement. Finally, the ‘Stick’, on the bottom, reflects a high duty factor, and thus a relatively long stance time, corresponding to a lower vertical displacement. Thus, according to the dual-axis model, modulating cadence or duty factor in a specific direction allows for modulating one’s running style. In this study, we combined acoustic pacing to modulate cadence and verbal instruction to modulate duty factor PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 2 / 13 under Grant P16–28 (Project 3). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. There was no additional internal or external funding received for this study. Competing interests: The authors have declared that no competing interests exist. in order to modulate running style (e.g. by combining pacing at the preferred cadence with a verbal instruction to decrease stance time, thereby guiding the runner in the direction of the ‘Bounce’ style). This study had three aims. 1) To examine the effects of the acoustic pacing and verbal instructions regarding stance time on cadence, duty factor, and speed. In this regard we focused on both the targeted effects and potential co-varying effects of the instructions on speed and on the non-targeted variable (i.e. duty factor with acoustic pacing and cadence with verbal instructions). We hypothesized that in overground running both acoustic pacing and verbal instructions result in the targeted effects, but to a lesser extent than on a treadmill, due to co-varying effects on running speed. 2) To examine the effectiveness of combined acoustic- pacing and verbal-instruction conditions in guiding participants towards a certain running style, as defined by the dual-axis model (Fig 1). Also in this context, possible effects on speed were considered. We expected the combined instructions to be more effective than the individ- ual instructions, due to a smaller chance of co-varying effects of acoustic pacing on the duty factor and verbal instructions regarding stance time on the cadence. 3) We aimed to assess the user experience of a 4-week guided-exploration training program with these acoustic instruc- tions in the field. Materials and methods For this study a convenience sample of five healthy adult recreational runners (Table 1) was recruited. All participants had multiple years of running experience (Table 1), and ran multiple times per week (Table 1). They routinely used a sports watch with an accessory (e.g., heart rate Fig 1. Visual representation of the dual-axis model. https://doi.org/10.1371/journal.pone.0295423.g001 PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 3 / 13 belt or footpod) measuring cadence and stance time, as required for the present purpose to assess the effects of pacing and instructions. The recruited number of participants was rela- tively small due to this requirement. Participants were recruited via social networks and athlet- ics clubs. Recruitment was stopped when after 6 months of active recruitment (between December 2021 and May 2022), only five applicants met the requirements. In view of this small sample of suitable volunteers, we opted for nonparametric statistical testing within each individual participant to best answer the research questions. All participants provided written informed consent before the start of the study. The study protocol was in compliance with the Declaration of Helsinki and approved by the Scientific and Ethical Review Board (VCWE) of the Faculty of Behavioural and Movement Sciences of the Vrije Universiteit Amsterdam. All participants were given a participant number. Only author AN had access to the key for this pseudonymization. To personalize acoustic pacing and verbal instructions, baseline values for speed, cadence and duty factor were required. Therefore, participants shared their running data collected on their personal sports watch for at least the month before the training period. This data was shared through the online platform of Move-Metrics (Ede, the Netherlands), and parameter- ized per training in terms of date, duration, and distance, and mean, standard deviation, median, inter quartile range (IQR), minimum, maximum, and lower (5%) and upper (95%) limit of the confidence interval for speed, cadence, and stance time while running. Based on these data, the baseline speed was determined as the rounded median speed over the training sessions. The median instead of the mean was used to reduce the effect of possible outliers when for example a training was labelled incorrectly. The baseline cadence and stance time were determined as the median cadence and stance time over the training sessions of which the rounded median speed was at the baseline speed. Duty factor was calculated based on the stance time and cadence according to: duty factor ¼ stance time ð60=cadenceÞ∗2 ; ðEq1Þ where stance time is expressed in seconds and (60/cadence) represents the step time in sec- onds, rendering the duty factor a dimensionless variable (Table 1). An audio file (see S1 Audio for an example) was created with verbal instructions to keep speed constant at the baseline speed and explore cadence and stance time relative to the base- line values guided by acoustic pacing and verbal instructions regarding stance time. Partici- pants trained according to the instructions twice a week for a period of four weeks by playing the audio file on a device of their choice while running. They were instructed to start the audio file and the measurement on their sports watch simultaneously. Short walking blocks were Table 1. Participants’ age, years of running experience, training frequency, determined baseline speed, number of runs at baseline speed, and baseline cadence, stance time, and duty factor. Age (years) Sex (male /female) Experience (years) Training frequency (training /week) Number of valid runs Baseline speed (km/h) Number of runs at baseline speed Baseline cadence (steps /min) Baseline stance time (ms) Baseline duty factor 1 19 f 3 2 11 11 5 182 260 0.39 2 56 m 15 3 154 11 75 170 260 0.37 3 60 m 40 4 209 10 123 166 260 0.36 4 50 m 20 2.5 144 11 65 182 250 0.38 5 49 m 9 4 30 9 13 158 325 0.43 https://doi.org/10.1371/journal.pone.0295423.t001 PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 4 / 13 included at specific times in the audio file, allowing synchronization of the training data to the audio file instructions. For the resultant eight training sessions, we received speed, cadence and stance time for each second of training. The data during the eight training sessions were normalized relative to the set target values. We then calculated the median speed, cadence and duty factor for each instruction. Cadence and duty-factor modulation To assess the effects of the individual instructions, we calculated the slope of the change in cadence, duty factor and speed as a result of the increase in pacing frequency for each training. For each participant, we then calculated the 95% confidence interval and used a one-sample Wilcoxon signed rank test over the eight training sessions to examine if the slopes were signifi- cantly different from zero (p<0.05), which would indicate an effect of the acoustic pacing. Effect size r was calculated as r = |Z/pN|, where Z is the standardized test statistic and N is the number of training sessions. Cadence, duty factor and speed were also compared between the instruction to increase stance time and the instruction to decrease stance time using a Wil- coxon signed rank test to assess the effect of the verbal instructions to change stance time. Running-style modulation Since the dual-axis model has been introduced only recently, no population reference values for the axes are yet available. We therefore decided to modulate running style relative to partic- ipant’s baseline running style and regarded that as the ‘Sit’ style in the center of the model. By giving combined acoustic pacing and verbal stance-time instructions we then aimed to guide participants away from their baseline running style towards one of the four other running styles (Table 2). The same audio file was used in all eight training sessions, and the order in which the instructions were given was the same for all participants. We defined cadence, duty factor and speed for each participant’s ‘Sit’ baseline running style, and compared cadence, duty factor and speed observed for each of the other four modu- lated running styles to these baseline values using Wilcoxon signed rank tests. The 95% confi- dence intervals for the mean differences were also calculated. Subjective user experience After the four-week training period, participants filled out a questionnaire on their experience with the guided-exploration training. For this purpose, the Dutch version of the User Experi- ence Questionnaire (UEQ) was used ([23]; translated by Adriaan Dekker according to [24]; obtained from www.ueq-online.org). The UEQ consist of 26 pairs of opposing terms on the Table 2. Overview of the instructions given to explore each running style. Instruction Speed Cadence Stance time Run at the constant baseline speed 1.00 * baseline cadence Increase stance time (‘Stick’) 1.10 * baseline cadence (‘Hop’) Stop increasing stance time 1.00 * baseline cadence (Baseline running style) 0.90 * baseline cadence (‘Push’) 1.00 * baseline cadence Decrease stance time (‘Bounce’) https://doi.org/10.1371/journal.pone.0295423.t002 PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 5 / 13 extremes of a 7-point Likert scale. Answers range from -3 (completely agree with the term on the left side) to +3 (completely agree with the term on the right side). Each of the 26 items belongs to one of six scales: (1) Attractiveness, what is the user’s overall impression of the product?; (2) Perspicuity, is it easy to understand?; (3) Efficiency, is the interaction deemed efficient?; (4) Dependability, does the user feel in control and safe?; (5) Stimulation, is it excit- ing to use the product?; and (6) Novelty, does it capture the users’ attention? For each individ- ual, the UEQ scores were grouped according to the six scales of the UEQ. These scores were then compared to the established benchmark values for the UEQ [23]. Results Cadence and duty factor modulation The slope for cadence over the increasing acoustic pacing frequencies was positive and signifi- cantly different from zero for all participants, indicating that they were able to change cadence when modulated by acoustic pacing (Fig 2A; Table 3). For the duty factor, acoustic pacing did not have a consistent systematic effect. The slope was only significant for participants 2 (posi- tive) and 3 (negative), indicating an increase and decrease in duty factor with increasing acous- tic pacing frequency, respectively (Fig 2C; Table 3). The slope for speed was positive and significant for all participants, indicating an increase in speed with increasing pacing fre- quency (Fig 2E; Table 3). Fig 2. Mean cadence, duty factor, and speed over the eight training sessions, relative to baseline values. The left panels (a, c, and e) show the values and regression lines for cadence modulation with acoustic pacing. The right panels (b, d, and f) show the values for the two verbal instructions to change the duty factor. The vertical lines represent the standard deviation between the eight training sessions. https://doi.org/10.1371/journal.pone.0295423.g002 PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 6 / 13 Duty factor was significantly different between the two verbal stance-time instructions for all participants, varying in the instructed directions (Fig 2D; Table 3). Cadence was signifi- cantly different between the instructions for participants 1, 2, and 5 and speed was significantly different for participants 1, 3, and 4, with a higher cadence and speed after the instruction to decrease stance time (aimed at a lower duty factor; Fig 2B and 2F; Table 3). Running-style modulation Manipulations towards the ‘Stick’ running style (increasing duty factor with verbal instruc- tions to increase stance time) led to a significant increase in the targeted duty factor compared to baseline for all participants (Fig 3B; Table 4). Co-varying effects were minimal: cadence was significantly higher for participants 1 and 4 while speed was significantly higher for participant 2 only (Fig 3A and 3C; Table 4). Manipulations towards the ‘Hop’ running style (increasing cadence with faster acoustic pacing) led to a significant increase in the targeted cadence compared to baseline for partici- pants 1, 2, 4, and 5 (Fig 3A; Table 4). Again, co-varying effects were limited: duty factor did not differ significantly from baseline for any of the participants, while speed only increased sig- nificantly for participants 4 and 5 (Fig 3B and 3C; Table 4). Manipulations towards the ‘Push’ running style (decreasing cadence with slower acoustic pacing) led to a significant decrease in the targeted cadence for all participants (Fig 3A; Table 4). Co-varying effects were more pronounced: duty factor decreased significantly for all participants and speed decreased significantly for participants 3 and 4 (Fig 3B and 3C; Table 4). Manipulations towards the ‘Bounce’ running style (decreasing duty factor with verbal instructions to decrease stance time) did not lead to a significant reduction in the duty factor compared to baseline in any of the participants (Fig 3B; Table 4). In the absence of an effect on the targeted variable, also co-varying effects were largely absent: only the speed of participant 4 was significantly higher compared to baseline (Fig 3A and 3C; Table 4). Table 3. Mean slope of the targeted variable cadence and the two potentially co-varying variables duty factor and speed as a function of acoustically paced cadence modulations (center rows) and mean difference between the verbal instructions to increase and decrease stance time (increase–decrease) in the targeted variable duty factor (positive values indicate a change in the modulated direction) and potentially co-varying variables cadence and speed (lower rows). Slope and difference data are complemented by 95% confidence intervals and the results of the Wilcoxon Signed Rank test. Expected changes for targeted (T) and potentially co-varying (C) var- iables are shown at the top of each block; results in line with these expectations are highlighted in green. modulated variable Cadence Duty factor Speed Cadence expectation Positive slope (T) no slope (C) no slope (C) Participant slope 95% confidence Sig. r slope 95% confidence Sig. r slope 95% confidence Sig. r 1 0.799 0.696 0.903 0.012* 0.893 0.025 -0.077 0.127 0.401 0.297 0.686 0.402 0.970 0.012* 0.891 2 0.940 0.895 0.984 0.012* 0.893 0.149 0.018 0.280 0.050* 0.693 0.471 0.322 0.620 0.012* 0.891 3 0.259 0.028 0.490 0.012* 0.893 -0.082 -0.136 -0.027 0.025* 0.792 0.615 0.345 0.884 0.012* 0.891 4 0.562 0.422 0.703 0.012* 0.893 0.058 -0.006 0.122 0.093 0.594 0.685 0.412 0.957 0.012* 0.891 5 0.453 0.348 0.557 0.012* 0.893 -0.098 -0.222 0.026 0.327 0.346 0.563 0.209 0.916 0.012* 0.891 Duty factor expectation no difference (C) increase (T) no difference (C) Participant difference 95% confidence Sig. r difference 95% confidence Sig. r difference 95% confidence Sig. r 1 -0.062 -0.085 -0.039 0.012* 0.893 0.048 0.029 0.067 0.012* 0.891 -0.055 -0.099 -0.012 0.012* 0.891 2 -0.026 -0.041 -0.010 0.012* 0.893 0.053 0.032 0.073 0.012* 0.891 -0.004 -0.036 0.027 0.575 0.198 3 -0.007 -0.015 -0.000 0.066 0.651 0.051 0.031 0.072 0.012* 0.891 -0.025 -0.045 -0.006 0.025* 0.792 4 -0.003 -0.010 0.004 0.340 0.337 0.050 0.030 0.070 0.012* 0.891 -0.022 -0.034 -0.011 0.012* 0.891 5 -0.089 -0.125 -0.052 0.011* 0.897 0.044 0.027 0.062 0.012* 0.891 -0.040 -0.094 0.015 0.161 0.495 https://doi.org/10.1371/journal.pone.0295423.t003 PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 7 / 13 Subjective user experience The scores for the six scales of the UEQ are presented in Fig 4 for all participants. The estab- lished benchmark scores for the different scales are in the background of the figure [23]. As can be seen in the figure, the scores were on average excellent for Perspicuity, good for Nov- elty, on the border between above average and good for Stimulation, above average for Attrac- tiveness, Efficiency, and bad for Dependability. The participants rated the guided-exploration training differently, as shown by the relatively large range in the ratings between participants on most scales. Discussion In this study, we examined the effects of acoustic pacing for cadence modulation and verbal instruction for duty-factor modulation on cadence, duty factor and speed in overground run- ning in the open field, as well as the effect of running-style modulation towards ‘Stick’, ‘Hop’, ‘Push’, and ‘Bounce’ running styles, relative to the participant’s baseline running style, inter- preted for the sake of the study as the ‘Sit’ running style. In addition, we examined the subjective evaluation of these instructions aimed at exploring different running styles during training. We expected acoustic pacing to be effective at modulating the cadence and verbal stance- time instructions to be effective at changing the duty factor, although to a lesser extent than on Fig 3. Difference in the relative cadence (a), duty factor (b), and speed (c) over the eight training sessions under the combined manipulation of acoustic pacing and verbal stance-time instructions to modulate participant’s baseline running style towards ‘Stick’, ‘Hop’, ‘Push’ and ‘Bounce’ running styles. The symbols between brackets indicate the targeted change, where the arrows up and down respectively represent a targeted increase and decrease, and the equal sign represents the targeted absence of a change. The vertical lines represent the standard deviation over the eight training sessions. The horizontal dotted line represents participant’s baseline running style. https://doi.org/10.1371/journal.pone.0295423.g003 PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 8 / 13 Table 4. Mean change in cadence, duty factor and speed as a result of the combined manipulations of acoustic pacing and verbal stance-time instructions to modu- late participant’s baseline running style towards ‘Stick’, ‘Hop’, ‘Push’ and ‘Bounce’ running styles, as well as 95% confidence intervals and the results of the Wil- coxon Signed Rank test. Expected changes for targeted (T) and potentially co-varying (C) variables are shown at the top of each running style block; results in line with the manipulations are highlighted in green. Running style Cadence Duty factor Speed ‘Stick’ expectation no change (C) Increase (T) no change (C) Participant mean 95% confidence Sig. r mean 95% confidence Sig. r mean 95% confidence Sig. r 1 0.015 0.003 0.027 0.026* 0.789 0.038 0.023 0.052 0.012* 0.891 0.020 -0.025 0.065 0.401 0.297 2 0.000 0.000 0.000 1.000 0.000 0.041 0.025 0.057 0.012* 0.891 0.043 0.000 0.087 0.036* 0.742 3 -0.001 -0.005 0.002 0.317 0.354 0.040 0.025 0.056 0.012* 0.891 -0.013 -0.048 0.022 0.528 0.223 4 0.003 0.001 0.006 0.042* 0.718 0.039 0.024 0.054 0.012* 0.891 -0.003 -0.021 0.016 0.779 0.099 5 -0.011 -0.033 0.010 0.276 0.385 0.035 0.021 0.048 0.012* 0.891 -0.034 -0.129 0.060 0.401 0.297 ‘Hop’ expectation Increase (T) no change (C) no change (C) Participant mean 95% confidence Sig. r mean 95% confidence Sig. r mean 95% confidence Sig. r 1 0.041 0.014 0.069 0.011* 0.897 -0.011 -0.029 0.008 0.093 0.594 0.066 -0.009 0.140 0.069 0.643 2 0.072 0.035 0.109 0.020* 0.825 -0.012 -0.032 0.008 0.093 0.594 0.016 -0.008 0.039 0.208 0.445 3 0.004 -0.005 0.013 0.257 0.401 -0.011 -0.031 0.008 0.093 0.594 0.012 -0.004 0.027 0.123 0.545 4 0.019 -0.004 0.042 0.027* 0.780 -0.011 -0.030 0.008 0.093 0.594 0.046 0.003 0.081 0.036* 0.742 5 0.043 0.021 0.065 0.011* 0.897 -0.010 -0.027 0.007 0.093 0.594 0.058 -0.002 0.118 0.050* 0.693 ‘Push’ expectation Decrease (T) no change (C) no change (C) Participant mean 95% confidence Sig. r mean 95% confidence Sig. r mean 95% confidence Sig. r 1 -0.071 -0.098 -0.045 0.011* 0.896 -0.022 -0.032 -0.012 0.012* 0.891 -0.026 -0.076 0.024 0.263 0.396 2 -0.092 -0.098 -0.085 0.011* 0.897 -0.024 -0.035 -0.013 0.012* 0.891 -0.039 -0.078 0.001 0.080 0.619 3 -0.015 -0.031 0.002 0.039* 0.728 -0.023 -0.034 -0.013 0.012* 0.891 -0.062 -0.110 -0.013 0.017* 0.843 4 -0.023 -0.036 -0.011 0.012* 0.892 -0.023 -0.033 -0.012 0.012* 0.891 -0.046 -0.080 -0.013 0.025* 0.792 5 -0.030 -0.044 -0.016 0.011* 0.898 -0.020 -0.029 -0.011 0.012* 0.891 -0.028 -0.096 0.041 0.441 0.273 ‘Bounce’ expectation no change (C) Decrease (T) no change (C) Participant mean 95% confidence Sig. r mean 95% confidence Sig. r mean 95% confidence Sig. r 1 0.005 -0.004 0.015 0.180 0.474 -0.017 -0.035 0.000 0.069 0.643 0.009 -0.043 0.061 0.889 0.049 2 0.000 0.000 0.000 1.000 0.000 -0.019 -0.038 0.000 0.069 0.643 0.017 -0.009 0.042 0.093 0.594 3 -0.001 -0.005 0.002 0.317 0.354 -0.019 -0.037 0.000 0.069 0.643 -0.023 -0.050 0.004 0.092 0.595 4 0.003 -0.003 0.008 0.271 0.389 -0.018 -0.036 0.000 0.069 0.643 0.038 0.027 0.048 0.012* 0.891 5 0.014 -0.013 0.042 0.236 0.419 -0.016 -0.032 0.000 0.069 0.643 0.033 -0.051 0.118 0.528 0.223 https://doi.org/10.1371/journal.pone.0295423.t004 Fig 4. The scores on the different user experience scales for all participants. The horizontal black line is the mean over the participants. Shaded areas represent the benchmark values of <25% (bad), 25%-50%, 50%-75%, 75%-90%, and >90% (excellent) performance. https://doi.org/10.1371/journal.pone.0295423.g004 PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 9 / 13 a treadmill [19] in view of the possibility to adapt the speed while running overground. Our results confirmed the expectation for acoustic pacing, as cadence did indeed increase with increasing pacing frequencies. Speed also increased with increasing pacing frequencies. Com- pared to a treadmill study, in which cadence relative to baseline cadence was 0.91, 1.00, and 1.11 in response to acoustic pacing at 90%, 100%, and 110% of baseline cadence [11], corre- sponding to a slope around 1.00, the change was indeed smaller for most participants in this study. Only participant 2 came close to this, with a mean slope of 0.94 (Table 3). For the duty factor, instructions to increase and decrease stance time led to changes in the targeted duty fac- tor, but also to changes in speed and cadence. Compared to another treadmill study [19], in which duty-factor changes of about 10% relative to baseline were found (i.e. a difference between instructions of around 20%), the instructions were considerably less effective, with differences between duty-factor instructions of around 5%. These relatively small changes in duty factor could be due to the simultaneous changes in cadence and speed (Table 3). Overall, we were able to modulate the targeted variables with acoustic pacing or stance-time instruc- tions in the right directions, but with lower magnitude of effect compared to fixed-speed tread- mill conditions given the observed co-varying effects on the non-targeted variables. If these instructions are used in practice, it is important to keep the covarying effects in mind. We expected the combined use of acoustic pacing and stance-time instructions to have a lower chance of co-varying effects of acoustic pacing on the duty factor and verbal stance-time instructions on the cadence, and thus to be more effective. In terms of the magnitude of the changes in the targeted variables of the running-style modulation, the combined duty-factor modulation was slightly more effective (‘Stick’: |0.0386 | + ‘Bounce’: |-0.0178| = 0.056) com- pared to verbal stance-time instructions alone (0.049). The combined cadence modulation was slightly less effective (‘Hop’: 0.036 and ‘Push’: -0.046) compared to acoustic pacing alone (110%: 0.054 and 90%: -0.059). More importantly, in line with our expectation, co-varying effects were smaller, with fewer changes in cadence as a result of the instructions towards the ‘Stick’ and ‘Bounce’ (which required a change in duty factor). Likewise, co-varying effects on duty factor were smaller for the ‘Hop’ (which required an increase in cadence), but not for the ‘Push’ (which required a decrease in cadence): duty factor decreased significantly for ‘Push’ running-style modulation for all participants. We further expected a co-varying effect on speed as a result of the acoustic-pacing variations to modulate running styles towards ‘Hop’ and ‘Push’ in a similar manner as for the acoustic-pacing conditions alone. However, this was not the case; in fact, there were fewer significant changes in speed for all running styles com- pared to the acoustic-pacing conditions alone (Tables 3 vs 4). Overall, the combined instruc- tions for running-style modulation were effective at changing the targeted variables, with fewer and smaller co-varying changes in the non-targeted variables, most notably speed. This indicates that it is better to use the combined instructions in practice. In general, user experience of the guided-exploration training was positive for all domains, except for the scale Dependability. Dependability reflects the extent to which the user feels in control and safe [23]. In this study, an audio file was used to administer the instructions during the training. As a result, the instructions were exactly the same for every training, and the users had no control over the instructions at all, which could explain the lower scores on this scale. If the guided-exploration training tested in this study was to be implemented in a mobile application, where the user can change settings and have more autonomous control over which instructions are provided, we expect this score to go up. In that situation, the already higher experience scores on the other scales might follow suit. We defined the participants’ preferred running style as the baseline running style and induced and evaluated any changes relative to this running style. However, the preferred run- ning style of a participant is not necessarily the ‘Sit’ running style according to the dual-axis PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 10 / 13 model and hence could influence their ability to change in certain directions, because there is a limit to these variables (e.g. when someone is a pronounced ‘Bounce’ runner (a relatively long flight phase), it could be more difficult to decrease duty factor even further, as the relative flight phase cannot be increased indefinitely). Unfortunately, there are currently no population reference values available for the different running styles on the basis of which the running style of an individual runner could be determined. For this, further research and speed-depen- dent reference values for different running styles are necessary. In this study only five runners participated, mainly due to the limiting requirement of using a sports watch and accessory that measured stance time. As stance time and duty factor have only recently been recognized as important running variables [17, 22], not many wearables or mobile apps currently can measure and report these variables. For this study, the measurement of this variable was necessary to assess the effect of the instructions on the duty factor. How- ever, this is not a requirement for the use of the instructions per se for a guided exploration towards different running styles, as one would only need a device to play the audio file. Furthermore, participants used their own sports watch to measure the data. While sports watches are used in research and generally seem to provide valid measurements [9, 15, 16, 25], it should be noted that different devices were used and the accuracy of the data could depend on the specific software and hardware and could vary between participants. Because the data was analysed per participant, effects of low validity should be limited, but low reliability could have affected the results. In conclusion, our results show that when acoustic pacing or verbal stance-time instruc- tions are provided in overground running, the targeted variable (cadence or duty factor, respectively) can be successfully modulated, but with co-varying effects on the non-targeted variables (speed and respectively duty factor or cadence) due to mutual dependencies among these variables. Combining acoustic pacing and verbal stance-time instructions for running- style modulation largely mitigated the number and magnitude of co-varying effects. Our results indicate that these combined instructions are especially effective when increasing the cadence (modulation towards a more ‘Hop’-like running style) and increasing the duty factor (modulation towards a more ‘Stick’-like running style). Overall, users were generally positive about the 4-week guided-exploration training, except for the degree of Dependability. We expect that increasing the autonomy of the user by implementing the instructions in an appli- cation with self-control options will help enhance user experience of guided exploration of running styles. Supporting information S1 Dataset. (CSV) S1 Audio. (MP3) Acknowledgments We would like to thank Ben van Oeveren and Move-Metrics for their technical support. This study was part of a larger consortium project with Dopple B.V. (Assen, The Nether- lands) and foundational for the joint development of the Running Buddy application for guided exploration of running styles using their instrumented earbuds to measure and modify running parameters. PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 11 / 13 Author Contributions Conceptualization: Anouk Nijs, Melvyn Roerdink, Peter Jan Beek. Data curation: Anouk Nijs. Formal analysis: Anouk Nijs. Funding acquisition: Peter Jan Beek. Investigation: Anouk Nijs. Methodology: Anouk Nijs, Melvyn Roerdink, Peter Jan Beek. Project administration: Anouk Nijs, Peter Jan Beek. Resources: Peter Jan Beek. Software: Anouk Nijs. Supervision: Melvyn Roerdink, Peter Jan Beek. Validation: Anouk Nijs, Melvyn Roerdink. Visualization: Anouk Nijs, Melvyn Roerdink. Writing – original draft: Anouk Nijs. Writing – review & editing: Anouk Nijs, Melvyn Roerdink, Peter Jan Beek. References 1. Hulteen RM, Smith JJ, Morgan PJ, Barnett LM, Hallal PC, Colyvas K, et al. Global participation in sport and leisure-time physical activities: A systematic review and meta-analysis. Prev Med (Baltim). 2017; 95: 14–25. https://doi.org/10.1016/j.ypmed.2016.11.027 PMID: 27939265 2. Oja P, Titze S, Kokko S, Kujala UM, Heinonen A, Kelly P, et al. Health benefits of different sport disci- plines for adults: Systematic review of observational and intervention studies with meta-analysis. Br J Sports Med. 2015; 49: 434–440. https://doi.org/10.1136/bjsports-2014-093885 PMID: 25568330 3. Janssen M, Walravens R, Thibaut E, Scheerder J, Brombacher A, Vos S. Understanding different types of recreational runners and how they use running-related technology. Int J Environ Res Public Health. 2020;17. https://doi.org/10.3390/ijerph17072276 PMID: 32230999 4. Van Hooren B, Goudsmit J, Restrepo J, Vos S. Real-time feedback by wearables in running: Current approaches, challenges and suggestions for improvements. J Sports Sci. 2020; 38: 214–230. https:// doi.org/10.1080/02640414.2019.1690960 PMID: 31795815 5. Willy RW, Buchenic L, Rogacki K, Ackerman J, Schmidt A, Willson JD. In-field gait retraining and mobile monitoring to address running biomechanics associated with tibial stress fracture. Scand J Med Sci Sport. 2016; 26: 197–205. https://doi.org/10.1111/sms.12413 PMID: 25652871 6. de Ruiter CJ, Verdijk PWL, Werker W, Zuidema MJ, de Haan A. Stride frequency in relation to oxygen consumption in experienced and novice runners. Eur J Sport Sci. 2014; 14: 251–258. https://doi.org/10. 1080/17461391.2013.783627 PMID: 23581294 7. Schubert AG, Kempf J, Heiderscheit BC. Influence of stride frequency and length on running mechan- ics: A systematic review. Sports Health. 2014; 6: 210–217. https://doi.org/10.1177/1941738113508544 PMID: 24790690 8. Adams D, Pozzi F, Willy RW, Carrol A, Zeni J. Altering cadence or vertical oscillation during running: Effects on running related injury factors. Int J Sports Phys Ther. 2018; 13: 633–642. https://doi.org/10. 2519/jospt.2010.3166 PMID: 30140556 9. de Ruiter CJ, van Daal S, van Diee¨n JH. Individual optimal step frequency during outdoor running. Eur J Sport Sci. 2019; 0: 1–9. https://doi.org/10.1080/17461391.2019.1626911 PMID: 31284825 10. van Oeveren BT, de Ruiter CJ, Beek PJ, van Diee¨n JH. Optimal stride frequencies in running at different speeds. PLoS One. 2017; 12: 1–12. https://doi.org/10.1371/journal.pone.0184273 PMID: 29059198 11. Nijs A, Roerdink M, Beek PJ. Cadence modulation in walking and running: Pacing steps or strides? Brain Sci. 2020; 10: 273. https://doi.org/10.3390/brainsci10050273 PMID: 32370091 PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 12 / 13 12. Heiderscheit BC, Chumanov ES, Michalski MP, Wille CM, Ryan MB. Effects of step rate manipulation on joint mechanics during running. Med Sci Sports Exerc. 2011; 43: 296–302. https://doi.org/10.1249/ MSS.0b013e3181ebedf4 PMID: 20581720 13. van Dyck E, Moens B, Buhmann J, Demey M, Coorevits E, Dalla Bella S, et al. Spontaneous entrain- ment of running cadence to music tempo. Sport Med—Open. 2015; 1–14. https://doi.org/10.1186/ s40798-015-0025-9 PMID: 26258007 14. Gray A, Price M, Jenkins D. Predicting temporal gait kinematics from running velocity. J Strength Cond Res. 2021; 35: 2379–2382. https://doi.org/10.1519/JSC.0000000000003198 PMID: 31268993 15. te Brake M, Stolwijk N, Staal B, Van Hooren B. Using beat frequency in music to adjust running cadence in recreational runners: A randomized multiple baseline design. Eur J Sport Sci. 2022; 1–10. https://doi. org/10.1080/17461391.2022.2042398 PMID: 35176971 16. Musgjerd T, Anason J, Rutherford D, Kernozek TW. Effect of increasing running cadence on peak impact force in an outdoor environment. 2021; 16: 1076–1083. https://doi.org/10.26603/001c.25166 PMID: 34386286 17. Malisoux L, Gette P, Delattre N, Urhausen A, Theisen D. Spatiotemporal and ground-reaction force characteristics as risk factors for running-related injury: A secondary analysis of a randomized trial including 800+ recreational runners. Am J Sports Med. 2022; 50: 537–544. https://doi.org/10.1177/ 03635465211063909 PMID: 35049407 18. Nummela A, Kera¨nen T, Mikkelsson LO. Factors related to top running speed and economy. Int J Sports Med. 2007; 28: 655–661. https://doi.org/10.1055/s-2007-964896 PMID: 17549657 19. Nijs A, Roerdink M, Beek PJ. Running-style modulation: Effects of stance-time and flight-time instruc- tions on duty factor and cadence. Gait Posture. 2022; 98: 283–288. https://doi.org/10.1016/j.gaitpost. 2022.10.002 PMID: 36242910 20. Morin JB, Samozino P, Zameziati K, Belli A. Effects of altered stride frequency and contact time on leg- spring behavior in human running. J Biomech. 2007; 40: 3341–3348. https://doi.org/10.1016/j.jbiomech. 2007.05.001 PMID: 17602692 21. Moore IS, Ashford KJ, Cross C, Hope J, Jones HSR, McCarthy-Ryan M. Humans optimize ground con- tact time and leg stiffness to minimize the metabolic cost of running. Front Sport Act Living. 2019; 1: 1– 10. https://doi.org/10.3389/fspor.2019.00053 PMID: 33344976 22. van Oeveren BT, de Ruiter CJ, Beek PJ, van Diee¨n JH. The biomechanics of running and running styles: A synthesis. Sport Biomech. 2021. https://doi.org/10.1080/14763141.2021.1873411 PMID: 33663325 23. Schrepp M, Hinderks A, Thomaschewski J. Construction of a benchmark for the User Experience Ques- tionnaire (UEQ). Int J Interact Multimed Artif Intell. 2017; 4: 40. https://doi.org/10.9781/ijimai.2017.445 24. Rauschenberger M, Schrepp M, Perez-Cota M, Olschner S, Thomaschewski J. Efficient measurement of the user experience of interactive products. How to use the User Experience Questionnaire (UEQ). Example: Spanish language version. Int J Interact Multimed Artif Intell. 2013; 2: 39. https://doi.org/10. 9781/ijimai.2013.215 25. van Oeveren BT, de Ruiter CJ, Hoozemans MJM, Beek PJ, van Diee¨n JH. Inter-individual differences in stride frequencies during running obtained from wearable data. J Sports Sci. 2019; 37: 1996–2006. https://doi.org/10.1080/02640414.2019.1614137 PMID: 31079578 PLOS ONE Exploring running styles in the field through cadence and duty factor modulation PLOS ONE | https://doi.org/10.1371/journal.pone.0295423 December 7, 2023 13 / 13

Exploring running styles in the field through cadence and duty factor modulation.

12-07-2023

Nijs, Anouk,Roerdink, Melvyn,Beek, Peter Jan

eng