Even Pacing Is Associated with Faster Finishing Times in Ultramarathon Distance Trail Running—The “Ultra-Trail du Mont Blanc” 2008–2019

cent years, there has been an increasing number of investigations analyzing the e ects of sex, performance level, and age on pacing in various running disciplines. However, little is known about the impact of those factors on pacing strategies in ultramarathon trail running. This study investigated the e ects of age, sex, and performance level on pacing in the UTMB ® (Ultra-trail du Mont Blanc) and aimed to verify previous ndings obtained in the research on other running disciplines and other ultramarathon races. Data from the UTMB ® from 2008 to 2019 for 13,829 race results (12,681 men and 1148 women) were analyzed. A general linear model (two-way analysis of variance (ANOVA)) was applied to identify a sex, age group, and interaction e ect in pace average and pace variation. A univariate model (one-way ANOVA) was used to identify a sex e ect for age, pace average, and pace variation for the fastest men and women. In our study, pace average and a steadier pace were positively correlated. Even pacing throughout the UTMB ® correlated with faster nishing times. The average pace depended signi cantly on sex and age group. When considering

model (one-way ANOVA) was used to identify a sex e ect for age, pace average, and pace variation for the fastest men and women. In our study, pace average and a steadier pace were positively correlated. Even pacing throughout the UTMB ® correlated with faster nishing times. The average pace depended signi cantly on sex and age group. When considering the top ve athletes in each age group, sex and age group also had signi cant e ects on pace variation. The fastest women were older than the fastest men, and the fastest men were faster than the fastest women. Women had a higher pace variation than men. In male competitors, younger age may be advantageous for a successful nish of the UTMB ® . Faster male runners seemed to be younger in ultramarathon trail running with large changes in altitude when compared to other distances and terrains. Keywords:pacing; trail running; sex di erence; ultramarathon 1. Introduction Ultra-trail running has previously been de ned as any running event that ful lls the following two criteria: The distance must be longer than the classical marathon distance of 42.195 km, and the race must take place in a natural environment in an open country that is mainly o -road [1,2]. In order to further clarify the terms used to describe the distinct o -road running categories, Scheer et al. recently Int. J. Environ. Res. Public Health2020,17, 7074; doi:10.3390 /ijerph17197074 /journal/ijerph

Int. J. Environ. Res. Public Health2020,17, 7074 2 of 11 published a consensus statement [3]. In their publication, ultramarathon running is de ned by distance as any race longer than the traditional marathon distance with no restrictions regarding the terrain. The authors recommend avoiding the term ultramarathon without precise de nition of further details such as surface, elevation change, and level of support as each ultramarathon is unique. Trail running, on the other hand, is de ned by the running surface with 20–25% of paved or asphalted road [3]. Elevation and race distance are not speci ed. In the last years, the various o -road running disciplines have become more popular with an increasing number of athletes participating, and a steady increase in the number of races held each year [4,5]. Additionally, trail and ultramarathon running have become more accessible to non-professional runners [6], despite the high risk of extreme fatigue and/or exhaustion and other potential medical issues that may occur during each race [7]. Among the di erent ultramarathon races, the Ultra-Trail du Mont Blanc (UTMB ® ), with its 171 km distance, over 10,000 m of vertical gain, and a time limit of 46 h, is one of the most important ultramarathons held in Europe [2]. With six International Trail Running Association (ITRA) points, the UTMB ® is in the highest category of trail running worldwide [2,8]. Every year, approximately 2300 quali ed runners participate in the race. Accordingly, we consider the UTMB ® as highly suitable for the analysis of pacing in trail running over the ultramarathon distance. Several studies have shown the importance of an athlete's pacing strategy on performance [9–11]. Furthermore, optimal pacing strategies for ultramarathons have been investigated in previous publications [9,10,12]. For example, Ho man et al. [10] were able to show that an even pacing over the entirety of a 161 km long ultramarathon resulted in the fastest nishing times. This nding concurs with the ndings of Knechtle et al. [12] about pacing strategies during a 100 km ultramarathon. Furthermore, a review published by Abbiss et al. [9] suggested an even pacing

Ho man et al. [10] were able to show that an even pacing over the entirety of a 161 km long ultramarathon resulted in the fastest nishing times. This nding concurs with the ndings of Knechtle et al. [12] about pacing strategies during a 100 km ultramarathon. Furthermore, a review published by Abbiss et al. [9] suggested an even pacing during a prolonged period of power output to be favorable. In recent years, there has been an increasing number of studies analyzing the e ect of sex, performance level, and age on pacing [13,14]. It could be shown that, on average, ultramarathon runners are older than marathon runners. They also have a higher weekly training volume, but train at a lower speed. The experience in ultramarathon events seems to have a great in uence on success. Furthermore, a low BMI and a low body fat percentage proved to be advantageous [15]. However, little is known about the e ects of sex, age and performance level on pacing strategies speci cally in ultramarathon running. Aging seems to have a signi cant e ect on pacing in endurance sports, and many studies have shown varying pacing strategies for di erent age groups in marathon and half-marathon runners [16–19]. Runners with age older than 30 years appear to have a more even distribution of pacing throughout a marathon when compared with younger contestants [17]. These ndings, however, have not yet been explored for ultramarathon distance running. Varesco et al. [20] investigated the impact of age on performance in the UTMB ® and found that the average speed progressively decreased with age. Another variable impacting the pacing strategy of runners is their sex. It is well known that, in general, male athletes achieve faster nishing times in running competitions than female athletes [21]. Regarding pacing, women have shown to be better pacers with a more even pace throughout a race [17,22,23]. This e ect could also be shown in two studies on ultra-triathlons [24,25] and one on a 100-km marathon [22]. However, to our knowledge, there has not yet been a publication observing the

nishing times in running competitions than female athletes [21]. Regarding pacing, women have shown to be better pacers with a more even pace throughout a race [17,22,23]. This e ect could also be shown in two studies on ultra-triathlons [24,25] and one on a 100-km marathon [22]. However, to our knowledge, there has not yet been a publication observing the correlation between sex and pacing in an ultramarathon with such great elevation gain as the UTMB ® . Furthermore, the performance level of athletes seems to play an essential role in pacing during running competitions. More experienced, faster runners are more constant pacers independent of the running distance, as shown in several studies [16,18]. Competing in the UTMB ® involves running on a rougher terrain than in most half-marathons, marathons, or even ultramarathons with over 10,000 m of vertical gain. Based on existing ndings and a lack of knowledge for ultramarathon running on rough, mountainous terrain, the present study aimed to analyze how age, sex, and performance level of

Int. J. Environ. Res. Public Health2020,17, 7074 3 of 11 athletes a ect their pacing during the UTMB ® , and to identify ideal pacing strategies for ultramarathon distance trail running considering these variables. With rising average participant age and an increasing number of non-professional participants, providing insight into possible pacing strategies for di erent groups of runners seems to be of particular importance. Considering previous ndings [10,12], we hypothesized that a more constant pacing throughout the UTMB ® would lead to better results (i.e., nal ranking) and, therefore, a higher performance level. The rough terrain and high vertical gain of the UTMB ® were potential factors in uencing pacing strategies. Taking these distinct conditions into account, we hypothesized the distribution of pacing over the race to vary more when compared to other race categories. Furthermore, we considered a successful pacing strategy to potentially be of even more critical importance for the outcome in a longer, more strenuous race, possibly accentuating previous ndings. 2. Materials and Methods 2.1. Ethical 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 participant as the study involved the analysis of publicly available data (EKSG 01-06-2010). The study was conducted in accordance with the recognized ethical standards according to the Declaration of Helsinki (2013). 2.2. Race Description The Ultra-trail du Mont-Blanc (UTMB ® ) has grown since its rst realization in 2003 to be one of the most important ultramarathons in Europe. The race is held annually in France in the region of Chamonix during the last week of August. Participants run along the Tour du Mont Blanc hiking path, a 170 km long circular trail with over 10,000 m of vertical gain, and cross three countries (France, Italy, and Switzerland), seven valleys, 71 glaciers and approximately 400 summits (see race pro le of the 2019 edition in Figure). While the path usually takes hikers seven to ten days to nish, elite runners, achieve nishing times of ~21 h. The time limit for completion of the race is

10,000 m of vertical gain, and cross three countries (France, Italy, and Switzerland), seven valleys, 71 glaciers and approximately 400 summits (see race pro le of the 2019 edition in Figure). While the path usually takes hikers seven to ten days to nish, elite runners, achieve nishing times of ~21 h. The time limit for completion of the race is 46 h and 30 min. All athletes must pass the 15 speci c checkpoints within a de ned time to be allowed to continue the race. Participants not nishing the entire race or not passing through the checkpoints within the determined times are excluded.Int. J. Environ. Res. Public Health 2020, 17, x FOR PEER REVIEW 3 of 11 ultramarathon distance trail running considering these variables. With rising average participant age and an increasing number of non-professional participants, providing insight into possible pacing strategies for different groups of runners seems to be of particular importance. Considering previous findings [10,12], we hypothesized that a more constant pacing throughout the UTMB ® would lead to better results (i.e., final ranking) and, therefore, a higher performance level. The rough terrain and high vertical gain of the UTMB ® were potential factors influencing pacing strategies. Taking these distinct conditions into account, we hypothesized the distribution of pacing over the race to vary more when compared to other race categories. Furthermore, we considered a successful pacing strategy to potentially be of even more critical importance for the outcome in a longer, more strenuous race, possibly accentuating previous findings. 2. Materials and Methods 2.1. Ethical 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 participant as the study involved the analysis of publicly available data (EKSG 01-06-2010). The study was conducted in accordance with the recognized ethical standards according to the Declaration of Helsinki (2013). 2.2. Race Description The Ultra-trail du Mont-Blanc (UTMB ® ) has grown since its first realization in 2003 to be one of the most important ultramarathons in Europe. The race is held annually in France

analysis of publicly available data (EKSG 01-06-2010). The study was conducted in accordance with the recognized ethical standards according to the Declaration of Helsinki (2013). 2.2. Race Description The Ultra-trail du Mont-Blanc (UTMB ® ) has grown since its first realization in 2003 to be one of the most important ultramarathons in Europe. The race is held annually in France in the region of Chamonix during the last week of August. Participants run along the Tour du Mont Blanc hiking path, a 170 km long circular trail with over 10,000 m of vertical gain, and cross three countries (France, Italy, and Switzerland), seven valleys, 71 glaciers and approximately 400 summits (see race profile of the 2019 edition in Figure 1). While the path usually takes hikers seven to ten days to finish, elite runners, achieve finishing times of ~21 h. The time limit for completion of the race is 46 h and 30 min. All athletes must pass the 15 specific checkpoints within a defined time to be allowed to continue the race. Participants not finishing the entire race or not passing through the checkpoints within the determined times are excluded. Figure 1. Race profile of the 2019 edition (Ultra-trail du Mont Blanc (UTMB ® ) Website (20)). As an ultra-trail held in exposed, mountainous terrain, the UTMB ® requires the ability to manage extreme conditions as well as autonomy in the mountains. The race considers the principle of semi- autonomy, defined as autonomy between two aid stations. Aid stations are equipped with food and drink supplies as well as medical facilities in some cases. Participants must at least carry all the items listed on the official UTMB ® website [26]. Since 2006, the number of participants is restricted to 2300 Figure 1.Race pro le of the 2019 edition (Ultra-trail du Mont Blanc (UTMB ® ) Website (20)). As an ultra-trail held in exposed, mountainous terrain, the UTMB ® requires the ability to manage extreme conditions as well as autonomy in the mountains. The race considers the principle of semi-autonomy, de ned as autonomy between two aid stations. Aid stations

2300 Figure 1.Race pro le of the 2019 edition (Ultra-trail du Mont Blanc (UTMB ® ) Website (20)). As an ultra-trail held in exposed, mountainous terrain, the UTMB ® requires the ability to manage extreme conditions as well as autonomy in the mountains. The race considers the principle of semi-autonomy, de ned as autonomy between two aid stations. Aid stations are equipped with food and drink supplies as well as medical facilities in some cases. Participants must at least carry all the items listed on the o cial UTMB ® website [26]. Since 2006, the number of participants is restricted to 2300 runners. To register as a participant, athletes must have scored a minimum of 10 ITRA (International Trail Running Association) points in two years and a maximum of two races. As ultra-trail running has become increasingly popular since the launch of the UTMB ® in 2003, there is a rising number of runners ful lling the inclusion criteria. A lottery system was established in 2007

Int. J. Environ. Res. Public Health2020,17, 7074 4 of 11 to prevent overcrowding. In the interest of equitable recognition of top athletes, the race organizers reserve places for elite runners. Requirement for inclusion in this category is an ITRA performance index of over 790 points for males and over 670 points for female runners, respectively. 2.3. Data Acquisition For this study, we collected publicly available data from the o cial UTMB ® website [26]. We included the split times of all participants, independent of age group, sex, and performance level. In addition to the split times, data regarding participants' name, age, country of origin, nishing time, sex, performance level, and average running speed was gathered from the website of the Deutsche Ultramarathon Vereinigung e.V. (German Ultramarathon Union, DUV) [5]. The database from DUV exclusively consists of nisher data. In contrast, the UTMB ® database includes nishers as well as non- nishers but lacks information on the exact age of athletes and their performance level. As the aim of this study was to analyze the impact of age, sex, and performance level on pacing, it was necessary to combine data from both databases and therefore exclude all non- nishers from our evaluation. The data was manually recorded in Microsoft EXCEL sheets for later statistical analysis. To calculate the pacing of the individual athletes over time, we determined the distance and vertical gain between the di erent time stations. Time stations seemed to vary between the years due to several di erent reasons, such as weather conditions, and only the route data for the year 2019 was publicly available at the moment of acquisition. Thus, we were forced to exclude sometime stations and even data from whole years from the nal analysis. This excluded data was comprised of all-time stations that we could not compare directly and all years, which showed less than 18 time stations directly comparable with the year 2019 (cf. Table). The cut-o of a minimum of 18 comparable time stations was de ned based on a detailed evaluation of the available data. We reached for the optimal trade-o

This excluded data was comprised of all-time stations that we could not compare directly and all years, which showed less than 18 time stations directly comparable with the year 2019 (cf. Table). The cut-o of a minimum of 18 comparable time stations was de ned based on a detailed evaluation of the available data. We reached for the optimal trade-o between included years and time stations. Table according to di erent cut-o values for the minimum number of time stations. Including all available time stations would have resulted in an analysis of merely a single year. In contrast, the inclusion of all years would have reduced the amount of comparable time stations to two. Since our study focused on pacing, we decided that a higher quantity of time stations and hence an improved observed ability of variation in pacing would be of great value for our statistical analysis. At the same time, we aimed to include as many years as possible without reducing the validity of our results. In synopsis of these considerations, we decided on the cut-o of 18 comparable time stations, hence including nine years in our nal analysis. 2.4. Statistical Analysis Data were tested for normality with the K-S test, and after showing parametric (p>0.05) distribution data were then expressed as mean and standard deviations. Pace variation was determined as the coe cient of variation (CV%) of pace average between di erent time station throughout the race. CV% was calculated individually for each athlete based on the pace throughout the race. A general linear model (two-way ANOVA) was applied to identify a sex, age group, and interaction e ect in pace average and pace variation. A univariate model (one-way ANOVA) was applied to identify a sex e ect for age, pace average, and pace variation for the fastest men and women. Pearson's correlation coe cient was applied. Correlation coe cient descriptors were considered as small: 0.1–0.3; moderate: 0.3–0.5; and large:>0.5 [27]. Statistical signi cance was setp<0.05. Statistical Software for the Social Sciences (IBM SPSS Statistics version 26.0 for Windows, IBM Corp, Armonk, NY, USA, 2018) and GraphPad

ect for age, pace average, and pace variation for the fastest men and women. Pearson's correlation coe cient was applied. Correlation coe cient descriptors were considered as small: 0.1–0.3; moderate: 0.3–0.5; and large:>0.5 [27]. Statistical signi cance was setp<0.05. Statistical Software for the Social Sciences (IBM SPSS Statistics version 26.0 for Windows, IBM Corp, Armonk, NY, USA, 2018) and GraphPad Prism (GraphPad Prism version 8.4.0 for Windows, GraphPad Software, La Jolla, CA, USA, 2018) were used to carry out the analysis.

The study analyzes pacing strategies in ultramarathon trail running based on data from the UTMB.