American Masters Road Running Records—The Performance Gap Between Female and Male Age Group Runners from 5 Km to 6 Days

ent studies investigating elite and master athletes in pool- and long-distance open-water swimming showed for elite swimmers that the fastest women were able to outperform the fastest men, and for master athletes that elderly women were able to achieve a similar performance to elderly men. The present study investigating age group records in runners from 5 km to 6 days aimed to test this hypothesis for master runners. Data from the American Master Road Running Records were analyzed, for 5 km, 8 km, 10 km, 10 miles, 20 km, half-marathon, 25 km, 30 km, marathon, 50 km, 50 miles, 100 km, 100 miles, 12 h, 24 h, 48 h and 144 h, for athletes in age groups ranging from 40 to 99 years old. The performance gap between men and women showed higher e ects in events lengthening from 5 km to 10 miles (d=0.617) and lower e ects in events lengthening from 12 to 144 h (d=0.304) running. Both other groups showed similar e ects, being 20 km to the marathon (d=0.607) and 50 km to 100 miles (d=0.563). The performance gap between men and women showed higher e ects in the age groups 85 years and above (d=0.953) followed by 55 to 69 years (d=0.633), and

ects in events lengthening from 12 to 144 h (d=0.304) running. Both other groups showed similar e ects, being 20 km to the marathon (d=0.607) and 50 km to 100 miles (d=0.563). The performance gap between men and women showed higher e ects in the age groups 85 years and above (d=0.953) followed by 55 to 69 years (d=0.633), and lower e ects for the age groups 40 to 54 years (d=0.558) and 70 to 84 years (d=0.508). In summary, men are faster than women in American road running events, however, the sex gap decreases with increasing age but not with increasing event length. Keywords:athlete; running; ultra-endurance; endurance; marathon 1. Introduction Sport performance di erences between men and women have been previously reported in several di erent modalities, such as in open-water swimming, in marathon running and in Ironman triathlon [1–4]. The most likely explanation of the sex gap in sports is possibly due to di erent human skeletal muscle gene expression and their interaction with sex-speci c hormones [5,6]. This fact is leading to larger muscles and more functional muscle bers in men, which elicits greater muscle strength/power and/or endurance [7–9]. However, in endurance sports the sex gap has been changing over the last decades [10]. For instance, for elite athletes, Nikolaidis et al. [4] reported that women were faster than men in open-water swimming in the English Channel Crossing. Furthermore, in long-distance pool swimming [11] and in open-water swimming like the English Channel Crossing [12] women were able to achieve a similar performance Int. J. Environ. Res. Public Health2019,16, 2310; doi:10.3390 /ijerph16132310 /journal/ijerph

Int. J. Environ. Res. Public Health2019,16, 2310 2 of 9 to men. In some ultra-distance swimming events, women were even able to outperform men [4]. Moreover, for other sports disciplines elite women seemed to have improved more than men in marathon running [2] and in Ironman triathlon [3]. Another report showed that the sex gap decreased with increasing age in marathon runners [1]. And in ultra-endurance road races, women seem to have a higher range of peak performance than men (30–54 years vs. 30–49 years) but the peak of age-related performance decline was similar between them (60–64 years) [13]. For master athletes, it seems that elderly women can reach the performance of elderly men, especially in pool and open-water swimming. Recent studies investigating master pool swimmers in freestyle [14], backstroke [15], butter y [16], breaststroke [17], individual medley [14], and open-water [18] swimming showed that women in the older age groups (i.e., older than 75 years) achieved a similar performance to men. Men have shown to have better values of physiological performance determinants such as higher VO2max, lower external load at anaerobic threshold, and better running economy [19]. However, women normally accumulate more body fat and also have an increased e ciency to produce energy through oxidation of this substrate during endurance performance [20], which could be an advantage in events with a longer length. Furthermore, the natural decline in sex-speci c hormones in men while aging [21] may a ect their performance, possibly leading to the recently observed sex di erence reduction as athletes age [1]. Aging is a natural, progressive and inevitable biological process [22]. As humans age, a decrease in anabolic hormones may lead to the well-known age-related aspects such as sarcopenia, osteopenia, increased body fat, and loss of function in several other tissues [22]. Although master athletes have a good life-style capable of improved health and attenuated aging [23,24], they are not immune to the biological age-related e ects, which may re ect in a performance decrease while aging. We therefore aimed to investigate the possible e ect of age and race length in endurance American

loss of function in several other tissues [22]. Although master athletes have a good life-style capable of improved health and attenuated aging [23,24], they are not immune to the biological age-related e ects, which may re ect in a performance decrease while aging. We therefore aimed to investigate the possible e ect of age and race length in endurance American master road runners of di erent lengths, which to the best of our knowledge has not yet been investigated for age group runners. We hypothesized, based on recent ndings for elite and master pool- and long-distance swimmers, that the longer the length and/or higher the age the smaller the sex di erence would be. 2. Materials and Methods To test our hypothesis, the world records in each distance and age-group of both men and women were included. Dispersion plots with linear regression analysis may show if the performance pattern of men is di erent from women with increasing age and increasing race length. 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 participants as the study involved the analysis of publicly available data (EKSG 01/06/2010). 2.2. Data All data were the o cial USA Track and Field race records from master athletes, publicly available from their web site [25]. The American records from all road running events for men and women were collected for further analysis. Collection of data ranged the years of 1970 to 2017. The last time the database was checked for new records was March 2019. The included events were: 5 km, 8 km, 10 km, 10 miles, 20 km, half-marathon, 25 km, 30 km, marathon, 50 km, 50 miles, 100 km, and 100 miles for distance-limited races, and 12 h, 24 h, 48 h and 144 h for time-limited races (for all age-groups). Data included age groups in 5-year-intervals ranging from 40–44 to 95–99 years old (for all events). The data sample included 214 men and 200 women (i.e., total sample sizen=414). Race times in the time-limited

50 miles, 100 km, and 100 miles for distance-limited races, and 12 h, 24 h, 48 h and 144 h for time-limited races (for all age-groups). Data included age groups in 5-year-intervals ranging from 40–44 to 95–99 years old (for all events). The data sample included 214 men and 200 women (i.e., total sample sizen=414). Race times in the time-limited races and achieved distance in the distance-limited races were all converted to running speed for results comparison.

Int. J. Environ. Res. Public Health2019,16, 2310 3 of 9 2.3. Statistical Analysis Data analysis was conducted with pooled groups by events, pooled groups by age-group and without any grouping. An analysis of variance adjusted by age (ANCOVA) was applied to compare men and women performances when pooled by event or age group. Furthermore, the e ect size [26] between men and women was calculated for each event and age group and then pooled in di erent groups. Finally, linear regressions with individual values were conducted between relative sex di erence (%) and age or event length (km). The signi cance level was 5% (p<0.05). All procedures were performed using Statistical Software for the Social Sciences (IBM, SPSS v21.0, Chicago, IL, USA) and GraphPad Prism (Graph Pad Prism v7.0, San Diego, CA, USA). 3. Results The comparisons of running speed between men and women were signi cantly di erent in all endurance events for all age groups. Men have a signi cantly faster average running speed in endurance events lengthening from 5 km to 144 h running (Table). Table 1. Average running speed (km h 1 ) of the American masters road running records by event. Data expressed as mean and standard deviation ( ). Events Men Women p-Value 5 km (n=22) 15.5 5.2 13.4 4.4<0.00001 8 km (n=18) 15.9 3.7 14.5 2.7<0.00001 10 km (n=18) 14.9 4.2 13.7 3.2 0.00009 12 km (n=16) 15.9 3.2 14.8 2.1 0.00012 15 km (n=20) 15.7 4.2 10.0 3.1 0.00041 10 miles (n=22) 15.0 4.1 13.0 3.6 0.00037 20 km (n=16) 16.2 2.6 14.3 2.8<0.00001 Half-marathon (n=16)15.5 4.4 12.6 3.0 0.00003 25 km (n=18) 15.1 3.2 13.2 3.2 0.00001 30 km (n=18) 14.0 4.0 12.6 3.5 0.00002 Marathon (n=22) 14.1 4.0 12.0 3.6<0.00001 50 km (n=18) 12.9 3.3 11.0 3.3 0.00047 50 miles (n=18) 12.0 2.5 9.6 3.4<0.00001 100 km (n=18) 11.2 2.8 8.7 3.8 0.00015 100 miles (n=18) 8.0 3.2 7.6 2.7 0.01730 12 h (n=14) 9.2 2.3 8.9 2.3 0.00052 24 h (n=18) 8.2 2.3 7.4 2.3 0.00009 48 h (n=12) 5.8 1.7 5.0 2.0 0.00023 144 h (n=8) 4.6 1.4

50 km (n=18) 12.9 3.3 11.0 3.3 0.00047 50 miles (n=18) 12.0 2.5 9.6 3.4<0.00001 100 km (n=18) 11.2 2.8 8.7 3.8 0.00015 100 miles (n=18) 8.0 3.2 7.6 2.7 0.01730 12 h (n=14) 9.2 2.3 8.9 2.3 0.00052 24 h (n=18) 8.2 2.3 7.4 2.3 0.00009 48 h (n=12) 5.8 1.7 5.0 2.0 0.00023 144 h (n=8) 4.6 1.4 4.3 1.0 0.01795 p-value: univariate model adjusted by age; Sample size was equal for men and women. The comparisons between men and women were not statistically di erent in almost every age group for all events. Men have a signi cantly faster average running speed in the age group 60–64 years (Table).

Int. J. Environ. Res. Public Health2019,16, 2310 4 of 9 Table 2. Average running speed (km h 1 ) of the American masters road running records by age-group. Data expressed as mean and standard deviation ( ). Age Groups Men Women p-Value 40–44 years (n=36)17.4 4.3 14.5 4.4 0.20427 50–54 years (n=36)16.3 4.2 13.7 4.3 0.07979 55–59 years (n=36)15.0 3.9 13.6 3.5 0.31009 60–64 years (n=36)15.5 3.9 12.1 3.8 0.00368 65–69 years (n=34)14.3 3.4 12.0 3.4 0.05812 70–74 years (n=38)12.8 3.9 11.7 3.0 0.26013 75–79 years (n=32)12.2 3.7 10.7 2.8 0.11431 80–84 years (n=30)11.3 3.7 8.8 3.4 0.29139 85–89 years (n=10) 8.7 3.9 7.3 3.1 0.20891 90–94 years (n=6) 7.9 2.6 7.1 0.7 0.42614 95–99 years (n=4) 7.1 1.3 5.3 1.1 0.04698 p-value: univariate model adjusted by age. Sample size was equal for men and women. The performance gap between men and women showed higher e ects in events lengthening from 5 km to 10 miles (d=0.617) and lower e ects in events lengthening from 12 to 144 h (d=0.304) running (FigureA). Both other groups showed similar e ects, being from 20 km to the marathon (d=0.607) and from 50 km to 100 miles (d=0.563). The performance gap between men and women showed higher e ects in the age groups 80–85 years and above (d=0.953) followed by 55–59 to 65–69 years (d=0.633), and the lower e ects for the age groups 40–44 to 50–54 years (d=0.558) and 70 to 84 years (d=0.508) (FigureB).Int. J. Environ. Res. Public Health 2019, 16, x 4 of 9 Table 2. Average running speed (km·h −1 ) of the American masters road running records by age-group. Data expressed as mean and standard deviation (±). Age groups Men Women p-value 40–44 years (n = 36) 17.4 ± 4.3 14.5 ± 4.4 0.20427 50–54 years (n = 36) 16.3 ± 4.2 13.7 ± 4.3 0.07979 55–59 years (n = 36) 15.0 ± 3.9 13.6 ± 3.5 0.31009 60–64 years (n = 36) 15.5 ± 3.9 12.1 ± 3.8 0.00368 65–69 years (n = 34) 14.3 ± 3.4 12.0 ± 3.4 0.05812 70–74 years (n = 38) 12.8

± 4.3 14.5 ± 4.4 0.20427 50–54 years (n = 36) 16.3 ± 4.2 13.7 ± 4.3 0.07979 55–59 years (n = 36) 15.0 ± 3.9 13.6 ± 3.5 0.31009 60–64 years (n = 36) 15.5 ± 3.9 12.1 ± 3.8 0.00368 65–69 years (n = 34) 14.3 ± 3.4 12.0 ± 3.4 0.05812 70–74 years (n = 38) 12.8 ± 3.9 11.7 ± 3.0 0.26013 75–79 years (n = 32) 12.2 ± 3.7 10.7 ± 2.8 0.11431 80–84 years (n = 30) 11.3 ± 3.7 8.8 ± 3.4 0.29139 85–89 years (n = 10) 8.7 ± 3.9 7.3 ± 3.1 0.20891 90–94 years (n = 6) 7.9 ± 2.6 7.1 ± 0.7 0.42614 95–99 years (n = 4) 7.1 ± 1.3 5.3 ± 1.1 0.04698 p-value: univariate model adjusted by age. Sample size was equal for men and women. The performance gap between men and women showed higher effects in events lengthening from 5 km to 10 miles (d = 0.617) and lower effects in events lengthening from 12 to 144 h (d = 0.304) running (Figure 1A). Both other groups showed similar effects, being from 20 km to the marathon (d = 0.607) and from 50 km to 100 miles (d = 0.563). The performance gap between men and women showed higher effects in the age groups 80–85 years and above (d = 0.953) followed by 55–59 to 65– 69 years (d = 0.633), and the lower effects for the age groups 40–44 to 50–54 years (d = 0.558) and 70 to 84 years (d = 0.508) (Figure 1B). Figure 1. Performance gap between men and women in American masters road running records. CI: confidence interval. Greater the effect size (>0), greater the difference between men and women. Dispersion data of ‘performance sex difference ratio’ vs. ‘age’ was plotted in Figure 2 for the linear regression analysis. Linear regressions showed an increase in the sex gap with increasing age Figure 1. Performance gap between men and women in American masters road running records. CI: con dence interval. Greater the e ect size (>0), greater the di erence between

Dispersion data of ‘performance sex difference ratio’ vs. ‘age’ was plotted in Figure 2 for the linear regression analysis. Linear regressions showed an increase in the sex gap with increasing age Figure 1. Performance gap between men and women in American masters road running records. CI: con dence interval. Greater the e ect size (>0), greater the di erence between men and women.

Int. J. Environ. Res. Public Health2019,16, 2310 5 of 9 Dispersion data of `performance sex di erence ratio' vs. `age' was plotted in Figure regression analysis. Linear regressions showed an increase in the sex gap with increasing age for all endurance events excepts 144 hours running, with data only available for age groups up to 60–64 years old (Figure).Int. J. Environ. Res. Public Health 2019, 16, x 5 of 9 for all endurance events excepts 144 hours running, with data only available for age groups up to 60–64 years old (Figure 2). Figure 2. Linear regression between performance sex difference (%) and age in American masters road running records. With the positive slopes, higher the age, higher the sex difference (%) in each event. (A) events 5 km to 10 miles; (B) 20 km to Marathon; (C) 50 km to 100 miles; (D) 12 to 144 h. Dispersion data of ‘performance sex difference ratio’ vs. ‘event length’ was plotted in Figure 3 for the linear regression analysis. Linear regressions with sex gap and event length did not present a pattern of an increase or a decrease with increasing distance in different age groups (Figure 3). Age groups 60–64, 75–79 and 80–84 years showed an increasing sex gap with increasing distance and/or duration, whereas age groups 45–49, 50–54, 70–74 and 85–89 years showed a decreasing sex gap with increasing distance and/or duration. 4. Discussion In this study we hypothesized to find that women would be able to close the gap to men in higher ages and longer race distances as has been recently reported for pool- and open-water swimmers. The main finding of this investigation was that the sex gap in sports performance for American road runners decreased with increasing age but did not seem to change with increasing distance and/or duration. Nonetheless, men remained faster than women regardless of age and race distance and/or duration. The overall sex difference (men outscored women) in running performance in the examined distances (5 km–6 days) might reflect the corresponding sex differences in the physiological determinants of performance in these events, i.e., maximal oxygen

but did not seem to change with increasing distance and/or duration. Nonetheless, men remained faster than women regardless of age and race distance and/or duration. The overall sex difference (men outscored women) in running performance in the examined distances (5 km–6 days) might reflect the corresponding sex differences in the physiological determinants of performance in these events, i.e., maximal oxygen uptake, anaerobic threshold and running economy [27]. For instance, men marathon runners had higher VO 2max than their women counterparts [28]. On the other hand, women had more economical running due to their smaller body mass [29]. Figure 2. Linear regression between performance sex di erence (%) and age in American masters road running records. With the positive slopes, higher the age, higher the sex di erence (%) in each event. (A) events 5 km to 10 miles; (B) 20 km to Marathon; (C) 50 km to 100 miles; (D) 12 to 144 h. Dispersion data of `performance sex di erence ratio' vs. `event length' was plotted in Figure for the linear regression analysis. Linear regressions with sex gap and event length did not present a pattern of an increase or a decrease with increasing distance in di erent age groups (Figure). Age groups 60–64, 75–79 and 80–84 years showed an increasing sex gap with increasing distance and/or duration, whereas age groups 45–49, 50–54, 70–74 and 85–89 years showed a decreasing sex gap with increasing distance and/or duration.

The study analyzes age group records in American road running events.