Could women outrun men in ultramarathon races?

Abstract

To test a hypothesis that the differences in running performances between men and women narrow with increasing race distance, recent best running speeds(m·s^-1) of 28 female ultramarathon runners over racing distances from 5 to 90 km were compared with those of male runners matched for 56 km race time, age and training. These comparisons revealed that the men ran faster than the women over 5 to 42.2 km (P < 0.01 -0.05), but not in a 90-km race (2.82 ± 0.04 vs 2.90 ± 0.09 m·s^-1). Declines in the men's and women's running speeds with increasing race distances were best fitted by Y = 2.9·exp·(-0.01·X) + 1.7 and Y = 2.1·exp·(-0.01·X) + 2.0 equations (r = -0.99). Differences between the exponential declines in the men's and women's running speeds over longer race distances were best fitted by a linear Y = -0.006(X-66) equation (r = -0.98, P < 0.001). The negative slope and the X-axis intercept of this equation at 66 km supports the hypothesis that women ultramarathon runners have greater fatigue resistance than do equally trained men whose performances are superior up to the marathon distance.

In 1972, women were finally allowed to participate in marathon(42.2 km) and, shortly thereafter, ultra-marathon (> 42.2 km) running races^(8,13). Since that time there have been rapid improvements in women's distance running world records^(8,13,18). Some believe that such improvements in performance are simply a result of the relatively recent introduction of more talented female athletes into long distance running and are unlikely to be maintained ^(8,13). However, others have argued that the world class running performances of women may equal those of men in the marathon by the end of this decade⁽¹⁸⁾.

Accordingly, the purpose of this study was to determine the running speeds over shorter and longer racing distances of men and women matched for age, training, and performance in a 56-km ultramarathon. The question addressed was whether the slower falls in running speed with increasing race distance in women than in men could result in an ultramarathon race distance at which the performances of the world's best female runners might match or exceed those of the best male runners.

METHODS

The possibility that female ultramarathon runners may have a fatigue resistance superior to that of male ultramarathon runners was examined in a sample of 50 female runners below the age of 35 yr who completed the Two Oceans 56-km ultramarathon in Cape Town, South Africa, within 4-5 h. Each of the 50 female runners was matched for finishing time (± 1 min) and age(± 1 yr) with four different male runners in the same race. This sample population of 50 women and 200 men was taken from a total of 770 women and 7,469 men who entered the race and of whom 7,189 women and men finished.

Of the selected runners, 28 women (56%) and 103 men (52%) responded to a questionnaire. In that questionnaire, the runners filled in their body mass, height, average number of kilometers run per week, number of high intensity training sessions per month, number of years of road running, and most recent(within 1 yr) and best-ever running times for 5-, 10-, 15-, 21.1-, and 42.2-km and time for the Comrades 90-km marathon, for those who competed in that event in that year. From these data, we initially further matched each female runner to the male runners who were most similar in terms of weekly training distance, kg·m^-2 body mass index (BMI), training, and BMI or least comparable in these respects. However, the type of matching had no significant effect on the differences in the running speeds of the women and men over the increasing race distances. Accordingly the male runners' data were subsequently combined so that each female runner could be matched to a male runner who had a recent (within 1 yr) best running time over that distance and who was most similar in terms of weekly training volume. Combining the male runners' data increased the number of matched pairs from 13 to 17 at 5 km, from 20 to 25 at km, and from 5 to 10 at 90 km. Whereas 61%, 89%, and 36% of the women had recent best running times over 5, 15, and 90 km, only 47%, 75%, and 22% of the men had equivalent times over these distances.

Results presented are the means and SE of the m·s^-1 running speeds calculated from the subjects' recent (≤ 1 yr) available best finishing times for 5-, 10-, 15-, 21.1-, 42.2-, 56-, and 90-km races. Exponential declines in male and female running speeds (Y) with increasing race distances (X) were fitted to Y = A·exp·(-B·X) + C equations by nonlinear regressions, using the Institute of Scientific Information “Graph Pad” program. The same program was also used to fit a linear regression through the differences between the men's and women's running speeds over increasing race distances.

Because only ten of the 28 women completed the Comrades 90-km marathon and only 17 and 25 of the women had recent best 5 and 15 km running times, statistical significance could not be assessed with a conventional analyses of variance for repeated measures. Instead, the differences between the women's and men's running speeds over each race distance were examined with a matched pairs t-test to determine if the 95% confidence intervals were significantly different from zero.

RESULTS

The characteristics of the 28 female ultramarathon runners and the 28 male ultramarathon runners best matched for 56-km race time (± 1 min), age(± 1 yr), and average weekly training volume are presented inTable 1. While there were no statistical differences in age or training, the women were significantly shorter and lighter than the men.

Comparisons between the exponential declines in the average m·s^-1 running speeds of the women and men (Y) with increasing km race distances (X) are shown in Figure 1. The running speeds of the men were best fitted with a Y = 2.9·exp·-0.01·X) + 1.7 equation (r = 0.99) and the running speeds of the women were best fitted with a Y = 2.1·exp·(-0.01·X) + 2.0 equation (r = 0.99). The statistical significances of the faster running speeds of the men than of the women over race distances from 5 to 42.2 km, but not at 56 and 90 km, were determined from the data in Figure 2, as described in Methods.

Figure 2 shows that the differences in the men's and women's exponential declines in average m·s^-1 running speeds (Y) over longer km race distances (X) were best-fitted by a linear Y = -0.006(X-66) equation. The negative slope of this equation (r = -0.98, P< 0.001) and its intercept with the X axis at 66 km suggests that the greater fatigue resistance of female ultramarathon runners could lead to their performances becoming comparable with, or better than, those of equally trained male ultramarathon runners at distances in excess of ≈70 km.

DISCUSSION

The finding that the female ultramarathon runners were able to negate the faster performances of the larger and presumably stronger male ultramarathon runners (Table 1) at race distances beyond the 42.2-km marathon (Figs. 1 and 2) suggests that female ultramarathon runners are more resistant to fatigue than male ultramarathon runners matched for 56 km race time, age, and training. A greater endurance in female than in male runners has also been indicated from analyses of world running records ⁽¹⁰⁾ and of individual racing performances of ultramarathon runners ⁽⁶⁾. Recently, Speechly et al. ⁽¹⁴⁾ showed that 10 women who ran standard (42.2 km) marathons at the same (3.2 ± 0.2 m·s^-1) speeds as 10 men, were significantly faster than the men in a 90-km Comrades ultra-marathon race. Whereas the lighter (57 vs 72 kg) women completed the 90-km race at an average speed of 2.85 ± 0.20 m·s^-1, the men covered the same distance at a speed of only 2.59± 0.25 m·s^-1 (P < 0.05).

In our study the slightly superior performances of the lighter (56 vs 71 kg) female runners than of the male runners over the 90-km Comrades ultramarathon were unlikely to have been a result of the course terrain. That year the Comrades ultramarathon was a downhill race from ≈600 m above sea level to the coast and the course was no more hilly than the earlier 56-km Two Oceans ultramarathon from which the subjects were matched. Both races contained hills of up to 200 m in height.

It is also doubtful whether the greater speeds of the male runners than of the female runners over 5-42.2 km might have been a result of the men previously specializing in shorter distance races before competing in ultraendurance events. Based on the 15-km race time distributions of the 3126 male and 1314 female runners in the 1984 Cincinnati Mini-Marathon⁽⁴⁾, the men in this study would probably have been less successful than the women over shorter running distances. Whereas the 61-62 min 15-km race times of the men would have placed them in the top 24% of the male competitors in the 1984 Cincinnati Mini-Marathon, the 65-67 min 15-km race times of the women would have placed them in the top 13% of the female competitors in that race.

In contrast, the observation that the male runners out-performed the female runners at shorter (≤ 42.2 km) race distances but not at longer (> 42.2 km) race distances (Figs. 1 and 2) may have been owing to a greater capacity of the women to oxidize fat than the^(9,15,16). Although not all studies have shown more fat oxidation in women than in men during prolonged exercise at given percentages of maximum work rate ^(5,11), an increased rate of fat oxidation would enhance endurance by sparing muscle glycogen ⁽¹⁾.

The greater endurance of female than of male runners might also result from their generally smaller sizes and lighter body masses (Table 1). It is a common observation that elite male distance runners are smaller than elite male middle distance runners ⁽³⁾, but why smallness is an asset in distance running remains to be established. In our previous comparison of elite, male distance and middle distance runners, we found that although the lighter (56 ± 5 kg vs 70 ± 6 kg) distance runners performed less absolute work at any given speed, their O₂ consumption per kg body mass was the same as the heavier middle distance runners at equivalent speeds ⁽³⁾. The faster running times of the distance runners than of the middle-distance runners over> 5 km were a result of their ability to sustain higher percentages of˙VO_2max⁽³⁾. Others have also noted that the greater relative exercise intensities of smaller runners progressively compensates for their lower ˙VO_2max values over increasing race distances irrespective of gender ^(2,14).

Alternatively, the muscles of women may have a fatigue resistance, which is unrelated to any metabolic differences, superior to those of men. When subjected to repeated maximal isometric contractions for even short periods of time, the muscles of women fatigue less rapidly than those of men^(7,12). However, whether the greater time to fatigue in women than in men is a result of differences in gender or strength remains to be determined. Previously, we have shown that less rapid rates of quadriceps isometric contraction fatigue in male distance runners than in male middle distance runners were a result of lower initial peak isometric contraction torques ⁽³⁾. Even after corrections for differences in estimated lean thigh volume, there was still a significant inverse correlation between initial peak torque and time to fatigue in both groups of runners, which was unrelated to any differences in muscle fiber-type composition.

Finally, psychological factors beyond the scope of this study might be operative. Van Arken ⁽¹⁷⁾ has postulated that women may be more resistant to pain than men.

Although we can only speculate as to why female ultramarathon runners are more resistant to fatigue than comparable male ultramarathon runners, these data show that the superior endurance capacity of smaller female runners compensates for their poorer performances in shorter (≤ 42 km) races. Unfortunately, differences in world record running performances of men and women are greater in 24-h races than in standard 42.2-km marathons (27% vs 12%) and probably reflect the lack of elite female runners competing in ultra-endurance races ⁽⁸⁾. If the world's best male and female runners could be attracted to participate in such events, there may be an ultralong distance over which the women would outperform the men.

Figure 1-Declines in running speeds over increasing race distances of men and women, matched for 56 km race times, age, and training. Results are the means and SEMs of the running speeds calculated from the subjects' recent(within 1 yr) best times for the indicated race distances. The number of paired female and male runners at race distances of 5, 10, 15, 21.1, 42.2, 56, and 90 km were 17, 28, 25, 28, 28, 28, and 10, respectively. Statistical significance was assessed from the data presented in: Figure 2 .

Figure 2-Differences in running speeds of matched men and women racing at increasing distances. Results are means ± SEM. The 95% confidence intervals of the differences between the men's and women's running speeds were significantly greater than zero over distances of 5 to 42.2 km: (Fig. 1) .