Effects of Heat Stress and Sex on Pacing in Marathon Runners : The Journal of Strength & Conditioning Research

Secondary Logo

Journal Logo

Original Research

Effects of Heat Stress and Sex on Pacing in Marathon Runners

Trubee, Nicholas W.1; Vanderburgh, Paul M.1; Diestelkamp, Wiebke S.1,2; Jackson, Kurt J.1

Author Information
Journal of Strength and Conditioning Research 28(6):p 1673-1678, June 2014. | DOI: 10.1519/JSC.0000000000000295
  • Free



The effect of ambient temperature on athletic performance has been extensively investigated (2–4,8,11,12,14,21,22). Some studies examined the impact of environmental heat on variability in running velocity (10–12) and the increase in internal body temperature (2,8,14) during competition. In general, as environmental and internal body temperatures increased, the ability of endurance runners to maintain a consistent running velocity was impaired, leading to a significant decrease in race velocity (10–12) or drop out (3,22). Research has suggested that a consistent running velocity throughout the race produced faster overall performances (1,12,17,19). This was found to be most effective during longer distances, such as the marathon, in both cooler and warmer temperatures (12).

Pacing strategies in general have been examined for both cycling and running events. For cycling, research findings suggest that adopting a slower road velocity during the early stages of the race led to faster finish times for the 20-km distance (20) and even the 2-km distance (13). In these cases, slower velocities were those when compared with the mean race velocity. Interestingly, for the 5-km run race, Gosztyla et al. (15) determined that a 6% faster first mile (1.6 km), when compared with mean run velocity for the entire race, led to faster overall run times than slower first miles. Ely et al. (12) researched pacing for 219 elite women runners over multiple years of different Asian marathons (42.2 km). Results indicated that race winners displayed a close to even pacing profile throughout the initial 40 km and that the 25th, 50th, and 100th place finishers showed a nonlinear decline in running velocity from their initial 5 km. They also reported that an increase in race temperature was accompanied by a decrease in running velocity for both faster and slower runners. Conversely, in races that were held in cooler weather conditions, faster runners were able to maintain a more even running velocity compared with their slower competitors.

Although elite men outperform elite women in distance runs, little is known about sex effects on pacing in such races for nonelite runners. March et al. (19) examined pacing in 319 nonelite marathoners who ran the same race under cool ambient conditions on a flat 1.6-km loop course. Pacing in this case was defined as the mean velocity of the last 9.7 km divided by the mean velocity of the first 32.5 km. Results suggested that nonelite women were better pacers than nonelite men, even after controlling for age, sex, and finish time.

The research literature has offered some insight as to mechanisms by which women might be better pacers than men though the key causal link has yet to be established. Speechly et al. (25) found, when grouping men and women based on equal finish times from a previous marathon, women were able to perform at a higher percentage of their V[Combining Dot Above]O2max than their men counterparts at marathon velocities. Loftin et al. (18) corroborated this trend for 20 middle-aged marathoners such that women ran at 76.3% of their V[Combining Dot Above]O2max compared with 67.7% for men on a 1-hour treadmill run at recent marathon pace. All V[Combining Dot Above]O2 values were measured with the use of a metabolic cart in the aforementioned investigations. Others (7,23,26) reported that women tended toward a lower respiratory exchange ratio (RER) than men during submaximal endurance exercise, suggesting a preference for oxidizing fat for energy whereas sparing glycogen. This, in turn, can delay “hitting the wall,” a phenomenon where glycogen stores in the body are depleted (5,9,23,26), thus contributing to the characteristic precipitous decrease in run velocity. Women also tend to have a larger surface area-to-mass ratio than men (8,16,24), allowing them to dissipate a larger percentage of heat produced because of running. To our knowledge, no empirical research has evaluated the sex effect on marathon pacing in cold vs. hot ambient conditions, when controlling for age and finish time. Therefore, the purpose of this investigation was to examine the influences of sex, age, and finish time on marathon pacing in nonelite marathon runners across 2 different dates, corresponding to 1 cool and 1 hot race temperature. We hypothesized that the sex advantage favoring nonelite women would be magnified in the hot race condition when compared with the cold. A secondary purpose of this investigation was to compare the pacing profiles of the elite men and women runners, in addition to the pacing profiles of the elite and nonelite runners. We hypothesized that there would be no difference in pacing between the elite men and women marathoners and that elite runners would be superior pacers compared with novice runners.


Experimental Approach to the Problem

All data for this study were obtained from the Bank of America Chicago Marathon Web site (http://www.chicagomarathon.com/cms400min/chicago_marathon/), which included subject age, sex, 5-km split time, and overall finish time. This marathon was chosen for its large sample size, flat course, and large mean ambient temperature difference between its 2007 and 2009 events: 26.67° C and 2.77° C for the 2007 and 2009 marathons, respectively (http://weather.org/weatherorg_records_and_averages.htm).


The sample included 11,581 and 20,540 runners (18 – 75 years of age) from the 2007 and 2009 marathons, respectively, not counting runners who ran both races, a necessary condition for multiple regression analysis. The descriptive data for these subjects are shown in Table 1. The University’s institutional review board granted approval for this study and approved the waiver of informed consent because these data are in the public domain. For a participant to be included in this study, all 5-km split times and the overall finish time must have been recorded, and finish times were less than 5 hours. This time cutoff was selected because a 5-hour marathon correlated to an average run velocity of 2.33 m·s−1 (approximately 11:30 min·mile−1 or 5.2 mph), a pace described by March et al. (19) to be a pace difficult to walk. This method ensured that all runners included in the study ran the majority of the race.

Table 1:
Runners descriptive statistics (mean ± SD).


The course consisted of 42.2 km (26.2 miles) with digital clocks set at every mile marker to facilitate accurate self-pacing as desired. At each 5-km checkpoint, each runner’s shoe-worn chips crossed a digital receiver, recording split times. The cooler mean ambient temperature (2.77° C) and flat landscape of the 2009 marathon was similar to the study by March et al. (19) in that the change in pacing because of hyperthermia was not likely. However, the warmer temperature in 2007 (26.67° C) would likely impair pacing and overall run performance (12,22).

Pacing was defined as the mean velocity of the last 12.2 km (7.6 miles) divided by the mean velocity of the first 30 km (18.6 miles). March et al. (19) used a similar pacing index where the mean velocity of the last 9.7 km was divided by the first 32.5 km. This method was used because glycogen depletion during the marathon often occurs at approximately 3 hours or 30 km for the nonelite marathoner, leading to a noticeable decline in running velocity (6,9,26). By calculating the change in running velocity over time, we were able to focus on the change in pacing during the latter stage of the race.

Statistical Analyses

Stepwise multiple linear regression was used on the 32,121 nonrepeat runners to evaluate the impact of age, sex, heat stress, and overall finish time on pacing. Key 2-way interactions involving heat stress were assessed to account for the effect that an independent variable on pacing may depend on the level of another independent variable. Independent samples t-tests were also conducted between the elite (top 25) and the nonelite (those outside of the top 25) men and women runners. No analyses were conducted on the repeat runners because the hot race occurred first and a significant learning effect on overall run velocity and pacing would have been likely.


In the multiple regression analysis, age, sex, finish time, and heat stress were each statistically significant (p < 0.01) determinants of pacing with all 4 in the model. Figures 1–3 depict the effects of age, sex, finish time, and heat stress on pacing (age and finish time shown by tertiles) by 5-km split times. Figure 1 shows the marked decrease in running velocity in men is greater than that of women. A larger decrease in pacing is indicated in Figure 2, as those runners with slower finish times (T2: finish time of 3:40:18–4:07:35; hot n = 1923 and cool n = 4,215; T3: finish time of ≥4:07:36; hot n = 9,134 and cool n = 10,195) seem to be less able to sustain a consistent running velocity in the latter part of the race compared with those with faster finish times (T1: finish time of ≤3:40:17; hot n = 1,708 and cool n = 5,898). Figure 3 suggests that the age effect on pacing is small but the difference in run velocity between younger and older runners is consistent throughout both races.

Figure 1:
Mean running velocity by 5 km and sex.
Figure 2:
Mean running velocity by 5 km and finish time tertile (T1, T2, and T3).
Figure 3:
Mean running velocity by 5 km and age tertile (T1, T2, and T3).

The equation developed from the multiple regression analysis can be beneficial when interpreting the magnitudes of effects. Using the same method as that of March et al. (19) and the resulting prediction equation with pacing as the dependent variable, we calculated each term (coefficient × independent variable difference) for the following independent variable differences: men vs. women, 25 vs. 55 years of age, 3 hours vs. 4.5 hours finish time, and hot vs. cool running temperature. As an example, because 25 vs. 55 years of age is an age difference of 30 years, we multiplied 30 by the age coefficient of 0.000223 to produce an effect of 0.0067, or 0.67% on pacing, independent of the effects of sex, heat stress, and finish time. Described differently, the effect of being 30 years older rendered a 0.67% increase in pacing. Similarly, the effect of being women and faster translated to a 5.09 and 6.40% increase in pacing, respectively. The greatest effect on pacing was heat stress, which caused an independent decrease of 9.18% in pacing by the runners.

Including all key heat stress interactions in the model, all 3 were significant (p ≤ 0.05): finish time × heat stress, age × heat stress, and sex × heat stress. The direction of the difference in pacing among all interactions was consistent and can be seen in Table 2. A finish time by heat stress interaction denotes that heat had an overall deleterious effect on pacing such that, as runners get slower, heat had a slightly greater effect on pacing. The interaction between age and heat stress indicates that the age difference on pacing increased from cool to warm temperatures. However, this difference was quite small. The third interaction, sex × heat stress, indicated that the women’s superiority in pacing over men increased from the cold to the hot racing conditions.

Table 2:
Two-way interactions involving heat stress (% difference).

The independent samples t-tests concerning pace index indicated statistical differences (p ≤ 0.05) between elite runners and nonelite runners for both men and women and can be seen in Table 3. However, comparisons between elite men and elite women displayed no difference in pacing (p > 0.05). Elite men and women tended to have enhanced pacing over their nonelite counterparts for both hot and cool racing conditions.

Table 3:
Independent samples t-tests between elite and nonelite runners.


The principle finding of this study is that nonelite women’s pacing superiority over nonelite men magnified from the cool to the warm race temperatures, when age and finish time are controlled for. In fact, women runners among all ages and finish times show a trend for better pacing than men in both the 2007 and 2009 marathon. This may be because of the finding that women have shown a tendency to spare more glycogen and oxidize fat for energy because of a lower RER at submaximal intensities (7,23,26,27).

The larger difference in pacing between novice men and women in warmer temperatures may also be because of the finding that women generally have a larger surface area-to-mass ratio than men (8,16,24). Because heat production via exercise is proportional to body mass and heat loss is a function of body surface area, women runners should be able to dissipate a higher percentage of excess heat generated owing to running. Men, then, would likely fatigue earlier because of the relatively greater thermoregulatory challenge (14) and display the markedly slower running velocities in late stages of the race.

As indicated in Tables 1 and 2, finish time and pace exhibited significantly worse scores from cold to hot race conditions, congruent with previous findings (2–4,8,14). Of note, however, is the fact that the overall effect of the heat in the 2007 race cannot be characterized because race officials called off the race for those runners who did not reach the halfway point by 3 hours and 35 minutes, thus leading to just over 10,000 runners not finishing the race. As a result, because the nonfinishers are the slower runners who would have been on the course longer as the ambient temperature continued to rise, the effects of heat stress on marathon performance, including pacing, are likely underestimated in this study. In fact, Roberts (22) concluded that the 2007 Chicago marathon should have been cancelled before the start of the race because of high temperatures on the morning of the race. Although heat seems to be the likely cause for the drop in performance, other variables such as running experience, training level, and weather acclimation may have an additive effect to this phenomenon as well. Based on the archival nature of the data, we could not determine if these variables played a role in the decreased pace indexes and increased finish times. On the contrary, the topography of the race course was deemed unfit as a potential variable possible of affecting pacing because of minimal elevation changes, particularly during the last 12.2 km. The remaining 12.2 km included 2 minor climbs with the larger of the 2 being 5.5 m of elevation change spanning one-half mile. The largest decent during the finishing 12.2 km was 4.3 m ranging nearly one-half mile as well. The final 4.8 km was the most constant in terms of elevation with ±1.2 m of grade change. These data were collected via the USA Track and Field course mapping Web site (http://www.usatf.org/routes/map/).

Comparisons of the 2009 cooler race data with those of March et al. (19) suggest some similar findings. The current study found that, when running in cooler temperatures, nonelite women had a 4.14% increase in pacing, whereas the aforementioned study found a 4.06% increase in pacing by being women. Being faster in cooler temperatures also translated to a 5.94% increase in pacing, whereas March et al. found that being faster lead to a 10.71% increase. The effect of age, however, differed in magnitude with March et al. reporting a 7.3% increase in pacing for being older and this study, indicating a smaller 0.91% increase for the same age range. Because the race for the study by March et al. was considered a qualifying race for the popular Boston marathon, one might speculate that its runners gave a more maximal effort, thus eliciting a greater chance of an age effect on pacing, if present. The Chicago marathon is much larger and may have many more recreational runners more interested in finishing than competing.

The decrease in running velocity in the heat is seen in Figures 1–3, negatively affecting age and sex similarly. This decline is seen as early as the 15-km mark and is more severe as the race progresses. There was, however, a trend for an increase in pace among all groups in both temperatures from the 40-km mark to the finish. This may have been caused by additional motivation to complete the task when runners are nearly finished with the race. A similar occurrence was studied by Ely et al. (12) where the term “end spurt” was used to explain the increase in pace in the latter stages of a race.

No statistical difference (p > 0.05) was found in pacing between elite men and women runners for both temperature conditions. In elite marathoners, pacing between men and women were virtually identical, with indexes of 0.91 and 0.90 in the heat and between 0.95 and 0.97 in cool race temperatures, respectively. However, when pacing was compared between elite runners and their nonelite counterparts, statistical differences were discovered (p ≤ 0.05) and can be seen in Table 3. The use of these comparisons was important when making conclusions concerning the pacing differences between men and women runners. When interpreting pacing superiority between sexes, differences can only be seen between the nonelite runners and should not be extrapolated to the elite runner.

In conclusion, the results from the current investigation match those of March et al. (19) in that nonelite women tend to be better marathon pacers than nonelite men when age and finish time are controlled for. The addition of heat stress in this study magnified this sex difference. Although this finding is associated with physiologic mechanisms involving sex, glycogen sparing, and thermoregulation, future studies should elucidate causal effects and the extent to which the psychosocial effects of sex may influence marathon pacing.

Practical Applications

The results of this investigation indicate that a consistent running velocity has been shown to produce faster performances by marathoners. The findings of this study also show that the detrimental effect of heat on pacing should be taken into consideration when preparing for a marathon. Nonelite and novice men and women runners may consider using the pacing strategies of the elite runners when planning a race strategy for hotter temperatures to enhance performance. In hotter temperatures, coaches of these runners should advise their athletes to implement a slightly slower initial velocity in an attempt to maintain or increase running velocity in the latter stages of the race. Running at a slightly slower initial velocity may be used as a technique to maintain internal body temperature and to spare glycogen in an effort to improve pacing and running velocity late in the race. Novice runners attempting to enhance marathon performance should use consistent pacing as this has produced faster running times at both the elite and nonelite marathon levels.


The current study was not supported by any funding sources. The authors have no conflicts of interest to report. The results of this study do not constitute endorsement of any product by the authors or the National Strength and Conditioning Association.


1. Abbiss CR, Laursen PB. Describing and understanding pacing strategies during athletic competition. Sports Med 38: 239–252, 2008.
2. Adams WC, Fox RH, Fry AJ, MacDonald IC. Thermoregulation during marathon running in cool, moderate, and hot environments. J Appl Physiol 38: 1030–1037, 1975.
3. Armstrong LE, Casa DJ, Millard-Stafford M, Moran DS, Payne SW, Roberts WO. American college of sports medicine position stand: Exertional heat illness during training and competition. Med Sci Sports Exerc 39: 556–572, 2007.
4. Armstrong LE, Epstein Y, Greenleaf JE, Haymes EM, Hubbard RW, Roberts WO, Thompson PD. American college of sports medicine position stand: heat and cold illnesses during distance running. Med Sci Sports Exerc 28: 39–60, 1996.
5. Buman MP, Brewer BW, Cornelious AE, Van Raalte JL, Petitpas AJ. Hitting the wall in the marathon: phenomenological characteristics and associations with expectancey, gender, and running history. Psychol Sport Exerc 9: 177–190, 2008.
6. Cade R, Packer D, Zauner C, Kaufmann D, Peterson J, Mars D, Privette M, Hommen N, Fregly MJ, Rogers J. Marathon running: physiological and chemical changes accompanying late-race functional deterioration. Eur J Appl Physiol 65: 485–491, 1992.
7. Carter SL, Rennie C, Tarnopolsky MA. Substrate utilization during endurance exercise in men and women after endurance training. Am J Physiol Endocrinol Metab 280: E898–E907, 2001.
8. Cheuvront SN, Haymes EM. Thermoregulation and marathon running. Sports Med 31: 743–762, 2001.
9. Coyle EF. Physiological regulation of marathon performance. Sports Med 37: 306–311, 2007.
10. Ely MR, Cheuvront SN, Montain SJ. Neither cloud cover nor low solar loads are associated with fast marathon performance. Med Sci Sports Exerc 39: 2029–2035, 2007.
11. Ely MR, Cheuvront SN, Roberts WO, Montain SJ. Impact of weather on marathon-running performance. Med Sci Sports Exerc 39: 487–493, 2007.
12. Ely MR, Martin DE, Cheuvront SN, Montain SJ. Effect of ambient temperature on marathon pacing is dependent on runner ability. Med Sci Sports Exerc 40: 1675–1680, 2008.
13. Foster C, Snyder AC, Thompson NN, Green MA, Foley M, Schrager M. Effect of pacing strategy on cycle time trial performance. Med Sci Sports Exerc 25: 383–388, 1993.
14. Gonzalez-Alonso J, Teller C, Anderson SL, Jensen FB, Hyldig T, Nielsen B. Influence of body temperature on the development of fatigue during prolonged exercise in the heat. J Appl Physiol 86: 1032–1039, 1999.
15. Gosztyla AE, Edwards DG, Quinn TJ, Kenefick RW. The impact of different pacing strategies on five-kilometer running time trial performances. J Strength Conditioning Res 20: 882–886, 2006.
16. Kaciuba-Uscilko H, Grucza R. Gender differences in thermoregulation. Curr Opin Clin Nutr Metab Care 4: 533–536, 2001.
17. Lambert MI, Dugas JP, Kirkman MC, Mokone GG, Waldeck MR. Changes in running speeds in a 100 km ultra-marathon race. J Sports Sci Med 3: 167–173, 2004.
18. Loftin M, Sothern M, Tuuri G, Tompkins C, Koss C, Bonis M. Gender comparison of physiologic and perceptual responses in recreation marathon runners. Int J Sports Physiol Perform 4: 307–316, 2009.
19. March D, Vanderburgh P, Titlebaum P, Hoops M. Age, sex, and finish time as determinants of pacing in the marathon. J Strength Conditioning Res 25: 386–391, 2011.
20. Mattern CO, Kenefick RW, Kertzer R, Quinn TJ. Impact of starting strategy on cycling performance. Int J Sports Med 22: 350–355, 2001.
21. McCann DJ, Adams WC. Wet bulb globe temperature index and performance in competitive distance runners. Med Sci Sports Exerc 29: 955–961, 1997.
22. Roberts WO. Determining a “do not start” temperature for the marathon on the basis of adverse outcomes. Med Sci Sports Exerc 42: 226–232, 2010.
23. Ruby BC, Robergs RA. Gender differences in substrate utilization during exercise. Sports Med 17: 393–410, 1994.
24. Shvartz E, Saar E, Benor D. Physique and heat tolerance in hot-dry and hot-humid environments. J Appl Physiol 34: 799–803, 1973.
25. Speechly DP, Taylor SR, Rogers GG. Differences in ultra–endurance exercise in performance–matched male and female runners. Med Sci Sports Exerc 28: 359–365, 1996.
26. Tarnopolsky MA. Gender differences in substrate metabolism during endurance exercise. Can J Appl Physiol 25: 312–327, 2000.
27. Venables MC, Achten J, Jeukendrup AE. Determinants of fat oxidation during exercise in healthy men and women: a cross sectional study. J Appl Physiol 98: 160–167, 2005.

ambient temperature; velocity; split time; pace index

Copyright © 2014 by the National Strength & Conditioning Association.