European Journal of Applied Physiology

, Volume 92, Issue 4, pp 524–532

Hyperthermia and maximal oxygen uptake in men and women

  • Sigurbjörn Á. Arngrímsson
  • Darby S. Petitt
  • Fabio Borrani
  • Kristie A. Skinner
  • Kirk J. Cureton
Original Article

DOI: 10.1007/s00421-004-1053-1

Cite this article as:
Arngrímsson, S.Á., Petitt, D.S., Borrani, F. et al. Eur J Appl Physiol (2004) 92: 524. doi:10.1007/s00421-004-1053-1


To compare the effect of hyperthermia on maximal oxygen uptake (O2max) in men and women,O2max was measured in 11 male and 11 female runners under seven conditions involving various ambient temperatures (Ta at 50% RH) and preheating designed to manipulate the esophageal (Tes) and mean skin \( (\ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}})\) temperatures atO2max. The conditions were: 25°C, no preheating (control); 25, 35, 40, and 45°C, with exercise-induced preheating by a 20-min walk at ~33% of controlO2max; 45°C, no preheating; and 45°C, with passive preheating during which Tes and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) were increased to the same degree as at the end of the 20-min walk at 45°C. Compared toO2max (l·min−1) in the control condition (4.52±0.46 in men, 3.01±0.45 in women),O2max in men and women was reduced with exercise-induced or passive preheating and increased Ta, ~4% at 35°C, ~9% at 40°C and ~18% at 45°C. Percentage reductions (7–36%) in physical performance (treadmill test time to exhaustion) were strongly related to reductions inO2max (r=0.82–0.84). The effects of hyperthermia onO2max and physical performance in men and women were almost identical. We conclude that men and women do not differ in their thermal responses to maximal exercise, or in the relationship of hyperthermia to reductions inO2max and physical performance at high temperature. Data are reported as mean (SD) unless otherwise stated.


Aerobic power Body temperature Gender Heat stress Physical performance 


Physical performance in prolonged strenuous exercise is reduced in the heat (Rowell et al. 1969; MacDougal et al. 1974; Sakate 1978; Nielsen et al. 1993). However, the literature is conflicting as to whether the reduction in performance is related to reduced maximal oxygen uptake (O2max) at high temperature. Some studies suggest that high ambient temperature (Ta) reducesO2max slightly (3–8%) (Taylor et al. 1955; Klausen et al. 1967; Rowell et al. 1969; Sakate 1978; Sawka et al. 1985), whereas other studies reported no effect (Williams et al. 1962; Rowell et al. 1965, 1966). In studies that found little or no change inO2max, exposure to the heat was for relatively short durations, and/or rectal temperature (Tre) often was not elevated to high levels (Rowell et al. 1965). Pirnay et al. (1970) were the first to observe a marked decrease (27%) inO2max when measured in the heat (46°C). Their protocol differed from those employed in the earlier studies in thatO2max was measured following 20 min of low-intensity exercise that elevated core temperature (Tc). Only a 7% reduction was observed without preheating in another set of subjects (Pirnay et al. 1970). Nybo et al. (2001) observed a 16% decrease inO2max following preheating with a water-perfused jacket and water-proof pants that elevated Tc and mean skin temperature \( (\ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}).\) However, the effect of different combinations of Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) onO2max has not been studied. In addition, there are no data on the effects of high Tc and\( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) onO2max in women. Whether the effect of high Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) onO2max and performance in women is qualitatively or quantitatively different from that in men is unknown.

The purpose of this study was to compare the effect of hyperthermia onO2max and physical performance in men and women. We hypothesized that there would be no gender difference in the effect of hyperthermia onO2max and physical performance.



Twenty-two healthy, male (n=11) and female (n=11) runners and triathletes were recruited as subjects. Their physical characteristics are presented in Table 1. The men and women had averaged 72±43 and 60±40 km·week−1 of running, respectively, for at least 6 weeks and were accustomed to exercising in a hot environment. The study was approved by the University’s Institutional Review Board with the stipulation that exercise would be terminated if Tes or Tre reached 41°C. Written informed consent was obtained before testing.
Table 1

Physical characteristics of the subjects. Data are mean (SD). V̇O2max Maximal oxygen uptake


Age (years)

Height (cm)

Mass (kg)

O2max (l·min−1)

O2max (ml·kg−1·min−1)

Males (n=11)

23.1 (4.7)

178.2 (4.4)

70.1 (8.8)

4.52 (0.46)

64.7 (5.3)

Females (n=11)

23.8 (3.8)

164.8 (5.7)

56.0 (4.9)

3.01 (0.45)

53.9 (7.5)

The sample size of 11 per group was sufficient to detect a 5% decrease inO2max (0.2 l·min−1) using a two-tailed t-test for dependent samples at alpha=0.05 and statistical power of 0.8, assuming individuals have a meanO2max of 55 ml·kg−1·min−1 with a SD of 7 ml·kg−1·min−1 and that the test-retest correlation forO2max is 0.95 (Lipsey 1990).


The study was conducted in an environmental chamber at 50% RH under the following seven conditions in which Ta, and pretest Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) were varied: (1) 25°C with no warm-up (25 N); (2) 25°C with a 20-min walking (active) warm-up at ~33% of controlO2max (25A); (3) 35°C with a 20-min walking warm-up at ~33% of controlO2max (35A); (4) 40°C with a 20-min walking warm-up at ~33% of controlO2max (40A); (5) 45°C with no warm-up (45 N); (6) 45°C with a 20-min walking warm-up at ~33% of controlO2max (45A); and (7) 45°C with passive heating prior to the test for a period of time sufficient to elevate the Tc to the same extent as after the 20-min walk in condition 6 (45P). Condition 1 was completed first and served as the control (baseline) condition. Then, conditions 2–6 were performed, with the order randomly assigned without replacement using a table of random numbers. Condition 7 was performed last. Using this approach, there was no effect of treatment order onO2max. Conditions 2–4 and 6 were designed to elevate Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) to different degrees using exercise-induced heating prior to theO2max test, and to alter the rate of rise during the test. Condition 5 was designed to investigate whether there was any reduction inO2max in the heat without preheating. Finally, condition 7 was designed to determine whether there was any effect of prior exercise onO2max, beyond the effect of heat. Each subject was tested at the same time of the day to minimize the effects of circadian rhythms on heart rate (HR) and Tc, and a minimum of 2 days intervened between tests of the same subject.

Menstrual status was assessed but it was not possible to control for phase of the menstrual cycle in scheduling women for testing. Six of the women were eumenorrheic (regular menstrual cycle for the last 2 years), two were oligomenorrheic (missed 4–6 menses in the last year), and three used oral contraceptives. However, there was no difference among the seven conditions for the number of women in the luteal and follicular phase. To further control for any effect menstrual status might have upon Tc during maximal exercise, resting Tc was held constant across all conditions during the statistical analysis of the data.


Subjects reported to the laboratory following a 3-h fast, but well hydrated. They were instructed not to consume alcohol or drugs 48 h prior to testing, not to consume caffeine 12 h prior to testing, and to drink water and other non-caffeinated beverages liberally. Urine specific gravity ranged from 1.0015–1.029 and there was no difference among conditions or between men and women. Pretest body mass also did not differ among conditions. On the morning of the test, subjects filled out a 24-h history questionnaire designed to determine adherence to pretest instructions. Then, subjects measured their nude body mass. Next, subjects inserted esophageal and rectal thermistors for measurement of Tc, thermistors for measurement of \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) were attached, and a strap containing the electrodes and transmitter for a HR monitor was placed around the chest. While being prepared for testing, subjects ingested water at room temperature to compensate for the estimated mass loss that would occur. The amount of water ingested was estimated from pilot studies of mass loss by male and female runners who performed the protocol prior to the study. Exercise was performed wearing shorts, shoes, socks (men) and a halter top (women).

Gas exchange, cardiorespiratory, temperature, and perceptual measures were obtained every 5 min during the preheating prior to the graded running test in 25A, 35A, 40A, 45A, and 45P (no gas exchange data collected during preheating in 45P). During the passive preheating in condition 7 (45P), body weight was monitored and fluid was ingested to compensate for mass lost. The subjects then completed a graded running test to exhaustion during which oxygen uptake (O2), oxygen (O2) pulse, respiratory exchange ratio (R), ratings of perceived exertion (RPE), HR, and treadmill test time to exhaustion (physical performance) were measured. During the test, a SensorMedics Vmax 29 metabolic cart was used to measure the metabolic variables over a sampling period of 30 s. Two consecutive 30-s values were averaged for all metabolic measures. Esophageal temperature (Tes), Tre, and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) were continuously monitored throughout the test and recorded every 2 min. Then, subjects dried off and measured their nude body mass to determine the amount of mass lost (dehydration).

Maximal oxygen uptake

To elicitO2max, subjects ran to exhaustion on a treadmill at a constant speed, with the grade increasing 2% every 2 min. A speed was chosen to exhaust subjects in 6–15 min. In the control test, after completion of the graded test, all subjects rested for 20 min and then ran to exhaustion at a grade 2% higher than the grade at the end of the graded test. The same protocol was used under all thermal conditions, except that the follow-up run to exhaustion was only done in the control condition, because of concern for possible heat injury.

Attainment ofO2max in the control condition (25 N) was determined by using a modification of the plateau criterion described by Taylor et al. (1955). The criterion for determining a plateau was an increase inO2 (ml·kg−1·min−1) between the last two stages of less than 50% of the expected increase, based on the American College of Sports Medicine metabolic equation (American College of Sports Medicine 1995);
$$ \dot{V}{\text{O}}_{{\text{2}}} {\text{=}}{\left( {{\text{0}}{\text{.2}}S} \right)}{\text{ + }}{\left( {{\text{0}}{\text{.9}}SG} \right)}{\text{ + 3}}{\text{.5}} $$
where S is speed in m·min−1 and G is the percentage grade expressed as a fraction. The criterion varied depending on treadmill speed and ranged from 1.3 ml·kg−1·min−1 (5.5 mph) to 2.2 ml·kg−1·min−1 (9 mph). Using this protocol, all subjects demonstrated plateau inO2, 11 during the graded exercise test and 11 during the subsequent run.

Because it was not possible to do follow-up tests in the heat due to the threat of heat injury,O2max (average of the two highest consecutive 30-s values) was assumed to be reached in the other conditions ifO2 was equal to theO2max in the control condition (within the margin of the plateau criterion, criterion 1) or if HR was within 5 bpm of that during the control condition (criterion 2). If neither criterion was met, the test was repeated on another day (five cases) during which one of the above criteria was satisfied. The number of subjects who achieved criterion 1 (or both criteria) at Ta of 25, 35, 40 and 45°C were 18, 9, 2, and 0, respectively, with the remaining subjects satisfying criterion 2. Physical performance was measured as the running time to volitional exhaustion during the graded exercise test.


Tes was measured using a thermistor (YSI model 4491E) inserted through the nasal cavity into the esophagus a distance equal to one-fourth of the standing height. \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) was calculated according to the formula of Burton (1935);
$$ \bar{T}_{{{\text{sk}}}} {\text{=0}}{\text{.5}}T_{{\text{1}}} {\text{ + 0}}{\text{.36}}T_{{\text{2}}} {\text{ + 0}}{\text{.14}}T_{{\text{3}}} $$
where T1, T2, and T3 are back, thigh, and forearm skin temperatures, respectively, measured with thermistors (YSI, model 409B). Mean body temperature \( (\ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{b}) \) was calculated from Tes and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) with the formula of Baum et al. (1976);
$$ \bar{T}_{{\text{b}}} =0.87T_{{{\text{es}}}} + 0.13\bar{T}_{{{\text{sk}}}} $$

This combination of Tes and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) was developed on highly trained runners and has been used for describing thermoregulatory responses during strenuous exercise (Schmidt and Bruck 1981; Olschewski and Bruck 1988; Lee and Haymes 1995). Temperature probes were connected to telethermometers (YSI, models 44TD or 4600). The accuracy of the thermistors was verified using water baths of various temperatures prior to use.

HR was measured using a Polar Vantage XL HR monitor (model 145900). O2 pulse was calculated by dividingO2 by HR. Ratings of perceived exertion were measured by the Borg 15-point category scale (Borg 1971). Body mass was measured to within 10 g with an electronic scale (A&D, Tokyo, Japan; model FW-150KA1).

Statistical analysis

Statistical analyses were done with SPSS 10 for Windows (SPSS, Chicago, Ill., USA). Data are reported as means (SD). A two-way (gender×condition) mixed-model repeated-measures ANOVA, or ANCOVA with resting temperature held constant (Tes and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{b}), \) was used to determine the significance of differences among conditions. ANCOVA was used to adjust for baseline temperature variation due to the menstrual cycle in women. Paired-sample t-tests (simple contrasts) were used to determine differences between conditions. A two-tailed α level of 0.05 was used for all significance tests. The significance level was adjusted using the modified Bonferonni correction for the family of contrasts performed.


Treatment effectiveness

The experimental treatments were successful in manipulating body temperatures prior to the graded exercise tests designed to measureO2max (Table 2). With minor exceptions, the pattern of differences among conditions was the same in men and women. Tes, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{b} \) at the start of the graded exercise tests (initial) ranged from 36.4 to 37.9, 32.1 to 37.6 and 35.9 to 37.8°C, respectively. Compared to the condition with no exercise-induced heating at the same Ta (25 N or 45 N), exercise-induced heating (25A and 45A) caused small significant increases in Tes-initial and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-initial}}}} \) (0.4–0.5°C), and no effect on \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-initial}}}} \) at 25°C, but caused larger significant increases in Tes-initial and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-initial}}}} \) (0.8–1.5°C) and a large significant increase in \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-initial}}}} \) (2.4–3.0°C) at 45°C. The effects of passive preheating on Tbs at the onset of the graded exercise test at 45°C were the same as exercise-induced heating by design. Increasing Ta during exercise-induced heating had no significant effect on Tes-initial until it reached 45°C, whereas \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-initial}}}}\) increased progressively as expected. Compared to 25 N, Tes-initial was not significantly elevated without exercise-induced or passive preheating at 45° (45 N) C, but \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-initial}}}} \) was significantly elevated by 2.3–2.7°C.
Table 2

Body temperatures at the start of the graded exercise test and at exhaustion. Data are mean (SD). Tes-initial Esophageal temperature at the start of the graded exercise test, Tes-max esophageal temperature at exhaustion, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-initial}}}} \) skin temperature at the start of the graded exercise test, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-max}}}} \) skin temperature at exhaustion, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-initial}}}} \) mean body temperature at the start of the graded exercise test, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-max}}}} \) mean body temperature at exhaustion, 25N 25°C without preheating, 25A 25°C with exercise-induced (active) preheating, 35A 35°C with exercise-induced preheating, 40A 40°C with exercise-induced preheating, 45N 45°C without preheating, 45A 45°C with exercise-induced preheating, 45P 45°C with passive preheating










Tes-initial (°C)

36.4 (0.4)*

36.9 (0.4)*, **

36.8 (0.2)*

37.2 (0.5)*

36.6 (0.4)*, **

37.9 (0.3)

37.8 (0.3)

Tes-max (°C)

38.6 (0.6)*

38.6 (0.6)*

38.7 (0.6)*

39.2 (0.4)*, **

39.2 (0.5)*, ***

39.7 (0.4)

39.6 (0.4)

\( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-initial}}}} \) (°C)

32.3 (0.6)*

32.3 (0.7)*

35.0 (0.4)*, **

36.0 (0.4)*, **

34.6 (0.6)*, **, ***

37.6 (0.3)

37.5 (0.4)

\( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-max}}}} \)(°C)

32.1 (0.8)*

31.9 (0.6)*

35.0 (0.5)*, **

36.6 (0.4)*, **

38.0 (0.4)*, **, ***

38.5 (0.4)

38.1 (0.4)*

\( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-initial}}}} \)(°C)

35.9 (0.3)*

36.3 (0.3)*, **

36.6 (0.2)*, **

37.0 (0.4)*, **

36.3 (0.3)*, **

37.8 (0.2)

37.8 (0.2)

\( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-max}}}} \)(°C)

37.8 (0.6)*

37.7 (0.5)*

38.2 (0.5)*, **

38.9 (0.3)*, **

39.0 (0.4)*, ***

39.5 (0.4)

39.4 (0.5)


Tes-initial (°C)

36.8 (0.4)*

37.3 (0.4)*, **

37.2 (0.4)*

37.3 (0.3)*

36.9 (0.3)*, **

37.7 (0.2)

37.6 (0.2)

Tes-max (°C)

38.8 (0.3)*

38.9 (0.3)*

39.0 (0.4)*

39.1 (0.4)*

39.1 (0.3)*, ***

39.4 (0.4)

39.3 (0.5)

\( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-initial}}}} \)(°C)

32.4 (0.4)*

32.1 (0.7)*

35.2 (0.4)*, **

36.3 (0.3)*, **

35.1 (0.6)*, ** ,***

37.5 (0.2)

37.6 (0.4)

\( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-max}}}} \)(°C)

31.4 (0.6)*

31.4 (0.9)*

34.9 (0.7)*, **

36.4 (0.4)*, **

37.9 (0.2)*, ** , ***

38.3 (0.4)

38.0 (0.4)*

\( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-initial}}}} \)(°C)

36.3 (0.4)*

36.6 (0.4)*, **

37.0 (0.4)*, **

37.2 (0.3)*

36.7 (0.4)*, ** , ***

37.7 (0.2)

37.6 (0.2)

\( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-max}}}} \)(°C)

37.8 (0.3)*

38.0 (0.3)*

38.5 (0.3)*, **

38.7 (0.3)*, **

38.9 (0.3)*, ***

39.3 (0.4)

39.2 (0.4)

*P<0.05 from 45A, **P<0.05 compared to the cell left of the value, ***P<0.05 45 N versus 25 N

Hyperthermia atO2max

Mean Tes at exhaustion (Tes-max) ranged from 38.6 to 39.7°C in men and from 38.8 to 39.4°C in women. It was significantly higher in 45A than in all conditions except 45P (Table 2). \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\)at exhaustion \( (\ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk - max}}}}) \) ranged from 31.9 to 38.5°C in men and from 31.4 to 38.3°C in women, and was significantly higher in 45A than in all the other conditions. Mean \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{\text{b}}} \) at exhaustion \( (\ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b - max}}}}) \) varied from 37.7 to 39.5°C in men and from 37.8 to 39.3°C in women, and was significantly higher in 45°C after exercise-induced or passive preheating (45A and 45P) than in the other conditions. There was a significant gender×condition interaction for Tes-max and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-max}}}}, \) but differences between men and women in the pattern of differences among conditions were too small to be meaningful.

Gas exchange, cardiorespiratory, perceptual and performance measures

Data on gas exchange, cardiorespiratory and perceptual measures assessed at exhaustion, and physical performance are contained in Table 3. There was a significant gender×condition interaction inO2max (l·min−1) and O2 pulse, but the pattern of differences among conditions was the same in men and women. Compared to control (25 N),O2max was decreased under all conditions except 25A. There was a significant, small reduction in 35A (3–5%); a modest reduction in 40A and 45 N (~9%); and a large reduction in 45A and 45P (17–19%). Similarly, O2 pulse at exhaustion was lower in the heat compared to the thermoneutral environment, with similar percentage reductions.
Table 3

Cardiorespiratory, perceptual, and performance variables at exhaustion. Data are mean (SD). V̇O2 Oxygen uptake, HR heart rate, respiratory exchange ratio, R ratings of perceived exertion, PTIME performance time (physical performance). For other abbreviations see Table 2










O2 (l·min−1)

4.52 (0.46)

4.52 (0.47)

4.30 (0.43)*

4.12 (0.44)*

4.13 (0.51)*, **

3.75 (0.48)*

3.70 (0.42)*

O2 pulse (ml·beat−1)

24.0 (3.7)

24.5 (3.7)

23.1 (3.4)*

21.9 (3.6)*

22.1 (4.1)*, **

19.6 (3.4)*

19.4 (3.3)*

HR (bpm)

189 (12)

186 (9)*

187 (10)*

190 (12)

189 (11)**

193 (12)*

192 (10)*


1.16 (0.03)

1.15 (0.03)

1.13 (0.03)*

1.12 (0.03)*

1.15 (0.03)**

1.09 (0.06)*

1.09 (0.04)*


19.3 (1.0)

18.9 (1.1)

18.8 (1.6)

19.3 (1.2)

19.0 (1.1)

19.2 (1.0)

18.9 (1.3)

Wt. loss (kg)

0.36 (0.33)

0.05 (0.11)*

0.28 (0.28)

0.38 (0.32)

0.31 (0.25)

0.45 (0.24)

0.78 (0.34)*, **

PTIME (min)

13.01 (1.18)

13.15 (1.41)

11.76 (1.22)*

11.33 (1.59)*

11.51 (1.15)*, **

8.77 (1.31)*

8.30 (1.25)*


O2 (l·min−1)

3.01 (0.45)

3.01 (0.44)

2.91 (0.47)*

2.75 (0.42)*

2.75 (0.44)*, **

2.52 (0.40)*

2.44 (0.40)*

O2 pulse (ml·beat−1)

15.9 (2.4)

15.9 (2.4)

15.4 (2.6)*

14.4 (2.2)*

14.5 (2.3)*, **

13.1 (2.0)*

12.8 (2.0)*

HR (bpm)

190 (8)

189 (7)

189 (7)

190 (8)

190 (7)

192 (8)*

190 (7)


1.18 (0.03)

1.16 (0.04)

1.14 (0.05)*

1.15 (0.04)*

1.14 (0.05)*

1.10 (0.04)*

1.06 (0.07)*


19.4 (0.7)

19.4 (0.5)

19.1 (0.8)

19.5 (0.5)

19.4 (0.7)

19.1 (0.8)

19.3 (0.6)

Wt. loss (kg)

0.32 (0.22)

0.14 (0.11)*

0.22 (0.20)

0.18 (0.18)

0.15 (0.13)*

0.40 (0.24)

0.48 (0.31)

PTIME (min)

13.00 (1.13)

12.64 (1.45)

12.10 (1.82)*

10.99 (1.48)*

11.38 (1.43)*, **

9.05 (1.79)*

8.32 (1.91)*

*P<0.05 from 25 N, **P<0.05 45A versus 45 N or 45P

Physiological and perceptual indicators of effort at exhaustion (HR, R, RPE) were similar across conditions (Table 3). A significant HR gender×condition interaction at exhaustion was found, but the pattern of differences among conditions was the same for men and women.

Loss of body mass during the test was small, less than 1.1% (Table 3). Body mass loss was not significantly greater in the heat compared to the control condition, except in 45P in men. Body mass losses were somewhat different in men and women, resulting in a significant gender×condition interaction. Men tended to have greater weight loss in 40A compared to 25 N, and in 45 N compared to 35A, whereas the pattern in women for these conditions was the opposite. The differences were small, however, and unlikely to have affected the results.

Significant decreases in physical performance on the graded exercise test were observed in the heat in men and women (Table 3). The percentage reductions in physical performance were larger than the percentage reductions inO2max. In 35A, the reduction in performance was 7–10%; in 40A, 13–15%; in 45 N, ~12%; in 45A, 30–33%; and in 45P ~36%. There were no significant interaction or main effects for gender for physical performance.

The relation between the reduction inO2max from control and Tes, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}},\) and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{\text{b}}} \) at exhaustion is shown in Fig. 1A–C. Similar relationships were found for men and women. When mean Tes-max was 38.6–39.0°C (25 N, 25A, 35A),O2max did not differ much (<5%) from the control value. When mean Tes-max was 39.1–39.2°C (40A, 45 N), moderate (~9%) reductions inO2max were observed. When mean Tes-max was 39.3–39.7°C (45A, 45P), large (17–19%) reductions inO2max were observed. Figure 1B demonstrates that in general the higher the mean \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-max}}}}, \) the larger the reduction inO2max, but high mean \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-max}}}} \) was not necessarily associated with a large (>10%) reduction inO2max (e.g., 45 N versus 45A and 45P). Only when both mean Tes-max and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-max}}}} \) were high, as reflected by a mean \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{b-max}}}} \) of 39.2–39.5°C (Fig. 1C), were marked reductions inO2max observed. The pattern of reductions in O2 pulse was the same, and there was a strong relation (r=0.98 for men, r=0.99 for women) between individual percentages changes from control inO2max and O2 pulse.
Fig. 1

Relation between changes in maximal oygen uptake (O2max) from 25°C with no warm-up (25 N) and A esophageal temperature (Tes), B mean skin temperature \( (\ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}})\)and C mean body temperature \( (\ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{\text{b}}}). \) Filled symbols are men, open symbols are women. Values are means ± SE

There was a strong linear relationship between the decrease inO2max and the decrease in performance time (physical performance) in the heat in men (r=0.84) and women (r=0.82) (Fig. 2). Reduced physical performance was not the result of elevatedO2 during the submaximal stages of the graded exercise test.O2 during the final minute of submaximal walking prior to the graded exercise test, and at individual minutes during the graded exercise test, was the same or slightly lower than that during 25A (data not shown).
Fig. 2

Relation between the reduction in performance time (physical performance) and the reduction inO2max from 25N in men (filled symbols; r=0.84, SEE=1.08 min) and women (open symbols; r=0.82, SEE=1.09 min). Solid line indicates the regression for men, dashed line indicates the regression line for women. ΔPTIME=reductions in performance time from 25 N


The primary finding of this study was thatO2max was reduced by the same relative extent in men and women under conditions in which Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) were progressively elevated to very high levels. The reduction inO2max was strongly related to reduced physical performance (treadmill test time). The effects of hyperthermia onO2max and physical performance were very similar in men and women, indicating that gender did not influence the effects of hyperthermia onO2max and performance in high-intensity exercise.

There is strong evidence thatO2max was attained in each of the conditions in the present study. All subjects met the traditional plateau criterion (control condition) or attained either the sameO2max as in the control condition or a HR within 5 bpm (experimental conditions) on all tests. There were no meaningful differences in means for HR, R and RPE at the point of exhaustion during the graded exercise tests. Mean peak HR was near 190 bpm, mean R was ~1.1 or higher, and mean RPE was near 19, indicating a similar, near-maximal effort was given under all conditions. The very similar means for peak HR strongly suggest that the cardiovascular capacity was reached under all conditions. Similar state of hydration before the tests and small change in body mass during the tests indicate that the reductions inO2max in the heat were not due to dehydration.

To our knowledge, this is the first study to describe the relation of changes inO2max to various combinations of Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) that included conditions in which both Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\)reached very high levels. Previous studies in which the effect of high ambient temperatures onO2max have been studied either did not report Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) atO2max (Taylor et al. 1955; Williams et al. 1962; Rowell et al. 1969; Pirnay et al. 1970; Sakate 1978; Sawka et al. 1985), did not investigate the effects of a range of hyperthermic conditions (Taylor et al. 1955; Williams et al. 1962; Pirnay et al. 1970; Sawka et al. 1985; Nybo et al. 2001), or did not use study conditions in which Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) increased to high levels (>39°C and 38°C, respectively) (Saltin et al. 1972). We found thatO2max was reduced in the heat in men and women in proportion to the increases in mean Tc, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}},\) and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{\text{b}}}. \) The largest decreases inO2max were achieved with very high Tes-max and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-max}}}}; \) high \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk-max}}}} \) alone, such as that which resulted from brief heat exposure without preheating (45 N) was insufficient to cause a large reduction inO2max.

This pattern of findings is consistent with available literature, although there are no previous data on the effects of hyperthermia onO2max in women, and only a few studies in men, in which the effects of high Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) have been reported. Most previous studies that have investigated the effect of heat stress onO2max have not used conditions that have elevated Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) to very high levels. Many studies investigated the effects of only relatively brief exposure to heat at rest prior to measuringO2max that did not elevate Tc (Taylor et al. 1955; Williams et al. 1962; Rowell et al. 1965, 1966; Klausen et al. 1967; Rowell et al. 1969; Pirnay et al. 1970; Saltin et al. 1972; Sen Gupta et al. 1977; Sakate 1978; Dimri et al. 1980; Sawka et al. 1985). Under these conditions, as shown in this study,O2max is either not reduced or is reduced only modestly. Even at very high levels of heat stress, such as the 45 N condition in the present study,O2max is reduced less than 10% if Tc is not substantially elevated. Elevation in Tc typically requires either prolonged exposure to high levels of heat stress (passive preheating), prolonged strenuous exercise (exercise-induced heating), or a combination of the two.

The only other studies that have found relatively large reductions inO2max in the heat are those that have employed exercise-induced heating prior to measuringO2max. Rowell et al. (1969) reported reductions inO2max of 10–14% in two individuals who were preheated such that they were unable to perform the intensity of exercise needed to elicit the controlO2max. After several minutes rest at 26°C, one of these individuals was able to controlO2max even though Tre>39.0°C. This observation is consistent with our conclusion that high Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) are needed to cause a large reduction inO2max at high temperature, because \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) would have been reduced during the follow-up test at 26°C, despite high Tre. They also noted that as long as the workload that normally elicited aO2max in cool conditions was achieved in a hot environment, decreases inO2max were minor. This finding is consistent with our data indicating that reducedO2max in the heat is closely linked to reduced treadmill test time (physical performance). Pirnay et al. (1970) observed a marked decrease (27%) inO2max when measured in the heat (46°C) following 20 min of low-intensity exercise that elevated Tc (preheating). Only a 7% reduction inO2max was observed at the same temperature without preheating in another set of subjects. Nybo et al. (2001) observed a 16% decrease inO2max following preheating, when Tes was 39°C and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) was ~37°C, and performance time during the constant-load test was reduced by half. They concluded thatO2max is only markedly compromised when both Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) are elevated. In our study, the 17–19% reductions inO2max are similar in absolute magnitude to the reductions found by Nybo et al. (2001), but smaller than reported by Pirnay et al. (1970).

Our results show that in the conditions in which exercise-induced warm-up was used, it was the rise in Tc in conjunction with increased \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) that was associated with the decrease inO2max, not some other consequence of prior exercise. The identicalO2max in 25 N and 25A indicates that, in absence of environmental heat stress and a large increase in Tc prior to the exercise test, the 20-min walk prior to the graded exercise test had no effect onO2max. In contrast, the significantly lowerO2max in 45A than in 45 N indicates that the 20-min walk prior to the graded exercise test, which was accompanied by increases in Tc, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{\text{b}}} \) at the start and end of the graded exercise test, did contribute to a reduction inO2max. The lack of a significant difference inO2max in 45A and 45P, in which the increases in Tc, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{\text{b}}} \) at the start and end of the graded exercise test were not different except for slightly higher mean \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) in 45A than 45P at the end of graded exercise test, strongly suggests that the significantly lowerO2max under these two conditions was related to the higher body temperatures and not to some other effect of prior exercise. Exposure to high levels of heat stress during the graded exercise test, but without preheating (45 N), was accompanied by more moderate increases in Tc, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{\text{b}}} \) and resulted in modest reductionsO2max compared to 25 N. Thus, the experimental manipulations strongly indicate that increases in Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) cause reduction inO2max.

The mechanism through whichO2max was reduced in the heat is uncertain. Maximal HR was not reduced and, in fact, was significantly higher by an average of 3–4 bpm (men) and 0–2 bpm (women) during 45A and 45P. Therefore, the decrease inO2max in the heat was due to lower arteriovenous oxygen difference [(a-v)O2 diff] or stroke volume, as reflected by the strong relation between reduction inO2max and reduction in oxygen pulse. Assuming that the (a-v)O2 diff was not reduced (Williams et al. 1962; Rowell et al. 1966), the lower O2 pulse accompanying reductions inO2max would reflect reduced stroke volume and cardiac output. Reduced stroke volume in the heat could be caused by (1) cutaneous vasodilation and increased venous volume, which reduces central blood volume, ventricular filling pressure and end-diastolic volume (Rowell et al. 1966); (2) a slightly greater degree of dehydration, which would lower blood volume and also contribute to lowered ventricular filling pressure and end-diastolic volume (Gonzalez-Alonso et al. 1995); and (3) increased sympathetic nervous system activity (Kim et al. 1979), increasing systemic vascular resistance and HR. The slightly higher HR in 45A and 45P would decrease ventricular filling time and can decrease end-diastolic volume (Fritzsche et al. 1999). Lower maximal cardiac output would be associated with lower maximal skeletal muscle blood flow and reduced oxygen delivery to the active muscles, limitingO2max. Our data showing large reductions inO2max when Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) are both high are consistent with the mechanism proposed by Rowell (1986) in which some combination of Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) “eventually overwhelms this adjustment [cutaneous vasoconstriction] as a consequence of too much cutaneous vasodilation” (p 384).

As expected, physical performance as reflected by treadmill time to exhaustion also was decreased in the heat in this study, consistent with the findings of others (Craig et al. 1954; Rowell et al. 1969; Saltin et al. 1972; Shvartz and Benor 1972; MacDougal et al. 1974; Sakate 1978; Nielsen et al. 1993; Febbraio et al. 1996; Gonzalez-Alonso et al. 1999; Nybo et al. 2001). The new data in the present study is the finding that the reductions in physical performance were strongly related to reductions inO2max. The strong relation between the reductions inO2max and physical performance is not surprising, although it differs from studies that have reported reductions in performance in the heat without reductions inO2max (Rowell et al. 1966, 1969; Saltin et al. 1972). Whether the relation means that a reduction inO2max caused the reduction in performance or vice versa cannot be determined from the data in this study. The data of Nybo et al. (2001), however, in which reducedO2max was observed with high Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) using a constant-rate test in which the power output was the same, suggest that reducedO2max with hyperthermia is not caused by a reduced work rate secondary to hyperthermia. Thus, reduced physical performance during the graded exercise test probably reflected reducedO2max caused by reduced maximal stroke volume, cardiac output and active skeletal muscle blood flow. The practical implications of our findings are that under conditions in which Tc and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) become elevated, reducedO2max contributes to a reduction in the maximal rate of exercise that can be sustained. This finding is important in prescribing exercise intensity for individuals who might exercise under these conditions (Arngrimsson et al. 2003).

To our knowledge, this is the first study to investigate the effect of high Ta onO2max in women. It also is the first to report Tes during high-intensity, exhaustive exercise at high temperature in women. Previous studies on thermoregulation in women have been limited to investigations of physiological responses to prolonged, low- and moderate-intensity, submaximal exercise (Stephenson and Kolka 1993). In general, these studies have found that for similarly trained men and women exercising at the same percentage ofO2max, thermoregulation is similar, although there are quantitative differences in sweating and effects of the menstrual cycle on the Tc thresholds for sweating and skin vasodilation and on Tc at rest and during submaximal exercise. The effects of the menstrual cycle on Tc during exhaustive, maximal exercise such as that in the present study have not been investigated, and there is no effect of the menstrual cycle onO2max (Casazza et al. 2002). We were unable to experimentally control for phase of the menstrual cycle but statistically controlled for variations in resting temperature across conditions. Since approximately the same number of women were in the follicular and luteal phases under each experimental condition, our data reflect an average across both phases of the cycle. We have extended previous data on thermoregulation in women by showing that thermal responses to shorter-duration, graded exercise to exhaustion in the heat following exercise-induced or passive pre-heating are similar in men and women. We also found that the percentage reductions inO2max with increasing Ta, and the relationships of Tes, \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{{\text{sk}}}}\) and \( \ifmmode\expandafter\bar\else\expandafter\=\fi{T}_{{\text{b}}} \) toO2max, were the same in men and women. Likewise, the relation between change inO2max and change in physical performance during exercise in the heat was the same in men and women. Our findings suggest that men and women are subject to the same limitations inO2max and physical performance during exercise in the heat.

We conclude that men and women do not differ in their thermal responses to maximal exercise, or in the relation of hyperthermia to the reductions inO2max and physical performance in the heat.


We thank the subjects for their enthusiasm and willingness to participate. We also thank Monika Strychova, Justin Shepard, Tom Rogozinski, and Derek Hales for invaluable help with the data collection.

Copyright information

© Springer-Verlag 2004

Authors and Affiliations

  • Sigurbjörn Á. Arngrímsson
    • 1
    • 2
  • Darby S. Petitt
    • 1
  • Fabio Borrani
    • 1
  • Kristie A. Skinner
    • 1
  • Kirk J. Cureton
    • 1
  1. 1.Department of Exercise ScienceUniversity of Georgia AthensUSA
  2. 2.Division of Sport and Physical EducationIceland University of Education LaugarvatnIceland

Personalised recommendations