Heatstroke During Endurance Exercise: Is There Evidence for Excessive Endothermy?

Dale E. Rae; Gideon J. Knobel; Theresa Mann; Jeroen Swart; Ross Tucker; Timothy D. Noakes


Med Sci Sports Exerc. 2008;40(7):1193-1204. 

In This Article


We describe three deaths in cyclists competing in the same 109-km cycle race in South Africa. We also present a case of a runner who experienced heatstroke during a 56-km footrace but made a full recovery. Cases 1 and 2 had cycled for between 4.5 and 5 h in an average ambient temperature of 28.7°C and relative humidity of 62%, and both fulfilled all criteria for a diagnosis of heatstroke because they were comatose with core body temperatures of 42.0 and 41.2°C, respectively. Tragically, the diagnosis seems not to have been entertained at the time of collapse, so that neither was initially treated appropriately with immersion in ice-cold water. Perhaps, as a consequence, both were dead within 24 h of their collapse, despite intensive in-hospital medical care. The third cyclist experienced a cardiac condition that may have been related to a more normal core temperature of 39°C when first measured 2h after his collapse. The runner (case 4) had run slowly for almost 7 h in cooler conditions (average ambient temperature of 18.1°C with a relative humidity of 61.4%) before he collapsed, also comatose, with a core temperature of 41.8°C. Unlike cases 1 and 2, case 4 survived after receiving immediate cooling including a 50-min immersion of his torso in ice-cold water that lowered his temperature by only 1.7°C. His core temperature, however, only returned to normal after a further 9 h of continuous cooling. Because the only case to survive was also the only case to be cooled appropriately, it seems reasonable to surmise that delayed cooling contributed to the fatal outcomes of at least two and perhaps as many as five cyclists in a single race.

Four intriguing questions arise from these case reports. First, if heatstroke is due principally to the environmental conditions in which the exercise is undertaken, why do so few athletes develop the condition even when the conditions are hot and humid? In the warmer 2002 Argus Cycle Tour, only 5 (0.02%) of 28,753 cyclists were hospitalized for heatstroke when exposed to the same environmental conditions. In milder conditions, only 1 (0.02%) of 6526 runners in the 2006 Two Oceans Marathon developed heatstroke and another was treated for hyperthermia. This paradox has been frequently reported ( Table 5 ).[16,19,38,40,43,50] Second, can exercise alone cause metabolic heat to be produced so rapidly and for so long that body temperature rises enough to cause exertional heatstroke and death? All the subjects reported here were recreational athletes who progressed slowly in these races, finishing in the last 30% of their fields. In contrast, many thousands of other athletes running or cycling very much faster, and therefore having higher metabolic races and producing more heat, were unaffected by this condition. Although the slower participants were exposed to the hotter conditions that developed in the latter stages of the races, the faster athletes were required to lose more heat to offset their greater exertional heat production. Third, why do the rectal temperatures of patients with heatstroke remain elevated even after they terminate exercise? If the raised core temperature observed in athletes experiencing heatstroke is due solely to excessive rates of heat production by the exercising muscles, spontaneous cooling should occur once exercise terminates. Yet, the rectal temperatures of cases 1 and 2 were still profoundly elevated on hospital admission 1 to 2 h after they had stopped exercising, as is usual in cases of heatstroke that are not cooled adequately. Case 4 required 10 h of active cooling after his collapse before his body temperature reached normal levels. Fourth, all these athletes had exercised regularly without ever developing heatstroke. Thus, one or more initiating factors must have been present during these races to spark the development of heatstroke. Next, we attempt to address these questions.

Already in 1938, Nielsen[34] reported that when exercise is performed at a constant workload, rectal temperature rises to a plateau value, which is determined by the workload, and is largely independent of the environmental conditions. This increase in rectal temperature is part of the normal thermoregulatory response to exercise; it is not life threatening and, within this thermoregulatory range, may not affect performance. For example, Pugh et al.[39] reported that the rectal temperatures of 42-km marathon runners increased to 39.0 ± 1.03°C during the race. Race winners almost always have the highest temperature at the end of the race, sometimes exceeding 41.0°C[8,39,51] and are usually the most dehydrated.[13,30] More recently, Byrne et al.[9] have shown that postrace rectal temperatures in excess of 40.0°C do occur in asymptomatic 21-km runners. Therefore, in healthy individuals and in the absence of compromised thermoregulatory function, it seems that the thermoregulatory centers in the brain allow the metabolic demands of exercise to elevate the body temperature to a new set point that is maintained within a narrow range for the duration of the event, presumably as a result of the matching of the rates of heat production and heat dissipation.[46]

Excessive Exertion-related Heat Production. It is usually assumed that exertional heatstroke occurs when there is a mismatch between the elevated rate of heat production during exercise and the rate at which that excess heat can be dissipated into the environment.[42] Nevertheless, it is retrospectively possible to postulate whether the abnormally elevated rectal temperatures in cases 1, 2, and 4 were due to the normal rates of heat production incurred during their respective races. Because the environmental conditions during the 2002 Argus Cycle Tour and the 2006 Two Oceans Marathon and the sizes of the three athletes are known, it is possible to use standard heat balance equations[35] to determine the amount of work each would have had to sustain to increase their core body temperatures to 42.0, 41.2, and 41.8°C, respectively. On the basis of a modification of these same equations by Brotherhood,[7] we were also able to more accurately estimate the skin temperature component of these equations. The heat exchange equation is expressed as:

where H = metabolic heat production, C = convective heat loss/gain, R = radiative heat loss/gain, E = evaporative heat loss/gain, and S = body heat storage. We have previously used these equations to calculate the effect of body mass on performance during exercise in the heat.[15]

Each component of the equation is a value in watts (W) and can be calculated as follows:

where v = airflow over the skin ≈ exercise speed (m·s-1), m = mass (kg), 12 = convection coefficient for running (W·m-2·°C-1), Tsk = skin temperature (°C), Ta = ambient temperature (°C), 0.92 = heat loss adjustment for clothing, Ab = body surface area (m2), 7.6 = convection coefficient for cycling, 5.5 = radiation coefficient (W·m-2·°C-1), Tr = mean radiant temperature (°C), 133 = evaporation coefficient for running (W·kPa-1·m-2), Psk = skin water vapor pressure (kPa), Pa = ambient water vapor pressure (kPa), 117 = evaporation coefficient for cycling (W·kPa-1·m-2), ΔTb = change in mean body temperature (°C) calculated from Tb = 0.87Tre + 0.13Tsk, 3474 = the heat storage capacity of humans (J·kg-1·°C-1), and t = duration of exercise (s).[7,35] A clear limitation in the application of these equations to this field setting is that they do not account for hills climbed and descended, or pack cycling. These limitations are accepted, because these factors would both increase and decrease the estimated heat production and heat loss capacities of the cyclists and runner. Yet, it seems unlikely that the inclusion of all these modifying factors would substantially alter the calculations.

The maximum capacity for heat loss in our first case (case 1; Table 5 ) would have been 1628 W during the 2002 Argus Cycle Tour and his rate of heat production was estimated to be 1191 W. Because this rate of heat production was less than his potential rate of heat loss, these calculations suggest that he should not have developed heatstroke during the 4 h in which he cycled in the environmental conditions on that day. The same was true for our second case (case 2; Table 5 ). His predicted rate of heat production was 1233 W, which was lower than his potential to lose heat (1664 W). Similarly, the estimated total heat production of case 4 (case 3; Table 5 ) who completed the 56-km footrace in 6 h 52 min 10 s would have been 720 W, substantially less than his maximum capacity for heat loss in the prevailing conditions (1502 W). Thus, these equations show that, in theory, the two cyclists and the runner should not have developed exertional heatstroke at the exercise intensities they sustained and in the environmental conditions in which they exercised. Therefore, either some other (unrecognized) source of heat production must have been present or else each must have had an unexplained impairment in their ability to lose heat at the normal rates.

We next reviewed the literature describing cases of exercise-related heatstroke in a range of environmental conditions ( Table 5 ).[1,16,19,25,27,33,38,40,41,43,50] Fatal heatstroke in this series occurred in environmental temperatures ranging from 4.1 to 30.6°C (mean ± SD = 19.1 ± 7.1°C) and relative humidities ranging from 42% to 88% (mean ± SD = 65 ± 14%). The cases occurred at distances ranging from 2.0 to 42.1 km (mean ± SD = 17.3 ± 12.7 km) and all patients had rectal temperatures in excess of 40.0°C (mean ± SD = 41.3 ± 0.8°C) when they collapsed. We applied these same equations (2a-5a) to predict the rates of heat production and heat loss in these cases ( Table 5 ). Figure 4 shows the ratio of calculated heat production to maximal heat loss capacity in all 17 cases. A ratio above 1 indicates the potential for heatstroke to develop. In only one case (case 17) did the calculated rate of exertional heat production exceed the potential for heat loss. Our calculations suggest that case 17 would have developed exertional heatstroke (rectal temperature of 41.5°C) after covering only 8.4 km. Interestingly, case 17 exercised in the most severe environmental conditions (30.6°C, humidity 80%) of the cases reviewed. We therefore suggest that, of these 17 cases, only one would have been expected to develop exertional heatstroke on the basis of the usually described pathogenesis: specifically high rates of heat production sustained in unfavorable environmental conditions causing progressive heat accumulation. Rather, these data suggest that an alternative explanation may be necessary to understand the development of heatstroke in subjects exercising in moderate environmental conditions at relatively low intensities.

Ratio of Calculated Heat Production to Maximal Heat Loss Capacity in 17 Cases of Heatstroke. Most of the cases have a ratio of less than 1 (open triangles). This positions them below the heat storage threshold because their estimated rates of heat production are less than their calculated heat loss potentials. Only one case (solid triangle) substantially exceeded the heat storage threshold. This individual had run 10 km in the most extreme conditions (ambient temperature = 30.6°C, relative humidity = 80%).

A Failure of Heat-losing Mechanisms. Another possible explanation for the development of heatstroke in these three cases that we have presented may be an inability to dissipate adequately their relatively low rates of heat production through exercise. In warm environmental conditions, evaporation is the principal mechanism for heat loss in exercising humans as the potential to lose heat via convection and radiation becomes negligible when the gradients between the skin and the surrounding air are reduced while the transfer of heat to the body by radiation increases.[35] It is generally assumed that either a very humid environment or a failure of sweating must impair evaporative heat loss to that point at which progressive heat gain occurs, leading ultimately to exertional heatstroke. Given that the average relative humidity levels during the races were not excessively high (approximately 60%), it seems that a failure of sweating would have been the more probable explanation for the development of heatstroke. Indeed, the most common mechanism for heatstroke development in the British Army in Iraq in 1917 was the sudden failure of sweating in susceptible soldiers exposed to extreme dry heat (temperatures of 40.6-50.0°C; average relative humidity approximately 20%) for extended periods of weeks to months.[20] However, it is not known if this failure can occur acutely without continuous daily exposure to dry heat for prolonged periods. Case 4 was sweating profusely on admission to the medical facility at the race finish, indicating that a failure of sweating was an unlikely cause of heatstroke in his case. Because there is no record of the clinical state of cases 1 and 2 when first treated, we are unable to absolutely exclude this mechanism as a cause of heatstroke in these two cases.

It is possible to calculate the minimal sweat rates that would have been necessary to prevent abnormal heat accumulation and heatstroke in these athletes. The maximal sweat rate capacities for cases 1, 2, and 4 were determined to be 2150, 2200, and 1514 mL·h-1, respectively. For the intensities at which they exercised, sweat rates of 1690, 1747, and 542 mL·h-1, respectively, should have dissipated sufficient heat to maintain normal rectal temperatures in cases 1, 2, and 4, respectively (Fig. 5). Therefore, for these three cases, the environmental conditions were not severe enough to limit the capacity for evaporative heat loss leading to a progressive heat accumulation during exercise. Figure 5 also includes the same data for the other cases of heatstroke that we studied. Only for case 17 were the environmental conditions sufficiently taxing for the maximal evaporative capacity to be less than the required sweat rate for the individual.

Required and Maximum Evaporation Rates in 17 Cases of Exercise-related Heatstroke. The left-hand column of each pair shows the rate of evaporation (sweating) required to maintain core body temperature at 37.5°C given each cases' exercise intensity and prevailing environmental conditions. The maximum capacity for evaporative heat loss for each case is represented by the right-hand column of each pair. In only one case (asterisk) was the maximal capacity for evaporative heat loss insufficient to maintain normothermia.

Excessive Endogenous Heat Production. However, if the heat-losing mechanisms of the cyclists and runner were not impaired and the environment was not impairing heat loss, so that each was able to sustain these quite normal sweat rates during exercise, then some other factor, secondary to exercise, must have stimulated endogenous heat production to the extent that it overwhelmed the normal capacity of the heat-losing mechanisms of these three athletes. Several possibilities could explain the development of an explosive endogenous heat production. One possibility is that the athletes experienced a form of malignant hyperthermia.[2,3,16,17,22,23,26,31,45,49] Although genetic testing was not conducted on these cases to rule out the presence of the mutated R1RY gene,[31] none had a family history suggestive of this condition. However, autopsy evidence of muscle necrosis was present in case 1 (Fig. 2), compatible with the theory that abnormal muscle metabolism could have contributed to this effect.[17,31,49] Alternatively, excessive sympathetic activation,[10,22,48] perhaps in the presence of a metabolic myopathy such as a glycogen storage disease, a fatty acid oxidation defect, or a mitochondrial myopathy,[44] may have triggered the development of hyperthermia.

The evidence linking susceptibility to malignant hyperthermia and exercise-related heatstroke has been reviewed.[21,22,31] Included in the most recent reports[22,31] are cases of heatstroke in football players in one of whom genetic testing revealed the presence of an altered RYR1 gene sequence identified as a cause for malignant hyperthermia. These authors also refer to the paper by Wappler et al.,[49] which found that 11 of 12 patients who had previously experienced exercise-induced rhabdomyolysis had positive in vitro contracture test responses indicating susceptibility to malignant hyperthermia. Interestingly, only 3 of these 12 subjects had an altered RYR1 gene considered causative for malignant hyperthermia. Thus, susceptibility to malignant hyperthermia must be conferred by several additional genes other than the RYR1 gene. Hopkins et al.[24] have concluded that although there are clear differences between malignant hyperthermia and exertional heatstroke and between those who are susceptible to either condition, there is yet "a high incidence of abnormal in vitro contracture test responses in muscles from individuals with a history of exertional heatstroke, suggesting that one or more skeletal muscle abnormalities are responsible for a significant proportion of heatstroke cases" (p. 100). They further conclude that: "These studies suggest that heatstroke victims have an underlying skeletal muscle abnormality that is probably distinct from malignant hyperthermia but which involves a similar deregulation of control of myoplasmic calcium ion concentration, which leads to both the in vitro contracture in response to drugs and the clinical heatstroke syndrome in response to extreme exertion" (p. 97). More recently, Hopkins and Wappler[22] has concluded that "a minority of episodes of exertional heat illness occur in individuals with an underlying predisposing skeletal muscle defect. Individuals who experience exercise heat illness, especially repeated episodes, in a temperate climate and without other predisposing factors are most likely to have such a muscle defect" (p. 284).

A second possibility is that an exogenous factor such as the ingestion of a drug or supplement may have activated an excessive thermogenic response. Neurololeptics, sympathomimetics, and drugs with anticholinergic properties have been associated with the development of heatstroke in warm environmental conditions.[29] Case 1's wife reported that her husband was taking a supplement that contained approximately 0.3 g of ephedra and 0.1 g of caffeine per capsule and had ingested some on the morning of the race. There are case reports of individuals taking ephedra who subsequently developed heatstroke during exercise.[12,37] The family members of case 2 reported that he did not use any drugs or supplements before the race, and case 4 reported taking only a carbohydrate gel during his ultramarathon.

A final possibility is the presence of an infection before the race. Notably, the runner reported by Roberts[40] had been ill 1 wk before developing heatstroke during a 42.1-km marathon he completed in 3 h 15 min. A study of 58 patients hospitalized with classic heatstroke during a heat wave in Chicago revealed that 33 (57%) had evidence for a bacterial infection on admission. The authors suggest that infection may contribute to the development of heatstroke either by inducing dehydration or by an increased thermal load.[14] Alternatively, infection could activate the immunologic and coagulative disturbances that are now believed to play an essential role in heatstroke.[5]

Clinical Evidence For Abnormal Endogenous Thermogenesis. Unpublished data from this laboratory show that asymptomatic cyclists participating in the 109-km Argus Cycle Tour finish the race with an average rectal temperature of 38.8°C. Spontaneous cooling, presumably because of the reduced metabolic heat production on cessation of exercise, sees a rapid decrease in the rectal temperature of these cyclists by 0.5°C in the first 10 min and normothermia is usually achieved within 40 to 60 min. That all three cases presented here had raised rectal temperatures for hours after they stopped exercising strongly suggests that abnormal endogenous thermogenesis was continuing. For the sudden removal of the heat source provided by muscular work should have produced cooling even if there was some impairment of sweating. However, more convincing evidence is provided by case 4 who showed extreme resistance to cooling, requiring 10 h of active cooling to reestablish normothermia (Fig. 3). A strong body of evidence has established that immersion in an ice water bath should reduce core body temperature by 0.15°C·min-1,[11] although the range may be wider. This predicts that case 4 should have become normothermic within 30 min as occurred in the comparison subject with hyperthermia who was treated with the same ice water immersion on the same day (Fig. 3) and whose rate of cooling was indeed 0.12°C·min-1. In reality, case 4's rate of cooling in the first 55 min when he was in the ice water bath was only 0.03°C·min-1, that is, 20% of the expected rate. During the second and third hours, when his body was surrounded by ice packs, his rate of cooling slowed to 0.01°C·min-1. We suggest that all these findings strongly suggest that case 4 exhibited marked endogenous thermogenesis that lasted for up to 8 h after he completed the ultramarathon that he ran at a very slow pace. In addition, it seems unlikely that he would have been unable to produce a sweat rate of only 126 mL·h-1 that would have maintained his core temperature at 37°C in the mild conditions in which he exercised and at the slow pace at which he was running.

Recently, Broessner et al.[6] have reported the case of a 38-yr-old man who developed postexertional heatstroke, resistant to conventional cooling techniques. For the first 20 h after his collapse, his temperature remained between 39 and 40°C despite appropriate cooling. Within 7 h of the placement of a heat exchange catheter in his superior vena cava, his temperature stabilized at 37°C where it remained for 5 days while active cooling with the IV catheter continued. Three attempts to remove the catheter resulted in an immediate increase in the core temperature, which was reversed only when IV cooling was restarted. Only after 111.5 h of IV cooling was it possible to remove the heat exchange catheter without an immediate increase in core temperature. Because sweating is not required to maintain the core temperature at 37°C at rest, this profoundly abnormal response can only be due to excessive endogenous heat production, similar to that which we believe was present in our fourth case (Fig. 4).

Support for Immediate Cooling of Heatstroke Patients. A crucial difference between cases 1 and 2 and case 4 was that only in case 4 was the diagnosis of heatstroke made at the point of collapse with the immediate initiation of active cooling. In contrast, there is no evidence that cases 1 and 2 were cooled before admission to the hospital, some 1 and 2 h after their respective collapses. It seems likely that the failure of early and adequate cooling in patients with heatstroke due to high rates of continuous endothermic heat production may contribute to a fatal outcome. Indeed, Casa et al.[11] provide good evidence that immediate cooling of patients is vital to their recovery. In particular, they show that cold water (14-20°C) and ice water (1-8°C) immersion is able to reduce rectal temperature by approximately 0.15-0.25°C·min-1.[11] It is tempting to speculate that the delayed cooling of cases 1 and 2 played a significant role in their subsequent demise. We also suspect that the other two fatal cases of heatstroke in this race may not have been properly cooled at the time of collapse, with grave implications for the outcome. However, there is also evidence that early and adequate cooling does not always ensure survival, indicating the complexity of this condition.[4]

Case 3. Case 3 seems to be somewhat different from cases 1 and 2, because he may have experienced a cardiac arrest during the race and then lived for a further 17 d before dying from septicemia. The autopsy failed to find a cause for the cardiac arrest experienced by the cyclist during the race. However, the fact that he was obese (BMI of 36), cycling his first Argus Cycle Tour, had mild to moderate atherosclerosis of major coronary vessels, and evidence of previous episodes of microinfarction places him in a higher-risk category during exercise. Therefore, this case adds to the pool of anecdotal evidence warning individuals who are overweight and undertrained that they may be at increased risk of cardiac damage during unaccustomed endurance exercise.[32]

Finally, to alter slightly the conclusion of Hopkins and Wappler,[22] we propose that individuals who develop heatstroke "in a temperate climate and without other predisposing factors" should be assumed to have a muscle defect causing excessive endothermy, mandating early, aggressive, and sustained cooling,[11,40] especially in those whose temperatures remain elevated despite apparently reasonable attempts at cooling (case 4; Fig. 3).[6] Hopkins and Wappler[22] further proposes that patients with heatstroke who are resistant to cooling or who show overt hyperkalemia or muscle rigidity should be treated with dantrolene at a dose of 2.5-3.0 mg·kg-1. The evidence that this treatment may enhance survival in patients with heatstroke due to acute Ecstasy (4,5-methylenedioxymethamphetamine) toxicity has been presented.[18]


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.