Cold-water immersion is a very effective way with which to extract heat. However, some have urged caution concerning its implementation in hyperthermic people. Recently, Casa et al. (8) presented a rebuttal to some arguments raised against this treatment for exertional heatstroke. One emphasis was to contest the possibility that cold-induced thermogenesis elevates heat production and storage, thereby delaying cooling. Certainly, shivering will occur, albeit with a reduced sensitivity in preheated people (4), so its contribution to heat production is not very powerful. Furthermore, such low-intensity shivering will accelerate heat loss due to its effect on the insulating boundary layer (12,30). If shivering is only moderate, as is the case when hyperthermic individuals are immersed in cold water, the impaired formation or subsequent disturbance of an established boundary layer can increase heat loss beyond the heat production associated with shivering. The second major emphasis of Casa et al. (8) was to challenge the possibility that an acute cold-induced cutaneous vasoconstriction might delay heat loss or even transiently elevate heat storage, as evident from the paradoxical and often protracted elevation in core temperature frequently observed during the cold immersion of thermoneutral individuals. The authors focused more on the latter aspect of this argument. However, it is believed that the more significant physiological aspect may be the dramatic decrease in heat loss associated with the powerful suppression of peripheral blood flow and the consequential reduction in convective heat delivery from the core to theperiphery. Indeed, this is a focus of the current article, inwhich experimental evidence will be presented to show that the immersion of hyperthermic individuals (rectal temperature ∼40°C) in temperate water will elicit very rapid cooling.
From a thermodynamics perspective, the greater the temperature difference between an object and its surrounding environment, the faster its temperature will change. Thus, when a heated inanimate object is immersed in cold water, its rate of heat loss will be proportional to the thermal gradient established between the surface of the object and the layer of water adjacent to that surface. When exposed to different ambient media, heat loss will also be a function of the physical properties of each medium. For instance, differences in the thermal conductivity (26.2 vs 630.5 mW·m−1·K−1), specific heat capacity (1.007 vs 4.1885 J·g−1·K−1), and density (0.0012 vs 0.9922 g·cm−3) (18) between air and water, respectively, dictate the rate at which heat is lost when exposed to each medium. The product of specific heat capacity and density yields a volume-specific heat capacity, which quantifies the thermal energy necessary to raise the temperature of a given volume of water by 1 K. Across the physiologically relevant temperatures, the volume-specific heat capacity of water is >3400 times that of air. Without question, water does not just have a greater capacity to accept thermal energy, but this energy transfer will proceed at a much greater rate. Not surprisingly, immersion in ice-cold water (2°C) has been shown to be a most effective means of rapidly reducing rectal temperature in hyperthermic individuals (24,25).
It is not our intention to challenge first principles in biophysics but to evaluate how changes in physiological function (cold-induced vasoconstriction) can modify thermal energy transfer during water immersion because less powerful vasoconstriction during an immersion may actually enhance heat loss, as it can when auxiliary cooling is used to extract heat (9,29). Furthermore, sudden immersion in cold water precipitates potentially lethal cold-shock responses (31), including supraventricular ectopic arrhythmias (32) and reduced cerebral blood flow (19). These responses are associated with the sudden, powerful, and simultaneously activated discharge from the cold-sensitive cutaneous thermoreceptors and the inhibition of warm-sensitive receptors. The intensity of this feedback is a function of the rate and magnitude of the change in skin temperature (23), resulting in a dramatic elevation of sympathetic activity (17) and the cold-shock responses (31). Thus, although cold-water immersion is very effective for removing heat, in some circumstances, its use may result in undesirable side effects, and one such condition may relate to the immersion of hyperthermic individuals.
Therefore, we tested the hypothesis that a rapid and effective heat loss, as reflected within esophageal temperature, could still be achieved in hyperthermic individuals during a temperate-water immersion (26°C) while simultaneously avoiding the cold-shock responses. This more conservative approach would also satisfy pragmatic issues related to the availability of cold water in the field. If effective, such immersions would perhaps be more appropriate for asymptomatic hyperthermic individuals and could perhaps also be efficacious for those with exertional heat illness.
Eight healthy and physically active males participated in this study [23 yr (SD 4.99); 182.5 cm (SD 7.72); 77.3 kg (SD 6.77); surface area 1.99 m2 (SD 0.010); surface area to mass ratio 0.025 m2·kg−1 (SD 0.002)]. Subjects were screened to eliminate those with a history of cardiovascular, respiratory, or thermoregulatory problems contraindicative of participation in this experiment, and all subjects provided written informed consent to procedures approved by the Human Research Ethics Committee (University of Wollongong).
Subjects acted as their own controls and completed three trials, each consisting of 5 min of preheating rest, 90 min of exercise with heating, 5 min of precooling preparation, and then a supine cooling phase (commencing at 100 min; Fig. 1). Subjects were first heated to an esophageal temperature of 39.5°C using the controlled hyperthermia technique (11,22). This was achieved using a combination of exercise and exogenous heat [climate chamber (36°C, 50% relative humidity) and water-perfusion garment (40°C)]. Subjects wore running shorts and shoes and performed semirecumbent cycling for 90 min. Exercise followed a four-phase pattern to achieve three target esophageal temperatures (38.5, 39.0, and 39.5°C). Within each phase, the work durations remained constant, but the work rate was varied to achieve and hold the target temperature. Thus, the targets dictated the work rate for each individual, resulting in the following average work rates: 30 min at 147.7 W (SD 23.9; phase 1), 30 min at 96.9 W (SD 14.7; phase 2), and 20 min 92.4 W (SD 4.2; phase 3). The final stage of exercise involved relatively low-intensity intermittent work (10 min averaging 67.8 W, SD 13.5) and rest and was designed solely to clamp esophageal temperatures at 39.5°C for 10 min. Iso-osmotic drinks (100 mL at chamber air temperature) were consumed at the end of the first three work phases. The work rates used in the first trial for each subject were replicated within subsequent trials. This protocol was designed to ensure uniform heat distribution throughout the body tissues before cooling. This was a critical experimental objective because comparisons among the cooling methods would be less valid if subjects commenced the cooling treatments with variations in thermal energy content.
After heating, subjects were transferred (wheelchair) to the immersion tank (adjacent to the chamber). The time between the termination of heating and the commencement of cooling was standardized across all trials [5 min: 3.5 min in chamber (removal of water-perfusion garment) and 1.5 min to transfer from chamber to supine cooling]. Cooling commenced at 100 min and was performed in a supine posture for all trials, with subjects placed on a wide-mesh litter (Fig. 2; feet ∼5° below the horizontal plane). A supine posture facilitated rapid immersion and replicated the posture of a person being treated for hyperthermia. Immersion was to chin depth, with the head raised sufficiently to enable total immersion without a breathing impediment. The three trials differed only in the form of postexercise cooling: 1) control (nonimmersion or air cooling), subjects lay supine (litter) in an air-conditioned laboratory (20-22°C); 2) cold-water immersion cooling (water temperature = 14°C); and 3) temperate-water immersion (water temperature = 26°C). The former water temperature was used by Proulx et al. (24,25), thereby providing a means for interlaboratory data comparison, whereas the latter is a reasonable approximation of the temperature of water that may be available in the field in some hot climatic regions. Proulx et al. (24,25) also used colder water temperatures (2 and 8°C), but these temperatures were not used because such water is unavailable in the field and because cooling data already existed for these temperatures (25). In each trial, cooling continued until an esophageal temperature of 37.5°C was achieved.
Testing for each subject was conducted at the same time of day, using hydrated subjects (2), and with the trial sequence balanced across subjects. Preexperimental urine-specific gravities were as follows: control, 1.018 (SD 0.009); cold, 1.013 (SD 0.006); and temperate, 1.021 (SD 0.006; clinical refractometer no. 140; Shibuya Optical Co. Ltd., Tokyo, Japan). Subjects were asked to refrain from strenuous exercise and consumption of alcohol and tobacco during the 12 h before each trial. Subjects were also instructed to drink 15 mL·kg−1 of additional water in the evening before testing and to eat an evening meal and breakfast high in carbohydrate and low in fat. An abstinence from caffeine for 2 h before testing was also required. On arrival at the laboratory, subjects provided a urine sample to check hydration state and were provided with supplementary water (10 mL·kg−1) if not adequately hydrated (urine specific gravity >1.020; S8 in control trial). Water was also consumed during the insertion of the esophageal probe (∼400 mL). Before leaving the laboratory, subjects were rehydrated, consuming an iso-osmotic drink equivalent to 150% of the body mass change (100% in the laboratory and 50% taken away).
Core temperature was measured continuously from the esophagus (inserted transnasally; Mekjavic and Rempel (20); Edale Instruments Ltd., Cambridge, UK), the auditory canal (insulated to minimize auditory canal gradient effects; Edale Instruments Ltd.), and the rectum (10 cm beyond anal sphincter; Edale Instruments Ltd.). Data were sampled at 15-s intervals using a portable data logger (1206 Series Squirrel; Grant Instruments Ltd., Cambridge, UK). Skin temperatures were measured (15-s intervals) using thermistors taped to the following eight skin sites (Type EU; Yellow Springs Instruments Co. Ltd., Yellow Springs, OH): forehead, right scapula, right chest, right upper arm, left forearm, left dorsal hand, right anterior thigh, and left posterior calf. Mean skin temperature was derived using standard surface area coefficients (after (15,16)). All thermistors were calibrated in a stirred water bath against a certified reference thermometer (Dobros total immersion; Dobbie Instruments, Sydney, Australia). Heart rate was monitored from ventricular depolarization throughout each trial (15-s intervals; Vantage NV Sports Tester; Polar Electro Oy, Kempele, Finland).
Design and analysis
This project was based upon a fully crossed, repeated-measures experimental design with subjects participating in all trials. Between-trial differences were analyzed using two-way, repeated-measures analyses of variance (with Tukey HSD post hoc procedure) and paired t-tests. Alpha was set at the 0.05 level for all statistical comparisons. When the sphericity assumption was not satisfied, nonparametric tests were conducted [Friedman chi-square for main effects (χ 2) and the Wilcoxon signed rank test for paired comparisons]. For multiple comparisons using the Wilcoxon signed rank test, an adjusted alpha level for each comparison was computed using the Boole-Bonferroni inequality adjustment. Data are presented as means ± SEM, unless otherwise stated (SD). Graphs are referenced to the commencement of the resting phase of the heat exposure (0 min).
The heating protocol was designed to both elevate and ensure relatively homogeneous body tissue temperatures. The extent that this was achieved is evident in Figure 3, from which it is clear that the esophageal temperature profiles, upon which the heating protocol was based, were faithfully reproduced across all trials. The esophageal temperature targets (38.5, 39.0, and 39.5°C) were successfully achieved, and the final esophageal temperatures did not differ significantly among conditions: control, 39.4°C (SD 0.08); cold immersion, 39.3°C (SD 0.05); and temperate immersion, 39.3°C (SD 0.09; P = 0.359). The similarity of these temperatures is very important because the rate of conductive, radiative, and convective heat transfers is a function of the magnitude of the body core to ambient thermal gradient. Auditory canal temperatures peaked at 39.9°C (control; SD 0.27), 39.7°C (cold; SD 0.25), and 39.7°C (temperate; SD 0.11; P = 0.184). The corresponding rectal temperatures were higher and slightly more variable: 40.4°C (SD 0.41), 40.1°C (SD 0.16), and 40.1°C (SD 0.67; P=0.060). The final mean skin temperatures averaged 38.1°C (SD 0.50; control), 37.7°C (SD 0.40; cold immersion), and 37.9°C (SD 0.40; temperate immersion), with between-trial differences also being nonsignificant (P = 0.217). Over the last 10 min of heating, heart rates averaged 146.2 beats·min−1 (SD 0.78; control), 152.6 beats·min−1 (SD 1.11; cold immersion), and 151.6 beats·min−1 (SD 2.35; temperate immersion). Although drinking was permitted, it was controlled and limited (Fig. 3). Thus, subjects progressively dehydrated to experience fluid deficits (corrected for fluid consumption) between 2.5% and 3%, but differences among conditions were not significant (P = 0.334). From these data, it may be concluded that all subjects were moderately (38.5-39.5°C) to profoundly hyperthermic (>39.5°C), although none displayed overt signs or symptoms of heat illness. It was also concluded that the controlled hyperthermia technique provided an effective means for ensuring that subjects had an equivalent thermal energy content across trials and before cooling commenced.
When analyzed over the entire cooling period, the mean skin temperatures averaged 33.8°C (SD 1.31; control), 21.2°C (SD 1.20; cold immersion), and 26.8°C (SD 1.33; temperate immersion). These averages all differed significantly from one another (P < 0.001). When the response curves were compared, significant time-by-treatment interactions were evident between the control and the cold trials (P = 0.002) and between the control and the temperate trials (P = 0.001) but not between the two immersion treatments (P = 0.199).
The time taken for esophageal temperature to reach 37.5°C from the start of cooling (100 min) was derived (Fig. 4), and these data are in Table 1 for each subject, rounded to the nearest 15 s. These data did not meet the sphericity assumption and were reanalyzed using the Friedman χ (χ = 14.25, P < 0.05). Each comparison was statistically significant at the adjusted alpha level (P < 0.017). Table 2 shows the esophageal temperature cooling rates for each subject. These rates also differed significantly between experimental conditions (χ = 13, P < 0.05), with each of the between-trial differences being significant at the adjusted alpha level (P < 0.017). Cooling rates for auditory canal temperatures averaged 0.10°C·min−1 (SD 0.01; control), 0.53°C·min−1 (SD 0.05; cold), and 0.31°C·min−1 (SD 0.01; temperate; P < 0.001), whereas those for rectal temperatures were 0.07°C·min−1 (SD 0.01; control), 0.18°C·min−1 (SD 0.04; cold), and 0.10°C·min−1 (SD0.02; temperate; P = 0.016). Comparisons with the nonimmersed (control) condition are largely academic because the laboratory was air-conditioned, and this state may not be relevant in the field. Nevertheless, cooling was still reasonably rapid albeit much more variable among subjects. Notwithstanding the statistical differences across all treatments, it is clear that whole-body cooling to an esophageal temperature of 37.5°C in temperate water took only marginally longer to be achieved, 2.91 versus 2.16 min (Table 1). Because one purpose of this experiment was to evaluate methods that may be used in the field, one must conclude that temperate water is more than adequate to rapidly cool hyperthermic individuals. Indeed, one cannot imagine that the time difference of 0.75 min (45 s) would have any meaningful physiological or clinical implications.
The current experiment has established, from the esophageal temperatures, that rapid and effective heat removal can be achieved during a temperate-water immersion (26°C) in moderately to profoundly hyperthermic but asymptomatic individuals. There is little doubt that colder water immersions should elicit slightly faster cooling, but what was in doubt was the physiological and clinical significance of this difference. Because cooling during a life-threatening heat illness aims to rapidly reduce central nervous system temperature, observations based upon esophageal temperature reductions are more pertinent to this objective than those based upon rectal temperature because the former provides a much closer approximation of the temperature of blood flowing to the brain (33). If one accepts this, then one may question the need to use water cooler than 14°C, or perhaps even 26°C, because the respective cooling times at these water temperatures were still <4 min and 6 min across every subject (Table 1). These observations, although apparently not being previously described within the literature, have a considerable practical significance, although some would suggest that they were predictable on a first-principle basis.
Elementary heat transfer equations inform us that the reduction in the skin-water thermal gradient and the corresponding lengthening of the transcutaneous conductive distance result in the predicted total heat loss in temperate water being only ∼15% of that predicted for 14°C water. However, these calculations assume a constant rate of convective heat delivery to the muscles and skin in both immersions. For muscle tissue, this assumption appears invalid, with muscle temperature being rapidly and powerfully influenced by water temperature in resting thermoneutral individuals (5,17), presumably due to differences in local blood flow. Indeed, sympathetic discharge to skeletal muscles is unaffected at immersion temperatures >21°C but is elevated at temperatures <15°C and is particularly powerful in water <10°C (17). For the current experiment, one could predict muscle temperatures to be ∼24°C (cold) and 30°C (temperate). From actual skin temperatures (cold, 21.2°C; temperate, 26.8°C), one can predict skin blood flow for each immersion using linear extrapolation, assuming maximal flow (7.5 L·min−1) occurs at ∼42°C (27), whereas minimal flow obtains at about 10°C (26). From these assumptions, one can approximate convective heat delivery (mass flow) to the skin to be about 115% greater during the temperate immersion. Thus, it is assumed that the rapid heat loss during that trial was due to the less powerful peripheral vasoconstrictor responses, with greater heat being transported to the muscles and skin. If this less constricted state can be maintained in warmer individuals and this has been established for blood flow to the extremities (6,10), then for a given ambient temperature, heat loss will be accelerated. Hence, a sustained physiological mechanism (peripheral blood flow) appears to have countered the impact of a smaller thermal gradient, resulting in physiologically and clinically insignificant differences in the rate of heat extraction in hyperthermic individuals between trials.
It is of note that the immersion time presently required to reach a core temperature of 37.5°C in 14°C water differed dramatically from that reported by Proulx et al. (25) for the same water temperature, 2.16 min (current study) versus 17.3 min. However, Proulx et al. (25) determined cooling times from the point at which rectal temperature reached 37.5°C. In the current study, these times were based upon esophageal temperature, which is a more relevant index of core temperature in circumstances where heat illness is a concern (33). Although it may be correct that most clinicians can only measure rectal temperature, rectal tissue survival should not take precedence over that of the central nervous system, regardless of the tools that one possesses. Thus, a more appropriate comparison between investigations is to evaluate cooling time differences based upon esophageal temperatures. Proulx et al. (24) only measured rectal temperatures, whereas Proulx et al. (25) measured three core indices but chose only to report cooling times based on rectal temperatures. In Figure 1 of that report, cooling data for each index are shown for one individual. Ifone digitizes these data, required cooling times of ∼3minare achieved at water temperatures of 2 and 14°C and ∼7 min for water at 8 and 20°C. Proulx et al. (25) did, however, report cooling rates for each of three core temperatures (Table 2). During immersion in 14°C water, esophageal temperature cooled at 0.77°C·min−1 (SD 0.25) and was slightly slower than the current cooling rate [0.86°C·min−1 (SD 0.17); Table 2], despite possible differences in the total thermal energy content of the subjects across studies due to differences in the heating protocol [90min (current study) vs 45.4 min] and body mass (9 kg heavier in this experiment). Indeed, none of the esophageal temperature cooling rates reported by Proulx et al. (25) differed significantly from one another across any of the water temperatures investigated-a fact that appears to have escaped the attention of the ice-cold-water enthusiasts.
This difference in methodological emphasis across studies is really quite important because if Proulx et al. (25) had reported cooling times based on esophageal temperatures, then others would perhaps have been less enthusiastic in the use of ice-cold water to cool hyperthermic individuals. That is, if one sees that a 17.3-min immersion at 14°C is necessary to cool the body, then one is absolutely encouraged to explore every opportunity to shorten this time, and this could most easily be achieved by using colder water. However, if one sees that an immersion of <4 min in water at 14°C is all that is necessary to drop esophageal temperature to 37.5°C or 1.6 min to achieve 38.5°C (Fig. 4B), then one's enthusiasm for more uncomfortable and even painful (17) water temperatures will perhaps be viewed with less urgency. Nevertheless, these esophageal versus rectal temperature cooling time differences do provide important information. First, rapid and effective central nervous system cooling will occur in water at both 14°C and 26°C. Second, cooling will not be uniform across body tissues, with less well-perfused tissue beds (e.g., the rectum) being cooled more slowly. It must therefore be assumed that the brain, although well perfused, may not necessarily cool within 2.16 min because blood leaving the heart will possibly gain thermal energy en route to the brain, and it will take some time for thermal equilibrium to be achieved between the blood and the brain. Thus, although esophageal temperature provides a better index of brain cooling than does rectal temperature, brain cooling rates remain unknown. However, it can be assumed, based on the evidence from esophageal temperatures, that 14°C water will provide a more than adequate condition for rapid brain cooling to occur.
Thus far, we have considered only asymptomatic, yet moderately to profoundly hyperthermic, individuals. Patients and animals suffering severe hyperthermia, heat exhaustion, and heatstroke will often experience a dramatic reduction in skin blood flow, relative to the asymptomatic hyperthermic state, to defend mean arterial pressure. Indeed, some may even suffer peripheral circulatory failure (13,21). Although this is widely accepted, such a failure is by no means a universal observation, with a significant number of such patients sustaining vasodilatation (1,14) albeit at a reduced level. Nevertheless, it may be argued that, in the presence of circulatory failure, it is critical to use the coldest water possible. This view is not contested, particularly when it has been established that peripheral circulation is no longer viable. However, if one accepts that cutaneous venules and arterioles are likely to be maximally constricted when normotensive normothermic individuals are immersedin water at 14°C or colder (3,7,28), then it is not unreasonable to assume that such subjects would possibly cool at a rate quite similar to that expected in patients with peripheral circulatory failure. Indeed, the nonsignificant differences in esophageal temperature cooling rates reported by Proulx et al. ((25); Table 2) across water temperatures from 2 to 20°C are consistent with this hypothesis. Thus, if one agrees with these assumptions and accepts the empirical evidence, then one should perhaps be less anxious over the need to use water cooler than 14°C because the required esophageal temperature cooling time at this water temperature was <4 min in each of the current subjects. These observations clearly show that immersed hyperthermic humans do not lose heat at rates equal to that predicted for inanimate objects. Although the mechanisms that explain this phenomenon remain unexplored, it is assumed, when immersed in less cold water, that a sustained blood flow to the less superficial tissues of hyperthermic individuals continues to support the convective delivery of heat to the periphery, which is then transferred to the skin surface via tissue conduction.
It is concluded from the esophageal temperature measures that for hyperthermic, but asymptomatic, individuals, temperate-water immersion will more than adequately facilitate rapid brain cooling due to the maintenance of a greater peripheral blood flow. Indeed, a case may be promoted that cold-water immersion should be avoided for such individuals because this is not only very unpleasant, but it may result in cardiovascular failure (cold shock) in some high-risk individuals. Furthermore, for heat-exhausted and heatstroke patients, it appears that water any cooler than 14°C may not be required. Thus, in a true emergency situation, one must immediately immerse the patient in the most readily available cool-temperate water and then seek cooler water. Data from the current experiment, and also from Proulx et al. (25), demonstrate that rapid central heat removal will be achieved via this initial immersion, and sufficient life-saving cooling may perhaps be achieved before one has time to organize immersion within colder water.
This project was funded by a grant from the Defence Science and Technology Organisation (Australia). The opinions expressed in this article are those of the authors and do not reflect the official policy or position of the Defence Science and Technology Organisation or the Australian Government. The results of the present study do not constitute an endorsement by the ACSM.
1. Al-Khawashi MI, Mustafa MKY, Khogali M, El-Sayed H. Clinical presentation of 172 heat stroke cases seen at Mina and Arafat. In: Khogali M, Hales JRS, editors. Heat Stroke and Temperature Regulation
. Sydney: Academic Press; 1983. p. 99-108.
2. Armstrong LE, Maresh CM, Castellani JW, et al. Urinary indices of hydration status. Int J Sport Nutr
3. Barcroft H, Edholm OG. The effect of temperature on blood flow and deep temperature in the human forearm. J Physiol (Lond)
4. Benzinger TH. Peripheral cold reception and central warm reception, sensory mechanisms of behavioral and autonomic thermostasis. In: Hardy JD, Gagge AP, Stolwijk JAJ, editors. Physiological and Behavioral Temperature Regulation
. Springfield (IL): CC. Thomas; 1970. p. 831-55.
5. Booth JD, Wilsmore BR, MacDonald AD, Zeyl A, Storlien LH, Taylor NAS. Intramuscular temperatures during exercise in the heat following pre-cooling and pre-heating. J Thermal Biol
6. Brajkovic D, Ducharme MB, Frim J. Influence of localized auxiliary heated on hand comfort during cold exposure. J Appl Physiol
7. Cannon P, Keatinge WR. The metabolic rate and heat loss of fat and thin men in heat balance in cold and warm water. J Physiol (Lond)
8. Casa DJ, McDermott BP, Lee EC, Yeargin ECL, Armstrong LE, Maresh CM. Cold water immersion: the gold standard for exertional heatstroke treatment. Exerc Sports Sci Rev
9. Cheuvront SN, Kolka MA, Cadarette BS, Montain SJ, Sawka MN. Efficacy of intermittent, regional microclimate cooling. JAppl Physiol
10. Ferris BG, Forster RE, Pillion EL, Christensen WR. Control of peripheral blood flow: responses in the human hand when extremities are warmed. Am J Physiol
11. Fox RH, Goldsmith R, Kidd DJ, Lewis HE. Acclimatization to heat in man by controlled elevation of body temperature. J Physiol (Lond)
12. Golden F, Tipton M. Essentials of Sea Survival
. Champaign (IL): Human Kinetics; 2002. 27 p.
13. Hales JRS. The redistribution of cardiac output in animals during heat stress. In: Tromp SW, editor. Progress in Animal Biometeorology. Volume 1, Part 1
. Amsterdam: Swetz Zeitlinger; 1976. p.285-94.
14. Hales JRS. Limitations to heat tolerance. In: Fregly MJ, Blatteis CM, editors. Environmental Physiology. Handbook of Physiology. Volume 1
. New York: Oxford University Press; 1996. p. 285-355.
15. Hardy JD, DuBois EF. The technic of measuring radiation and convection. J Nutr
16. ISO 9886. Evaluation of Thermal Strain by Physiological Measurements
. Geneva: International Standard Organisation; 1992. p. 9-11.
17. Kregel KC, Seals DR, Callister R. Sympathetic nervous system activity during skin cooling in humans: relationship to stimulus intensity and pain sensation. J Physiol (Lond)
18. Lide DR. CRC Handbook of Chemistry and Physics
. Section 6. New York: CRC Press; 1997. p. 1-3.
19. Mantoni T, Belhage B, Pedersen LM, Pott FC. Reduced cerebral perfusion on sudden immersion in ice water: a possible cause of drowning. Aviat Space Environ Med
20. Mekjavic IB, Rempel ME. Determination of esophageal probe insertion length based on standing and sitting height. J Appl Physiol
21. O'Donnell TF, Clowes GHA. The circulatory abnormalities of heatstroke. N Engl J Med
22. Patterson MJ, Stocks JM, Taylor NAS. Humid heat acclimation does not elicit a preferential sweat redistribution towards the limbs. Am J Physiol
23. Pierau F-K. Peripheral thermosensors. In: Fregly MJ, Blatteis CM, editors. Environmental Physiology. Handbook of Physiology. Volume 1
. New York: Oxford University Press; 1996. p. 85-104.
24. Proulx CI, Ducharme MB, Kenny GP. Effect of water temperature on cooling efficiency during hyperthermia in humans. J Appl Physiol
25. Proulx CI, Ducharme MB, Kenny GP. Safe cooling limits from exercise-induced hyperthermia. Eur J Appl Physiol
26. Roddie IC. Circulation to skin and adipose tissue. In Shepherd JT, Abboud FM, editors. The Cardiovascular System. Volume III. Peripheral Circulation and Organ Blood Flow Part 1
. Handbook of Physiology
. Bethesda (MD): American Physiological Society; 1983. p. 285-317.
27. Rowell LB. Human cardiovascular adjustments to exercise and thermal stress. Physiol Rev
28. Spealman GR. Effect of ambient air temperature and of hand temperature on blood flow in hands. Am J Physiol
29. Stephenson LA, Vernieuw CR, Leammukda W, Kolka MA. Skin temperature feedback optimizes microclimate cooling. Aviat Space Environ Med
30. Stocks JM, Taylor NAS, Tipton MJ, Greenleaf JE. Human physiological responses to cold exposure. Aviat Space Environ Med
31. Tipton MJ. The initial responses to cold-water immersion in man. Clin Sci
32. Tipton MJ, Kelleher PC, Golden FS. Supraventricular arrhythmias following breath-hold submersions in cold water. Undersea Hyperb Med
33. Whitby JD, Dunkin LJ. Cerebral, oesophageal and nasopharyngeal temperatures. Br J Anaesth