Human temperature regulation under heat stress in health, disease, and injury

Abstract

The human body constantly exchanges heat with the environment. Temperature regulation is a homeostatic feedback control system that ensures deep body temperature is maintained within narrow limits despite wide variations in environmental conditions and activity-related elevations in metabolic heat production. Extensive research has been performed to study the physiological regulation of deep body temperature. This review focuses on healthy and disordered human temperature regulation during heat stress. Central to this discussion is the notion that various morphological features, intrinsic factors, diseases, and injuries independently and interactively influence deep body temperature during exercise and/or exposure to hot ambient temperatures. The first sections review fundamental aspects of the human heat stress response, including the biophysical principles governing heat balance and the autonomic control of heat loss thermoeffectors. Next, we discuss the effects of different intrinsic factors (morphology, heat adaptation, biological sex, and age), diseases (neurological, cardiovascular, metabolic, and genetic), and injuries (spinal cord injury, deep burns, and heat stroke), with emphasis on the mechanisms by which these factors enhance or disturb the regulation of deep body temperature during heat stress. We conclude with key unanswered questions in this field of research.

CLINICAL HIGHLIGHTS

During heat stress, human temperature regulation depends on the ability to adequately increase heat loss via sweating and cutaneous vasodilation. Physiological control of these thermoeffector responses varies due to several intrinsic factors (e.g., age, biological sex), diseases, and injuries that can alter thermal afferent signaling, central integration of thermal afferents, efferent signaling, and/or thermoregulatory end-organ function. Morphological features add to this variability through passive effects on heat loss (surface area) and heat storage (mass, tissue composition). Understanding the independent and interactive effects of physiological and morphological factors on human temperature regulation has important implications for identifying the risk of heat-related illness in different populations.

1. INTRODUCTION

1.1. Background

During exercise and with exposure to hot climatic conditions, there is tendency for the human body to store heat and for body temperatures to rise. Changes in deep body temperatures reflect the interplay between the passive and active thermoregulatory systems. The passive system comprises the body’s anatomical structures, the size and composition of which influence rates of heat flow between tissues and between the skin and environment. The active system consists of the regulatory pathways and effector organs that elevate skin blood flow and sweat rate, which potentiate heat loss and thereby regulate body temperature. Physiological regulation of cutaneous vasodilation and sweat production relies on appropriate communication of thermal afferents to the central nervous system (CNS), integration of thermal afferent information within the CNS, efferent signaling to cutaneous arterioles and eccrine sweat glands, and thermoeffector end-organ function. Structural or functional alterations to components of the active thermoregulatory system due to, for example, different physiological traits (e.g., age and biological sex), disease, or injury, can profoundly alter thermoeffector function, the potential for heat loss, and, ultimately, deep body temperature control. Importantly, physiological traits like the aforementioned examples of age and biological sex also affect the passive system due to morphological disparities between old and young individuals and between males and females, respectively.

The practical importance of understanding how differences in the active and passive thermoregulatory systems affect deep body temperature is borne of the recognition that heat stress can negatively impact health and performance (1–5). Beginning in the 19th century, much of the research into human thermoregulatory function was driven by military and industry interests in the effects of heat stress on the effectiveness and safety of service members and workers. As a result, much of the literature on this topic focused on thermoregulatory responses in young, healthy, able-bodied, males. However, the need to better understand the heat stress response among other segments of the population has grown. For example, since regular physical activity has become a tool for rehabilitation and disease prevention/management, the consequences of various injuries and diseases on temperature regulation during exercise need to be identified. Similarly, as females are now accepted into combat military units, and as the labor force becomes more diverse with regard to age, sex, and health, it is necessary to determine how these factors impact heat illness risk to ensure occupational safety in hot environments. Perhaps most importantly, global surface temperatures are rising and heat waves are becoming increasingly frequent, long-lasting, and intense, trends that are expected to continue in coming decades (6). A J- or U-shaped association between air temperature and mortality is commonly observed (7, 8), with elderly individuals and those with cardiometabolic diseases particularly vulnerable to higher temperatures (1, 9–14). Rising temperatures will also impact the safety of individuals across the life span during recreational exercise or competitive sport. Substantial effort has been devoted toward understanding thermoregulatory impairments associated with age and clinical disorders (e.g., Refs. 15–17), with the ultimate goal of identifying and protecting those most susceptible to heat-related health complications. Therefore, a summary of current knowledge pertaining to human temperature regulation under heat stress, with specific reference to disordered temperature regulation, is timely.

1.2. Scope

Our discussion begins with an overview of the biophysical aspects of human temperature regulation under heat stress. The effects of heat production and the thermolytic requirements for heat balance on human thermoregulatory responses must be understood and accounted for when investigating the influence of a particular physiological factor or interpreting reported findings. Next, we review the architecture and function of the regulatory pathways governing heat loss thermoeffectors. In the remaining sections, we describe how intrinsic morphological and biological factors, diseases, and injuries alter thermoregulatory function during heat stress. For each, we note the anatomical/functional origin of the thermoregulatory disturbance and examine the resultant effects on thermoeffector responses and deep body temperature control. Finally, we conclude with a summary of key points and a series of questions that should be addressed with future research. Topics that will not be covered include behavioral temperature regulation, nonthermal factors affecting thermoeffector responses, the influence of clothing, cooling strategies, diagnosis and treatment of heat illness, and physical performance. The reader is directed elsewhere for excellent reviews on these topics (3, 5, 18–21).

2. BIOPHYSICS OF HUMAN TEMPERATURE REGULATION DURING HEAT STRESS

Temperature is a fundamental physical property with profound effects on the structure and function of biological tissues, such as plasma membrane fluidity, rates of transmembrane transport and enzymatic catalysis, and the three-dimensional configuration of proteins. Humans are homeothermic, maintaining tissue temperatures at relatively constant values. For deep tissues of the head and thorax (i.e., the deep body), resting values of ∼36–38°C are typical. Maintaining deep body temperature requires constant metabolic activity as a heat source (endothermy), as well as control of metabolic heat flow from deep tissues to the environment. During exercise or ambient heat stress, elevations in deep body temperature to 38–40°C are normal and generally well tolerated, particularly in heat-acclimatized or aerobically trained individuals. Deep body temperature values exceeding 40°C increase the risk of heat injury and heat stroke (22, 23). Denaturation of proteins in mammalian cells occurs in the range of 40–45°C, causing inactivation of the protein, cell injury, and cell death (24, 25). Reaching deep body temperatures slightly above 40°C does not guarantee heat illness or injury (26). For instance, esophageal temperatures of 41.6–42.0°C have been induced in unacclimatized, sedated humans during thermal therapy (27) without any major hyperthermia-related sequela. Moreover, elite endurance athletes exhibit deep body temperatures between 41.1 and 41.9°C following competitive races without clinical symptoms of heat illness (28–30). In the case of athletes, repeated episodes of hyperthermia during training may induce thermotolerance by increasing the expression of heat-shock proteins, which preserve the configuration of proteins under heat stress (31).

In a physical sense, “heat stress” represents a heat load that tends to increase body heat storage and elevate deep body temperature. The magnitude of heat stress is dictated by six properties. Heat generated as a byproduct of metabolism, particularly during physical exertion, is typically the predominant source of heat. Four climatic factors, air temperature, mean radiant temperature, ambient vapor pressure, and air velocity, govern rates of heat loss between the skin and respiratory tract and the external environment. Finally, when clothing is worn, the rate of heat loss to the environment is reduced. Heat stress differs from “heat strain” in that the latter represents the physiological response to a heat load, such as elevations in deep body temperature or heart rate.

2.1. Heat Balance

In line with the first law of thermodynamics, energy exchange between the body and the external environment is summarized by the conceptual heat balance equation:

$$\text{S}=\text{M}-\text{Wk}-\text{K}-\text{R}-\text{C}-\text{E}-{\text{C}}_{\text{res}}-{\text{E}}_{\text{res}}\text{\hspace{0.17em}[W]}$$ (1)

The rate of heat storage (S) equals the difference between metabolic rate (M), the external work rate (Wk), the rates of dry heat transfer via conduction (K), radiation (R), convection (C), and evaporation (E) from the skin surface, and the rates of heat transfer via convection (C_res) and evaporation (E_res) from the respiratory tract. Since metabolic energy is transformed into work or heat only, M – Wk represents the rate of metabolic heat production. For details pertaining to the measurement and estimation of each component, the reader is referred to other publications (32–35). If the rates of metabolic heat production and total heat loss are equivalent, there is no net heat storage and deep body temperatures do not change. If the rate of metabolic heat production exceeds the rate of total heat loss (e.g., non-steady-state exercise, extremely hot and humid environmental conditions), the rate of heat storage is positive and deep body temperature rises. Ultimately, total heat storage, or the change in body heat content, is the product of time and the cumulative difference between rates of metabolic heat production and heat loss. In line with the International System of Units, rates of energy exchange are expressed in watts; however, heat balance parameters, particularly the metabolic rate, have been reported in other units, such as kilocalories per hour (1 kcal/h = 1.163 W) or kilojoules per minute (1 kJ/min = 16.67 W).

2.1.1. Metabolic heat production.

The metabolic rate represents the rate of free energy released from the oxidative and nonoxidative breakdown of stored macromolecules. Resting metabolic rate is conventionally taken as 1 MET, which is equivalent to the consumption of 3.5 mL of O₂ per kilogram of total body mass per minute and is also taken as 58.2 W/m² of energy expenditure for a standard adult male. At this rate, energy supply and tissue temperatures are adequate for normal physiological function (e.g., cell transport, enzymatic reactions, membrane fluidity). Several factors influence whole body metabolic rate at rest, including exposure to higher air temperatures (36, 37), the postprandial state (38, 39), stimulants (40, 41), hyperthyroidism (42), fever (43), posture (44), and pregnancy (45). Elevations in deep body temperature ≥1°C induced via passive heating also increase whole body metabolic rate by ∼11–23% (46–48). Part of this increase is attributed to the Q₁₀ effect, the ratio of the rate of a physiological process at a particular temperature to the rate at a temperature 10°C lower, indicating the sensitivity of enzymatic reaction rates to temperature (49). Greater energy demands associated with increased heart rate and pulmonary ventilation also contribute to the hyperthermia-induced elevation in the metabolic rate (46).

During physical activity, metabolic rate increases in proportion to exercise intensity. While some of this metabolic energy is transformed into mechanical energy for work (e.g., pedaling a bicycle against a resistance, running up a hill, lifting items), most is liberated as heat, with muscles contributing ∼90% of all metabolic heat production during intense exercise. Typical values of heat production associated with various tasks/exercises can be found in TABLE 1. For a given work rate or running speed, the rate of metabolic heat production varies with mechanical efficiency or movement economy. Gross mechanical efficiency, which is the percentage of metabolic energy transformed to external work, is ≤25% in cycling and rowing (56, 57) and ≤9.5% in freestyle swimming (55). During level-grade walking or running, there is no net displacement of body mass, so the external work performed is negligible (a small quantity of work is performed to overcome ground friction) (58). Movement economy during weight-bearing activities is expressed as the mass-specific oxygen uptake required per kilometer (ml O₂·kg⁻¹·min⁻¹) and ranges from ∼180 to 220 mL O₂·kg⁻¹·min⁻¹ (59). Better mechanical efficiency or movement economy translates into a lower rate of metabolic heat production for a given absolute work rate or speed and therefore a smaller stimulus for heat storage (60).

Table 1.

Rates of metabolic heat production at rest and during various physical activities

Activity	Heat Production, W
Rest
Sleeping	75
Sitting	85
Standing	100
Occupational tasks
Low	180
Moderate	295
High	415
Very high	520
Military tasks
Light	250
Moderate	425
Heavy	600
Walking	300
Cycling
Work rate = 75 W	300
Work rate = 250 W	1,000
Running
Speed = 8 km/h	750
Speed =16 km/h	1,500

Heat production values for rest (44, 50), desk work (51), occupational tasks (52, 53), military tasks (54), walking (50), cycling (assuming gross efficiency of 20%), and running (assuming a running economy of 220 mL·kg⁻¹·km⁻¹). Modified and reproduced from Ref. 86, with permission from Autonomic Neuroscience.

2.1.2. Heat loss from the skin.

Metabolic energy that is not transformed into mechanical work is released as heat, which must be dissipated to the external environment to prevent overheating. The rate of heat and mass loss to the surrounding air follows the general equation:

$$\mathrm{Rate of heat or mass flow}\mathrm{\alpha }\mathrm{gradient/resistance}·\mathrm{surface area }\left[\mathrm{W or g/min}\right]$$ (2)

The gradient for dry heat flow (conduction, radiation, convection) is a temperature difference between the skin and external environment in accordance with the second law of thermodynamics. If skin temperature exceeds ambient temperatures, the thermal gradient is positive and heat flows from the body to the environment. If ambient temperature exceeds skin temperature, the thermal gradient is negative and heat flows from the environment to the body (i.e., heat is gained). The gradient for evaporative mass transfer and heat loss is the skin-air vapor pressure differential. Unlike avenues of dry heat transfer, evaporation contributes only to heat loss. Resistance to dry and evaporative heat losses is imposed by the boundary air layer surrounding the body and clothing. Surface area refers to the skin area available for heat loss, which determines the absolute thermolytic capacity. Total body surface area may be assessed via three-dimensional scanning or estimated using various formulas from height and mass (61–64). However, for a given height and mass, these formulas may underestimate body surface area in individuals with high levels of adiposity. Due to the lower density of fat (0.9 kg/L) compared with lean tissue (e.g., muscle 1.1 kg/L), a higher body fat percentage for a given mass results in a greater body volume and thus surface area. Havenith (65) estimated that, for an individual of 75 kg and 10% body fat, replacing 10% of muscle mass with an equivalent amount of fat would lead to a 1.9% increase in body volume, with a resultant change in surface area of +3.5% or +0.5%, depending on whether fat was added to the limbs or torso. It should be noted that the surface area available for each avenue of heat loss may be somewhat lower than total body surface due to body posture, damage to the skin, or the denervation of particular dermatomes. For example, the surface area available for evaporative heat loss is reduced following a spinal cord injury as skin areas below the lesion are no longer able to secrete sweat (see sect. 4.1.2).

Conduction involves heat transfer through direct contact. The absolute rate of conductive heat transfer is directly related to the temperature gradient between the objects, the thermal conductivity of the materials, insulation between the objects (e.g., clothing), and the surface area of contact. Investigations of thermoregulatory responses to passive heating occasionally use conductive heating methods such as water-perfusion suits and hot-water immersion (e.g., hot tub or foot bath) (66). During exercise or nonencapsulated passive heating, conductive heat transfer is generally considered negligible since the skin area in contact with surfaces of different temperatures is very small.

Thermal radiation is heat exchanged via electromagnetic waves between bodies of unequal surface temperatures. The most important radiant heat source is the sun, although other potent sources of radiant heat can be found in industrial (e.g., boilers, engines), residential (e.g., furnaces), and clinical settings (e.g., radiant lamps). Sunlight reaching the Earth falls in the ultraviolet (∼5%), visible (∼40%), and near-infrared (∼40%) regions of the electromagnetic spectrum (67), whereas most terrestrial objects, including human skin, emit radiation at infrared wavelengths (35). The thermal gradient that drives radiant heat exchange is the difference between mean skin and mean radiant temperatures, the latter of which represents the area-weighted average surface temperature of all radiating objects surrounding the body. Other factors that influence the absolute rate of radiant heat transfer include skin color, with darker skin absorbing more energy from the visible part of the spectrum (35, 68); emissivity of the body surface (the ability for a surface to emit radiation compared with a black body), which is taken as 0.98 for bare skin; the effective radiant surface area (the surface area available for radiant heat exchange, which depends on the body’s position relative to other radiating bodies); and clothing insulation.

Convection is heat transfer by conduction between the skin and surrounding fluid (e.g., air) and movement of the fluid away from the skin. The gradient for convective heat transfer is the skin-air temperature gradient, and resistance is imposed by clothing and the boundary air layer. With natural convection, air warmed by the skin expands and diffuses upward due to buoyancy forces. If air movement over the skin is self-generated (e.g., running through stationary air) or driven by an external source (e.g., wind or fan), resistance imposed by the boundary air layer declines, enabling a higher rate of “forced” convective heat transfer that is directly related to the air velocity. Natural and forced mechanisms occur simultaneously, but forced convection predominates once air velocity exceeds 0.2 m/s (35).

Evaporation is the vaporization and subsequent diffusion of moisture from the skin surface into the external environment. The thermal energy needed to vaporize sweat, the latent heat of vaporization, is drawn from the skin. The rate of evaporative heat loss is directly related to the vapor pressure difference between the saturated skin surface and surrounding air, air velocity, and the level of skin wettedness; of note, resistance to evaporation is imposed by clothing and the boundary air layer. For a given sweat rate, higher skin temperature raises the vapor pressure at the skin surface. Greater air velocity over the skin surface increases evaporation in a mechanistically analogous fashion to convection (35) since both processes involve heat transfer into surrounding air and subsequent diffusion away from the body. Skin wettedness represents the fraction of the sweat-saturated skin surface available for evaporation (69), expressed as the ratio between the actual rate of evaporative heat loss and the maximum capacity for evaporative heat loss (E_max). For a given set of ambient conditions and mean skin temperature, E_max is related to body surface area and maximum skin wettedness (ω_max), the latter of which represents the highest proportion of total skin area that can be saturated with sweat. The value of ω_max reflects an individual’s level of heat adaptation, with ω_max values increasing from 0.7 in sedentary and non-heat-acclimated individuals to 0.85 after aerobic training and to 1.0 with heat acclimation (70–72).

Evaporative heat loss and skin wettedness can be expressed in terms of heat balance requirements. Through a rearrangement of the heat balance equation to isolate the evaporation term, the rate of evaporative heat loss required for heat balance (E_req) represents the difference between the rates of metabolic heat production, total dry heat loss, and total respiratory heat loss:

$${\text{E}}_{\text{req}}=\text{M}-\text{Wk}-\text{K}-\text{R}-\text{C}-{\text{C}}_{\text{res}}-{\text{E}}_{\text{res}}\text{[W]}$$ (3)

Greater heat production and reduced dry heat loss increase E_req. Since greater requirements for evaporative heat dissipation from the skin demand a higher rate of sweat production, whole body sweat rate is strongly associated with the absolute value of E_req (73, 74), as well as the external work rate or absolute heat production at a fixed rate of dry heat loss (75–79), assuming all secreted sweat evaporates. This concept is presented graphically in FIGURE 1. Absolute E_req and the corresponding whole body sweat rate are lowest when metabolic heat production and ambient temperatures are low (point A). At the same metabolic heat production, but a higher ambient temperature (point A to point B), E_req and whole body sweat rate rise to compensate for the reduction in dry heat loss (80). Similarly, at higher metabolic heat production but the same ambient temperature (point A to point C), E_req and whole body sweat rate rise to dissipate the additional metabolic heat (81). This explains why whole body sweat rate under fixed ambient conditions is strongly correlated with work rate (79, 82, 83) and the absolute rate of heat production (76, 84, 85). The conditions at points B and C produce the same E_req despite different rates of metabolic heat production and air temperature (i.e., dry heat loss). Although similar whole body sweat rates would be expected at points B and C, point C would be accompanied by a higher elevation in deep body temperature due to a greater rate of metabolic heat production (811).

FIGURE 1.

Whole-body sweat rate and the corresponding absolute evaporative requirement for heat balance (E_req) at different combinations of metabolic heat production (M − Wk) and air temperature (T_a), indicated by A, B, and C.

The skin wettedness required for heat balance (ω_req, calculated as E_req/E_max) is a theoretical term indicating the fraction of the skin surface that would need to be saturated with sweat to yield E_req (69). There are two important uses for ω_req. First, ω_req can define the physiological compensability of the prevailing heat stress conditions (FIGURE 2). Conditions in which ω_req was ≤1.0 (i.e., E_req ≤ E_max) are “compensable”; that is, heat gained from metabolism and the environment is ultimately offset through increased evaporative heat loss. Once evaporative heat loss equals E_req, the rate of heat storage is zero and deep body temperature remains relatively stable. In contrast, conditions producing ω_req >1.0 (i.e., E_req > E_max) are physiologically “uncompensable.” Constraints on evaporative heat loss imposed by high humidity, low air velocity, and clothing limit E_max to a level below E_req, resulting in persistent heat storage [equal to the difference between E_req and E_max (87, 88)] and an unrestrained increase in deep body temperature. Two points are worth noting about ω_req and physiological compensability: 1) since ω_max increases with the level of heat adaptation (72), the threshold for physiological compensability can be ω_req <1.0; and 2) ω_req <1.0 may be physiological uncompensable if E_max is limited by one’s ability to secrete sweat at a sufficiently high rate and achieve ω_max, owing to disordered sweating (e.g., grafted skin) or to extremely hot and windy environments (89–91).

FIGURE 2.

The relation between maximum evaporative heat loss (E_max), the evaporative requirement for heat balance (E_req), and the actual rate of evaporative heat loss (E_sk). The skin wettedness required for heat balance (ω_req) defines the compensability of heat stress. Under compensable heat stress (left), E_req is less than E_max, yielding a ω_req ≤1. E_sk rises until equaling E_req, at which point the rate of heat storage is zero. The area between the E_req and E_sk curves indicates heat storage. During uncompensable heat stress (right), E_req exceeds E_max, resulting in a ω_req >1. E_sk rises throughout heat stress up to the level of E_max, after which the rate of heat storage equals the E_req-E_sk (i.e., E_req-E_max) difference. Reproduced from Ref. 86, with permission from Autonomic Neuroscience.

Another use of ω_req is in describing sweat evaporative efficiency, which is defined as the fraction of secreted sweat that actually evaporates from the skin surface (sometimes expressed as a percentage). As ω_req increases, a greater fraction of the skin surface must become saturated in sweat to achieve E_req. If ω_req is low, all secreted sweat evaporates such that sweat evaporative efficiency is 100%. It is these conditions that produce the very strong association between E_req and whole body sweat rate. As ω_req increases above some threshold value, sweat rate exceeds that needed to achieve E_req, leading to some dripping sweat. This “inefficient” quantity of sweat does not contribute to evaporative cooling, only to body water loss. FIGURE 3 provides an example of the relation between skin wettedness and sweat evaporative efficiency during ambient heat stress. The threshold ω_req above which sweat evaporative efficiency declines has been shown to vary widely between studies, ranging from 0.24 to 0.74 (70, 71, 92). This variability likely reflects differences in clothing, activity, ambient conditions, and differences in local sweat rates between skin regions.

FIGURE 3.

Relation between skin wettedness and sweating efficiency under ambient heat stress at rest. Reproduced from Ref. 71, with permission from the American Physiological Society.

In conditions of dripping sweat, estimating sweat evaporative efficiency is necessary to more accurately determine the rate of evaporative heat loss. Furthermore, even if E_req is matched between groups or individuals, differences in whole body sweat rate may be observed due to differences in sweat evaporative efficiency, secondary to differences in E_max and thus ω_req.

2.1.3. Heat loss from the respiratory tract.

As inspired air travels down the pulmonary tree, heat exchange occurs between the inspired air and the surface of the respiratory tract by convection and evaporation. During expiration, some of this heat is recovered by the respiratory tract, but most is dissipated to the environment. The same physical principles that govern convective and evaporative heat transfer at the skin surface apply to the respiratory tract. Convective heat exchange in the lungs is directly related to the rate of pulmonary ventilation and the thermal gradient between the deep body and inspired air, the latter of which promotes body heat gain when ambient air temperature exceeds deep body temperature. Evaporative heat loss in the lungs is driven by the vapor pressure gradient between the respiratory tract surface and the inspired air, as well as the rate of pulmonary ventilation. The contribution of respiratory heat losses to whole body heat loss is far greater in cold and dry environments than warm/hot and humid environments due to the vastly greater thermal and vapor pressure gradients (93).

3. PHYSIOLOGY OF HUMAN TEMPERATURE REGULATION DURING HEAT STRESS

3.1. Architecture of the Human Thermoregulatory System

Physiological temperature regulation is a homeostatic feedback control system, in which a displacement of the regulated variable is detected by sensors that transmit information to a controller that generates a command signal to activate effectors thereby minimizing displacement of the regulated variable (94, 95). The regulated variable of the thermoregulatory system is deep body temperature, which constitutes the temperature of the internal organs, including the brain. In humans, deep body temperature is often measured within the rectum, intestines, and esophagus but it can also be measured within arterial/venous blood. There is no single measurement of deep body temperature that can be taken as the regulated variable, as each measurement will represent the temperature dynamics of the site of measurement under the conditions studied (96, 97). Furthermore, it is likely that thermal afferents are located within each of these tissues that, combined, contribute to overall thermoafferent flow during heat stress. Therefore, the regulated variable within the human thermoregulatory system is a spatial summation of temperature within the various internal organs that possess thermal afferents. When deep body temperature increases, thermoafferent flow is transmitted to the brain via the spinal cord where an efferent command is generated to activate the heat loss thermoeffectors of cutaneous vasodilation and sweating. The activation of heat loss thermoeffectors is not immediate, requiring a certain load error to the system in order to be initiated. Once activated, heat loss thermoeffector output increases proportionally with the increase in deep body temperature until steady-state or maximal values are attained. The activation of heat loss thermoeffectors serves to alter heat exchange between the body and the environment, thereby minimizing the increase in deep body temperature that is continuously fed back to fine-tune thermoeffector output. In addition to deep body temperature, skin temperature contributes to thermoafferent flow during heat stress. However, skin temperature is considered an auxiliary variable, in that it modifies the response time and the increase in thermoeffector output for a given thermoafferent flow due to changes in deep body temperature. During heat stress, warm skin temperatures will result in a faster activation and greater output of heat loss thermoeffectors, whereas cool skin temperatures will delay thermoeffector activation and output for a given increase in deep body temperature. It should be noted that, in humans, the modulating effect of skin temperature cannot solely be ascribed to a neural effect. For example, although an increase in skin temperature can elicit a reflex increase in sweating (e.g., Ref. 98), other studies have shown that eccrine sweat gland sensitivity is modulated by local skin temperature (99, 100). Furthermore, reduced skin blood flow contributes to the effect of local skin cooling on sweat output (101). Lastly, skeletal muscle may contain thermosensitive receptors that contribute to thermoafferent flow during heat stress, especially during exercise. As described in the next section, this possibility is supported by indirect evidence, and therefore, it is unknown whether skeletal muscle temperature should be considered a regulated or auxiliary variable within the human thermoregulatory system. In the following sections, the thermoeffector pathways that control heat loss thermoeffectors are presented. An emphasis is placed on studies performed in humans, although studies employing animal models are discussed when needed as these pathways have been almost exclusively studied in such models.

3.2. Neural Components of the Human Thermoregulatory System

3.2.1. Peripheral thermoreceptors and afferent pathways.

Peripheral warm thermosensitivity was identified relatively early by studies in animal models that examined the influence of skin temperature on thermoeffector output during an increase in brain temperature (102, 103). Direct heating of the brain within a thermoneutral environment increased thermoeffector output, and the level of thermoeffector output for a given brain temperature was shifted upwards when these experiments were performed in a warm environment. These studies led to early conclusions that thermoeffector output is likely controlled by an integration of both “internal” (deep body tissues, including the brain) and “external” (skin) temperatures (104). Nonetheless, it is interesting to note that skin temperature was considered, by some, as the primary thermoafferent input driving heat loss thermoeffectors. This contention motivated studies performed by Benzinger, published in 1959 and 1961 (105, 106), that demonstrated a primary role for deep body temperature (in this case tympanic) in driving heat loss thermoeffectors in humans, with skin temperature modulating effector output for a given deep body temperature. Benzinger (105, 106) further suggested that skin thermoreceptors served no role in temperature regulation and only played a role in thermal sensation for cold avoidance. This contention is likely explained by the fact that, at the time, the only direct evidence of cutaneous thermal afferents came from neural recordings of skin fibers that increased their firing frequency when the skin was cooled and decreased firing frequency during skin warming (104). Cutaneous receptors that increased their firing activity during skin warming were later identified in cats and primates (107) and eventually in humans (108–111). In parallel, studies employing animal models also established the presence of thermosensitive afferents within the spinal cord, intestines, stomach, and skeletal muscle (112). However, supporting evidence in humans remains scarce. Bain et al. (113) observed that whole body sweat production increased or decreased with the ingestion of hot (50°C) or cold (10°C and 0°C) water, respectively, relative to when body temperature (37°C) water was ingested during moderate intensity exercise in a 24°C environment. To determine if the modulation of sweat production by water temperature was driven by changes in oral and/or stomach temperature, Morris et al. (114) subsequently measured local sweat production during moderate intensity exercise in a 24°C environment during which cold (1.5°C) or hot (50°C) water was either swilled or administered via a nasogastric tube. No difference in sweat rate was observed between water temperature conditions when the water was swilled, whereas a transient increase and decrease in sweat rate was observed when hot or cold water was delivered via the nasogastric tube, respectively. Importantly, deep body (rectal and aural canal) and skin temperatures were unaffected by swilling or water delivery with the nasogastric tube. These findings suggest that the stomach contains warm- and cold-sensitive afferents that modulate thermoeffector output for a given thermoafferent flow from the skin and other deep body tissues. Skeletal muscle may also contain thermosensitive afferents in humans, as suggested by a study that employed sinusoidal variations in exercise intensity to determine the phase lag between changes in sweat rate, deep body temperature (esophageal), and active skeletal muscle temperature (115). Using this approach, a change in sweat rate followed a change in both deep body and active skeletal muscle temperatures. However, a shorter time delay was observed between changes in sweat rate and active skeletal muscle temperature, relative to the delay with deep body temperature. The shorter time delay was ascribed to thermosensitive afferents within skeletal muscle. However, it should be noted that the potential contribution of other exercise-related factors known to modulate sweating, such as central command and/or muscle metaboreceptors (116), was not ruled out in this study. Taken together, studies performed in animal models and humans support the existence of several thermosensitive areas within the body that contribute to thermoafferent flow during heat stress.

The thermosensitivity of various tissues is mediated by a specific class of warm-sensitive neurons distributed within the peripheral and central nervous systems. Peripherally, the axons of these neurons are unmyelinated C fibers (108–111) and the general properties of these fibers were largely described by studies performed between the 1950s and 1970s, including their static discharge at constant temperatures, a dynamic response to temperature change, and an insensitivity to mechanical stimuli (117). The molecular basis for temperature sensitivity of warm-sensitive neurons is largely attributed to the family of transient receptor potential (TRP) channels. When stimulated by temperature changes, these ion channels allow an influx of cations that depolarize the cell, resulting in action potential firing. A role for these channels in mediating thermoafferent feedback was first identified in vitro within cells expressing the vanilloid receptor subtype 1 (TRPV1) that respond to noxious (≈45°C) heat (118). Subsequently, several other TRP channels from the vanilloid (TRPV3, TRPV4), melastanin (TRPM2), and ankyrin (TRPA1) subfamilies were shown to respond to heat stimuli and/or to be implicated in warmth sensation (119, 120). Combined, these receptors cover the range of innocuous warmth (≈30–40°C) to noxious heat (>42°C) stimuli (95). In considering the channels that may drive autonomic heat loss thermoeffectors within the physiological range of (innocuous) warm temperatures, most of the evidence implicating TRP channels comes from animal models that evaluated thermal preference rather than autonomic effectors. From such studies, there is evidence both for and against each of these channels in mediating innocuous warmth transduction (119, 121). Notwithstanding these considerations, TRPV1 and TRPM2 are considered the most likely candidates to mediate thermoafferent flow during innocuous warmth stimulation (119). The case for TRPV1 is especially complex. Noteworthy evidence against a role for TRPV1 in mediating innocuous warmth transduction includes its high activation threshold in vitro (>42°C), its low expression in the brain and absent expression within the preoptic area (POA) of the hypothalamus (POAH), and the fact that TRPV1 knockout mice do not display altered body temperature or temperature regulation (121). Nonetheless, application of the TRPV1 agonist capsaicin on the skin of mice stimulates warm-sensitive neurons within the POA that drive autonomic heat-defense responses during heat stress (122). Furthermore, a relatively small proportion (≈10%) of trigeminal ganglion neurons expressing TRPV1 respond to innocuous warmth (35–42°C) and in vivo pharmacological inhibition of these neurons impairs warmth discrimination in mice (123). Evidence supporting a role for TRPM2 in mediating afferent innocuous warmth signaling comes from a study that identified neurons expressing this channel within the dorsal root ganglion (and superior cervical and pterygopalatine ganglia) of mice that are activated within the range of 34 to 42°C (124). In this same study, TRPM2 knockout mice also displayed impaired avoidance of a warm environment (38°C).

An afferent pathway through which peripheral warm receptors relay information to the CNS to stimulate autonomic heat-defense responses was identified in rats by Nakamura and Morrison (125). These authors had previously identified that the afferent pathway responsible for initiating autonomic cold-defense responses occurred through a spinoreticulohypothalamic tract, rather than the well-described somatosensory spinothalamocortical pathway (126), but it remained unclear if heat-defense responses were activated by a specific warm sensory pathway or though inhibition of the newly identified cold afferent pathway. Nakamura and Morrison (125) identified clusters of neurons activated by warmth in the dorsal part of the lateral parabrachial nucleus (LPBd) of the brain stem that provide warm sensory input most densely to the median preoptic nucleus (MnPO) within the POA of the hypothalamus. These neurons increase their firing activity beyond a skin temperature threshold of ∼34.7°C, and their firing activity increases with further increases in skin temperature up to ∼38°C at which point firing frequency remains at a steady, elevated level. The activation of these neurons was paralleled by a decrease in sural (paw) sympathetic nerve activity leading to withdrawal of vasoconstriction (and therefore passive vasodilation) and greater tail skin temperature. Combined with previous evidence demonstrating that dorsal spinal cord lamina I neurons receiving input from unmyelinated C fibers increase their firing frequency during skin warming in cats (127), and that dorsal horn neurons project to the LPB (128), the study by Nakamura and Morrison (125) established the existence of an afferent pathway from primary somatosensory neurons in the skin that, when activated by innocuous warmth, relay afferent information to secondary order neurons in the dorsal horn of the spinal cord which, in turn, project to third-order neurons located within the LPBd that project to quaternary neurons within the MnPO nucleus of the POA (FIGURE 4). In the absence of supporting evidence, we can only speculate that a similar pathway mediates thermoafferent input from the skin and/or other thermosensitive tissues during heat stress in humans.

FIGURE 4.

Neural pathways involved in temperature regulation during heat stress. Temperature regulation is mediated by primary somatosensory neurons located in the skin and viscera that transmit afferent information to the brain via the spinal cord. In rodents, dorsal horn neurons project to the dorsal part of the lateral parabrachial nucleus (LPBd) in the brainstem. These neurons, in turn, activate (+) neurons within the median preoptic nucleus (MnPO) of the hypothalamus. Increased activity of warm-sensitive neurons within the MnPO results in greater inhibitory input (−) to the paraventricular hypothalamus (PVH) and medial preoptic area (MPA) that provide tonic excitatory input to the dorsomedial hypothalamus (DMH). Innocuous warming has also been shown to activate ventral lateral preoptic area (vLPO) neurons resulting in greater inhibitory input to the DMH. The greater inhibitory input directed to the DMH results in less excitatory drive to the raphe pallidus area (RPa) of the brainstem that normally sends an excitatory drive to preganglionic neurons controlling cutaneous vasoconstriction (VC) and thermogenesis. The inhibition of this pathway results in a passive increase in skin blood flow (SkBF) and a decrease in heat production. In contrast, the neural pathways mediating heat loss thermoeffector responses in humans remain largely unknown. Brain imaging studies have confirmed that the preoptic area (POA) of the hypothalamus and a juxtafacial area of the brainstem are activated during heat stress and that their activity correlates with sweating. Importantly, temperature regulation during heat stress relies on the activation of specific heat loss thermoeffectors in humans, namely active cutaneous vasodilation (VD) and eccrine sweat production, rather than the withdrawal of cold-defense responses in rodents. It is therefore unclear how the neural pathways for heat-defense responses identified in rodents can be translated to humans. Image created with BioRender.com with permission.

3.2.2. Integration of warm thermoafferent feedback.

The brain was possibly the first thermosensitive organ identified, and this finding is often credited to the work performed by Barbour during the early 1900s (129). This finding led to a series of studies that eventually identified the POA of the hypothalamus as the most, and up until the 1960s the only, thermosensitive area within the body (112). Noteworthy are studies that quantified thermoeffector output for a given change in hypothalamic temperature through direct manipulation of POA temperature (102, 103). These studies laid the foundation for considering the POA as a central site of thermoregulatory control. It is now accepted that the POA receives input from numerous thermosensitive areas within the body to generate an efferent command signal that activates heat loss thermoeffectors. How this occurs in humans remains largely speculative and has been the subject of healthy scientific debates (130–133). Initial models applied control systems theory to explain how the brain may integrate thermoafferent input to generate appropriate effector responses. These models introduced temperature regulation as a combination of on/off, proportional, and rate control components (102). Other models integrated neurophysiological concepts to describe temperature regulation in terms of afferent signaling, integration, and effector responses (103, 107). Common to most models was the independent control of thermoregulatory responses during cold and heat exposures. The main matter of debate was the existence of a (hypothalamic) temperature set point that serves as a reference signal to the thermoregulatory system. Arguments that temperature regulation can function without a reference signal were provided as early as 1980 (134), but the existence of such a set point was debated until at least the mid to late 2000s (130–133). The current consensus is to present temperature regulation as a collection of independent effector loops with their own neural networks (95). The concept of independent thermoeffector loops is perhaps best exemplified by the separate neural networks controlling cold- and warm-defense responses. Although this concept was recognized in early models of temperature regulation (107, 135), elegant studies in animal models have provided direct evidence of its existence. In rats, afferent signals from both cutaneous cold and cutaneous warm receptors activate neurons within the LPB. However, skin cooling activates neurons specifically within the external lateral portion of the LPB whereas skin warming specifically activates neurons within the dorsal area of the LPB (LPBd) (125, 126). Nonetheless, there is cross talk between these networks as evidenced by the observation that activating LPBd neurons inhibits autonomic heat-defense responses during cold stress (125). A second argument supporting the existence of independent effector loops relates to observations that the temperature threshold to activate thermoeffectors can be altered differentially in response to a given stimulus. This contention relies primarily on animal models of anapyrexia in the context of shock, starvation, and injury during which the threshold for cold-defense responses decreases, sometimes by several degrees, whereas the threshold for heat-defense responses (tail vasodilation) does not change (130). Equivalent evidence for a dissociation between cold- and heat-defense responses in humans appears limited to conditions of anesthesia (130), and we are unaware of studies that have provided strong evidence for a dissociation of thermoeffector thresholds in response to other more common modulators of temperature regulation (e.g., circadian rhythm, acclimation, menstrual cycle, dehydration, sleep deprivation, etc.). It is also interesting to note that while the neural networks controlling cold- and heat-defense responses are clearly separate from one another, it remains unclear, at least in humans, whether heat loss thermoeffectors are controlled by the same or different effector loops.

Our current understanding of the brain circuitry coordinating heat-defense responses has been shaped by studies performed in rodents over the past 10 years. The previously described study by Nakamura and Morrison (125) first established the existence of an LPBd-MnPO pathway that mediates heat-defense responses to skin warming. Importantly, this study revealed that activation of this pathway exerts an inhibitory influence on brainstem areas that control cold-defense responses. Specifically, heat-defense responses were inhibited when a glutamate receptor antagonist was injected into the LPBd during skin warming. Furthermore, injections of glutamatergic antagonists within the MnPO inhibited heat-defense responses caused by direct stimulation of LPBd neurons or during skin warming. Consequently, these findings demonstrate that, in rodents, skin warming activates a pathway that inhibits cold-defense effectors thus resulting in heat loss via withdrawal of cutaneous vasoconstriction and reduced thermogenesis. Subsequently, Song et al. (136) identified a population of neurons that express TRPM2 within the POA that project to corticotropin-releasing hormone neurons in the paraventricular nucleus of the hypothalamus. Activation of these neurons in the absence of heat stress (via designer receptors exclusively activated by designer drugs) decreased deep body temperature by an impressive 8°C. The activation threshold for these neurons was 45°C in vitro but 38°C in brain slice preparations. Furthermore, TRPM2 knockout mice displayed greater hyperthermia during fever leading to the suggestion that these neurons may place a “break” on fever-induced hyperthermia by activating heat-defense responses. Although activation of autonomic heat-defense responses was not considered during skin warming, it is interesting to speculate that these neurons may also play a functional role in limiting body hyperthermia during environmental heat stress. An additional study by Tan et al. (122) also identified neurons activated by warmth within the MnPO that send predominantly inhibitory projections to brain regions involved in autonomic control and behavior. The temperature sensitivity of these neurons overlapped with the range of innocuous temperatures at which heat-defense responses are activated and increased their firing activity in response to peripheral capsaicin application, suggesting that they are regulated by signals arising from sensory fibers expressing TRPV1. Activation of these neurons in the absence of heat stress (by photostimulation) reduced deep body temperature, increased tail temperature, reduced brown adipose tissue temperature, and stimulated cold-seeking behavior. Lastly, Zhao et al. (137) identified an additional cluster of neurons within the ventral part of the lateral preoptic (vLPO) area that responded to warmth. Activation of these neurons in vivo reduced deep body temperature and physical activity in freely behaving mice. In contrast, inhibition of these neurons caused hyperthermia, so much so that inhibition time had to be limited to prevent hyperthermia-induced death. This study also revealed that these neurons send inhibitory projections to the dorsal part of the dorsomedial hypothalamus, thereby inhibiting thermogenesis. Taken together, studies performed in rodents have established a dedicated neural network mediating autonomic heat-defense responses during skin warming (FIGURE 4). However, it is important to consider that autonomic heat-defense responses in rodents rely on the suppression of cold-defense effectors. Specifically, skin warming results in activation of POA warm-sensitive neurons that, in turn, inhibit brainstem areas involved in the control of cutaneous vasoconstriction and thermogenesis, the net result being a passive increase in skin vasodilation (via withdrawal of vasoconstriction) and reduced metabolic heat production. It is unknown to what extent the neural circuitries identified in rodents translate to humans, especially since humans rely on the activation of specific heat loss thermoeffectors (rather than inhibition of cold-defense responses) to defend against heat stress. It is assumed that warm-sensitive neurons within the POA also exert an inhibitory influence in humans (138). If so, additional unidentified relays within the hypothalamus or brainstem must occur for this inhibitory influence to eventually activate areas involved in the control of autonomic heat loss thermoeffectors (FIGURE 4). Alternatively, we are unaware of evidence to dismiss the possibility that hypothalamic warm-sensitive neurons are excitatory (rather than inhibitory) in humans.

In humans, studies have been performed to determine brain areas activated or deactivated during heat stress. Nunneley et al. (47) employed positron emission tomography (PET) imaging to quantify changes in brain metabolism during whole body hyperthermia (deep body temperature of 38.6°C). The results revealed brain areas that were associated with an increase (hypothalamus, thalamus, corpus callosum, cingulate gyrus, and cerebellum) and decrease (caudate, putamen, insula, and posterior cingulum) in metabolism. A subsequent study also employing PET imaging demonstrated that skin warming (to 38°C) in the absence of changes in oral temperature or sweating was associated with activation of somatosensory cortices, the insula, anterior cingulate, hypothalamus, and parahippocampus (139). A subsequent series of studies built upon these observations by relating changes in brain activity (measured by magnetic resonance imaging) with changes in sweating stimulated by psychogenic and thermal stimuli. The first study focused on the brainstem and demonstrated that no region showed deactivation during psychogenic or thermal sweating and that no area showed a greater response to one stimulus versus the other (140). The dorsal midbrain and rostral lateral medulla were the main areas activated by both stimuli. Although both regions showed similar patterns of activation during each stimulus, the ventral midbrain and the caudal ventral medulla were only activated during thermal sweating. A second study focused on the POA region and observed a midline cluster of sweating activation, inferior to the anterior commissure (141). The signals extracted from this cluster showed significant temporal changes during repeated sweating events and functional connectivity analyses revealed numerous brain regions whose activity correlated with that of the POA during thermal sweating (limbic cortices, paralimbic cortices, prefrontal regions, parietal cortices, thalamus, putamen, midbrain, pons, and cerebellum). The final study focused on regions above the midbrain and observed that sweating events were associated with activation in widely distributed regions of the brain during both thermal and psychogenic stimuli (142). Nonetheless, some areas were also preferentially activated during thermal sweating (lentiform nuclei and amygdalae), whereas others were preferentially activated during psychogenic sweating. Thermal sweating was notably associated with clusters of activation in the posterior and pregenual cingulate cortex, and only thermal sweating was associated with activation in the anterior hypothalamus/POA region. Taken together, imaging studies have shown that heat stress is associated with a widely distributed activation of several brain regions. These studies have notably confirmed activation of the POA region during heat stress, and some studies have identified areas preferentially related to autonomic thermoeffector (sweating) responses such as the ventral midbrain, the caudal ventral medulla, and the posterior and pregenual cingulate cortex. Furthermore, one study identified areas functionally correlated with the POA during heat stress that may provide some guidance for future studies to identify the brain networks involved in the control of heat loss thermoeffectors. Nonetheless, the interpretation of brain imaging studies remains challenging since it is difficult to determine whether changes in metabolism or activation patterns are specifically related to processes associated with temperature regulation rather than those associated with somatosensory input. Due to technical constraints, these imaging studies employed a water-perfused suit model of heat stress that elicits relatively high skin temperatures that would be expected to activate somatosensory pathways. Furthermore, the studies that related changes in brain activity to thermoeffector output only considered sweating. Presumably, cutaneous vasodilation also occurred during these studies and the changes in brain activity therefore may also be related to activation of this heat loss thermoeffector.

3.2.3. Efferent pathways.

The efferent neural pathways driving autonomic heat loss themoeffectors in humans remain unknown (FIGURE 4). As alluded to previously, the efferent pathways mediating autonomic heat-defense responses in rodents consist of a direct inhibitory input from hypothalamic neurons to neurons within brainstem areas that control cutaneous vasoconstriction and thermogenesis. It seems unlikely that a direct inhibitory pathway can underlie the activation of sweating and cutaneous vasodilation in humans. For this reason, it has been proposed that hypothalamic activation of heat loss thermoeffectors occurs indirectly via a relay within the brainstem (138). In cats, an electrical stimulation study demonstrated an excitatory pathway from the hypothalamus through to the ventrolateral brainstem (143). Subsequent studies demonstrated that paw sweating could be stimulated by microinjections of excitant amino acids (glutamate or an analog) within the ventral medulla (144), specifically within the rostral ventromedial medulla between the facial nucleus and pyramidal tract (145). Based on the premise that excitant amino acids stimulate cell bodies, but not axons, the identified brainstem areas were proposed to represent a synaptic relay in the efferent pathway that drives paw sweating in cats (145). An analogous juxtafacial area was shown to be activated during sweating in humans, and it was further suggested that this area may directly project to sympathetic preganglionic sudomotor neurons (140). In contrast to sweating, no studies have considered possible efferent pathways for the control of active cutaneous vasodilation in humans. Despite a lack of knowledge regarding efferent hypothalamic-brainstem pathways driving heat loss thermoeffectors, the efferent neural signal they generate has been studied. In humans, postganglionic skin sympathetic nerve activity (SSNA) can be recorded directly from peripheral nerves by microneurography (146). Most investigations have performed multiunit recordings of SSNA that can contain activity from cutaneous, sudomotor, and possibly piloerector neurons (147, 148). Nonetheless, it is reasonable to assume that an increase in multiunit SSNA during heat stress primarily reflects greater activity from cutaneous vasodilator and sudomotor neurons. During the transition from a thermoneutral to a hot environment (149, 150) or during the initial stages of whole body passive heat stress (148), a decrease in SSNA is often observed likely reflecting withdrawal of cutaneous vasoconstriction. With continued heat stress, SSNA activity can become silent before reappearing once deep body and skin temperatures reach a certain onset threshold (148). Beyond this onset threshold, SSNA increases linearly with further increases in deep body temperature until reaching a plateau or maximum response at which point no further increases in SSNA can be measured despite continued increases in deep body temperature (151–154). The increase in SSNA during heat stress ultimately leads to the release of neurotransmitters from postganglionic nerve terminals that synapse on heat loss thermoeffector organs. Accordingly, the onset threshold for an increase in SSNA precedes the onset threshold for cutaneous vasodilation and sweating during whole body passive heat stress (154, 155). It should be noted that one study performed single-unit SSNA recordings from sudomotor neurons during heat stress, demonstrating that sudomotor neurons are characterized by a low firing probability and that they display some cardiac and respiratory rhythmicity (156, 157). Interestingly, multiunit bursts of SSNA were predominantly followed by changes in skin resistance (reflecting sweat production), suggesting that multiunit bursts of SSNA are primarily driven by the firing of sudomotor neurons (156). However, later studies observed that most multiunit SSNA bursts during mild body heating are followed by a transient change in both skin blood flow and sweating (158, 159). One of these studies quantified these temporal relationships, reporting that 70% of multiunit SSNA bursts were followed by a transient change in skin blood flow and sweating, whereas 1% and 10% of bursts were followed by a transient change in skin blood flow or sweating alone, respectively (159). Although the presence of postganglionic sudomotor neurons has been established from single-unit SSNA recordings, it remains unclear whether active cutaneous vasodilation is driven by a separate and unique pool of postganglionic neurons or via the same nerves that innervate eccrine sweat glands. Studies of the temporal relationship between multiunit bursts of SSNA and thermoeffector output observed some multiunit bursts that are only followed by an increase in skin blood flow (158, 159). Such an observation may suggest that cutaneous vasodilation is activated by a separate pool of postganglionic neurons, although a caveat to this possibility is that such bursts represent a minority of multiunit SSNA activity during heat stress (159). That said, studies that examined the temporal relationship between multiunit SSNA bursts and thermoeffector output generally employed very mild levels of heat stress, raising the possibility that active cutaneous vasodilation may not have occurred and that the transient changes in skin blood flow reflected withdrawal of cutaneous vasoconstriction. During more moderate body heating that elicits active cutaneous vasodilation, a component of the multiunit SSNA signal has been shown to be preferentially associated with cutaneous vasodilation. Kamijo et al. (160) recorded multiunit SSNA during passive heat stress in young adults that were either normovolemic or hypovolemic. In both groups, passive heat stress increased multiunit SSNA, but the authors identified a component of the signal (representing ≈20% of total SSNA) that was synchronized with the cardiac cycle. The group that was hypovolemic displayed less of an increase in this component of the multiunit SSNA signal during heat stress, whereas the remainder of the multiunit SSNA signal (not synchronized with the cardiac cycle) was similar between groups. In parallel, the hypovolemic group displayed lower cutaneous vasodilation, but not sweating, during heat stress. In subsequent studies, the component of the multiunit SSNA signal synchronized with the cardiac cycle was shown to be reduced during combined heat stress and head-up tilt (161) and increased following rapid saline infusion performed during passive heat stress in hypovolemic individuals (162). In both studies, decreases and increases in the component of the multiunit SSNA signal synchronized with the cardiac cycle were paralleled by corresponding decreases and increases in cutaneous vasodilation. Although it remains to be confirmed, it was suggested based on these findings that active cutaneous vasodilation and sweating are controlled by distinct efferent neural pathways (162).

3.3. Effector Responses

3.3.1. Eccrine sweating.

Human temperature regulation during heat stress relies predominantly on sweat production and its evaporation. At air temperatures ≥34°C, a temperature at which there is little to no dry heat exchange between the body and the environment and above which heat will be gained from the environment, the evaporation of sweat becomes the only means of heat loss from the body. Sweat production during heat stress is mediated by the ∼2 million functional eccrine glands that are located within the dermis layer of the skin (163). Once initiated, sweating occurs over nearly the entire skin surface, though the distribution of sweat is not uniform. Differences in the mode (passive versus active) and intensity (evaporative requirements) of heat stress make it somewhat challenging to compare the pattern of sweat distribution observed between studies (164). Nevertheless, under potent thermal stimuli, sweat rates tend to be greatest on the forehead and posterior torso, followed by the anterior torso and shoulders, head and neck, the limbs, and finally, the volar surfaces of the hands and feet (164–178). Local sweat rates show marked variability even within the same body segment. On the head, sweat rates are substantially lower on the chin and cheeks than the forehead (164, 172), while sweat rates over the rest of the cranium are lower along the midline of the skull from top to rear compared with the sides (164, 179, 180). Sweating on the torso is highest over the spine and along the midline of the chest/abdomen but decreases in a medial to lateral direction (164, 167, 172, 181, 182). On the hands, sweating is lowest on the palm, highest on the dorsal hand, and higher in distal versus proximal phalanges on the volar side (164, 167, 172, 183, 184). On the lower extremities, sweat rates tend to be higher on the buttocks and anterior thigh versus the medial and anterior areas of the thigh; on the medial versus lateral aspects of the lower leg and ankle; and on the dorsal versus plantar surface of the foot (164, 167, 172, 185, 186).

Regional variation in sweat rate could arise from differences in the number of active eccrine sweat glands, gland size, or cholinergic sensitivity. Based on data compiled from 31 studies, Taylor and Machado-Moreira (163) reported that the number of active sweat glands is highest on the palmar surface of the hand and plantar surface of the foot (mean: ∼500–520 glands/cm²), followed by the face and dorsal hand (mean: ∼165–185 glands/cm²); the dorsal foot, forearm, back, abdomen, chest, and upper arm (mean: ∼100–120 glands/cm²); the axilla, thigh, and leg (mean: ∼60–85 glands/cm²); and the buttocks (mean: ∼40 glands/cm²). Juxtaposed against the pattern of sweat distribution described above, it can be seen that there is a poor relationship between the number of active sweat glands and local sweat rate. The most conspicuous example is the difference between the volar surface of the hand and feet and the back: despite having a fivefold greater active sweat gland density than the back, local sweat rates on the palms and soles of the feet are often the lowest, while those on the back are often the highest. This suggests that differences in gland size and cholinergic sensitivity better explain regional disparities in sweat distribution.

Eccrine sweat glands are surrounded by a rich network of capillaries, and they receive dual innervation from adrenergic and cholinergic nerves (187, 188). Nonetheless, thermoregulatory sweating is almost entirely driven by postganglionic sympathetic nerves that release acetylcholine (189). Upon release, acetylcholine binds to muscarinic (M3 subtype) receptors (190, 191) triggering a release of calcium (Ca²⁺) from intracellular stores (192–194) and an influx of extracellular Ca²⁺ (195–197). The net increase in cytosolic Ca²⁺ concentration stimulates an efflux of potassium (K⁺) into the interstitium and Cl⁻ into the lumen that, in turn, stimulates sodium-potassium-chloride (NKCC1) cotransporters leading to an influx of sodium (Na⁺), potassium (K⁺), and chloride (Cl⁻) into the cell (198, 199). The greater Cl⁻ concentration within the lumen of the secretory coil provides a negative electrochemical gradient that favors paracellular transport of Na⁺ leading to a greater sodium chloride (NaCl) concentration that creates the osmotic gradient necessary for water to move into the lumen through aquaporin channels (200, 201). As the primary fluid is secreted into the lumen of the duct, an increase in hydrostatic pressure propels the fluid through the duct toward the skin surface (202). During its passage through the duct, Na⁺ and Cl− are reabsorbed resulting in sweat discharged at the skin surface that is composed at 99% of water (203). The reabsorption of Na⁺ occurs due to the activity of Na⁺-K⁺-ATPase pumps (199, 204) that create an electrochemical gradient for Na⁺ to exit the duct though epithelial sodium channels (205, 206). The gradient for Cl⁻ to exit the duct is established by apical cystic fibrosis transmembrane conductance regulators (CFTR) (207). The ultimate concentration of ions within sweat discharged at the skin surface depends on the rate at which primary fluid is secreted by the secretory coil and the rate at which the contained ions are reabsorbed in the duct (208). Greater secretion rates of primary fluid result in greater sweat concentrations of Na⁺ and Cl⁻ (165, 209–213) due to a reduced transit time through the duct (213, 214).

A predominant role for acetylcholine in mediating thermoregulatory sweating is supported by studies showing that local (215–217) and systemic (218) cholinergic blockade nearly abolishes sweat production during heat stress. Furthermore, inhibition of acetylcholinesterase, the enzyme that breakdowns acetylcholine within the synaptic cleft, leads to an earlier onset threshold for sweating during passive heat stress (219). That said, thermoregulatory sweating is likely potentiated by transmitters co-released with acetylcholine. Calcitonin gene-related peptide potentiates sweating during intradermal infusion of cholinergic agonists but does not elicit sweating on its own (220, 221). Nitric oxide (222, 223) and cyclooxygenase (223) also potentiate thermoregulatory sweating. In contrast, adrenergic agonists play little (224–227) to no role (228, 229) in thermoregulatory sweating. For a more detailed overview of eccrine sweat gland function, the reader is referred to recent review articles (230–232).

3.3.2. Cutaneous vasodilation.

It is well established that active cutaneous vasodilation accounts for almost all of the increase in skin blood flow that occurs during heat stress (233). The remainder results from withdrawal of any vasoconstriction during the initial stages of heat stress. A caveat to this statement is that the increase in blood flow in glabrous skin areas such as the hands, feet, and ears occurs primarily via withdrawal of vasoconstriction (233). Although the contributions of glabrous skin to heat exchange should not be neglected (234), nonglabrous skin contributes to a greater absolute extent due to the large body surface area it covers (235). Initial studies employing measurements of skin temperature as an index of skin blood flow provided evidence for a neutrally mediated increase in blood flow within the extremities (feet and hands) during heat stress (236) and that this increase was mediated by vasodilator nerves (237). Building upon these initial studies, Roddie et al. (215) and Edholm et al. (238) provided more direct evidence for the existence of neurally mediated cutaneous vasodilation. These studies made use of the development of venous occlusion plethysmography to perform noninvasive measurements of blood flow through a limb (239, 240). A key advancement for the interpretation of these studies was initial evidence that the increase in forearm blood flow during passive heat stress is restricted to the skin (241, 242), thus providing more confidence to rule out contributions from increases in muscle blood flow. Roddie et al. (215) demonstrated that immersion of the lower limbs in hot water resulted in an initial small increase in forearm blood flow that was followed by a large and sustained increase as heating time progressed. They further demonstrated that the large and sustained increase in blood flow was delayed and attenuated, though not eliminated, when atropine was infused intra-arterially before and throughout heat stress. In contrast, no reduction in forearm blood flow occurred when atropine was infused during heat stress once the increase in forearm blood flow was established. Complementing these observations, Edholm et al. (238) demonstrated that cutaneous nerve block performed before and during hot water immersion either abolished or greatly reduced the increase in forearm blood flow that occurs during heat stress. Combined, these studies established that the increase in forearm blood flow during heat stress is mediated by vasodilator nerves and that the initial vasodilation is mediated through a cholinergic mechanism. Later studies demonstrated that blocking the local release of norepinephrine inhibits cutaneous vasoconstriction during cold stress without affecting cutaneous vasodilation during heat stress (243) and that presynaptic blockade of cholinergic nerves results in an absence of active cutaneous vasodilation during heat stress (244). These findings firmly established the existence of separate cutaneous vasoconstrictor and vasodilator nerves and that cholinergic transmission is required for active cutaneous vasodilation. In addition, the study by Kellogg et al. (244) confirmed previous studies that antagonism of postsynaptic muscarinic receptors partially attenuates active cutaneous vasodilation leading to the theory of cholinergic cotransmission. This set off a quest to determine the cotransmitter responsible for active cutaneous vasodilation (233, 235, 245). Briefly, this quest revealed complex and redundant contributions from several vasoactive molecules. The most studied of these is nitric oxide, with its contribution to active cutaneous vasodilation being ∼30% during passive heat stress (246–249). Studies have also established a role for histamine (250), neurokinin receptors (251), TRPV1 receptors/sensory nerves (252, 253), prostaglandins (254), as well as calcium-activated potassium channels and endothelial derived hyperpolarizing factors (255). Less conclusive evidence has been provided for vasoactive peptides (256–258) and adenosine receptors (259, 260). To date, the largest pharmacological reduction of active cutaneous vasodilation observed during passive heat stress (≈80%) occurred through combined nitric oxide synthase inhibition and neurokinin-1 receptor desensitization (251). The current consensus is that active cutaneous vasodilation during heat stress is mediated by cholinergic cotransmisison of several vasoactive substances (233, 235, 245). One consideration to this statement is that the signaling pathways mediating cutaneous vasodilation may differ between passive heat stress and exercise in the heat (261–264). For example, the neuronal isoform of nitric oxide synthase underlies cutaneous vasodilation during passive heat stress (265), whereas the endothelial isoform mediates cutaneous vasodilation during exercise (261).

When considering the mechanisms that mediate active cutaneous vasodilation, it is interesting to note that initial studies considered a potential link with sweating. This link has been termed the “bradykinin hypothesis,” positing that sweat glands release a bradykinin forming enzyme thereby leading to greater bradykinin formation and subsequent cutaneous vasodilation (266–268). This hypothesis was eventually ruled out by Kellogg et al. (269) who demonstrated that intradermal infusion of a bradykinin receptor antagonist does not alter cutaneous vasodilation during heat stress. This study argues against a role for bradykinin in mediating cutaneous vasodilation during heat stress, although it does not rule out the possible interaction between sweating and cutaneous vasodilation. Although active cutaneous vasodilation may not occur through bradykinin, an interaction between sweating and cutaneous vasodilation is supported by the observation that individuals with a congenital absence of eccrine sweat glands do not display sweating nor cutaneous vasodilation during heat stress (270). Although not in the hypothesized direction (i.e., sweating-related responses driving increases in skin blood flow), a functional relationship between skin blood flow and sweating was demonstrated by Wingo et al. (101). In that study, preventing the increase in skin blood flow during passive heating attenuated the increase in sweating. An important consideration to the interpretation of this observation is that the increase in skin blood flow was completely prevented. When skin blood flow (271, 272) or sweating (273) is manipulated within physiologically relevant conditions, the functional relationship between sweating and skin blood flow does not appear to have a meaningful impact on temperature regulation.

3.4. Thermoregulatory Control

Thermoregulatory control in humans is often studied by considering changes in thermoeffector output relative to changes in skin and/or deep body temperatures. A common analytical approach is to plot thermoeffector output as a function of a weighted average of skin and deep body temperatures. Commonly used skin:deep body temperature weightings during heat stress include 0.1:0.9 and 0.2:0.8, and these weightings reflect the relative influence of skin and deep body temperatures on sweating and skin blood flow (274–276). The validity of this approach is justified by studies that established the relative importance of skin and deep body temperatures on heat loss thermoeffectors. An important distinction should be made with the weightings used to calculate mean body temperature to estimate changes in body heat content. That is, estimating changes in body heat content during heating or cooling is conceptually different from estimating the relative contributions of skin and deep body temperatures on heat loss thermoeffector responses. For this reason, we have argued that it is inappropriate to justify the skin:deep body temperature weightings from studies that were intended to estimate changes in body heat content (277).

When thermoeffector output is plotted as a function of a weighted average of skin and deep body temperatures, the relation is typically characterized by an initial flat portion during which body temperature increases without a corresponding increase in thermoeffector output. As body temperature continues to increase, an abrupt increase in thermoeffector output eventually occurs representing the onset threshold of the response. Beyond this point, thermoeffector output increases proportionally with the increase in body temperature until eventually stabilizing at an elevated level despite further increases in body temperature. The linear portion of the increase in thermoeffector output is used to define the thermosensitivity of the response (change in thermoeffector output relative to the change in body temperature) whereas the elevated stable values reflect maximum thermoeffector output under the experimental conditions examined. The model provides three readily available values to subsequently determine whether a given variable of interest influences thermoregulatory control: the onset threshold, the thermosensitivity and the maximum output of the response. This model has proven useful to identify several modulators of heat loss thermoeffectors, including but not limited to age, sex, menstrual cycle, heat acclimation, circadian rhythm, aerobic fitness, and dehydration (277). A consideration for the use of this model is that it requires continuous measurements of thermoeffector output and body temperature to subsequently obtain adequate relationships to derive the onset threshold, thermosensitivity, and maximum output. This may not always be possible, and an alternative approach is to quantify the overall change in thermoeffector output relative to the overall change in body temperature. This last approach can also provide a general indication of whether a given variable of interest alters thermoregulatory control, although it can only provide an indication of the overall thermosensitivity of the response without providing information on the onset threshold.

Directly evaluating individual components of thermoeffector loops remains challenging in humans. Microneurography can be used to record warm-sensitive afferent signals from the skin, but obtaining such recordings is rare. For example, in one of the earliest studies that identified skin nerves responding to innocuous warmth, only 3 recordings were obtained out of 50 attempts (109). Such recordings are rare even in animal models. For example, Andrew and Craig (127) identified only 10 dorsal spinal cord neurons that specifically responded to heat out of a total of 474 neurons studied. Perhaps for this reason, a limited number of studies have reported on single-unit recordings from skin afferents responding to innocuous warmth (108–111). For integration of thermal afferents, brain imaging studies must be performed thus requiring logistical and financial resources not readily available. Furthermore, as discussed previously, such studies remain limited to providing associations between activated (or deactivated) brain regions relative to changes in thermoeffector output. For efferent signaling, SSNA can be measured, as previously described, but it nonetheless remains a measurement that is technically challenging and not widely accessible. Another limitation of multiunit SSNA recordings performed during heat stress is that they contain sudomotor and presumably active cutaneous vasodilator activity. To our knowledge, no single-unit recordings have been performed from putative active cutaneous vasodilator nerves. Although attempts have been made to delineate a component of the multiunit SSNA signal that is specific to active cutaneous vasodilation, as previously addressed, these attempts remain correlative, and a direct validation is lacking. Due to the challenges associated with directly evaluating individual components of thermoeffector loops, the thermoeffector output-body temperature relationship has been used to infer how various modulators of thermoregulatory control exert their effect on thermoeffector output. This interpretation relies primarily on the onset threshold and thermosensitivity parameters, with the onset threshold proposed as a marker of the neural control of body temperature and the thermosensitivity proposed as a marker of intrinsic properties of the thermoeffector organ (277). However, this interpretation is not without its limitations, and it does not benefit from direct validation. The reader is referred to a recent review article for more details (277). Briefly, the onset threshold has most often been quantified using absolute temperature values. Most modulators of thermoregulatory control induce a shift in resting deep body temperature (e.g., menstrual cycle, age, acclimation, exercise training, etc.) that parallels the change in the onset threshold. The net result is that the change in body temperature required to activate thermoeffector output is generally unchanged by the modulator under investigation. Although it remains unknown whether the absolute or change in onset threshold better reflects thermoregulatory control, studies should consider whether the effect of a given modulator on the onset threshold goes beyond any change in resting deep body temperature. It is also important to remember that the thermoeffector output-body temperature relationship encompasses all components of thermoeffector loops, from afferent signaling to effector output. It is therefore debatable that the relationship can be used to isolate a single component of thermoeffector loops. The relationship may best be used to determine whether a variable of interest modulates overall thermoregulatory control and motivate future, more focused studies to examine a specific component of the thermoeffector loop to determine how this modulation may occur (277).

4. MODULATORS OF TEMPERATURE REGULATION: INTRINSIC FACTORS

With the fundamental concepts related to energy exchange and human thermoregulatory control as background, our focus now shifts to the various intrinsic morphological and physiological traits that modulate the response to heat stress in healthy humans. Although these traits could conceivably alter temperature regulation at any point of thermoeffector loops, our current understating of their modulating influence relies almost exclusively on measurements of deep body temperature and thermoeffector output. For some traits, postjunctional responses of thermoeffector organs have been considered and, in rare instances, measurements of SSNA have been performed. These instances are highlighted when available. In contrast, there is a lack of research investigating how these traits may modulate thermoafferent signaling and integration of thermoafferent flow.

4.1. Morphological Characteristics

The dimensions and composition of the body’s passive system are important modulators of the heat stress response. In this section, we review the effects of body mass, body surface area, the surface area-to-mass ratio, and body composition and discuss how morphological factors should be considered when assessing the independent effects of various physiological traits (e.g., age, biological sex, disease, injury).

4.1.1. Body mass.

Mass has multiple thermodynamic effects. First, absolute metabolic rate at rest and during fixed-speed weight-bearing locomotion rises with body mass, resulting in a greater absolute rate of metabolic heat production in heavier individuals (278–280). Additionally, mass represents the body’s heat sink, such that the rise in deep body temperature for a given change in heat storage is inversely related to body mass. It follows that for a given absolute rate of metabolic heat production, a heavier mass should attenuate the rise in deep body temperature, provided similar regulatory control of thermoeffector responses. Several studies have found a strong negative correlation between body mass and end-exercise deep body temperature while cycling (i.e., non-weight-bearing exercise) at a fixed absolute work rate (281–283), absolute heat production (284), and fixed relative intensity (285). Cramer and Jay (286) later demonstrated the independent effect of body mass by comparing deep body temperature responses to cycling exercise at fixed absolute rates of heat production between groups of heavier (mean: 91.5 kg) and lighter (mean: 67.6 kg) individuals matched for age, sex, and heat acclimation status. At both 500 and 600 W of heat production, elevations in rectal temperature were 0.3–0.4°C lower in the heavier group due to their larger heat sink (FIGURE 5, top). Similar observations were made by Ravanelli et al. (287) in uncompensable conditions, with the rise in rectal temperature attenuated by an average of 0.7°C in larger (mean: 100.0 kg) compared with smaller (mean: 65.8 kg) individuals while cycling at a heat production of 520 W in a hot-humid environment.

FIGURE 5.

Change in rectal temperature (ΔT_re) between groups of larger (LG; 91.5 kg) or smaller (SM; 67.6 kg) body mass during exercise eliciting absolute (top: 500 W and 600 W) or mass-specific (bottom: 6.5 W/kg and 9.0 W/kg) rates of heat production (H_prod). *Significantly greater elevation in rectal temperature in the SM group. Redrawn from Ref. 286, with permission from the American Physiological Society.

Because heat production and mass independently influence deep body temperature responses, both factors must be considered when comparing deep body temperature between groups that cannot be matched for body size (e.g., children versus adults, males versus females, individuals with versus without obesity). To address this issue, Cramer and Jay (286) further compared deep body temperature responses in their heavier and lighter mass groups while cycling at fixed rates of heat production relative to mass (W/kg) in a compensable environment [∼25°C and 37% relative humidity (RH)]. Despite the 23.9-kg difference in mass and differences in the absolute rates of heat production, changes in rectal temperature were similar between the groups at 6.5 and 9.0 W/kg of heat production (FIGURE 5, bottom). When the 9.0 W/kg trial was repeated in a ∼35°C environment, similar elevations in deep body temperature were again observed irrespective of body mass, demonstrating the validity of this approach under different ambient conditions within the range compensable environments. Furthermore, no differences in sweat onset or thermosensitivity (i.e., thermoregulatory control) were observed between groups. Taken together, these findings suggest that deep body temperature comparisons between groups unmatched for body mass should be compared at the same mass-specific heat production in compensable conditions.

4.1.2. Body surface area.

As body mass (volume) increases, there is a proportional rise in surface area, the dimensions of which dictate the absolute rate of whole body heat exchange. That is, for given thermal and vapor pressure gradients, the absolute rate of heat exchange is greater among individuals of larger body surface area. However, with diseases and injuries that impair sweat production, the surface area that can be saturated with sweat and thereby participate in evaporative heat dissipation, the effective body surface area for evaporative heat loss, may be lower than the total body surface area, reducing E_max (86). The effect of a reduced effective body surface area on deep body temperature was demonstrated by Totel (288), who assessed deep body temperature and sweat rates during prolonged exposures to 38°C and 20% RH among individuals with ectodermal dysplasia (congenital absence of sweat glands), quadriplegia (denervation of sweat glands from the neck down), and able-bodied individuals (FIGURE 6). Individuals with ectodermal dysplasia and quadriplegia had the lowest sweat rates and highest elevations in deep body temperature, although individuals with ectodermal dysplasia could tolerate the heat for only half the duration of the other individuals.

FIGURE 6.

Deep body temperature responses in individuals with different effective body surface areas for evaporative heat loss. The ability to produce sweat is severely limited with ectodermal dysplasia (absence of sweat glands) and quadriplegia (denervation of most sweat glands). As such, greater reductions in effective surface area are associated with lower sweat rates and higher elevations in deep body temperature. Redrawn from Ref. 288, with permission from the American Physiological Society.

Differences in body surface area can also impact local sweat rates. As discussed earlier (sect. 2.1.2), whole body sweat rate (in g/min) is strongly influenced by absolute E_req (in W) under physiologically compensable conditions allowing 100% sweating efficiency, irrespective of body size (74, 286, 289). However, for a given E_req, the same rate of sweat secretion over a larger body surface area leads to lower local sweat rates, when expressed per squared centimeter of the measurement area. Indeed, between groups of larger (2.12 m²) and smaller (1.80 m²) body surface area, cycling exercise at E_req values of 340 and 400 W resulted in similar steady-state whole body sweat rates in each group of ∼7.5 and ∼10.0 g/min, respectively, although steady-state local sweat rates were ∼0.2 mg·cm⁻²·min⁻¹ lower in the larger group (286). It follows that comparing local sweat rates between groups of dissimilar body size requires consideration of both E_req and body surface area. In the aforementioned study (286), when local sweat rate responses were further assessed at exercise intensities eliciting equivalent area-specific E_req values of 165 and 190 W/m², no differences in local sweat rate were observed between the larger and smaller groups at each area-specific E_req, though whole body sweat rate was greater in the larger group due to a correspondingly greater absolute E_req (∼50 W) in that group. Similar findings were reported by Notley et al. (290) for the upper back (E_req values of 83 and 142 W/m²) and dorsal forearm (E_req of 142 W/m²); however, a significant negative association was found between local sweat rates on the forehead and dorsal hand and body size at both area-specific E_req values. Since whole body sweat rate was associated with the absolute E_req as in previous studies, the latter observations raise the possibility that these local sweat rates simply do not reflect the whole body response; that is, comparing local sweat rates at a given area-specific E_req may be valid only at certain sites. This is supported by the observation that exercise at the same area-specific E_req leads to a disproportionately greater local sweat rate on the forehead, but not the forearm, in individuals with a low aerobic capacity [i.e., maximum oxygen uptake (V̇o_2max)] due to a higher relative exercise intensity (177).

It is also possible that individual variability in body surface area can yield disproportionately greater whole body sweat rates in smaller-sized individuals because of their lower absolute E_max, which raises ω_req and thereby lowers sweating efficiency (291). Such an effect could at least partly explain the inverse relationship between whole body sweat rate and E_max for a given E_req (292–294). However, whether body surface area independently affects whole body sweat rate and sweating efficiency at higher ω_req remains to be determined. Until then, the prudent approach to comparing whole body sweat rate between groups is to ensure the experimental conditions facilitate 100% sweat evaporative efficiency (86).

4.1.3. Surface area-to-mass ratio.

Beyond the independent effects of mass and surface area, the ratio of surface area-to-mass has also been considered. Since the absolute rate of heat loss is proportional to body surface area, smaller individuals with a greater surface area-to-mass ratio can theoretically dissipate more heat per unit of mass. When skin temperature exceeds air temperature, the mass-specific rate of dry heat loss is higher in individuals with a higher surface area-to-mass ratio, minimizing the requirement for sweating and evaporative heat loss (i.e., a lower E_req). With a reversal of the skin-air temperature gradient in hot conditions, the mass-specific rate of dry heat gain is higher in those with a larger surface area-to-mass ratio (i.e., E_req is higher). Because of this effect, a higher surface area-to-mass ratio has been described as a thermoregulatory liability in hot environments (295). However, if evaporation is not completely restricted [e.g., by immersion in hot water or encapsulation in vapor-impermeable clothing (86)], and heat stress persists long enough for steady-state sweating to occur, a larger surface area-to-mass ratio also enables a higher mass-specific E_max, which should favor smaller individuals in extreme heat by raising the limit of physiological compensability (296).

Supporting the notion that mass-specific E_max is higher in those with a larger surface area-to-mass ratio, several studies have noted significant negative correlations between surface area-to-mass ratio and deep body temperature (297, 298) or the rate of heat storage (299) during weight-bearing exercise (bench stepping, walking, or running) in conditions near or above the limits of physiological compensability, though the metabolic heat load was not matched across study participants. More recent studies have found the surface area-to-mass ratio to be a significant predictor of deep body temperature (73, 284), but the overall contribution to the total variance in deep body temperature is lower for surface area-to-mass ratio compared with metabolic and climatic effects (73, 300). For example, Cramer and Jay (73) found that, following 60 min of cycling exercise in temperate conditions, mass-specific heat production alone explained 49.6% of the variability in the change in temperature, with surface area-to-mass ratio explaining an additional 4.3% of variance. Bar-Or et al. (301) did not detect any significant differences in the change in temperature between smaller (73 kg, 250 cm²/kg) and larger women (91 kg, 219 cm²/kg) while walking at an estimated rate of metabolic heat production of 5.1 W/kg in extreme heat (∼48°C and ∼27% RH). Recently, Ravanelli et al. (287) compared deep body temperature responses to uncompensable heat stress, involving cycling in 35°C and 70% RH at the same mass-specific heat production (6 W/kg) and the same mass-specific heat production above E_max (3 W/kg above E_max), with the latter approach accounting for differences in E_max arising from variability in surface area-to-mass ratio between larger (100 kg, 226 cm²/kg) and smaller (66 kg, 271 cm²/kg) individuals. No differences in temperature were observed between groups at 6 and 3 W/kg above E_max, indicating a negligible role of surface area-to-mass ratio once accounting for heat production and mass and suggesting that either approach can be used to assess deep body temperature responses to uncompensable exercise-heat stress between groups unmatched for mass. Collectively, current data suggest that beyond the effect of mass itself, the influence of the surface area-to-mass ratio on the deep body temperature response to exercise-heat stress appears to be minimal (302).

4.1.4. Body composition.

Because of differences in water content, adipose and lean tissues have different thermodynamic properties. First, fat has a lower thermal conductivity than lean tissues [fat: 0.21 W·m⁻¹·°C⁻¹; skeletal muscle: 0.49 W·m⁻¹·°C⁻¹ (303–306)], meaning that fat provides greater thermal resistance (i.e., insulation) against internal conductive heat transfer, when blood flow through fat tissue is negligible (i.e., vasoconstriction). Additionally, fat tissue has a lower specific heat capacity than lean tissues; that is, a smaller quantity of thermal energy is needed to raise a fixed mass of fat by 1°C [fat: 2.065 kJ·kg⁻¹·°C⁻¹; skeletal muscle: 3.322 kJ·kg⁻¹·°C⁻¹ (306)]. Individuals with a higher body fat percentage have a lower average specific heat capacity, such that less thermal energy is required to raise 1 kg of mass by 1°C. Since the rise in deep body temperature is inversely related to the average specific heat capacity, a higher body fat percentage should exacerbate the rise in deep body temperature for a given mass-specific heat load unless compensated by an increase in whole body heat loss.

Despite the sound theoretical basis for fat-related effects on deep body temperature responses, observations of such an effect have been mixed. Several studies have reported paradoxically greater deep body temperature responses to exercise in individuals with lower versus higher relative body fatness. In some cases (307–310), exercise was performed at intensities corresponding to a fixed percentage of V̇o_2max or the same rate of heat production per kilogram of lean body mass. Given that leaner people tend to possess a greater aerobic capacity (when expressed relative to total body mass) and lower total body mass, higher deep body temperature responses in individuals of lower versus higher body fatness are most likely explained by higher rates of heat production per kilogram of total mass in leaner people (311). Haymes et al. (312) found that rectal temperature rose to a significantly greater extent (+0.3°C) in prepubertal girls of lower (mean: 13.3% body fat) compared with higher (mean: 24.7% body fat) adiposity while walking at the same mass-specific heat production in extreme heat (46.7–51.6°C, 13% RH). Both groups had reached similar peak sweat rates and mean skin temperatures, and the higher surface area-to-mass ratio and lower body fatness should have theoretically led to greater mass-specific evaporative capacity and greater average tissue heat capacity, respectively, in the leaner group, making this finding difficult to explain.

Other studies have reported no effect of relative body fatness on deep body temperature responses to heat stress. Several investigations using regression analysis have not detected a significant association between body fat percentage and deep body temperature. For example, Soule al. (313) found no correlation between relative body fat and rectal temperature between leaner (mean: 11.9% body fat) and fatter (mean: 30.5% body fat) individuals during intermittent walking in air temperatures of 24–49°C while wearing different military clothing ensembles. Havenith (281, 283) showed no significant association between deep body temperature and body fat percentage during cycling exercise at a fixed absolute work rate in hot-dry and warm-humid environments in large groups of individuals with heterogeneous morphological characteristics. In studies comparing deep body temperature responses between independent groups, differences in adiposity of ∼10–13 percentage points did not influence the deep body temperature response to moderate-intensity cycling eliciting a similar mass-specific rate of heat production in compensable and uncompensable heat stress conditions (177, 286, 287, 314, 315). Similar findings have been noted with even larger between-groups differences in body fat percentage and under hotter conditions. For instance, Miller and Blyth (316) did not find a difference in the rectal temperature between non-heat-acclimated lean males (mean: 10.0% body fat) and males with obesity (mean: 31.2% body fat) at rest and during low-intensity walking (3 mph) at 48–50°C and 20–30% RH, nor during moderate-intensity walking (4 mph, 10% grade) at 25°C. Bar Or et al. (301) reported that in lean women (mean: <15.5% body fat) and those with obesity (mean: 31.5% body fat), rectal temperature did not differ between groups in compensable (27–38°C, ∼13% RH) and uncompensable heat stress (42–49°C, ∼15% RH).

Finally, several investigations have detected an influence of relative body fatness on the deep body temperature response to heat stress. Using multiple regression analysis, Havenith and van Middendorp (300) examined the variance explained by factors related to metabolism, environment, anthropometry, and thermoeffector responses during cycling exercise at 25% and 45% of V̇o_2max in 21°C, 34°C, and 45°C. Beyond metabolic rate and environmental parameters, residual variance in esophageal and rectal temperature responses was found to have a significant positive association with body fat percentage. Cramer and Jay (73) reported similar findings following cycling exercise of wide-ranging intensity in 25°C, with body fat percentage explaining a significant, albeit minor (2.3%), amount of the residual variance in the rectal temperature response after accounting for the mass-specific rate of heat production (49.6%) and surface area-to-mass ratio (4.3%). When the influence of body fat was examined at different ambient conditions, Havenith et al. (282) demonstrated that body fat percentage explained a significant portion of the variance in deep body temperature at an air temperature of 21°C, but not in hot-dry (45°C, 20% RH) or warm-humid (35°C, 80% RH) conditions, concomitant with lower forearm blood flow, supporting an insulating effect of subcutaneous fat when the cutaneous vasculature is somewhat vasoconstricted. Recently, Dervis et al. (289) compared thermoregulatory responses during cycling exercise between mass-matched groups of high (mean: 32.0% body fat) or low (mean: 10.8% body fat) adiposity in a warm (28°C) and compensable environment. Metabolic heat production and total heat losses did not differ between groups, but the authors observed a significantly greater elevation in temperature by 0.21°C in the higher fat group after 60 min of cycling, suggesting an effect of a higher average specific heat capacity in this group (FIGURE 7). Similar findings were subsequently reported by Tucker et al. (317); that is, a greater change in deep body temperature by ∼0.5°C in higher fat individuals with no change in sweat rate during uncompensable heat stress (40°C, 33% RH) (FIGURE 7). Overall, it appears that the deep body temperature response to exercise in warm and hot environments is greater in individuals of higher versus lower adiposity (i.e., lower average tissue heat capacity) but only when the difference in relative body fatness is more than ∼20 percentage points.

FIGURE 7.

Time-dependent changes in rectal temperature (ΔT_re) during exercise in groups with high (HI-BF) or low (LO-BF) body fat percentages. *Significant difference between groups. Left: exercise elicited 500 W of heat production in a ∼28°C environment, with HI-BF (32.0 ± 5.6%) and LO-BF (10.8 ± 3.6% body fat) individuals pair-matched for mass. Reproduced from Ref. 289, with permission from the American Physiological Society. Right: exercise elicited 6.0 W/kg in a ∼40°C environment, with groups unmatched for mass but vastly different in body fat percentage (HI-BF: 30.2 ± 4.1% and LO-BF: 13.6 ± 3.8%). Reproduced from Ref. 317, with permission from the American Physiological Society.

4.2. Heat Acclimation/Acclimatization

The adaptations to repeated laboratory-based heat exposures (acclimation) and in response to seasonal changes in ambient temperature (acclimatization) have been studied extensively (318). Initially, these adaptations were primarily considered within the context of work performance in hot environments (319), whereas more recent research has focused on athletic performance in hot but also temperate environments (3, 320, 321). Regardless of the context, the classical adaptations to heat adaptation have been known since the 1960s, including a lower basal deep body temperature, a greater sweating response and a lower heart rate during exercise in the heat (322, 323). Repeatedly elevating deep body temperature is considered the main stimulus to induce these adaptations, but the method to induce optimal adaptation has been the subject of a reappraisal in recent years (324). The most widely used approach has been “traditional” heat acclimation, consisting of fixed intensity exercise in the heat repeated daily over shorter (5–7 days) or longer (≥14 days) periods (318). Initially, deep body temperature during exercise reaches high levels but it subsequently decreases over the course of the acclimation period. The reduction in deep body temperature is taken as evidence of heat acclimation. However, the progressive reduction in deep body temperature may result in a diminishing adaptation stimulus, potentially leading to suboptimal adaptations (318, 324). In contrast, “controlled hyperthermia” consists of increasing deep body temperature to a predetermined level (typically ∼38.5°C) and maintaining it at this level for a given duration (typically 60 minutes) (325). As the body becomes acclimated, exercise intensity and/or environmental conditions are adjusted to maintain the adaptation stimulus constant. Fixing the level of heart rate, rather than deep body temperature, has also been studied as a practical alternative (326). It is unknown if controlled hyperthermia (or controlled heart rate) stimulates greater physiological adaptations, particularly since only one study has directly compared it to traditional heat acclimation (327). In that study, traditional and controlled hyperthermia protocols led to similar reductions in end-exercise deep body temperature and improvements in total sweat loss during exercise in the heat. Although these findings suggest that both protocols may provide similar adaptations, it is important to consider that properties of heat loss thermoeffector loops were not measured.

Regardless of the protocol employed, countless studies have demonstrated that heat adaptation improves thermoeffector output during heat stress. Arguably, the most important adaptation for temperature regulation is a greater sweating capacity that is characterized by a greater whole body sweat rate (328), with peripheral limbs showing relatively greater improvements thereby resulting in a more uniform distribution of sweat across the body surface (329). Combined, these responses result in a greater ω_max (72) that raises the E_req value at which heat stress becomes uncompensable. Indeed, a greater sweat rate following heat acclimation is most evident at high E_req values (328) or within uncompensable environments (330). These last observations highlight the importance of considering whether pre/postacclimation assessments of temperature regulation provide a sufficiently challenging stimulus to observe potential improvements in sweating capacity. In contrast to sweating, improved cutaneous vasodilation following heat acclimation is less consistently observed (318). Nonetheless, recent studies employing the controlled hyperthermia protocol observed improvements in cutaneous vasodilation following heat acclimation (154, 331).

The underlying physiological adaptations mediating improvements in heat loss capacity following heat acclimation have received less attention. A greater postjunctional cholinergic sensitivity of heat loss thermoeffector organs has been observed in several studies (332–336). One exception is a study that did not observe an improved cholinergic sensitivity of heat loss thermoeffector organs in noninjured skin of burn survivors (336), despite these individuals displaying classical signs of heat acclimation during a heat stress test (337). Potential mechanisms that could contribute to an improved postjunctional sensitivity include morphological and functional adaptations within thermoeffector organs. For example, an increase in tubular length and volume of the eccrine sweat gland was observed in primates throughout 9 months of continuous exposure to a hot environment (338). A greater in vivo maximal sweating capacity of individual sweat glands was also observed following heat acclimation and this improvement correlated with sweat gland size (338). In contrast to adaptations at the level of the thermoeffector organ, few studies have considered potential improvements in the neural control of body temperature following heat acclimation. Recently, Barry et al. (154) observed that 7 days of passive controlled hyperthermia lowers the change in mean body temperature required to elicit an increase in SSNA during whole body passive heat stress. The lower onset threshold for the activation of SSNA was paralleled by a reduced onset threshold for cutaneous vasodilation and sweating. Importantly, a lower change in mean body temperature was required to activate SSNA and thermoeffector output suggesting more sensitive thermoeffector loops following heat acclimation. Potential mechanisms that could underlie an improved neural control of body temperature following heat acclimation include a greater sensitivity of peripheral warm-sensitive neurons leading to greater thermoafferent flow for a given heat stimulus and/or greater sensitivity of hypothalamic warm-sensitive neurons resulting in greater thermoefferent outflow for a given thermoafferent input (339–343).

4.3. Aerobic Capacity

4.3.1. Heat adaptation with training.

Aerobic exercise training evokes a suite of cardiovascular and metabolic adaptations that enable improvements in oxygen delivery and utilization (344, 345). Additionally, aerobic training, even in temperate conditions, also repeatedly elevates deep body temperature and sweat production, which are requisite stimuli for heat adaptation. Not surprisingly, heat-acclimated and aerobically trained individuals (characterized by a high V̇o_2max) display many of the same thermoregulatory adjustments to exercise in the heat (318).

Piwonka et al. (296) directly compared thermoregulatory responses to exercise-heat stress between groups distinguished by training status. During an 85-min treadmill walk in 40°C and 25% RH conditions, deep body temperature increased continuously in untrained individuals, reaching a terminal value of 39.4°C. In contrast, deep body temperature in highly trained runners reached a steady state at only 38.2°C. In a subsequent study from that laboratory, Gisolfi and Robinson (346) demonstrated that 6 weeks of interval training in temperate conditions led to a reduction in basal rectal temperature, as well as attenuated elevations in rectal temperature, mean skin temperature, and heart rate. This attenuation in heat strain among highly trained runners and following aerobic conditioning of previously untrained individuals led these authors to conclude that these athletes were “preacclimatized” by virtue of their “frequent stimulation of the temperature regulatory responses” encountered during regular training and competitions (296). Numerous longitudinal training studies and cross-sectional comparisons between individuals of high- versus low-V̇o_2max have since confirmed these findings (347–358).

In the 1970s, several laboratories began to investigate the mechanistic underpinnings of this preacclimatized state. Nadel et al. (359) demonstrated that 10 days of training (1 h daily at 75–80% V̇o_2max) in temperate conditions led to an augmentation in the thermosensitivity of the sweating response. Subsequent observations of greater posttraining whole body sweating for a given change in deep body temperature or E_req during exercise-heat stress support this observation (72, 296, 348–350, 360–363). One reason for this adaptation appears to be a peripheral effect at the level of the sweat gland. Greater responsiveness of eccrine glands to cholinergic stimulation has been noted among self-described “heavy sweaters” (364), as well as trained athletes (365, 366). This effect could suggest training-induced sweat gland hypertrophy, a higher density of muscarinic receptors on the sweat gland, and/or reduced acetylcholinesterase activity. Nonetheless, it should be noted that a potential contribution of improved neural control of body temperature that results from exercise training has not been investigated and therefore cannot be ruled out.

Greater sudomotor responsiveness in trained individuals translates into an enhanced potential for evaporative heat loss. Recently, Lamarche et al. (349) used direct calorimetry to assess whole body evaporative heat loss in males classified as high, moderate, and low fitness on the basis of V̇o_2max during three 30-min bouts of cycling at increasingly greater rates of metabolic heat production of 300, 400, and 500 W under hot-dry conditions. Compared with the low-V̇o_2max group, higher rates of evaporative heat loss occurred at 400 and 500 W of heat production and a significantly lower esophageal temperature was evident at 500 W in the high-V̇o_2max group. The moderate-V̇o_2max group demonstrated intermediate responses. A similar protocol was administered in high- and low-V̇o_2max females at 250, 325, and 400 W of heat production (350). The high-V̇o_2max female group demonstrated higher rates of evaporative heat loss at 325 and 400 W and an attenuated rise in esophageal temperature at 400 W. Collectively, these findings point to an enhanced sweat sensitivity at high heat loads in trained individuals.

Ravanelli et al. (72) investigated the impact of aerobic training on thermolytic capacity by assessing the critical environmental limit for heat balance during exercise in the heat before and after an 8-wk training regimen. Participants cycled for 45 min at a heat production of 450 W in 37.5°C and an ambient vapor pressure of 2.0 kPa (31% RH). Once esophageal temperature reached a steady state, ambient vapor pressure increased by 0.04 kPa/min for another 60 min. The critical environmental limit for heat balance was taken as the ambient vapor pressure at which an upward inflection in esophageal temperature occurred; this critical vapor pressure marks the transition from compensability to uncompensability as the imposed heat load overwhelms the capacity for heat loss. With training, the mean critical vapor pressure increased from 2.98 to 3.42 kPa, indicating a higher posttraining evaporative capacity. Since the skin-air vapor pressure gradient at the critical vapor pressure (i.e., the drive for evaporation) was lower posttraining (3.10 kPa) compared with pretraining (3.48 kPa), the enhanced evaporative capacity was mediated by an improvement in ω_max, which increased from 0.72 to 0.84. Because training increased the number of heat-activated sweat glands and sweat rate, these results suggest that training-related improvements in evaporative capacity (and thus the range of heat stress conditions in which heat balance can be attained) are mediated by an increased density and output of the sweat glands, which broadens the surface area of skin sweat coverage.

Training-related modifications to the cutaneous vascular response to heat stress were first assessed by Roberts et al. (367). Following a 10-day training program, the onset threshold for forearm blood flow shifted to a lower deep body temperature without any change in the thermosensitivity of the response (367). This shift in onset was later shown to be unrelated to changes in vasoconstrictor activity (368) but possibly due to an expansion of blood volume (369). With an earlier onset of skin blood flow, a more rapid increase in deep body-to-skin heat transfer would be expected. However, the magnitude of the shift in vasodilatory onset is typically matched by an equivalent reduction in basal deep body temperature (370), suggesting that the relative stimulus to increase skin blood flow was in fact unchanged by aerobic training.

Fritzsche and Coyle (370) demonstrated higher cutaneous and forearm blood flow responses in trained versus untrained individuals while cycling at relative intensities of 50, 75, and 90% of V̇o_2max. Although these findings are often cited as evidence of adaptation, it must be recognized that heat production and the requirements for heat loss were greater in the trained subjects at each relative intensity. It follows that the peak rate of skin blood flow during exercise would be greater in trained individuals given their higher capacity for heat production and thus a greater need for heat loss. When skin blood flow responses were compared at a matched rate of heat production of 9 W/kg, corresponding to 50% V̇o_2max in trained and 70% V̇o_2max in untrained subjects, no differences in cutaneous or forearm blood flow were evident.

Intrinsic microvascular adaptations associated with training include improved endothelium-dependent cutaneous vasodilation (371, 372) and enhanced nitric oxide bioavailability (373), evidenced by greater vasodilatory responsiveness to acetylcholine and local heating, respectively. Such adaptations can be stimulated by elevations in hemodynamic shear stress and temperature (374, 375), although the influence of the latter would presumably be negligible with aerobic training under temperate conditions. As noted by Simmons et al. (376), and given the findings of Fritzsche and Coyle (370), the functional importance of these microvascular adaptations may only be relevant in individuals with cardiometabolic diseases. In these populations, impairments in vasodilatory responsiveness, and the attendant reductions in vasodilation and skin blood flow, may have a negative impact on deep body temperature regulation (377–379).

As previously outlined, full expression of the heat-adapted phenotype is achieved if repeated exposures to exercise and an external heat load induce marked elevations in deep body and skin temperatures, as well as profuse sweating. Since exercise in temperate conditions does not appreciably elevate skin temperatures, even highly trained endurance athletes only reach a fraction of their adaptive potential. For example, when three of the runners from the study by Piwonka et al. (296) subsequently undertook a 4-day acclimation regimen in 50°C heat, further adaptations were observed, including an increase in sweat evaporation by ∼10% and decreases in end-exercise rectal temperature, mean skin temperature, and heart rate by 1.2°C, 0.7°C, and 30 beats/min, respectively (352). For sedentary individuals, more strenuous exercise programs produce the largest improvements in training-related thermoregulatory function due to the greater thermal strain incurred. Interval training that increased rectal temperature to ∼38.4–39.7°C and evoked sweat losses of 0.47–1.44 L/h during each session attenuated the rise in rectal temperature, mean skin temperature and heart rate by ∼0.6°C, ∼1°C, and 20 beats/min, respectively, during a standard heat stress test (50°C, metabolic rate of 293 W/m²). In contrast, low-intensity training that elevated rectal temperature to 37.7–38.3°C and sweat losses by 0.3–0.5 L/h led to minor reductions in deep body temperature and no change in sweat rate (380, 381), both of which may have been better explained by a training-induced reduction in metabolic rate (i.e., greater muscular efficiency) rather than a true thermoregulatory adaptation. Training that does not elevate deep body and skin temperatures (e.g., cool-water swimming) may improve V̇o_2max but does not produce any evidence of thermoregulatory adaptation (382, 383). Not only does this observation confirm the need for elevations in deep body temperature and sweating to achieve some level of thermal adaptation, it also demonstrates that it is training, not an improvement in V̇o_2max per se, that leads to heat adaptation (384).

4.3.2. Heat stress responses at a relative exercise intensity (%V̇o_2max).

In the 1960s, studies by Åstrand (385) and Saltin and Hermansen (79) reported that individual variability in deep body temperature was diminished when exercise was performed at the same relative exercise intensity (i.e., a fixed %V̇o_2max) rather than the absolute rate of oxygen uptake (V̇o₂). Based on seminal work by Christensen (46) and Nielsen (386), which showed that rectal temperature responses to exercise were graded with absolute intensity, one would expect higher deep body temperature responses in aerobically trained versus untrained individuals for a given %V̇o_2max due to a higher absolute V̇o₂ and thus heat production. However, it was argued that a higher rate of heat production in trained individuals is offset by greater thermolytic capabilities resulting from training-induced improvements in sweat rate and skin blood flow (387). As a result, it became accepted practice to compare thermoregulatory responses to exercise between groups unmatched for V̇o_2max at the same relative intensity.

The notion that %V̇o_2max influences deep body temperature responses, independently of heat production, has not stood up to careful investigations. Jay et al. (314) assessed changes in deep body temperature and sweat rate in individuals of high (∼60 mL O₂·kg⁻¹·min⁻¹) versus low (∼40 mL O₂·kg⁻¹·min⁻¹) V̇o_2max while cycling at 60% of V̇o_2max. Importantly, the groups were matched for body size (mass and surface area) to avoid the confounding effect of body size on the observed responses (discussed below), and the ambient conditions were conducive to 100% sweating efficiency. Heat production and E_req were greater in the high-V̇o_2max group and, accordingly, so too were the elevations in deep body temperature and whole body and local sweat rates (FIGURE 8). When the same groups exercised at a fixed absolute rate of heat production of ∼540 W, %V̇o_2max differed between the groups (40% vs. 58%), yet no differences were observed in deep body temperature and sweat rate (FIGURE 8). Nearly identical findings have been reported for treadmill exercise (60). In other studies, the role of %V̇o_2max on deep body temperature was assessed by using hypoxia to manipulate the within-subject relative intensity (388–390). Lowering the fraction of inspired oxygen reduces V̇o_2max (391), thereby raising the %V̇o_2max associated with a given absolute work intensity (i.e., heat production). Coombs et al. (388) had subjects cycle for 45 min at a work rate of ∼90 W (absolute heat production of ∼475 W) under normoxia and hypoxia (inspired oxygen fraction of 13%), representing 45% and 62% of normoxic and hypoxic V̇o_2max values. Despite the difference in %V̇o_2max, similar elevations in deep body temperature and sweat rate were found because heat production and E_req did not differ between trials. Conversely, when exercise was performed at ∼45% of normoxic and hypoxic V̇o_2max values, the rise in deep body temperature and sweat rate was lower in hypoxia versus normoxia despite an equal %V̇o_2max due to a much lower work rate (and heat production) required to achieve 45% of the hypoxic V̇o_2max.

FIGURE 8.

Changes in rectal temperature (ΔT_re; left) and whole body sweat loss (right) in groups of high (HI-V̇o_2max) and low (LO-V̇o_2max) aerobic capacity after 60 min of exercise at a relative intensity of ∼60% V̇o_2max and a fixed heat production of ∼540 W. Values within the bars represent the corresponding rate of heat production or relative exercise intensity. At 60% V̇o_2max, heat production was significantly higher in the HI-V̇o_2max group. At 540 W, relative intensity was significantly higher in the LO-V̇o_2max group. Redrawn from Ref. 314, with permission from the American Physiological Society.

It is important to note that, in these investigations, combinations of metabolic and ambient heat stress produced ω_req values low enough to ensure physiologically compensable environments and 100% sweat evaporative efficiency. Thus, the observed deep body temperature and sweat rate responses could not be explained by training-related differences in E_max and, thus, ω_req.

Collectively, these studies show that thermoregulatory responses are not set according to the relative exercise intensity (%V̇o_2max) under modest exercise-heat stress, but rather the rate of heat production, confirming the seminal findings of Christensen (46) and Nielsen (386) and those of subsequent investigators (300, 392, 393). It follows that comparisons of the thermoregulatory responses to dynamic exercise between groups of equivalent body size, irrespective of V̇o_2max, should be performed at the same absolute rate of heat production.

4.4. Biological Sex

4.4.1. Males versus females.

Whether biological sex modulates temperature regulation during heat stress depends on the question of interest. If one’s interest relates to sex as an independent modulator of temperature regulation, attempts must be made to control for typical biophysical differences between males and females. As a population, females display lower body weight, height, and absolute V̇o_2max compared with the male population. Each of these factors can impact the parameters of the heat balance equation, especially during exercise. In contrast, if one’s interest resides in determining whether females, as a population, display different thermoregulatory responses during heat stress relative to males (or vice versa), then accounting for these biophysical factors may not be necessary. However, potential sex-related differences in thermoregulatory responses cannot be ascribed to biological sex per se when using this approach.

Our initial understanding of potential sex-related differences in temperature regulation during heat stress came primarily from studies that employed exercise at a fixed percentage of V̇o_2max (394). When employing this approach, males and females will exercise at different metabolic rates unless absolute V̇o_2max (in L/min) is matched between groups. Most often, females will exercise at a relatively lower metabolic rate, and therefore metabolic heat production, compared with males. According to the heat balance equation, a lower rate of metabolic heat production requires less evaporative heat loss for a given air temperature (and therefore dry heat exchange). This explains initial observations of lower sweat rates during exercise in females, and such differences are entirely attributable to the lower rate of metabolic heat production at which they are exercising (167, 395). A second biophysical factor important to consider when studying males and females is body mass. If body mass is not matched, females are more likely to display a greater change in deep body temperature during exercise even if the absolute rates of metabolic heat production and total heat loss are the same as males (396). When biophysical factors are considered, sex-related differences in temperature regulation are minimized (397). Nonetheless, sweat distribution differs between males and females with males displaying a relatively greater distribution toward the torso compared with females and females displaying a relatively greater distribution toward the arms and hands compared with males (167). Furthermore, females display a lower maximal sweating capacity compared with males. This evidence comes from experiments that studied either males and females matched for body mass and/or fixed rate of metabolic heat production during exercise relative to body surface area. Importantly, females display similar sweating and evaporative heat loss as males at low to moderate E_req and only display lower sweat rates at a relatively high (≈300 W/m²) E_req (398, 399). To place these findings into perspective, a female with a V̇o_2max of 3 L/min would have to exercise on a cycle ergometer (20% efficiency) at combinations of air temperatures and percent V̇o_2max of 25°C/96%, 30°C/85%, 35°C/78%, 40°C/67%, and 45°C/61% to attain the E_req at which sweat rate is lower than males. This example highlights that although maximal sweat rate is lower in females, a combination of warm/hot air temperature and intense exercise is needed to observe a lower sweat rate in females compared with males (FIGURE 9). A lower postjunctional cholinergic sensitivity of the sweat glands may contribute to the lower maximal sweat rate observed in females (400–406), although we are unaware of a study that considered potential sex differences in the neural control of heat loss thermoeffectors. It is also worth noting that no differences in cutaneous vasodilation were observed between males and females in these studies, including cutaneous microvascular sensitivity to endothelial-dependent and -independent vasodilators (406).

FIGURE 9.

Sex differences in sweat rate relative to the required evaporation for heat balance (E_req) during exercise. Note how sex differences in sweat rate only occur at relatively high E_req (approximately ≥300 W/m²). Image created with BioRender.com with permission.

4.4.2. Menstrual cycle.

The fluctuation in female sex hormones that accompanies the menstrual cycle or hormonal contraceptive use is well known to alter basal deep body temperature (407–409). Specifically, basal deep body temperature is lower during the midfollicular phase of the menstrual cycle, or placebo phase of oral contraceptive use, when circulating estrogen concentration is high and unopposed by progesterone, whereas it increases during the midluteal phase of the menstrual cycle, or high hormone phase of oral contraceptive use, when progesterone and estrogen concentrations are high. These fluctuations in basal deep body temperature are likely mediated through the action of estrogen and progesterone on hypothalamic neurons involved in body temperature regulation. Animal studies have shown that progesterone inhibits (410), whereas estrogen enhances (411), the activity of warm-sensitive POA neurons. Furthermore, estrogen modulates the activity of MnPO neurons that control heat-defense responses in rodents (412). In humans, the fluctuation of basal deep body temperature over the course of the menstrual cycle is paralleled by similar fluctuations in the onset threshold for activation of heat loss thermoeffectors (174, 413–418). Although the shift in onset threshold for thermoeffector activation may be interpreted as an alteration in temperature regulation, this shift parallels the change in basal deep body temperature. Furthermore, the subsequent increase in thermoeffector output for a given change in deep body temperature is generally unaffected by fluctuations in female sex hormones (414–416, 418). Consequently, there is little evidence to suggest that temperature regulation is altered by the menstrual cycle (409, 419). Indeed, studies have demonstrated that whole body heat exchange is unaffected by menstrual cycle phase during passive heat stress (420) and exercise in the heat (421). Furthermore, Lei et al. demonstrated that thermoeffector (forearm blood flow and sweating) to mean body temperature relations are unaffected by menstrual cycle phase (422) or oral contraceptive use (423) in trained females during fixed-intensity exercise in warm/humid and warm/dry environments. They also demonstrated that exercise performance was unaffected by menstrual cycle phase or oral contraceptive use in both environments (422, 423). Taken together, the fluctuations in female sex hormones that occur over the course of the menstrual cycle do not appear to alter thermoregulatory control, even though they are associated with a shift in basal deep body temperature.

4.4.3. Pregnancy.

The effect of pregnancy on temperature regulation during exercise in the heat has received attention of late, motivated by recommendations that pregnant females should avoid exercise in warm/humid environments to minimize the risk of hyperthermia-related complications on the fetus (424). A recent study demonstrated that pregnant females display similar deep body temperature and sweating responses compared with nonpregnant counterparts during short (45 min) bouts of cycling exercise at fixed rates of metabolic production (350 W and 5 W/kg of body mass) in a warm/humid (32°C/45% RH) environment (425). A key feature of the study design was that body temperature and sweating responses were compared during two exercise bouts; one performed at a fixed rate of metabolic heat production and a second performed at a rate of metabolic heat production normalized to body mass to account for the greater body mass associated with pregnancy. Importantly, deep body temperature did not exceed 38°C in any of the pregnant female participants, a temperature that is at least 1°C lower than the hypothesized threshold (39°C) at which hyperthermia-related complications to the fetus may occur (424).

4.5. Age

Humans undergo numerous structural and functional changes across the life span. From birth until adulthood, these changes involve growth and development; with primary aging, there is a general deterioration of physiological function. In this section, we provide an overview of age-associated thermoregulatory function in the pediatric population and older individuals, using young adults as the reference point.

4.5.1. Pediatric population.

Pediatric age is defined as the period of life spanning birth to adulthood and comprises several subpopulations, including newborns (birth to 1 month of age), infants (1 month to 2 years), children (2–12 years), and adolescents (12–21 years); adolescents may be further divided into early (11–14 years), middle (15–17 years), and late (18–21 years) stages (426, 427). Throughout this period of growth and development, numerous anatomical and physiological changes take place, some of which could theoretically influence thermoregulatory function under heat stress (295, 428). Here, we examine whether these changes collectively yield actual child-adult differences in the thermoregulatory response to heat stress.

4.5.1.1. efficiency, economy, and heat production.

Inferior movement economy or mechanical efficiency translates into a higher rate of metabolic heat production for a given running speed or work rate (60). Notably, during weight-bearing exercise, children possess poorer movement economy compared with adults. Mass-specific V̇o₂ during treadmill walking or running is up to 24% greater in prepubertal children (4–13 years) compared with adolescents (14–18 years) and young adults (429–431). Consequently, prepubertal children performing weight-bearing activities at a fixed speed generate more heat per unit mass and have greater thermolytic requirements than adolescents and adults. The origin of this difference is likely biomechanical rather than physiological. Due to their shorter legs, children must maintain a faster stride rate to run at a particular velocity, which is energetically more demanding. If the metabolic cost is expressed per stride, no differences in running economy are evident between children and adults (432).

The conventional view is that gross efficiency during cycling (i.e., the percentage of total metabolic energy transformed into mechanical work) is not appreciably different between children, adolescents, and adults. Taylor et al. (433) reported mean gross efficiency values of 13.0%, 15.7%, and 12.9% for male children aged 6–8, 9–11, and 12–15 years of age at a work rate of ∼25 W. Similar observations were made in female children, albeit with slightly lower gross efficiency values than males (434). Unfortunately, no statistical analyses were performed to compare age groups in these two studies. Rowland et al. (435) found that delta efficiency (i.e., the change in metabolic rate for a given change in work rate) was not different between adults and children as work rate increased from 65 to 105 W. However, based on the reported metabolic data, V̇o₂ values were 0.15–0.25 L/min lower, and calculated gross efficiency values were 1.4–2.7 percentage points higher, in children versus adults across the range of workloads assessed, potentially reflecting the positive association between leg mass on internal metabolic power (i.e., the energy cost to keep the limbs moving) (436). Similar data have been reported more recently (437). Additionally, V̇o₂ and gross efficiency values were not statistically analyzed. Nevertheless, the fact that gross efficiency at low-to-moderate intensities is at least similar, if not better, in children compared with adults suggests that children should have similar or lower rates of heat production and thermolytic requirements for a given absolute work rate.

4.5.1.2. morphological considerations.

Growth in body size is the most conspicuous aspect of childhood development (TABLE 2). Body mass and stature increase from birth, with more rapid rates of growth in the first 1–2 years of life and again during puberty. Body mass continues to increase to adulthood, but stature tends to plateau during midadolescence. Increases in mass and stature lead to a larger body surface area; however, because the increase in mass is of greater magnitude than body surface area, surface area-to-mass ratio decreases during this time. Additionally, body fat percentage rises from ∼10 to 15% at birth to ∼25% by 2 years of age (441, 442). This rise in relative adiposity continues to adulthood in females, whereas in males, body fat percentage decreases at puberty, only to rise again in adulthood.

Table 2.

Anthropometric and body composition values during childhood, adolescence, and young adulthood from the National Health and Examination Survey and U.S. Centers for Disease Control growth tables

Age	Body Mass, kg		Stature, cm		Surface Area, m2		SA/Mass, cm2/kg		Body Fat, %
	M	F	M	F	M	F	M	F	M	F
0–2 mo	5.0	4.9	50.0	49.3	0.24	0.24	484	485
2 yr	13.7	13.0	91.1	89.7	0.57	0.56	419	427
8 yr	29.5	28.3	131.8	129.8	1.04	1.01	352	357	26.5	31.6
12 yr	46.4	52.1	153.9	154.3	1.41	1.48	304	285	26.4	31.4
18 yr	71.0	62.7	175.5	162.3	1.86	1.67	262	266	22.1	34.3
20–29 yr	81.3	69.5	176.0	162.6	1.97	1.74	243	251	26.1	37.8

Data represent 50th percentile values (438–440). Surface area calculated according to DuBois and DuBois (61). SA/mass, surface area-to-body mass ratio; M, male; F, female.

The relationships between morphology, heat exchange, and heat storage were detailed earlier in sect. 4.1 but will be briefly recapped here. For children, a lighter body mass indicates a smaller heat sink, while a smaller body surface area leads to a lower absolute thermolytic potential, assuming equivalent skin-air thermal and vapor pressure gradients. Since the number of functional eccrine sweat glands is proposed to be fixed from the age of 2 years (443), an increasingly larger surface area also means that the density of sweat glands declines until adolescence. With a greater surface area-to-mass ratio, children have a higher mass-specific rate of dry heat loss (lower E_req) if skin temperature exceeds ambient temperatures and greater mass-specific dry heat gain (higher E_req) if ambient temperature exceeds skin temperature (444). Body fatness is greater in adults versus children for females only (TABLE 2), but the difference between these groups is unlikely to have a meaningful effect on the average specific heat capacity of body tissues and thus deep body temperature responses to heat stress (see sect. 4.1.4).

4.5.1.3. physiological considerations.

Numerous studies have reported attenuated sweat rates in children compared with adolescents and adults (437, 445–454). To our knowledge, the neural control of body temperature during heat stress has not been compared between children and adults. Nonetheless, the vast majority of studies have shown that the onset of sweating is not different, or may be slightly earlier, in children (443, 450, 455–459). Furthermore, Shibasaki et al. (456) noted that the frequency of sweat expulsions was not different between children and adults for a given change in mean body temperature, but the local sweat rate on the chest, back, and thigh (though not the forearm) for a given frequency of sweat expulsion was attenuated in children. These observations suggest that any differences in the physiological control of sweating are of peripheral rather than neural origin. Sweat gland output is lower in children due to smaller glandular size (460) and lower responsiveness to cholinergic stimulation (461). Duct length and secretory coil area are, on average, 27% and 52% smaller, respectively, in children compared with adolescents (295, 460); duct length in adolescents is comparable to that observed in adults (462). Based on the positive association between glandular volume and methacholine-induced glandular sweat rate (364), it is possible that smaller gland size in children could limit sweat rate. However, since the number of functional sweat glands may be set by 2 years of age (443), and children have a smaller body surface area, attenuated sweat gland output may be offset by greater heat-activated sweat gland density (456). Thus, it is possible that children, despite inherent limitations in sweat gland output, can achieve the same level of skin sweat coverage (and evaporation) as adults (449). Such an effect could also contribute to better sweat evaporative efficiency during more intense heat exposures (453).

The circulatory response to heat stress in children has been previously thought to be inadequate (295). Although absolute blood volume and cardiac output are lower in children, blood volume scaled to surface area and cardiac index do not differ between children and adults (463, 464). No conclusive evidence exists to suggest that vasomotor control differs to any meaningful extent between children and adults, and skin and forearm blood flow responses to heat stress tend to be equivalent or even higher in children relative to adults (448, 451, 458, 463–465). The onset time to increased skin blood flow is shorter in children at the forearm (450, 451), but not the chest or back (450), during exercise. Meanwhile, a different pattern of responses, i.e., earlier onset on the chest and back, but not the forearm and thigh, occurs during lower limb hot-water immersion (458). With regard to vasomotor thermosensitivity, elevations in skin blood flow and cutaneous vascular conductance for a given change in deep body and mean skin temperatures are greater in children on the chest, but not the forearm, thigh, and forehead, and contradictory findings are observed on the back (458, 465). Maximal forearm vascular conductance at 42°C is higher in children 5–15 years of age compared with individuals up to 76–86 years (466), but maximal cutaneous vascular conductance at 44°C does not differ between boys (mean age: 9 years) and men (mean age: 21 years) (467). Whether the origin(s) of these discrepancies is physiological, or perhaps due to different measurement techniques, modes of heating, and/or slight differences in local skin temperature adjacent to the measurement sites, is unknown. Regardless, the balance of evidence suggests that deep body-to-shell heat transfer via the circulatory system is not impaired in children.

4.5.1.4. child-adult differences in the responses to heat stress.

Numerous studies have compared the thermoregulatory responses of children to adults during exercise-heat stress, the results of which are summarized in TABLE 3. Two patterns emerge from these data: 1) deep body temperature during exercise does not appreciably differ between children and adults, and 2) whole body sweat rates are consistently lower in children. The first point is best explained by the fact that, in general, the mass-specific rate of heat production did not differ greatly between children and adults. Although exercise was prescribed at a fixed relative intensity, which can lead to vastly different mass-specific rates of heat production if groups are unmatched for V̇o_2max (see sect. 4.3.2), children and adults in these studies were closely matched for this characteristic. Regarding the second point, relatively attenuated sweat rates in children typically corresponded with lower absolute and surface area-specific rates of heat production and, thus, E_req values. Based on the investigations performed to date, a physiological basis for child-adult differences in sweating cannot be confirmed.

Table 3.

Summary of findings comparing deep body temperature responses to exercise across stages of puberty and between children and adult

Study	Protocol	Environment	Sex	Age, yr	Hprod*				WBSR†
					W	W/m2	W/kg	ΔTdeep body, °C	g/h	g/(m2·h)
Wagner et al. (444)‡	Level-grade walking, 5.6 km/h	49°C, 17% RH	M	11–14	256	194	6.3	1.1	649	492
			M	25–30	417	219	5.6	1.0	1,106	580
Drinkwater et al. (463)	Intermittent walking, 30% V̇o2max	28°C, 45% RH	F	12.0 ± 0.0	193	157	5.5	0.9	173	141
			F	20.6 ± 1.6	314	181	4.8	0.9	287	165
		35°C, 65% RH	F	12.0 ± 0.0	193	157	5.5	1.2	354	288
			F	20.6 ± 1.6	314	181	4.8	1.2	513	295
		48°C, 10% RH	F	12.0 ± 0.0	193	157	5.5	1.3	531	432
			F	20.6 ± 1.6	314	181	4.8	1.3	1,002	576
Davies (445)	Level-grade running, 68% V̇o2max	21°C, 67% RH	F	13.8 ± 0.7	592	456	15.2	1.1	226	174
			M	12.9 ± 0.8	548	436	14.5	1.1	225	179
			M	36.1 ± 6.7	1016	590	16.1	1.6	567	329
Falk et al. (446)	Intermittent cycling, 50% V̇o2max	42°C, 20% RH	M	11.8 ± 1.1	307	249	8.3	1.0	366	296
			M	17.0 ± 0.3	624	343	9.3	1.3	712	392
Falk et al. (449)	Intermittent cycling, 50% V̇o2max	42°C, 20% RH	M	10.8 ± 0.8	320	272	9.2	0.8	372	317
			M	16.2 ± 0.4	647	370	10.2	0.9	673	385
Meyer et al. (454)	Intermittent cycling, 50% V̇o2max	41°C, 19% RH	F	9.1 ± 1.4	231	206	6.9	0.7	188	168
			F	21.4 ± 0.7	459	276	7.6	0.8	548	330
			M	9.1 ± 1.3	257	218	7.6	0.9	212	180
			M	23.4 ± 2.0	628	343	8.2	0.8	637	348
Shibasaki et al. (450)	Cycling, 40% V̇o2max	30°C, 45% RH	M	10 to 11	240	203	6.6	0.5	285	241
			M	21 to 25	486	276	7.2	0.5	587	333
Inbar et al. (453)§	Intermittent cycling, 50% V̇o2max	41°C, 21% RH	M	9.4 ± 1.7	235	174	8.0	1.0	342	326
			M	22.7 ± 2.3	560	317	7.8	1.3	837	445
Rivera-Brown et al. (464)	Cycling, 60% V̇o2max	33.7°C, ∼55% RH	F	11.3 ± 0.9	314	265	9.1	0.8	647	546
			F	26.8 ± 5.7	498	309	8.5	0.9	1,160	720
Rowland et al. (468)	Cycling, 65% V̇o2max	∼20°C, 62% RH	M	11.7 ± 0.4	448	311	9.5	0.5
			M	31.8 ± 2.0	748	380	9.1	0.5
		31°C, ∼54% RH	M	11.7 ± 0.4	466	324	9.9	0.6
			M	31.8 ± 2.0	738	375	9.0	0.6
Leites et al. (437)	Intermittent cycling, fixed Hprod	35°C, 35% RH	M	11.5 ± 1.3	234	126	5.7	0.6	273	223
			M	22.9 ± 2.3	397	213	5.6	0.7	612	329
			M	22.9 ± 2.3	234	175	3.3	0.3	387	209
Woloschuk et al. (451)	Cycling, 60% V̇o2max	∼24°C, ∼31% RH	M	9.7 ± 1.2	276	230	7.9		246	208
			M	22.2 ± 2.0	602	305	7.5		592	298

Values are means ± SD. H_prod, rate of metabolic heat production; ΔT_{deep body}, change in deep body temperature from baseline to end-exercise (the minimum exercise time); WBSR, whole body sweat rate; RH, relative humidity; M, male; F, female. *If not explicitly reported, heat production was calculated from oxygen uptake and work rate using conventional equations (33). For several studies (449, 454, 463,464, 468), calculated values represent metabolic rate, not H_prod, since external work rates were not reported. †For consistency, sweating responses are reported as absolute and surface area-specific WBSR based on reported sweat responses or evaporation rates, as well as total exposure time. ‡Data from exercise in a cool environment were omitted since baseline deep body temperature was not reported, preventing a conclusive determination of ΔT_{deep body}. §The H_prod for children was reportedly 8.6 W/kg; however, closer scrutiny of these data suggests a value approximating 8.0 W/kg.

Leites et al. (437) provided important insight into possible differences in thermoregulatory function between children and adults during exercise. Responses to intermittent cycling at 35°C were assessed in children and adults at a fixed absolute heat production of 234 W, as well as a fixed mass-specific heat production of 5.7 W/kg. When exercise was performed at the same absolute heat production, the rise in rectal temperature was greater in children (0.6°C) than adults (0.3°C) due to their lower body mass and thus higher mass-specific heat production. Absolute E_req was not different between groups since absolute heat production was matched and dry heat exchange was minimal in the 35°C environment. However, despite the similar E_req, whole body sweat rate was 30% lower in children. A lower sweat output to meet the thermolytic requirement for heat balance (i.e., better sweat evaporative efficiency) implies a thermoregulatory advantage in children. Conversely, a lower sweat rate in the context of a higher change in mean body temperature could be viewed as a disadvantage. In the 5.7 W/kg trial, the rise in rectal temperature and mean skin temperature (and thus mean body temperature) was not different between groups (0.6°C). Because of differences in body size, E_req was greater in adults (353 W or 190 W/m²) compared with children (201 W or 109 W/m²), leading to higher absolute and area-specific whole body sweat rates in adults. Because mass-specific heat production and the rise in deep body temperature were similar, and dry heat losses were minimal in the 35°C environment, mass-specific evaporative heat loss was likely similar in children and adults. However, mass-specific whole body sweat losses were, in fact, lower in children (11.8 g of sweat/kg) compared with adults (15.8 g of sweat/kg), indicating enhanced sweating efficiency. This latter finding supports previous work by Inbar et al. (453), who estimated higher mass-specific evaporation rates in children for an equivalent mass-specific whole body sweat rate. Based on the similar mean body temperatures and mass-specific evaporative requirements, but lower mass-specific whole body sweat rates in children (437), it can be concluded that children do not exhibit any differences in temperature regulation and possess better sweat evaporative efficiency compared with adults during exercise-heat stress.

Differences in local sweat rates, and thus sweat distribution, have been investigated in a few studies and often only at a small number of measurements sites. Early work reported by Kuno (443) described a “scanty” level of sweating on the limbs that was more common in children than adults, as well as a higher palmar sweat response in children, based on colorimetric starch-iodine assessments of sweating. Decades later, Shibasaki et al. (456) and Inoue et al. (465), using ventilated capsules, reported lesser increases in local sweat rate on the chest and thigh, but no differences on the back, forearm, and forehead, between children and adults under passive heat stress. Recently, Arlegui and colleagues (469) produced a comprehensive map of sweat distribution in prepubertal children, compiled from measurements of local sweat rate at 31 sites. Following 30 min of intermittent exercise at a 30°C air temperature, local sweat rates were assessed using technical absorbent patches. Absolute local sweat rates (in g·m⁻²·h⁻¹) were then converted to ratio values (i.e., normalized to the area-weighted total sweat rate from all patches), to enable a comparison of sweat distribution against existing sweat maps from young adults (164). In children, the highest sweat rates were found on the forehead, followed by the feet, hands, forearms, and on the upper back. A slightly different pattern was observed in adults, with the highest local sweat rates observed on the forehead, upper back, and lower back in adults, and relatively lower sweat rates on the hands and feet compared with other sites. Notably, the absolute local sweat rates of the hands and feet were similar in children and adults, suggesting that developmental changes in sweat distribution occur in areas of the body outside of the hands and feet.

During passive heat stress, deep body temperature is generally comparable between prepubertal children aged 6–12 years and adults (455, 456, 458, 459). A noteworthy exception is the study by Tsuzuki-Hayakawa et al. (457), who noted an ∼0.15°C elevation in rectal temperature among young children aged 9 months to 4.5 years during a 30-min exposure to 35°C, 70% RH environment, whereas no change was observed in their mothers. There is no conspicuous explanation for this difference, since whole body sweat and evaporation rates were higher in the children and dry heat loss was likely negligible due to a narrow mean skin-air temperature gradient in both groups (metabolic rate was not reported). Area-specific whole body sweat and evaporation rates are similar or higher in children (455, 457, 459), arguing against a possible peripheral deficiency in sudomotor function associated with pediatric age during passive heat stress.

4.5.1.5. heat adaptation in children.

To our knowledge, the study by Wagner et al. (444) is the only one that directly compared the magnitude of adaptation to heat stress between children and adults. Prepubertal boys (11–14 years) and men (25–30 years) completed 45–90 min of treadmill walking in 49°C and 17% RH conditions for 8 consecutive days. On days 1 and 8, absolute rectal temperature values were higher in the boys at baseline and during exercise. However, the magnitude of reduction in rectal temperature at baseline and at 40 min (the minimum exercise time on day 1) following heat acclimation was −0.2°C and −0.5°C, respectively, in both groups. Similarly, mean skin temperature was higher in the boys than men, but the reduction in mean skin temperature at 40 min of exercise following acclimation was similar between groups (boys: −1.0°C; men: −0.9°C). Evaporative heat loss increased by an estimated 48 W/m² in the boys and 24 W/m² in the men. The difference in the magnitude of the improvement in evaporation may have been due, at least in part, to the improved walking economy in the men, resulting in a lower rate of metabolic heat production postacclimation (heat production was unchanged in the boys). The reduction in heart rate at 40 min of exercise (∼15–16 beats/min) was nearly identical between groups. The coefficient of deep body-to-skin heat conductance, which represents the requirement for skin blood flow, decreased with acclimation in both groups [boys: −42 W/(m²·°C); men: −24 W/(m²·°C)]. Since deep body and skin temperatures changed similarly in both groups with acclimation, the magnitude of the difference in conductance between boys and men reflects the reduction in heat production in the men. Although subtle differences in the acclimation response were found, the findings of Wagner et al. (444) suggest that prepubertal boys are able to adapt to a similar degree and at a similar rate as adult men. The absence of similar data in prepubertal girls needs to be addressed.

4.5.2. Older adults.

Heat-related hospital admissions and deaths occur predominantly in older adults (378, 470–472). The etiology underlying greater heat vulnerability in older adults is likely multifactorial (473). Nonetheless, several studies have considered the effect of age on temperature regulation as a contributing factor and these studies support the general contention that age is associated with reduced thermoeffector output during heat stress.

Since a seminal study published in the late 1980s (474), research has overwhelmingly demonstrated that older adults display lower cutaneous vasodilation during heat stress (475). The first mechanism considered to explain this observation was reduced nitric oxide-dependent dilation. Interestingly, the contribution of nitric oxide to cutaneous vasodilation during heat stress was shown to be greater in older adults, highlighting a potential role for impaired cotransmitter-mediated dilation within this population (476). Subsequent studies effectively demonstrated impaired acetylcholine-mediated, cyclooxygenase-dependent and platelet-mediated dilation in older adults (377). The greater reliance on nitric oxide also implies that impairments in nitric oxide availability that accompany aging play a large role in mediating reduced cutaneous vasodilation during heat stress. Studies have shown that cutaneous nitric oxide availability is reduced in older adults due to increased arginase activity (477, 478), greater oxidative stress (478), and reduced endothelial nitric oxide cofactor availability (479, 480). Interestingly, the mechanisms mediating lower nitric oxide-mediated cutaneous vasodilation in older adults may be specific to the heating modality. For example, Meade et al. (481) observed that arginase inhibition increases cutaneous vasodilation in older adults during passive heating but not during exercise in the heat. Furthermore, Fujii et al. (482) observed that intradermal ascorbate infusion does not augment cutaneous vasodilation in older adults during exercise in the heat. Nonetheless, cutaneous vasodilation during passive heat stress is completely restored in older adults when nitric oxide availability is augmented either locally via microdialysis (480) or systemically via oral supplementation of folic acid (483). In addition to these altered mechanisms of cutaneous vasodilation, other studies have shown that older adults display reduced maximal forearm vascular conductance during local heating, suggestive of alterations in cutaneous microvascular structure (466, 484). Finally, one study also observed that older adults display an attenuated increase in SSNA relative to younger adults during heat stress that correlates with lower levels of cutaneous vasodilation (153). Taken together, attenuated cutaneous vasodilation during heat stress in older adults can be ascribed to reduced efferent neural outflow, impaired cutaneous vasodilatory signaling, and altered cutaneous microvascular structure.

The functional impact of reduced cutaneous vasodilation for temperature regulation is a lower skin blood flow during heat stress in older adults. When considering such skin blood flow responses, it is important to appraise the contributing role of age-related differences in the cardiovascular response to heat stress (66, 485). Minson et al. (486) first demonstrated that older adults display a blunted increase in cardiac output during heat stress that, combined with reduced blood flow redistribution from the renal and splanchnic circulations, resulted in 3.1 L/min less blood flow directed toward the skin circulation. Recent studies attempted to determine whether a blunted increase in cardiac output contributes to a lower skin blood flow during heat stress in older adults or whether reduced cutaneous vasodilation contributes to a blunted increase in cardiac output. These studies have provided evidence both for (487) and against (488) the possibility that cardiac function limits the increase in skin blood flow that older adults can achieve during heat stress. However, the interpretation of these studies is difficult due to inherent limitations associated with trying to determine a potential relationship between cardiac output and cutaneous vasodilation/skin blood flow during heat stress (485). Regardless, the ultimate consequence of reduced cutaneous vasodilation and lower skin blood flow appears minimal in the context of whole body heat exchange. Although less skin blood flow will alter the distribution of heat between deep body tissues and the skin surface, older adults may display lower skin temperatures that result in a greater deep body-to-skin temperature gradient that offsets lower internal convective heat transfer (489, 490). However, this observation is not universal as several other studies observed no difference in mean skin temperature between young and older adults (491–495), whereas regional skin temperatures are greater in older adults in body segments associated with lower sweating (496). When interpreting these differences, it is important to consider that the magnitude of difference, if any, in skin temperature is small (typically ≤1°C). Consequently, such differences have a minimal impact on dry heat exchange (492–494, 497). That said, some studies suggest that small differences in heat exchange between younger and older adults may become important when cumulated over prolonged exposures. Specifically, older adults gain ∼10 W more heat from the environment during passive exposure to hot (36.5°C) or very hot (44°C) environments resulting in a greater accumulation of body heat and a potentially greater rise of deep body temperature (493, 494). Nonetheless, attempts to increase skin blood flow during heat stress in older adults have not resulted in improved temperature regulation (272). Taken together, these observations highlight the fact that a reduced sweating capacity (discussed below) is arguably the most important factor underlying impaired temperature regulation during heat stress in older adults.

It is generally accepted that older age is associated with a lower sweating capacity during both passive heat stress and exercise in the heat (473, 498–500). During passive exposure to a hot environment, a generalized reduction in sweat rate across the body surface is seen in older males (496). In contrast, a preferential reduction of sweating on the lower limbs contributes to the lower sweating capacity of older males during exercise in the heat (496). Recent studies have shown that age-related reductions in sweat rate are most evident at high E_req values. These studies employed an intermittent exercise model performed at progressively greater exercise intensities and therefore greater E_req, in warm or hot/dry environments (35 to 40°C/15 to 20% RH). With this model, whole body sweat rate is similar between younger (≈21 years), middle-aged (≈49 years), and older (≈65 years) males at a relatively low E_req (≈400 W), whereas whole body sweat rate was lower in sedentary middle-aged and older males at moderate to high (≈500 to 600 W) E_req (495). These studies also demonstrated that a lower whole body sweat rate is evident in males as young as 40–45 years of age (501) and that a linear reduction in whole body sweat rate is observed when a cross-sectional comparison is performed between males aged 20 to 70 years (502). Similar age-related reductions in whole body sweat rate have also been demonstrated in females (503, 504).

The reduction in sweating capacity is primarily the result of a lower sweat output per gland, rather than a reduced number of activated sweat glands (332, 505–507). Older adults generally display a reduced postjunctional sensitivity of the eccrine sweat gland to cholinergic agonists (332, 401–403, 505), albeit this is not always observed (508). Other studies have also provided evidence of a reduced contribution of nitric oxide and cyclooxygenase in mediating local sweat rate during exercise in older males (509, 510). Specifically, intradermal infusion of a nitric oxide synthase or cyclooxygenase inhibitor does not alter local sweat rate in older males (509–511) whereas these agents can reduce local sweat rate during single (222) or repeated (223, 511) exercise bouts in younger males. Lastly, it is likely that the aforementioned attenuated increase in SSNA observed in older adults during passive heat stress (153) also contributes to the age-related reduction in sweating capacity. However, this remains to be determined since sweat rate was not measured in that study. Taken together, older adults display a reduced sweating capacity compared with younger adults that may be attributed to reduced efferent neural outflow, reduced cholinergic sensitivity of eccrine sweat glands, absent contribution of cotransmitters that sensitize the sweating response, and lower sweat gland output. Practically, the reduced sweating capacity limits evaporative heat loss potential in older adults thereby resulting in a lower critical limit at which the combination of exercise intensity and environmental conditions (temperature, humidity, airflow) become uncompensable (512). A lower critical environmental limit for uncompensability in older adults may be especially important to consider within the context of establishing safe exposure limits for work in the heat within this population (513).

It should be noted that no study has considered whether alterations in thermoafferent signaling and/or integration of thermoafferent flow also contribute to the reduced thermoeffector output observed in older adults during heat stress. Some studies have reported a reduced perception of warmth stimuli applied to the skin surface in older adults (491, 514, 515). Considering the perception of warmth occurs through the same cutaneous thermoreceptors as those that mediate autonomic heat loss thermoeffectors (516), it is tempting to suggest that altered thermoafferent signaling and/or integration of thermoafferent flow contributes to the reduced thermoeffector output observed in older adults during heat stress. However, an important caveat to this possibility is that cutaneous thermal sensitivity mediates thermoregulatory behaviors via the spinothalamocortical pathway, whereas autonomic thermoeffector responses are mediated via the previously described LPBd-MnPO pathway (sect. 3.2). Furthermore, deep body temperature is the primary (regulated) variable that drives autonomic heat loss thermoeffectors, whereas skin temperature provides auxiliary feedback (as discussed in sect. 3.1). It therefore remains unknown whether age is associated with altered thermoafferent signaling and/or integration of thermoafferent flow in the context of autonomic thermoeffector responses.

When interpreting the literature pertaining to age-related changes in temperature regulation during heat stress, it is important to consider that almost all studies employed a cross-sectional design. Our current knowledge therefore relates to the effect of age, rather than aging per se. That said, the few longitudinal studies performed support the contention from cross-sectional studies that sweating capacity decreases with aging (507, 517). It is also important to consider the potential confounding effect of a reduction in fitness (V̇o_2max) that typically accompanies aging. The effect of fitness on age-related differences in temperature regulation has been primarily studied with the use of cross-sectional comparisons between younger and older fit/unfit participants during exercise in the heat. With the use of this approach, studies have demonstrated that fit older adults display improved heat loss thermoeffector output relative to unfit counterparts (518, 519), and/or that thermoeffector output correlates with V̇o_2max but not age (84, 519, 520). When interpreting such studies, it is important to consider whether metabolic heat production differed between groups during exercise given that several studies employed protocols based on exercise at a fixed percentage of V̇o_2max. Since older adults, as a population, display a lower V̇o_2max, this approach results in a lower rate of metabolic heat production during exercise in unfit older adults relative to their fit counterparts and relative to younger adults (518). Therefore, a greater thermoeffector output in fit older adults may simply reflect the fact that they exercised at a greater rate of metabolic heat production. Such a scenario makes it difficult to determine if a greater thermoeffector output in fit older adults reflects improved temperature regulation relative to their unfit counterparts. This scenario also makes it difficult to determine if fit older adults display preserved heat loss thermoeffector output relative to younger adults. This conclusion often relies on the observation that unfit older adults display a lower thermoeffector output relative to younger counterparts. However, the lower thermoeffector output observed in unfit older adults may simply reflect the lower rate of metabolic production at which they exercise. When rate of metabolic heat production is matched between young and older adults, one study observed that whole body sweat rate is preserved in middle-aged males (495), whereas other studies observed a lower whole body sweat rate in middle-aged females matched for fitness with younger counterparts (497, 503, 521). A retrospective analysis of these studies, nonetheless, suggests that greater fitness may curtail the reduction in sweating capacity associated with age, especially at high E_req (522). A greater postjunctional cholinergic sensitivity of the eccrine sweat gland in physically active older adults may underlie this possibility (403). Beyond cross-sectional comparisons, the few studies that considered the effect of exercise training have reported mixed results (368, 523, 524). One adaptation that may underlie the variable effects of fitness and exercise training on heat loss thermoeffector output is blood volume (525). Studies have observed that aerobic exercise training combined with carbohydrate and protein supplementation increases blood volume and improves heat loss thermoeffector output during exercise in older healthy (526) and hypertensive (527) males, whereas exercise training without protein supplementation does not alter these variables. Greater blood volume may improve thermoeffector output via attenuation of cardiopulmonary baroreceptor unloading that accompanies heat stress (528). That said, cardiopulmonary baroreceptors have only been shown to modulate cutaneous vasodilation (529) but not sweat rate (530). Consequently, the mechanism(s) underlying the improvement of sweat rate following exercise training combined with carbohydrate and protein supplementation in older adults remains to be identified.

4.5.2.1. menopause.

Considering the cyclic variations in deep body temperature over the course of the menstrual cycle, some studies have considered the impact of the loss of estrogens following menopause on temperature regulation during heat stress. A first study demonstrated that 2–3 weeks of estrogen replacement therapy reduces basal deep body temperature without affecting the change in deep body temperature and heat loss thermoeffector output during subsequent exercise in the heat (531). Subsequently, thermoeffector output during exercise in the heat was compared between groups of postmenopausal females not receiving hormonal replacement therapy, receiving estrogen replacement therapy, and receiving estrogen + progesterone replacement therapy for a minimum of 2 years (532). Basal deep body temperature was again observed to be lower in the group of females receiving estrogen replacement therapy, but not combined estrogen + progesterone therapy relative to females not receiving any hormone replacement therapy. Nonetheless, the changes in deep body temperature, cutaneous vasodilation, and forearm blood flow were similar between groups during subsequent exercise in the heat. Similar results were also observed during whole body passive heating with the use of the water-perfused suit model (533). Taken together, these results suggest that beyond resulting in a slightly elevated basal deep body temperature, the loss of estrogens following menopause does not modulate heat loss thermoeffector output during heat stress.

5. TEMPERATURE REGULATION IN DISEASE

In the remaining sections of this review, our attention turns to disordered human temperature regulation, beginning with the effects of various diseases. The diseases selected encompass a range of neurological, cardiovascular, and hereditary disorders that are commonly associated with heat intolerance or heat-related health complications. As for intrinsic modulators of temperature regulation, our current understanding of how disease modulates temperature regulation during heat stress also relies predominantly on measurements of deep body temperature and thermoeffector output. When available, we highlight studies that considered postjunctional responses of thermoeffector organs and SSNA. Whether the diseases discussed below modulate thermoafferent signaling and/or integration of thermoafferent flow remains generally unknown.

5.1. Neurological Disorders

Pathologies within autonomic structures of the central or peripheral nervous systems can disturb thermal sensory and effector responses, leading to heat intolerance and a higher risk of heat-related illness. More serious thermoregulatory impairments would be expected with damage to preganglionic segments of the autonomic neuroaxis or widespread pathology in postganglionic sudomotor neurons, as the skin surface area that could still participate in evaporative heat dissipation (i.e., the effective surface area) would be most profoundly reduced. In some conditions, this problem may be exacerbated by the use of certain medications (e.g., anticholinergics) (534).

5.1.1. Multiple sclerosis.

Multiple sclerosis (MS) is an autoimmune disease of the CNS that affects an estimated ∼2.5 million people worldwide. The pathophysiology of MS involves inflammatory demyelination and axonal loss within the white and gray matter regions of the brain and spinal cord. Myelin serves as an electrical insulator, raising the transverse resistance and lowering the capacitance of ensheathed axons to facilitate rapid and high-fidelity action potential propagation between nodes of Ranvier. Demyelination and plaque formation reduce axonal insulation, resulting in slower conduction velocity or even conduction block (535, 536). Disrupted neuronal signal transmission leads to physical disability, cognitive impairment, and autonomic dysfunction, including thermoregulatory disturbances.

To assess the extent of autonomic dysfunction in MS on thermoregulatory responses, early studies examined sudomotor responses using a “thermoregulatory sweat test” (537–539). During such a test, an individual is passively heated to evoke profuse sweating, after which an indicator powder (e.g., quinizarin or alizarin) is spread across the body surface to visualize the distribution of sweating and estimate the degree of anhidrosis (540, 541). Using this approach, Noronha et al. (538) demonstrated that a ∼1.1°C rise in oral temperature produced an abnormal distribution of sweating in 25/60 MS patients tested, with sweating completely absent (2/60) or absent on the limbs (14/60), below the waist (7/60), or everywhere except the face (2/60). Greater reductions in the sweat response were generally associated with more severe disease progression.

More recent studies have employed quantitative methods to assess sweating responses in individuals with MS. Saari et al. (542) reported that following a 15-min bout of torso heating (≤0.2°C rise in tympanic temperature), transepidermal water loss on the feet was attenuated in individuals with MS compared with control participants. Allen and colleagues (15), using a passive heating model (water-perfused suit with 48°C water), showed that local sweat rate on the forearm was attenuated by an average of 0.24 mg·cm⁻²·min⁻¹ (∼35%) in individuals with MS at gastrointestinal temperatures of 0.7°C and 0.8°C above baseline. In a subsequent exercise study (543), the same authors found that whole body sweat loss was ∼90 g (or ∼31%) lower in individuals with MS following 60 min of cycling at an external work rate of 70 W in 25°C and 35% RH conditions. The lower whole body sweat loss was explained by a reduced thermosensitivity of forearm and upper back local sweat rates but did not translate into significantly greater rectal and esophageal temperatures in the MS group.

Cutaneous vascular responses to heat stress have also been investigated in individuals with MS. In the study by Noronha et al. (538), the authors noted that erythema was almost always present in anhidrotic skin regions of passively heated MS patients. In the studies by Allen and colleagues (15, 543), MS and control groups displayed similar skin blood flow and cutaneous vascular conductance for a given rise in gastrointestinal temperature, suggesting intact vasomotor control. Furthermore, no differences in peak cutaneous vascular conductance were observed during local heating, together indicating that cutaneous perfusion was unaffected by MS.

The available evidence, albeit from few studies, suggests that in individuals with MS, sudomotor, but not vasomotor, responses to heat stress can be impaired, but this impairment does not necessarily lead to significantly greater elevations in deep body temperature under mild heat stress (e.g., low-intensity exercise in temperate ambient conditions). Whether sudomotor dysfunction associated with MS leads to greater elevations in deep body temperature with exposure to higher heat loads (e.g., intense exercise, warmer air temperatures), and whether sudomotor function can be enhanced via acclimatization/acclimation, are open questions. However, obtaining such information may be limited by Uhtoff’s phenomenon, characterized by episodes of symptom worsening associated with elevated deep body temperatures common in MS (discussed below).

An attenuated sweat response to heat stress in an individual with MS is presumed to be a consequence of impaired signaling along thermoeffector loops due to lesions in the CNS. Demyelination and axon loss have been documented in the hypothalamus, brainstem, and pathways that synapse on intermediolateral column cells of the spinal cord (544, 545), which could disrupt the transmission and integration of thermoafferent information, and/or efferent signaling. Quantitative sensory testing has revealed abnormal sensory responses to warm stimuli in up to 58% of individuals with confirmed MS (546–549), which could alter behavioral thermoregulatory responses (550). However, it should perhaps not be assumed that MS-associated CNS lesions affecting thermal sensation necessarily impair autonomic control of deep body temperature. This is because transmission of thermal afferents could conceivably be disrupted by lesions along signaling pathways between the thalamus and sensory cortex, affecting temperature perception, but not the parallel LPBd-MnPO pathway, which influences autonomic temperature regulation (126). Whether SSNA is altered by MS has not been assessed, but the observation of depressed resting muscle sympathetic nerve activity (551) suggests sympathetic outflow may be compromised in this population (552).

Thermoregulatory dysfunction may be particularly consequential for the 60–80% of individuals with MS who suffer from Uhtoff’s phenomenon, a transient worsening of symptoms during heat stress. Even slight elevations in body temperatures induced by exercise, hot air, sun exposure, or hot showers and baths can lead to sensorimotor impairments that most commonly include visual impairments (diminished acuity and color perception, double vision), sensory loss, fatigue, muscle weakness, or loss of motor control (553–555). Davis (535) proposed the presently-accepted hypothesis that elevations in temperature lead to conduction blockade in demyelinated axons by reducing the conduction safety factor, which represents the ratio between the current generated by a nerve impulse and the current required to maintain conduction (in other words, the reserve capacity of the current) (556). The safety factor is estimated to be 3–7 in myelinated axons (557) but declines if the amplitude of the stimulating current is reduced and/or if the current required to initiate an action potential increases. Following demyelination, greater axolemmal exposure disperses current and increases axonal capacitance, thereby reducing the safety factor (558). Compounding this problem is the heightened expression of Na⁺ channels in demyelinated areas of the axolemma. These Na⁺ channels close at higher temperatures, disrupting action potential generation (559). As a result, the combination of demyelination and an elevated temperature greatly increases the probability of conduction slowing or conduction block. It follows that the temperature at which conduction block occurs would be lower with greater demyelination (560).

5.1.2. Synucleinopathies.

Synucleinopathies are a group of adult-onset neurodegenerative diseases that includes Parkinson’s disease (PD) and dementia with Lewy bodies (DLB), as well as the rarer conditions of multiple system atrophy (MSA) and pure autonomic failure (PAF). The underlying neuropathology involves the abnormal aggregation of the protein α-synuclein within neurons (Lewy bodies or Lewy neurites) and oligodendroglial cells (561, 562), resulting in dysfunction and cell loss. Parkinsonism (tremor, bradykinesia, muscle stiffness, and postural instability), sleep disturbances, and cognitive impairment are key features of PD, DLB, and MSA, whereas orthostatic hypotension and urinogenital issues define PAF. Moreover, deposits of α-synuclein in central and/or peripheral autonomic structures can lead to sudomotor dysregulation.

As with other neurological disorders, the nature and extent of sudomotor dysfunction in synucleinopathies can be evaluated using a combination of the thermoregulatory sweat test (described earlier) and the quantitative sudomotor axon reflex test (QSART), which assesses the axon reflex in postganglionic sudomotor axons (541). Such tests can provide important insight into the potential for compromised deep body temperature regulation. For QSART, acetylcholine is iontophoresed at several skin sites on the forearm, leg, and foot. A sudomotor axon terminal is stimulated, creating a signal that travels antidromically down the axon and then orthodromically to a nearby sweat gland that leads to sweat secretion. An attenuated or absent sweat volume indicates postganglionic sudomotor axon damage or neuropathy. Combining the thermoregulatory sweat test and QSART can localize an autonomic lesion. Anhidrosis with a normal QSART result indicates a preganglionic lesion, whereas anhidrosis with a negative QSART result indicates a postganglionic lesion. A mixture of pre- and postganglionic lesions is represented by anhidrosis and an attenuated QSART volume or if pre- and postganglionic patterns occur at different QSART sites.

During thermoregulatory sweat testing (563–565) or exposure to hot-humid conditions (566), individuals with PD tend to display either a normal pattern of sweating or anhidrosis on the distal limbs. Larger areas of anhidrosis occur on the neck, trunk, and limbs, but anhidrosis rarely exceeds 40% of the frontal surface area (567–570), such anhidrosis could be due to anticholinergic medications common in these patients (567, 568, 571). Compensatory hyperhidrosis on the face has been observed in conjunction with larger areas of anhidrosis (567, 568). Reduced QSART volumes are typically observed in a length-dependent pattern, that is, reduced QSART volumes primarily in the most distal segments of the lower extremities (572). Coupled with a tendency for distal anhidrosis, the origin of sudomotor dysfunction in PD, at least during early stages, is a length-dependent neuropathy in postganglionic sudomotor axons. Indeed, sweat gland denervation is more prevalent in individuals with PD (573) and may be the result of α-synuclein aggregation in peripheral neurons (574) or other mechanisms (575). Although early stage sudomotor dysfunction appears to be relatively mild, the magnitude of sudomotor impairment tends to increase in later stages of the disease (564, 572). Such an observation may explain why approximately one-third of individuals with PD report sweating less than other people or feeling exhausted in hot weather (576). Moreover, there is some evidence to suggest that hospitalization of individuals with PD increases during heat waves (1, 577).

While autonomic dysfunction tends to be more severe in DLB, the pattern of sudomotor dysfunction is similar to that observed with PD. In a comparison of autonomic function tests in DLB and PD patients, Thaissetthawatkul et al. (563) found no differences in the number of abnormal thermoregulatory sweat test results (DLB: 12/13; PD: 5/7) and the mean percentage of anhidrosis (DLB: 26.4%; PD: 19.3%), and a distal pattern of sudomotor dysfunction was most common in both groups. Furthermore, the severity of the QSART-derived deficit correlated with the percentage of anhidrosis, suggesting the source of sudomotor dysfunction is also peripheral in DLB, which is supported by the presence of α-synuclein deposits in autonomic skin nerves of individuals with DLB (578).

In MSA, motor dysfunction and autonomic dysfunction (orthostatic hypotension, genitourinary symptoms) are the main clinical features (562, 579). The magnitude of anhidrosis is typically far greater in MSA than PD or DLB, with mean values of ∼65% or higher anhidrosis during thermoregulatory sweat tests (564, 565, 570, 580–582). Normal sweating responses occur in only a few individuals with MSA (582). Based on QSART results, anhidrosis may be caused by preganglionic lesions, postganglionic lesions, or a mixture of the two, with greater postganglionic involvement in later stages of the disease (582, 583). This pattern of sudomotor dysfunction is in line with α-synuclein accumulation in oligodendroglial cells and neurodegeneration across the CNS (584), as well as peripheral denervation of sweat glands (585). Widespread anhidrosis likely underlies reports of heat intolerance and heat stroke in this population (582, 586).

In contrast to the other synucleinopathies, PAF is characterized by autonomic dysfunction without motor impairment. The underlying pathology involves α-synuclein inclusions in sympathetic ganglia and autonomic neurons, as well as degradation of peripheral autonomic neurons (587). Anhidrosis over a large fraction of the skin surface is commonly observed (587–590). In a recent study (590), the median percentage anhidrosis on thermoregulatory sweat testing was 58% in individuals with stable PAF. Interestingly, 45% of these individuals exhibited a preganglionic pattern of sweat loss, indicative of a central lesion.

5.1.3. Peripheral nervous system disorders.

Disorders of the peripheral nervous system involving preganglionic nerves, sympathetic ganglia, or postganglionic nerves can disrupt efferent thermoregulatory signaling between the CNS and effector targets in the skin, resulting in severe anhidrosis and heat intolerance. The causes for such disease states are wide ranging and include metabolic (e.g., diabetic neuropathy), genetic (e.g., Fabry disease), autoimmune (e.g., Sjögren’s syndrome, autoimmune autonomic ganglionopathy), and bacterial or viral infection (e.g., human immunodeficiency virus, leprosy). Because several comprehensive reviews on peripheral nervous system disorders and thermoeffector dysfunction have been recently published (534, 591–593), we refer to the reader to these excellent publications for summaries on the topic.

5.2. Cardiovascular Diseases

Extreme heat is consistently associated with a greater risk of cardiovascular mortality, and to a lesser extent hospital admissions (378, 470–472, 594). Furthermore, individuals with cardiovascular disease are at a greater risk of adverse cardiovascular outcomes during extreme heat events (595). Considering these observations, it is somewhat surprising that only a few studies have examined whether cardiovascular diseases may alter temperature regulation. These studies have focused on the effect of hypertension, heart failure with reduced ejection fraction, and chronic coronary artery disease.

5.2.1. Hypertension.

The impact of hypertension on temperature regulation during heat stress was first considered within the context of exercise in the heat. An initial study (596–598) observed that middle-aged hypertensive males (≈44 years, unmedicated) display a blunted increase in forearm blood flow during moderate-intensity exercise in a 38°C environment, despite elevated levels of arterial blood pressure, suggestive of reduced cutaneous vasodilation. Furthermore, the slope of the forearm blood flow-deep body temperature relationship was reduced in hypertensive males (597). Although reduced forearm blood flow was associated with a lower deep body-to-skin temperature gradient (598), sweating and the increase in deep body temperature during exercise were similar between groups suggesting that lower forearm blood flow did not meaningfully alter heat exchange with the environment in hypertensive males. Subsequent studies delved further into the potential mechanisms mediating the blunted forearm blood flow observed in hypertensive individuals. With the use of whole body heating combined with intradermal microdialysis, these studies established that middle-aged adults with hypertension (≈57 years, unmedicated) display a reduced cutaneous vasodilation during heat stress that is mediated through reduced nitric oxide-dependent dilation (599). The reduced nitric oxide-mediated dilation was attributed to an upregulation of arginase activity (599) and greater oxidative stress (600). The functional implications of these findings in terms of temperature regulation remain uncertain, as recent studies demonstrate that middle-aged adults with hypertension (≈64 years, medicated) display similar whole body heat exchange and deep body temperature relative to normotensive counterparts during exercise in the heat (601). It is interesting to note that despite seemingly unaltered temperature regulation, hypertensive individuals nonetheless display a reduced tolerance time during simulated work in the heat relative to age-matched healthy controls (602). This observation may be explained by lower fitness (V̇o_2peak) levels in the hypertensive individuals.

5.2.2. Heart failure with reduced ejection fraction.

Studies have examined temperature regulation in individuals with heart failure with reduced ejection fraction during passive heat stress and exercise in the heat. A first study compared thermoregulatory responses between individuals with heart failure and age-matched healthy controls by employing the water-perfused suit model of heat stress (16). By design, the increase in deep body temperature was matched between groups. Individuals with heart failure displayed less cutaneous vasodilation and forearm blood flow but preserved local sweat rate during heat stress. The lower cutaneous vasodilation was paralleled by a lower maximal capacity of forearm skin to vasodilate, suggestive of altered vasodilator responsiveness and/or structure of the cutaneous microvasculature. A second study published nearly at the same time (603) demonstrated that individuals with heart failure also display less cutaneous vasodilation during a 90-min exposure to a hot/humid environment (38°C/50% RH). This study further demonstrated that cutaneous vasodilation was reduced to a greater extent by infusion of a nitric oxide synthase inhibitor (via microdialysis) in healthy controls compared with individuals with heart failure, suggesting that reduced nitric oxide-dependent dilation contributes to impaired cutaneous vasodilation during heat stress in individuals with heart failure. However, the increase in deep body temperature during heat stress was similar between groups. Combined, these studies provided evidence that individuals with heart failure display lower cutaneous vasodilation during heat stress due to peripheral alterations in cutaneous vasodilator signaling and/or structure of the cutaneous microcirculation. A later study employing the water-perfused suit model of heat stress ruled out a potential contribution of altered neural outflow (604). Specifically, a similar increase in SSNA between individuals with heart failure and healthy age-matched controls was observed during a fixed increase in deep body temperature. This observation was paralleled by less cutaneous vasodilation and local sweat rate in individuals with heart failure relative to healthy controls. The implications of these findings for temperature regulation during heat stress is uncertain considering that sweating was similar in all but one of these studies (604). Even when sweating was lower in individuals with heart failure, the magnitude of difference was small and restricted to a small area of forearm skin. Nonetheless, reduced cutaneous vasodilation in individuals with heart failure may affect internal heat distribution and ultimately deep body temperature during heat stress. Balmain et al. (605) demonstrated that individuals with heart failure display a greater increase in deep body temperature during moderate intensity exercise in a warm environment (30°C), despite exercising at the same rate of metabolic heat production and similar local (upper back) and whole body sweating responses compared with age-matched healthy controls. When interpreting the described studies, it should be considered that all of them studied individuals who maintained β-blocker therapy and most other classes of medication, except for diuretics in one study (16) and all other medications in another (604). A potential contribution of medications to the lower cutaneous vasodilation observed in individuals with heart failure with reduced ejection fraction can therefore not be ruled out.

5.2.3. Chronic coronary artery disease.

Three studies have considered the effect of heat stress on individuals with chronic coronary artery disease (606–608). However, only one of these (608) included a healthy control group to determine whether chronic coronary artery disease alters temperature regulation during heat stress. This study compared deep body temperature and sweat losses between individuals who had undergone coronary artery bypass graft surgery and healthy age-matched controls during moderate-intensity walking in warm/humid (35°C/65% RH) and cool (18°C/45% RH) environments. Individuals with chronic coronary artery disease displayed a lower change in deep body temperature and a similar sweat rate compared with healthy individuals. An important caveat to the interpretation of these findings is that exercise was performed at a fixed percentage of V̇o_2max (60%). Since individuals with chronic coronary artery disease had a lower V̇o_2max, they exercised at a lower external workload and therefore likely generated less metabolic heat during exercise. This caveat potentially explains why individuals with chronic coronary artery disease displayed a lower change in deep body temperature. For this reason, the effect of chronic coronary artery disease on thermoregulatory control during heat stress remains unknown.

5.3. Diabetes Mellitus

Diabetes mellitus refers to metabolic diseases characterized by impaired insulin-mediated glucose uptake, resulting in hyperglycemia. In type 1 diabetes mellitus (T1DM), the pancreas is unable to produce and secrete insulin due to destruction of pancreatic beta cells, secondary to an autoimmune response of unknown etiology. Consequently, individuals with T1DM are required to administer exogenous insulin to ensure proper glucose metabolism. T1DM is normally diagnosed in infancy or adolescence (609). It is estimated that ∼10% of the population with diabetes has T1DM (610). Type 2 diabetes mellitus (T2DM) is characterized by insulin resistance (cells respond poorly to insulin) and reduced insulin production (611). Although the cause of T2DM is multifactorial, lifestyle factors, such as physical inactivity, diet, and high body mass index or adiposity, have been implicated (612).

In addition to the immense disease burden of diabetes (613), individuals living with diabetes are disproportionately susceptible to heat stress, evidenced by higher rates of hospitalization and death during heat waves in this population (1, 471, 614–616). Although the mechanisms underpinning this vulnerability are still being investigated, it appears that impaired heat loss thermoeffector responses play a role.

5.3.1. Sudomotor responses in diabetes.

It has been known for some time that a common complication of diabetes mellitus is neuropathy in peripheral nerves (617), including sudomotor nerves, which can disrupt sympathetic control of sweating. For this reason, assessments of sudomotor function have been used to identify the existence and/or severity of diabetic neuropathy.

In patients with diabetic neuropathy, the impairment in sweat production is often observed as a “stocking-and-glove” distribution of hypohidrosis or anhidrosis in the distal extremities (618–624), which is consistent with the fact that length-dependent diabetic neuropathies are most common. However, other patterns of impairment exist, including a segmental pattern, in which anhidrosis occurs in large skin areas adjacent to regions with normal sweating, including the torso; a focal pattern, in which isolated dermatomes on the torso are affected; and the global pattern, in which sweating is absent across >80% of the skin surface (621). Distal, segmental, and focal types may present symmetrically or asymmetrically (621, 625). Paradoxically, hyperhidrosis can occur in upper body regions in diabetics with large areas of anhidrosis. This phenomenon may represent a compensatory response (619, 620); that is, sweat production increases in areas with intact sudomotor function to compensate for the reduced effective surface area for sweating elsewhere (626). Pharmacological stimulation of sweating in hypohidrotic or anhidrotic areas often reveals a reduction in the density of activated sweat glands and sweat gland output (627–629). Sweat volumes during QSART are attenuated or normal in these areas, indicative of postganglionic or preganglionic sudomotor neuropathy, respectively (621). Histological analysis of sympathetic nerves has shown fiber thickening and fragmentation, with the most severe abnormalities located near sweat glands (630). Reductions in sweat gland nerve fiber density have also been noted in skin biopsies from diabetic patients with neuropathy (631, 632). Additionally, the nitric oxide contribution to sweat production is absent in young individuals with T1DM performing exercise in the heat (633), suggesting reduced nitric oxide bioavailability. Similar findings have been reported in older individuals with T2DM but could simply reflect the absence of the nitric oxide contribution to sweat production with aging rather than an independent effect of T2DM (482).

5.3.2. Cutaneous vasomotor responses in diabetes.

In addition to the possibility of neuropathy within vasodilator nerves (634, 635), diabetes is associated with microvascular dysfunction (636), which could impair the skin blood flow response to heat stress. Under thermoneutral conditions, cutaneous vasodilation and skin blood flow at rest are elevated (637–639) or normal (634, 635, 640, 641) in individuals with diabetes. A heightened resting vasodilatory/skin blood flow response may be due to impaired vasoconstrictor tone (620, 637, 642) and/or elevated insulin levels (643, 644), the latter of which can have a nitric oxide-mediated vasodilatory effect in the absence of insulin resistance (645, 646) (FIGURE 10). In contrast, maximal cutaneous vasodilatory and skin blood flow responses to local heating or vasoactive agents are lower in diabetes (264, 635, 637–640, 647–649), which may reflect alterations in the structural (vascular stiffening) and/or functional (nitric oxide responsiveness) aspects of the cutaneous vasculature (636, 649).

FIGURE 10.

Effects of insulin on endothelial cells. In healthy individuals, insulin binding to the IR-IGF1-R receptor activates the PI3K-Akt pathway. Phosphorylation of eNOS leads to the production of nitric oxide, which stimulates vasodilation. With insulin resistance or lack of insulin, there is increased serine phosphorylation of IR substrate and metabolic signaling with uninhibited activation of mitogenic and growth pathways. Such responses may contribute to attenuated skin blood flow responses in those with type 2 diabetes depicted at bottom. eNOS, endothelial nitric oxide synthase; IGF1-R, insulin-like growth factor-1 receptor; IR, insulin receptor; PI3K, phosphatidylinositol 3-kinase; Akt, protein kinase B.

Early studies noted that the expected elevations in foot skin temperatures during body heating were delayed or absent in diabetic patients with symptoms of neuropathy, suggestive of impaired reflex cutaneous vasodilation (620, 623, 624, 642). Unfortunately, few studies have since investigated the mechanisms by which diabetes impairs reflex vasodilation. Wick et al. (637) compared cutaneous vasodilatory control responses to passive heating (+0.7°C elevation in sublingual temperature) between individuals with T2DM without clinical neuropathy and matched healthy controls. Although a significantly greater absolute temperature threshold for cutaneous vasodilation was found in individuals with T2DM, the onset threshold quantified as change in temperature was similar between individuals with T2DM and the control group due to a greater baseline temperature in those with T2DM. The heat-induced increase in cutaneous vascular conductance (expressed as percentage of maximum) was similar between groups. Given that baseline and maximal cutaneous vascular conductance values were higher and lower in individuals with T2DM, respectively, the change in absolute cutaneous vascular conductance was lower in T2DM, and the absolute value of cutaneous vascular conductance was unlikely to be different between groups. In a similarly designed follow-up study by the same group (639), assessments of forearm blood flow and vascular conductance revealed similar values between individuals with T2DM and healthy controls at baseline but attenuated elevations in these parameters during body heating, indicating an impaired response due to T2DM and thus less of an ability to increase core-to-skin convective heat transfer (FIGURE 10). Notably, the contribution of nitric oxide to the reflex vasodilatory response was not different between groups, suggesting that the vasodilatory impairment did not relate to nitric oxide signaling.

5.3.3. Whole body thermoregulatory consequences of diabetes during heat stress.

Research conducted over the last 15 years has sought to better understand the whole body thermoregulatory consequences of diabetes during heat stress, including thermoeffector control, whole body heat exchange, and deep body temperature responses. In contrast to many previous studies involving local sudomotor and vasomotor assessments, participants included in these studies had well-controlled blood glucose and no overt symptoms of neuropathy.

Among individuals with T1DM, deep body temperature during exercise-heat stress is similar or greater than matched healthy controls. Stapleton et al. (650) observed no differences in esophageal and rectal temperatures in individuals with T1DM and age-, sex-, and morphology-matched controls during 60 min of cycling at an intensity eliciting 400 W (∼220 W/m²) of heat production in a 35°C, 20% RH environment. Dry heat gain was significantly higher in individuals with T1DM at the end of exercise, a perplexing finding given that mean skin temperature was nearly identical between groups, but nonetheless of minor consequence given the very low rate of dry heat exchange at 35°C. More importantly, evaporative heat loss, representing the only avenue of heat loss in this environment, was not different between groups at any time point. Furthermore, local thermoeffector responses (local sweat rate and laser-Doppler skin blood flow on the upper back) and their control properties (onset, thermosensitivity) were not different between groups.

Given the modest heat load imposed in the study by Stapleton et al. (650), subsequent studies investigated whether a T1DM-mediated impairment in temperature regulation may only arise under a more robust heat stress. Carter et al. (651) and Notley et al. (652) had individuals with and without T1DM perform three 30-min bouts of cycling at progressively increasing intensities to elicit 200, 250, and 300 W/m² of heat production in a 35°C and 20% RH environment. Although deep body temperatures were not different between groups at the lower heat loads, the elevation in deep body temperatures was greater in individuals with T1DM by ∼0.3–0.4°C at 300 W/m² of heat production in both studies. Carter et al. (651) reported varied responses among local sweat rates and their control properties. Local sweat rate on the upper back was not different between groups at any intensity, although forearm and chest local sweat rates were attenuated in individuals with T1DM at 250 and 300 W/m². No differences in the onset time for sweating were observed, but the thermosensitvity of local sweat rate on the chest at all intensities, as well as the forearm and back at 300 W/m² of heat production, was attenuated. Skin blood flow (percentage of maximum) and skin blood flow onset on the forearm and upper back were not different between groups, whereas the thermosensitivity of skin blood flow was attenuated in individuals with T1DM only at 250 W/m² of heat production. In the direct calorimetry study by Notley et al. (652), total heat loss was ∼20 and ∼30 W/m² lower in the T1DM group at 250 and 300 W/m² of heat production, respectively. Dry heat loss was not different between groups throughout exercise, but evaporative heat loss tended to be lower at 250 W/m² and was significantly lower at 300 W/m² of heat production by ∼30 W/m² in the T1DM group, coinciding with a greater deep body temperature. Collectively, these findings suggest that deep body temperature regulation in T1DM is impaired due to attenuated sweat evaporation. However, this impairment is heat load dependent, with a threshold E_req above which T1DM attenuates evaporative heat loss (i.e., the heat load that must be actively dissipated through sweating) between 250 and 300 W/m².

A similar pattern of heat exchange responses has been observed in T2DM. Evaporative heat loss and rectal temperature did not differ between individuals with T2DM and matched healthy controls during a prolonged passive exposure to 44°C (E_req ≈100 W/m²) (653). However, during exercise at intensities evoking E_req values greater than ∼165 W/m², the rate of evaporative heat loss is attenuated, resulting in greater body heat storage (654, 655). Thus, the impairment in heat loss is also heat load dependent in T2DM. The lower threshold for attenuated evaporative heat loss in T2DM versus T1DM may relate to the older age of those with T2DM, which could independently affect whole body heat loss (see sect. 4.5.2). Despite greater heat storage in T2DM, deep body temperature responses are not different relative to healthy controls (602, 655). Whether this observation is the result of diabetes-related effects on systemic blood flow and thus internal heat distribution (655), methodological reasons [i.e., clothing effects (602)], or other factors is not clear.

To summarize, thermoregulatory dysfunction in diabetes manifests in a heat-load dependent manner, with the principal effect being an attenuated rate of evaporative heat loss. Notably, these effects were reported among subjects without symptoms of neuropathy, indicating a possible subclinical pathology affecting thermoeffector responses to heat stress, as in previous work (639, 641). It is reasonable to expect that individuals with more severe T1DM and/or T2DM, inclusive of neuropathies, may exhibit a greater state of thermoregulatory dysfunction relative to the findings in the aforementioned studies.

5.3.4. Heat adaptation in diabetes.

Repeated exposure to passive heat stress has beneficial effects on weight loss, fasting plasma glucose, and HbA1C levels in individuals with T2DM (656). To our knowledge, only one study has assessed thermoregulatory adaptations to repeated heat stress in diabetes. Notley et al. (654) compared changes in whole body heat loss, heat storage, and mean body temperature responses between individuals with well-controlled T2DM and matched healthy controls before and after 7 days of daily 90-min exercise sessions at 50% of V̇o_2peak in a 40°C environment. Although preacclimation evaporative heat loss was lower in individuals with T2DM, the improvement in evaporative heat loss was significantly greater in the T2DM group (T2DM: +37 W/m²; CON: +17 W/m²), such that no between-groups differences were found postacclimation. Although the mechanism(s) underpinning a more robust adaptive response in the T2DM group is not clear, individuals with T2DM may have been under greater thermal strain during the early days of heat acclimation, due to the diabetes-related attenuation in evaporative heat loss, leading to a stronger initial adaptive stimulus. Ultimately, heat acclimation attenuated heat storage and the mean body temperature response in both groups, although the magnitude of change in these parameters did not differ between groups. Adding to previous findings of metabolic adaptations following repeated passive heat stress, these observations suggest exercise in the heat can induce heat adaptations in T2DM as in nondiabetic individuals.

5.4. Genetic Disorders

Inherited disorders that lead to structural malformations or functional deficits within eccrine sweat glands have been associated with a higher risk of heat-related illness. Two such genetic disorders are discussed below.

5.4.1. Cystic fibrosis.

Cystic fibrosis (CF) is a lethal autosomal recessive disease that occurs in ∼1 in 3,000 live births among those of European descent but is far less common in those of African or Asian descent (657). The disease is caused by a mutation in the gene that encodes the CFTR Cl⁻ channel, which enables salt and water transport through the apical membrane of epithelial cells, and also regulates sodium transport (see sect. 3.2.1) (658). In CF, abnormal expression or function of CFTR disturbs electrolyte flux through epithelial cells, affecting secretions from exocrine tissues and leading ultimately to respiratory infections, as well as various gastrointestinal and endocrine disorders (659). Without a functional CFTR, the sweat duct becomes impermeable to Cl⁻ and Na⁺, resulting in abnormally elevated sweat Cl⁻ and Na⁺ concentrations (212, 660). Such “salty” sweat ([Cl⁻] >60 mmol/L) is one of the main symptoms of CF.

That sweat salt concentrations are abnormally high in individuals with CF was discovered following the observation that a remarkably high number of hospital admissions for heat illness during a heat wave were individuals with CF (661). Symptoms of such individuals included hyperthermia, dehydration, and circulatory collapse, with hypochloremia occurring in the two assessed patients. Based on this latter finding, follow-up studies were performed to examine sweat rate and electrolyte concentrations from abdominal skin in patients with CF and healthy controls during mild passive heating (32°C, 50% RH) (662, 663). Although sweat rate was not different between groups, mean sweat concentrations of Cl⁻ (106 mmol/L, range: 60–160 mmol/L) and Na⁺ (133 mmol/L, range: 80–190 mmol/L) were distinctly higher in CF relative to sweat Cl⁻ (32 mmol/L, range: 4–80 mmol/L) and Na⁺ (59 mmol/L, range: 10–120 mmol/L) concentrations in non-CF patients (663). In nine patients with CF, hyponatremia (range: 127–130 mmol/L), and hypochloremia (range: 72–89 mmol/L) were noted. In two patients, assessments of renal and adrenal function showed normal Na⁺ and Cl⁻, indicating that abnormally low serum values were unlikely caused by excessive excretion rates. From these results, di Sant’Agnese et al. (663) proposed that in CF, excessive salt and fluid losses from sweating can ultimately produce hyponatremia (i.e., serum Na⁺ <135 mmol/L) and a large reduction in vascular volume, leading to circulatory collapse and hyperthermia.

To address this possibility, Orenstein et al (664) investigated the thermoregulatory and electrolyte responses in CF and healthy controls during moderate-intensity exercise in the heat. Although no differences in end-exercise rectal temperature or whole body sweat loss occurred, high sweat salt losses led to significant decreases in serum Cl⁻ and osmolality, and a small but nonsignificant decrease in serum Na⁺ was observed in individuals with CF but not healthy controls. Given the short duration of the heat stress in that study (90 min), it is conceivable that much longer exposures could precipitate harmful reductions in serum Na⁺ in individuals with CF. Consistent with that possibility, several case reports have documented hyponatremic dehydration among individuals with CF, or later diagnosed with CF, who collapsed following longer duration heat exposures such as heat waves, football practice, and military deployment to a hot climate (665–668). Therefore, the detrimental effects of CF on temperature regulation may only manifest under long-lasting and/or very intense levels of heat stress. For this reason, current physical activity guidelines for individuals with CF emphasize electrolyte replacement during exercise and heat stress (669).

5.4.2. Ectodermal dysplasia.

In 1848, Thurnam (670) reported observations of two male cousins with an absence of hair, few teeth, and an inability to perspire. Darwin (671) later described a family from India with 10 males, but no females, across four generations exhibiting few teeth, little body hair, and “excessive dryness of the skin” during hot weather. This triad of hypodontia (missing teeth), hypotrichosis (sparse hair), and hypohidrosis would later come to clinically define hypohidrotic ectodermal dysplasia (HED) (672), also known as anhidrotic ectodermal dysplasia or Christ-Siemens-Touraine Syndrome, which occurs in an estimated 1 in 20,000 newborns worldwide (673).

Most commonly (65–75% of cases) (674), HED is caused by a mutation in the ectodysplasin-A (EDA) gene and inherited in an X-linked recessive manner (675). Less common are the autosomal dominant and autosomal recessive patterns of HED inheritance involving mutations in the ectodysplasin-A receptor (EDAR; 10–15% of cases) or ectodysplasin-A receptor-associated death domain (EDARADD; 1–2% of cases) genes, as well as a WNT10A gene mutation inherited in an autosomal recessive pattern (5–6% of cases) (674). Since EDA, EDAR, EDARADD, and WNT10A genes code for proteins involved in signaling pathways essential for the development of ectodermal tissues during embryogenesis (674, 676), mutations in these genes interfere with the development of eccrine sweat glands. Genotype-phenotype associations have revealed that anhidrosis (vs. hypohidrosis) is more common with certain types of EDA mutations (see Ref. 677).

In hemizygous males with X-linked HED, sweat glands are structurally abnormal, sparse, or absent (670, 678, 679), which can be confirmed microscopically following a punch biopsy or via subcutaneous injection of a sudorific agent. Consequently, sweat production during heat stress is often inadequate, resulting in an attenuated rate of evaporative heat loss and greater rise in deep body temperature. Sunderman (679) found that during a 30-min exposure to 43°C air, two HED patients did not exhibit any active thermoregulatory sweating and an ∼2°C elevation in deep body temperature, whereas control subjects achieved a sweat rate of ∼600 g/h without any appreciable elevation in deep body temperature. Under nearly identical experimental conditions, McGibbon et al. (680) similarly did not find any evidence of thermoregulatory sweating in two male patients with ectodermal dysplasia, and reported a greater rectal temperature in HED patients compared with controls. Using partitional calorimetry, Rietschel and Wilmore (681) found that the rate of evaporative heat loss was six times lower in two HED males during exposure to 33°C air. The authors note that the HED patients increased convective heat loss to compensate for the lack of sweating. However, this response is unlikely to be compensatory; rather, the higher convective heat loss with HED simply stemmed from the effects of negligible evaporative heat loss, resulting in a higher mean skin temperature (35.8–36.8°C vs. 34.2°C in controls) and thus a wider skin-air temperature gradient. Studies of exercise-heat stress responses in HED have produced similar findings as with environmental heating. Roskind et al. (682) found HED patients to be completely anhidrotic during treadmill walking at effective temperatures ranging from 15 to 30°C, and a similar response was observed by McGibbon et al. (680) in 36°C air. Even in neutral ambient conditions, children and adolescents with HED exhibit a three to four times greater rise in deep body temperature during graded exercise testing (683). During low-intensity exercise at 30°C and 36°C (representing increasingly greater cumulative heat loads), Shoenfeld et al. (684) found that two male HED patients secreted sweat only at the higher heat load, though at a much lower rate than expected from control subjects. Not surprisingly, heat intolerance has been subjectively reported or empirically demonstrated among individuals with HED (679, 680, 683).

Heterozygotic female carriers of X-linked HED exhibit a highly variable HED phenotype, often with mild symptoms of thermoregulatory dysfunction. That is, sweat production is less extensive in heterozygotic females, with a patchy distribution of hypohidrosis across the torso and head (685, 686). Other studies have revealed a mosaic pattern of hypohidrosis along the lines of Blaschko (687, 688), which mark the migration path of developing epidermal cells. Thus, it is conceivable that female carriers of X-linked HED may not have the same capacity for sweat secretion as non-HED females, potentially resulting in reduced evaporative cooling during heat stress. That female carriers can experience heat intolerance support this possibility (689).

Given that during embryogenesis, blood vessels are derived from the lateral plate mesoderm rather than the ectoderm, it is possible that cutaneous vasodilatory and skin blood flow responses to heat stress would be unaltered in those with HED. Brengelmann et al. (270) assessed forearm blood flow in three males and one “mosaic” female (i.e., clinical record reported an inability to sweat on the right side of her body) during passive whole body heating with a water-perfused suit. Despite 1.4–1.7°C elevations in oral temperature, males were anhidrotic and demonstrated only a slight upward drift in forearm blood flow during heating. The drift in forearm blood flow was consistent with a local heating-induced vasodilation due to warmed blood entering the cutaneous vasculature from deeper tissues, rather than neurogenic vasodilation. Further investigation demonstrated that in two males with HED, local heating produced forearm blood flow values comparable to previously tested control subjects. In the female, a 1.6°C rise in oral temperature resulted in normal sweat rate and forearm blood flow responses on the left side, but markedly attenuated responses on the right side. Somewhat conflicting data were reported by Massey et al. (690), who noted a two- to fourfold increase in forearm skin blood flow, assessed by laser-Doppler flowmetry, in two males with HED during exercise in a 30°C environment that resulted in a 1.0–2.5°C rise in tympanic temperature. Moreover, the relative increase in skin blood flow in patients with HED was similar to that previously reported in healthy individuals during exercise (370); however, no non-HED control subjects were tested for comparison (690). Notably, in contrast to Brengelmann et al. (270), Massey et al. (690) observed elevations in skin blood flow in both patients with HED despite one being able to sweat and the other being anhidrotic. Therefore, unlike the effects of HED on sweating, the impact on skin blood flow is less clear, possibly due to heterogeneity of the disease between the few participants evaluated in these studies.

6. TEMPERATURE REGULATION FOLLOWING INJURY

Damage to neural pathways and/or effector organs involved in heat dissipation can impair deep body temperature control during heat stress. Consequent limitations on the capacity for heat dissipation will exacerbate elevations in deep body temperature during exercise or ambient heat stress, which can potentially impair physical performance and/or exacerbate the risk of heat illness in individuals suffering from such injuries.

6.1. Spinal Cord Injury

A spinal cord injury (SCI) involves damage to the spinal cord or cauda equina, resulting in temporarily or permanently disrupted motor, sensory, and autonomic function in body regions innervated by spinal nerves originating below the level of damage. Tetraplegia (or quadriplegia) refers to an SCI in the cervical region of the spinal cord that leads to total or partial loss of motor and sensory function in all extremities and torso. Paraplegia refers to SCI in the thoracic region or below that results in complete or partial impairments of motor and sensory function in the lower extremities only. Because autonomic function is disturbed following SCI, profound thermoregulatory consequences have been documented in this population.

6.1.1. Deep body temperature responses to heat stress following SCI.

In 1897, Pembrey noted that while a “normal man” was able to match heat production and heat loss across moderate and hot external temperatures, a man with a cervical-level SCI would experience higher deep body temperature responses in hot temperatures due to, in part, diminished heat loss (691). Later, Gardiner and Pembrey (692), as well as Holmes (693), found that patients with SCI exhibited poikilothermia, whereby resting deep body temperature fluctuated with ambient temperature, an effect that had been associated with abolished vasomotor control (694). Unfortunately, since many patients with SCI at that time, particularly in the cervical region, succumbed to their injuries within weeks (695, 696), a paucity of information was available concerning the long-term thermoregulatory consequences of SCI in humans. By World War II, improvements in surgical techniques, nursing care, and access to antibiotics greatly enhanced SCI survivability (695, 696), enabling studies of thermoregulatory function following recovery.

In line with studies conducted using animal models (42, 79), Pollock et al. (695) reported a near-linear relationship between air temperature and resting oral temperature in individuals with cervical SCI, while subsequent work by Guttmann et al. (697) demonstrated that exposure to 35–37°C air temperatures led to greater elevations in rectal temperature, higher and more uniform local skin temperatures, and greater heat intolerance in individuals with higher level SCI (C7 and T4) compared with an individual with a lower level SCI (T8) and an able-bodied individual. In a more recent study by Griggs et al. (698), participants with tetraplegia and ∼60% of participants with paraplegia could not maintain heat balance during exposure to an air temperature of 37°C. Collectively, these findings indicate that 1) the magnitude of the increase in deep body temperature is greater in individuals with higher level lesions in hot environments, and 2) ambient temperatures above 35°C may be physiologically uncompensable for individuals with SCI, even at rest.

Beginning in the late 1940s, greater participation in rehabilitative exercise and competitive sport by individuals with SCI (699) led to several decades of inquiry into SCI-associated alterations in deep body temperature control during physical exertion (e.g., arm-crank and wheelchair activities) (700, 701). Although several studies have reported higher deep body temperature responses in individuals with SCI versus able-bodied individuals and/or in individuals with higher versus lower level lesions (702–706), others have reported that deep body temperature control is not adversely affected by SCI during such activities (707–712). One reason for the discrepancy could be that short-duration [e.g., 20 min (707)] or low-intensity [e.g., 20 W (710)] exercise may be insufficient to unmask potential differences in deep body temperature regulation. In other studies, similar deep body temperature responses were observed between individuals with SCI and able-bodied individuals despite a lower mass-specific rate of metabolic heat production in the SCI group (709, 711–713). The same change in deep body temperature for a lower rate of mass-specific heat production following SCI can be interpreted as evidence for impaired deep body temperature control. Studies that have employed exercise protocols that deliberately (or serendipitously) evoked similar mass-specific rates of metabolic heat production during exercise have consistently observed greater elevations in deep body temperature with higher lesion levels when compared with able-bodied individuals (702–704, 706). A notable example is the study by Forsyth et al. (703), in which esophageal temperature responses were monitored in able-bodied and individuals with tetraplegia (C5-C6) exercising at 4.0 W/kg of heat production, and between able-bodied individuals and those with high (T1-T5) or low (T6-L1) paraplegia at 6.0 W/kg of heat production under hot-humid ambient conditions. At the end of exercise, the increase in esophageal temperature from baseline was significantly greater in tetraplegic (+1.89°C) versus able-bodied (+0.29°C) individuals at 4.0 W/kg and in high paraplegic (+1.20°C) versus low paraplegic (+0.66°C) and able-bodied (+0.53°C) individuals at 6.0 W/kg. In accordance with the studies conducted under ambient heat stress discussed above, the results of Forsyth et al. (703) conclusively demonstrate that deep body temperature responses to exercise are exacerbated by SCI and that higher lesions, particularly in the cervical region, are more deleterious to deep body temperature control than lower level lesions.

6.1.2. Thermoafferent signaling following SCI.

Following SCI, central thermoreceptors in the spinal cord below the lesion, as well as peripheral thermoreceptors in the skin, muscle, and viscera that transmit thermosensory information via spinal nerves below the lesion, are neurally disconnected from the hypothalamus. Consequently, the volume of thermosensory information received by the hypothalamus is reduced in proportion to the lesion level.

Studies by Downey et al. (714) and Tam et al. (715) found that paraplegia delayed the deep body temperature onset threshold for sweating during whole body heat stress. This shift has been attributed to disrupted afferent signaling (716). Recently, however, Forsyth et al. (703) found no difference in the change in esophageal temperature at the onset of forehead and upper back sweating between high-paraplegic, low-paraplegic, and able-bodied individuals. The origin of this discrepancy is unclear but could relate to the metric used to define sudomotor onset (time vs. change in deep body temperature) or the relatively low sample sizes included in earlier studies.

Two other investigations have provided more compelling evidence for impaired thermosensory signaling following SCI. In a study by Rawson and Hardy (717), an individual with a T12-level SCI was heated in a 38–40°C chamber until a steady-state sweat response was observed, after which blood flow to the leg was occluded and the skin on the occluded leg was cooled. Using this approach, any observed cooling effect on thermoregulatory control will be of peripheral origin, as occluding the circulation in that leg prevented the return of cooled blood to the systemic circulation. Without SCI, cooling an occluded limb reduces the rate of sweat production in proportion to the fall in skin temperature (443). However, in the case of a paraplegic individual, cooling the leg skin by as much as 7°C during circulatory arrest of that leg did not reduce sweating, indicating impaired thermosensory signaling from dermatomes below the lesion. More recently, Shibasaki et al. (718) compared vasodilatory responses to a +1.1°C elevation in deep body temperature induced via lower leg hot-water immersion among abled-bodied individuals and those with thoracic- or cervical-level SCI. Cutaneous vascular conductance on the thigh was not different between groups, owing to the proximity of the measurement site to the heat source causing a local heating effect. In contrast, able-bodied individuals exhibited greater cutaneous vascular conductance on the chest (i.e., at greater distance to the heat source) compared with individuals with thoracic- and cervical-level SCI, with no vasodilation evident in individuals with a cervical-level SCI. Collectively, these findings suggest that for a given central stimulus (i.e., rise in deep body temperature), impaired thermoafferent signaling from the periphery following SCI attenuates sweating and reflex cutaneous vasodilation.

6.1.3. Efferent signaling and effector responses to heat stress following SCI.

Before presenting evidence of dysregulated thermoeffector responses following SCI, it is important to briefly revisit the anatomical layout of the sympathetic nervous system and note the effects of SCI on sympathetic traffic to its effector targets. Since preganglionic sympathetic fibers exit the spinal cord through spinal nerves T1-L2/3, the sympathetic nervous system becomes “decentralized” with complete transection of the cervical spinal cord, resulting in an absence of centrally mediated thermoregulatory responses to heat stress (697, 718–720). Additionally, increases in atrial and ventricular systolic function, which support elevations in heart rate and cardiac output in response to body warming, are abolished following a cervical-level SCI (718). Following a thoracic-level SCI, at least some sympathetic signaling, and thus thermoeffector control, is retained. High thoracic lesions (T1-T6) impede sympathetic outflow to skin on the head, neck, arms, and/or upper torso, as well as all skin regions innervated by neural pathways originating below the lesion level. Additionally, high thoracic lesions may impede sympathetic signaling to the heart (innervated by T1-T5), limiting an individual’s capacity to increase heart rate, cardiac output, and thus skin blood flow (709, 718). Low thoracic lesions (T6-L2) interfere with sympathetic outflow to skin on the lower torso and legs, as well as the splanchnic viscera, which may impair splanchnic and renal vasoconstriction during heat stress, and thus redistribution of blood to the skin (721). Individuals with SCI below L3 largely retain sympathetic innervation of skin areas in the lower limbs (720). It should be noted that incomplete lesions of the spinal cord can produce a less predictable and “patchy” pattern of sudomotor and vasomotor responsiveness (720), indicating retention of some sympathetic innervation below the lesion level.

The main cause of higher elevations in deep body temperature during exercise and ambient heat stress following SCI is impaired thermoeffector activation. Concerning a patient with a thoracic-level SCI, Pembrey (722) noted that “when the patient was hot she sweated on the nonparalysed parts, but the paralysed parts were very dry.” This absence of sweating below the lesion level was confirmed in subsequent studies conducted during ambient heat stress (695, 705, 715, 719–721, 723, 724) and exercise (703, 705). The attendant lack of cutaneous evaporative cooling leads to greater increases in skin temperature in skin areas below the lesion during exercise (703) and ambient heat exposures (697). Vasomotor responses of neural origin are also absent within insensate skin regions (710, 718, 725), but skin below the spinal cord lesion retains some ability to dilate in response to local heating (726). In general, the distribution of skin areas with intact vasomotor function is similar to those of sweating following SCI. However, Normell (720) noted that in ∼50% of SCI patients, sudomotor function was preserved across a wider surface area than vasomotor function, which could be ascribed to heterogeneity in the arrangement of sudomotor and vasodilator pathways.

Several studies have reported sweat production in areas innervated below a spinal cord lesion. Thermoregulatory sweat production within skin dermatomes slightly below the level of the lesion may arise due to the fact that preganglionic sympathetic fibers originating from above a spinal lesion can descend to and synapse on ganglia below the lesion level, preserving sympathetic signaling to skin areas below the lesion (695). Several investigators have noted generalized sweating responses following cervical-level SCI or sweating far below the lesion level in thoracic-level SCI, suggesting that the spinal cord, and not only the hypothalamus, can initiate sudomotor responses to heat stress. While this may be the case, it is important to note that studies supporting the existence of “spinal reflex sweating” report only trace amounts of sweat secretion that would contribute minimally to whole body thermoregulatory capacity (727, 728). Finally, episodes of profuse sweating within skin areas below the lesion commonly occur in response to noxious stimuli, such as bladder distention or pain, during the acute phase of an SCI (729–732). This sweating response is not of thermoregulatory origin, but rather due to post-SCI autonomic dysreflexia.

Following SCI below T1, skin areas with intact sympathetic innervation exhibit sweating and vasodilatory responses to heat stress; however, control of these responses may be altered following the injury. For example, thermosensitivity and peak forearm and skin blood flow responses are attenuated in skin areas above the lesion following whole body passive heat stress or exercise (710, 721, 733). Without any differences in arterial blood pressure, Freund et al. (721) suggested that the alteration in skin blood flow control following SCI may arise from a reduced ability to redirect blood flow from the viscera to the skin surface. With regard to sweating, Forsyth et al. (703) did not observe any difference in local sweat rate thermosensitivity on the forehead and upper back between low-level paraplegics, high-level paraplegics, and able-bodied individuals, suggesting that central and peripheral control of sweating remains intact in paraplegics.

While sudomotor control may be preserved, the inability to secrete sweat over a large surface area of skin means that individuals with SCI often cannot meet E_req during heat stress (705, 709, 713, 734), thus exacerbating the rise in deep body temperature (incidentally, this problem may be further worsened if sweat cannot evaporate from skin areas on the back with functional sweat glands when in a wheelchair). For instance, Petrofsky (705) assessed whole body sweat rate responses at rest and during exercise (50 W) in air temperatures of 30°C, 35°C, and 40°C. Although mean skin temperatures were not reported, standardization of the air temperature and the absolute work rate likely produced similar E_req between groups. Whole body sweat rate was found to be highest in able-bodied individuals and lowest in tetraplegic individuals with paraplegic individuals in between, indicating that the degree to which whole body sweat rate was attenuated was graded with the lesion level. Importantly, the rise in aural canal temperature reflected the attenuated whole body sweat rate response in the SCI groups (i.e., lower whole body sweat rates led to higher elevations in deep body temperature). In contrast, several studies have found that whole body sweat rate is unaffected by SCI (708, 711, 735). There are two possible explanations for this discrepancy. First, it is possible that individuals with SCI produce higher sweat rates above their lesion to compensate for the absence of sweating below the lesion. Compensatory sweating following SCI has been observed (697, 714), and represents a necessary response to achieve heat balance in compensable conditions (626). A second possible explanation relates to the level of the injury. In the study by Dawson et al. (708), participants with injuries from T12 to L3 did not show any differences in whole body sweat rate or deep body temperature during exercise in cool and hot conditions. However, Normell (720) found that at least some sweating was preserved below the lesion to the feet in individuals with T12-L3 SCI, indicating that perhaps the subjects of Dawson et al. (708) retained a sufficient amount of sudomotor function to appropriately regulate deep body temperature.

6.1.4. Heat adaptation following SCI.

Conceivably, repeated exposure to heat stress could augment sweat rate and thermolytic capacity in individuals with lower level SCI who retain some sudomotor function (i.e., paraplegics), while the absence of any sudomotor function with higher level SCI (i.e., tetraplegics) would prevent these heat adaptations. In individuals with thoracic-level SCI (T5-T12), Gass and Gass (736) found that 5 days of hot water immersion led to a small but nonsignificant reduction in sudomotor onset and an increase in sudomotor thermosensitivity, without any adjustments to deep body temperature or heart rate during heat stress. Using a combination of exercise and ambient heat stress (20-min exercise and 40-min rest in ∼33°C and ∼65% RH), Castle et al. (737) demonstrated lower baseline deep body temperature and end-exposure heart rate by ∼0.3°C and 25 beats/min, respectively, in individuals with C5-C6 to T9/10 SCI; however, no differences in the changes in deep body temperature and body mass (indicative of sweat losses) were observed. In contrast, both Price et al. (738) and Trbovich et al. (739) found no evidence of any physiological adaptations to 7 days of exercise-heat stress in tetraplegic and paraplegic individuals. Based on these results, it appears that even paraplegic individuals have a low capacity for thermoregulatory adaptation to heat stress. It is possible that the absence of robust thermoregulatory adaptations in paraplegic individuals may be explained by the heat impulse imposed, which was substantially lower than that used in traditional heat acclimation protocols. Future studies investigating these questions could consider using longer duration bouts of heat stress (e.g., 90 min) and/or longer term adaptation periods (>7 days).

Aerobic training can also improve the capacity for heat loss (72). Observations of similar deep body temperature responses during exercise between trained paraplegic and able-bodied individuals (735) has led to the conclusion that aerobic training confers a thermoregulatory advantage in individuals with SCI (716). However, the paraplegic individuals in the study by Price and Campbell (735) demonstrated similar elevations in deep body temperature while exercising at a mass-specific metabolic rate that was ∼28% lower than the able-bodied group, which in fact points to SCI-associated thermoregulatory dysfunction. Yaggie et al. (740) showed that endurance-trained individuals with SCI and able-bodied individuals achieve comparable upper extremity sweat rates following pharmacologically induced sweating via pilocarpine iontophoresis. This finding suggests that the sweat glands with intact sympathetic innervation can adapt to heat via physical training. Future studies should assess training-related thermoregulatory adaptations using a longitudinal experimental design with assessment of the upper limits of compensability.

6.1.5. Exposure limits.

As noted above, the challenge of maintaining heat balance following SCI arises from an inability to sufficiently increase sweat rate, resulting in a reduced effective body surface area and thus a reduced E_max. It follows that the critical environmental limits, the upper limits of metabolic heat production, air temperature, humidity, etc. above which thermal balance could no longer be achieved, would be lower for individuals with SCI than for able-bodied individuals, and for individuals with higher versus lower level injuries. This issue was addressed recently by Griggs and colleagues (698), who assessed how SCI affects the “critical” RH values in a hot environment. Individuals with tetraplegia (C5/6-C6/7), paraplegia (T3-T12), and able-bodied individuals were exposed to 37°C and 20% RH for 20 min, after which RH was increased from by 5% every 7 min until an upward inflection in gastrointestinal temperature was observed. As described earlier, the upward inflection in deep body temperature in this protocol signifies a transition from a physiologically compensable to uncompensable state; the “critical” RH at which the upward inflection in deep body temperature occurs serves as a proxy for an individual’s capacity for evaporative heat loss, with higher critical RH indicating a greater evaporative capacity. In able-bodied individuals, a clear inflection in gastrointestinal temperature was observed at a mean RH of 77%. However, an inflection in gastrointestinal temperature was observed in only three of eight paraplegic participants and at a significantly lower mean RH of 53%, while the tetraplegic groups exhibited a continuous rise in gastrointestinal temperature, which indicates that the baseline conditions (37°C, 20% RH) were uncompensable for those with tetraplegia. The lower thresholds for compensability with higher level SCI were attributed to greater deficits for evaporative heat loss potential, itself the result of lower whole body sweat rates and, likely, ω_max. Overall, these results indicate that individuals with SCI will not be able to maintain heat balance at rest in ambient air temperatures at or above resting deep body temperature with even moderate levels of humidity.

6.2. Burn Injury and Skin Grafting

Burns are traumatic injuries, resulting from skin contact with an intense heat source. Approximately, 500,000 burn injuries requiring medical attention occur each year in the United States (741) and burns are the fourth common cause of injury worldwide (742, 743). Improvements in acute and long-term burn care have reduced mortality rates in individuals with large burn injuries (744). However, due to permanent skin damage, deep burn injuries lead to significant thermoregulatory sequela and persistent heat intolerance (745).

6.2.1. Etiology of thermoregulatory dysfunction following a burn injury.

Burn injuries are classified by their depth (746) and thus the need for surgical intervention. “Superficial” or first degree burns involve only the epidermis and heal relatively quickly without the need for surgery. Conversely, some “partial-thickness” or second degree burns, which damage the epidermis and some portion of the dermis, as well as “full-thickness” or third degree burns, which involve the entire epidermal and dermal layers, require surgical intervention to remove necrotic tissue, reduce the risk of infection, and promote wound healing (50). Following a partial- or full-thickness burn injury, early excision of the damaged skin leads to the removal of part of or the entire dermal skin layer, which contains the sweat glands, vasculature, and associated neural connections required to facilitate heat dissipation. Next, a skin graft is transplanted from a noninjured “donor” site to the recipient bed of the wound. Most skin grafts are “split-thickness,” meaning they contain the epidermis and a superficial portion of the dermis (747). Because these thin split-thickness grafts have relatively little tissue requiring revascularization, these grafts are more likely to “take.” However, since important thermoregulatory end-organ structures, such as the secretory coils and proximal ducts of sweat glands and the reticular cutaneous vasculature, are located deep within the dermis, these structures are not transplanted with the graft onto the wound (748, 749). Therefore, restoration of normal thermoregulatory function within grafted sites would rely on regrowth and reinnervation of sudomotor and vasomotor end organs.

6.2.2. Impaired cutaneous vasodilatory and skin blood flow responses in grafted skin.

Reestablishment of blood flow through a skin graft is critical to its survival. Following transplantation of a skin graft, inosculation of capillary buds between the recipient bed and the underside of the graft begins within 48 h, angiogenesis and revascularization commences by ∼72 h, and blood flow through the graft is apparent within ∼5 days (750, 751).

Whether revascularization of skin grafts restores the vasodilatory and skin blood flow response to heat stress was first addressed by Freund et al. (752). In that study, forearm blood flow was measured via venous occlusion plethysmography in individuals with circumferential arm partial- and full-thickness burns treated with split-thickness skin grafts. With local heating to 42°C, elevations in forearm blood flow were evident in grafted areas, and mean forearm blood flow values were shown to be similar between grafted and noninjured skin in a subset of subjects tested. However, the response to whole body heating, which is a reflex vasodilatory response that requires revascularization and intact innervation, was heterogenous, with subjects demonstrating normal, attenuated, or completely absent elevations in forearm blood flow. In later studies, Davis et al. (753, 754) used laser-Doppler techniques to compare reflex vasodilatory responses within the cutaneous vasculature only (as opposed to whole limb responses obtained with plethysmography) between grafted skin and noninjured skin during whole body heating. Reflex cutaneous vasodilation was found to be appreciably attenuated in all skin grafts (753), with the effect persisting for more than 4 years postsurgery and likely for the life of the burn survivor (FIGURE 11) (754). Similarly, during local heating, cutaneous vasodilatory responses were attenuated in grafted relative to noninjured skin, and this difference was independent of the age of the graft up to 8 years postsurgery (754). The origin of the discrepancy between the results of Freund et al. (752) and Davis et al. (753, 754) is likely methodological, as plethysmographic estimates of forearm blood flow reflect perfusion of cutaneous and underlying vascular beds, whereas laser-Doppler measurements reflect perfusion of the cutaneous vasculature only. Collectively, these findings indicate that despite revascularization and some reinnervation of the cutaneous vasculature, the resultant vasodilatory and skin blood flow responses to heat stress are impaired in grafted skin.

FIGURE 11.

Graphical depiction of reduced sweating and cutaneous vasodilation in burned and subsequently grafted skin during a whole-body heat stress.

To elucidate the mechanisms underpinning the impaired cutaneous vasodilatory responses to heat stress in skin grafts, Davis et al. (755) compared the postsynaptic responsiveness of the microvasculature within grafted and normal skin to infusions of increasing concentrations of acetylcholine (an endothelium-dependent vasodilator) and sodium nitroprusside (an endothelium-independent vasodilator). The peak response to acetylcholine infusion was blunted in grafted skin, and the effective concentration that elicits 50% of the maximum response (EC₅₀) was shifted to a higher acetylcholine concentration, which points to a reduced endothelium-dependent vasodilator sensitivity in grafted skin. In contrast, the peak response and EC₅₀ for sodium nitroprusside were not different between grafted and noninjured skin, suggesting that attenuated endothelium-independent vasodilation in grafted skin (753, 754) may be unrelated to an attenuated vascular smooth muscle responsiveness to nitric oxide, but rather a reduction in nitric oxide release from the vascular endothelium (265, 745). The reason for diminished endothelial nitric oxide release in skin grafts is not known.

6.2.3. Impaired sweat production and evaporative heat loss in grafted skin.

During heat stress, sweat production from split-thickness skin grafts is severely attenuated or completely absent (680, 748, 749, 753–756). Early studies that qualitatively assessed sweat responses between graft types (748) or between grafted and noninjured skin (680, 757) reported anhidrosis in skin areas covered by split-thickness grafts. More recently, Davis et al. (753) showed quantitatively with capacitance hygrometry that elevations in sweat rate from grafted skin were profoundly attenuated and that impaired sweat production in grafted skin persists for at least 4–8 years postsurgery (754).

Since full-thickness skin grafts, which include the entire dermal layer from the donor site, retain some sudomotor function (748, 749, 758), impaired sweating in split-thickness skin grafts suggests that these grafts are either devoid of functional sweat glands or functional sweat glands within these grafts are not reinnervated. Accordingly, in the study by Freund et al. (752), a punch biopsy of well-healed grafted skin in one subject revealed an absence of sweat glands or nerve tissue, a finding confirmed by others (759). Davis et al. (753) investigated this issue by examining the dose-response relationship between the sudorific agonist acetylcholine and sweat rate. While a high concentration of acetylcholine evoked a robust sweating response in noninjured skin, the same dose produced no sweating in the grafted site. Taken together, these findings suggest that anhidrosis in split-thickness skin grafts likely stems from a complete absence of functional sweat glands.

Of course, the major implication of impaired sweat production in skin grafts is a reduced evaporative potential. Using direct calorimetry, Ganio et al. (760) compared the rate of evaporative heat loss between a burn survivor with 75% of the body surface covered in skin grafts and two control subjects matched for age, sex, aerobic capacity, and morphological characteristics during 60 min of exercise in a 35°C and 20% RH environment (E_req of ∼400 W). In the control subjects, evaporative heat loss increased throughout exercise, nearly matching E_req by the end of exercise, and deep body temperature increased by ∼0.6°C. In the grafted individual, evaporative heat loss peaked by 30 min of exercise at only ∼240 W. With E_req exceeding the peak rate of evaporative heat loss by an estimated 160 W, the conditions were clearly uncompensable for the individual with skin grafts, resulting in greater body heat storage and a higher elevation in deep body temperature (1.22°C) compared with their nongrafted counterparts (mean: 0.62°C). This finding not only demonstrates the attenuating effect of skin grafting on evaporative potential but also indicates that the range of physiologically compensable conditions for burns survivors with extensive skin grafts may be considerably narrower than for noninjured individuals (more on this below).

Sweat rates from nongrafted skin may be higher in burn survivors during exercise-heat stress (293, 680, 682, 761), possibly representing a compensatory sweating response. For a given heat load (i.e., evaporative requirement), the reduction in E_max due to extensive skin grafting leads to a correspondingly higher skin wettedness from noninjured skin to achieve heat balance (293). When the evaporative requirement for heat balance is high, the reduced E_max and higher corresponding skin wettedness requirement from noninjured skin reduce sweat evaporative efficiency from those noninjured areas (70, 71), resulting in greater whole body sweat production per square meter of nongrafted surface area and, potentially, total whole body sweat rate in burn survivors (293, 680, 682, 761). In such circumstances, higher total sweat losses borne of poor sweat evaporative efficiency impart no thermoregulatory advantage but rather contribute to more rapid fluid depletion.

6.2.4. Sweating and skin blood flow in nongrafted skin: donor sites.

As noted above, split-thickness grafts contain only epidermal and superficial dermal strata from donor skin. Although removal of these skin layers induces some superficial damage at the donor site, the remaining skin retains deeper components of thermoeffector appendages (e.g., secretory coils of sweat glands and reticular cutaneous arterioles) and their neural connections. As such, healed donor sites could retain normal thermoregulatory function. Accordingly, Davis et al. (753, 754) did not find any difference in local sweat rate (capacitance hygrometry) and cutaneous vasodilation (laser-Doppler techniques) between donor and noninjured sites during whole body passive heating. Similar findings were reported by Cramer et al. (762) during combined exercise and ambient heat stress using the Technical Absorbent sweat patch method (182, 763) and laser-Doppler imaging to assess local sweat rate and skin blood flow, respectively. These studies indicate that heat loss thermoeffector function is preserved in donor sites.

6.2.5. Deep body temperature control following a burn injury.

As discussed above, skin grafting and the resultant impairment in sweat function limit absolute E_max compared with noninjured individuals. However, the extent to which attenuated evaporative heat loss affects deep body temperature control and heat illness risk in burn survivors depends on several other factors that influence the balance between heat gain and heat loss.

With regard to heat loss, the most important factor affecting deep body temperature control following a burn injury is the size of the injury, which is typically characterized by the percentage of total body surface area (%TBSA) with partial- or full-thickness burns (superficial burns are excluded). In most studies, higher %TBSA injuries have been associated with greater elevations in deep body temperature during exercise in the heat (293, 680, 682). It should be noted, however, that it is not the %TBSA per se that dictates the deep body temperature response. Rather, since rates of heat exchange are strongly associated with absolute surface area (in square meters), it is the absolute surface area of noninjured skin that can participate in evaporative heat loss that is the strongest predictor of the deep body temperature responses for a given heat load. This distinction was highlighted by Ganio et al. (761). In that study, burn survivors with injuries spanning 17–75% TBSA and heterogeneous morphological characteristics exercised for 90 min at a fixed rate of heat production in a hot-dry environment. The %TBSA injured was significantly associated with the change in deep body temperature from baseline, explaining 24% of the variance. However, the strongest predictor, explaining 41% of the variance of the deep body temperature response, was the absolute surface area of nongrafted skin. Despite strong differences in sweat production between body areas (see sect. 3.3.1), the anatomical location of a burn injury does not likely alter temperature regulation during uncompensable heat stress, at least when the effects of torso versus limb injuries are compared (764).

Building off the work of Ganio et al. (761), a series of studies was initiated to investigate how the size of a burn injury and various factors interactively influence thermoregulatory capacity. In many of these experiments, a novel simulated burn injury model was used to overcome logistical challenges associated with the recruitment of burn survivors with specific morphological traits, %TBSA injured, and injuries in particular anatomical locations and also minimized intersubject variability that attends an independent-group experimental design. Absorbent material with a vapor-impermeable cover was cut to predetermined sizes and then applied to the skin surface of noninjured individuals. The absorption of secreted sweat in the material impeded evaporative heat loss from covered skin areas.

The first experiment in this series (765) addressed the role of body size; specifically, whether individuals of larger body size, who retain a greater surface area of noninjured skin and thus have a higher absolute E_max, have a thermoregulatory advantage over smaller individuals with the same simulated 40% TBSA injury during uncompensable exercise-heat stress. When exercise was performed at a similar absolute rate of metabolic heat production (in watts; typical of fixed-load tasks), elevations in deep body temperature were indeed attenuated in larger (∼96 kg) versus smaller individuals (∼65 kg), which reflected not only the greater noninjured surface area in larger individuals, but also the lower body mass and thus the heat storage capacity of the smaller group (286, 287) (FIGURE 12). In contrast, when exercise elicited the same mass-specific rate of metabolic heat production (typical of weight-bearing tasks), elevations in deep body temperature were similar between larger (∼96 kg) and smaller (∼61 kg) individuals (FIGURE 12). Although the larger individuals likely had a greater absolute E_max, their mass-specific E_max was probably similar to the smaller individuals due to a lower surface area-to-mass ratio. As such, the difference between the mass-specific rate of metabolic heat production and mass-specific E_max, the latter of which drives the increase in deep body temperature during uncompensable heat stress (287), was unlikely to have been appreciably different between larger and smaller individuals. These findings suggest that for a given %TBSA, deep body temperature responses are affected by morphological traits, but only during physical tasks of a fixed absolute intensity.

FIGURE 12.

Schematic summarizing the interactive effects of a burn injury and various modulators of the deep body temperature response to exercise. The effect of body size on the elevation in deep body temperature is shown for a 40% total body suface area (TBSA) injury during exercise at intensities eliciting a fixed absolute rate of metabolic heat production (H_prod), consistent with weight-independent activities such as cycling at a fixed work rate, and intensities eliciting a mass-specific rate of heat production, consistent with weight-bearing activities. The effect of air temperature (hot vs. temperate environments) and work intensity (moderate vs. low intensity) are shown across a range of %TBSA, from noninjured (0% TBSA injury) to 60% TBSA. Finally, the impact of 7 days of heat acclimation is shown. Adapted from Ref. 337, with permission from the American Physiological Society, and from Refs. 765–767, with permission from Medicine and Science in Sports and Exercise.

Since E_req is dictated by the total heat load imposed, and E_max is highly dependent on the surface area of intact sudomotor function, the boundaries of the “prescriptive zone” [the range of physiologically compensable conditions (768)] for burn survivors should depend on factors such as ambient temperatures and work intensity, as well as the size of a burn injury. To determine the interactive effect of air temperature and burn injury size on thermoregulatory capacity, Cramer et al. (766) compared elevations in deep body temperature during 60 min of exercise at a fixed rate of metabolic heat production in 40°C and 25°C air temperatures with simulated burn injuries of 20%, 40%, and 60% TBSA, plus a noninjured “Control” trial. The greater heat load associated with exposure to a 40°C environment led to a higher elevation in deep body temperature with 40% and 60% TBSA simulated burn injuries, though no difference in deep body temperature was apparent with a 20% TBSA injury versus control (FIGURE 12). In contrast, when the ambient heat load was reduced by exposure to a temperate 25°C environment, similar elevations in deep body temperature were observed across all %TBSA compared with control. In a conceptually similar investigation, Belval et al. (767) examined the interaction between exercise intensity and burn injury size on deep body temperature control during 60 min of exercise under fixed hot-dry ambient conditions with 20%, 40%, and 60% TBSA burn injuries and a noninjured control trial. When exercise was conducted at a moderate intensity evoking 6 W/kg of metabolic heat production, greater elevations in deep body temperature were observed with 40% and 60% TBSA simulated burns compared with control, with no difference in deep body temperature between 20% TBSA and the control trial (FIGURE 12). Meanwhile, low-intensity exercise that elicited only 4 W/kg of metabolic heat production yielded similar elevations in deep body temperature across all %TBSA versus control. Collectively, the findings of these studies indicate that burn injuries ≥40% TBSA substantially reduce the physiologically compensable range of air temperature and exercise intensities, but burns ≤20% TBSA, which represent over 90% of all burn injuries (741), do not significantly impair thermoregulatory capacity under modest heat loads.

6.2.6. Heat acclimation following a burn injury.

Despite the harmful effects of burn injuries and skin grafting on whole body temperature regulation, well-healed burn survivors can adapt to heat stress (337, 769). Schlader et al. (337) compared elevations in deep body temperature in noninjured individuals and burn survivors with injuries spanning 17–40% TBSA and >40% TBSA before and after a heat acclimation protocol in which participants exercised at ∼50% of V̇o_2max in 40°C and 30% RH for 90 min/day over 7 consecutive days. Elevations in deep body temperature during exercise were graded in relation to the size of the injury, and this observation was unaltered by heat acclimation. However, the magnitude of the elevation in deep body temperature was attenuated to an equal extent in all three groups following heat acclimation (FIGURE 12). Additionally, the temperature of nongrafted skin decreased and the sensitivity of the whole body sweat rate response increased with heat acclimation. Taken together, these findings demonstrate that burn survivors can adapt to the heat, and that the magnitude of the improvement is unrelated to the size of the injury.

6.3. Prior Heat Stroke

Heat-related illnesses, as outlined in TABLE 4, represent a continuum of life-threatening disorders that range in intensity and severity from mild cardiovascular (e.g., heat exhaustion) and CNS abnormalities to profound cellular damage (e.g., heat injury and heat stroke) due to a systemic inflammatory response syndrome (SIRS) in hyperthermic victims (5, 26). These cellular and systemic dysfunctions are often noted in key regulatory organs, such as the brain (5), kidneys (773), liver (774), skeletal muscles (775), and heart (776). Coagulopathies are also reported in more severe manifestations (777).

Table 4.

Definition and signs/symptoms of different manifestations of heat illness

	Heat Exhaustion (770)	Heat Injury (771)	Heat Stroke (772)
Definition	Failure to maintain cardiac output to match demands of blood flow for temperature regulation and exercising muscles and vital organs.	An intermediate condition between heat exhaustion and heat stroke with organ injury.	Life-threatening condition characterized by CNS dysfunction in hyperthermic subjects.
Signs/symptoms	Ataxia, anxiety, compromised cognition, dizziness, headache, nausea, tachycardia, hyperventilation.	Persistent mild confusion and disorientation with hyperthermia.	Delirium, agitation, improper aggressiveness, convulsions, or coma. Deep body temperature (i.e., rectal) >40°C, although reliance on a specific rectal temperature value is not advised.

CNS, central nervous system.

Heat stroke is one of the oldest known medical conditions and is considered the most severe manifestation of heat related illnesses. The common use of the term “dog days” of summer, which stems from the appearance of Sirius, the Dog Star constellation Canis Major (∼3,000 B.C.), and the mention of human deaths from heat exposure in the Bible in reference to the deaths of farmers as well as armored fighting forces ∼1,000 B.C. (778) indicate that heat has wreaked havoc on humans for millennia. Despite this long history, there is still a lack of effective strategies to prevent or properly treat heat stroke in a way that results in full recovery (779). To date, prompt and aggressive cooling upon collapse is the only effective measure to mitigate severe organ injury induced by severe heat illness. Nevertheless, while immediate cooling may be available at sports events such as races and organized competitions in more developed regions, it may not be promptly available during other scenarios such as military combat or sporting events in less developed countries or regions. In addition, successful short-term recovery from heat stroke may not be the end of problems for victims. Attention has been given to the long-term health consequences of heat stroke. A typical concern regarding long-term consequences of heat stroke is whether the hypothalamus, considered the thermoregulatory control center of the brain, becomes damaged or dysfunctional after an episode, which would hamper the ability to thermoregulate adequately after a heat stroke episode. Retrospective and epidemiological evidence also suggest that heat stroke may be linked to a greater risk of cardiovascular diseases following years after the heat stroke episode (780–782). The development of preclinical models of heat stroke (783, 784) has allowed hypothesis-driven research to be conducted, which adds to our knowledge about the pathophysiology of this condition, including cardiovascular and epigenetic alterations. In this section, we explore three aspects of recovery from heat stroke: the likelihood of hypothalamic damage, silent cardiovascular abnormalities, and epigenetic modifications following a heat stroke episode. Before exploring the literature to address these concerns, we must briefly categorize the different types of heat stroke and provide a succinct overview of their distinct pathophysiology and incidence.

6.3.1. Pathophysiology, incidence, and long-term consequences.

Heat stroke manifests under two different forms. Classic heat stroke (CHS), also known as passive heat stroke, occurs at rest and impacts vulnerable populations such as the elderly (785), in particular during annual heat waves, as well as the unfortunate instances where children are left entrapped inside sun-exposed vehicles (786). Exertional heat stroke (EHS) occurs during physical activity often, but not always, in hot and humid environments and affects physically active populations such as recreational and elite athletes as well as military personnel (5, 26). In addition, EHS is the third leading cause of mortality in athletes during physical activity behind cardiac disorders and head and neck trauma (787). A distinctive feature of EHS, in comparison to CHS, is the active participation of skeletal muscles, which has recently been thoroughly reviewed elsewhere (788). Nevertheless, it is important to highlight that rhabdomyolysis (e.g., rapid skeletal muscle breakdown) is normally present in EHS, and absent in CHS, and the subsequent leakage of intramuscular content results in kidney overload and failure (789).

The precise incidence of EHS is likely underestimated due factors such as treatments outside of the hospital setting without clinical reporting, temperature recordings at imprecise sites, misdiagnosis, or inaccurate intensive care unit coding by hospitals, among other factors (772). Furthermore, from the time of collapse and loss of consciousness to removal to a hospital, the high deep body temperature may have dropped. The victim is then typically diagnosed for the remaining clinical presentation at the hospital admittance, which normally includes hypoglycemia, low blood pressure, dehydration, and myoglobinuria, among others. The database from the U.S. Armed Forces is likely the best resource to approach the heat stroke incidence. According to the latest U.S. Armed Forces report published in 2020, there were >2,174 hospitalizations due to heat illness with >500 diagnoses of heat stroke (790). The surveillance period was from January 1, 2016 to December 31, 2020. The surveillance population included all individuals who served in the active component of the Army, Navy, Air Force, or Marine Corps at any time during the surveillance period. Nevertheless, caution must be taken when considering these epidemiological data as they are specific to a subset of the population that may be a higher risk to develop heat illnesses when compared with the nonmilitary population.

A concern regarding long-term consequences of heat stroke (e.g., exertional) involves the time required for recovery and safe “return to physical activity” following EHS. In response to acute cooling, fluid replacement, and intensive medical care, most patients with EHS appear to quickly recover overall health in a few days. This apparent recovery may be misleading. In the French military, 15.4% of the patients experiencing EHS reported a previous EHS episode (791). Another study reported an incidence of 4.1% of recurrent exertional heat illness episodes in a cohort of 145 patients (792). Likewise, US active military personnel, who experienced EHS in the previous year, had a 7.3 times greater chance of a second EHS episode (793). Despite the apparent return to homeostasis, underlying repair or regenerative processes may be ongoing after EHS that could impose an increased risk of a subsequent episode of EHS but are relatively invisible to normal clinical testing (see FIGURE 13). This is particularly true if such repair mechanisms are ongoing in the heart, because cardiovascular collapse is the hallmark of the final stages of heat stroke in animals (794) and in severely heat-stressed humans (795). At a cellular/molecular level, acute hyperthermia can result in long-term effects on cellular function and epigenetics (796) that may remain hidden until exposure to a secondary stress or in response to long-term processes associated with aging (797).

FIGURE 13.

Overview of the three aspects (damage to the hypothalamus, myocardial dysfunction, epigenetic alterations) putatively associated with long-term consequences of heat stroke covered in sect. 6.3.

6.3.2. Hypothalamic injury/dysfunction following heat stroke.

A widespread assumption is that a heat stroke episode leads to hypothalamic damage, but this paradigm is considered a misconception in the context of heat stroke (26). The origin of this paradigm emerged from early studies reporting transient hypothermia or recurrent fever following heat stroke episodes (798). These abnormalities were attributed to damage of the POAH, considered the thermoregulatory controlling hub of the central nervous system (799). Damage to the POAH (e.g., transient or permanent) would result in thermoregulatory failure and an inability to maintain the deep body temperature within a physiological range after collapse. However, a study performed necropsies on 125 fatal military EHS cases and did not observe structural changes to the POAH (798), suggesting no identifiable damage to this area. In support of this view, a study used computed tomography (CT) scan to view brain structures of classic and exertional heat stroke victims and did not find abnormalities to any regions of the hypothalamus (800). It is interesting to note that brain abnormalities were only detected by CT scan in fatal instances.

In addition to the POAH, other regions of the brain have been studied in the context of heat stroke. Dysfunction of the cerebellum accounts for central nervous system dysfunction before the episode and postrecovery issues (e.g., ataxia) in heat stroke victims (784). The precise mechanism underlying cerebellum dysfunction in heat stroke victims remains unknown but seems to involve a selective vulnerability of Purkinje cells to heat-induced injury (801). For instance, a case study series assessing classic heat stroke victims demonstrated severe loss of Purkinje cells as the predominant neuropathological change observed in MRI scans, which may also explain the ataxia early during a heat stroke episode (801). Other documented brain abnormalities in heat stroke include cerebral edema, loss of gray-white matter discrimination, and microhemorrhages (798, 800).

Despite the lack of apparent damage to the hypothalamus, studies have demonstrated elevations in proinflammatory mediators following heat stroke including increases in eosinophilic neurons (802). Studies in mice confirmed transient disturbances in the thermoregulatory feedback loop during the first day of CHS recovery, suggesting transient dysfunction of the efferent motor pathways, peripheral sensors of body deep body temperature, afferent sensory feedback pathways, or integration of these afferent sensory signals within the POAH. However, definite POAH cell or tissue damage was not evident in these animals (803). In most cases of heat stroke, in humans and rodents, hypothermia and fever during recovery are regarded a regulated response likely to have protective value in preventing overall organ injury and control of the SIRS.

It is reasonable to assume that upon collapse due to a heat stroke, transient neuronal dysfunction may occur in most regions of the brain. Nevertheless, as demonstrated by the resumption of consciousness in victims who are rapidly cooled, this is likely a reversible response and does not impact overall thermoregulatory control in the following recovery.

6.3.3. Long-term cardiovascular consequences of heat stroke.

Long-term cardiovascular dysfunction has been documented following heat stroke. A 14-yr follow-up study of military personnel hospitalized for heat stroke reported an ∼2.2 times greater risk of dying of ischemic heart disease and an ∼1.7 greater risk of dying from other forms of cardiovascular disease (780). Another 14-yr follow-up study of heat stroke victims reported an ∼4 times higher incidence of major cardiovascular events, an ∼5.5 times greater incidence of ischemic stroke (781). However, whether a heat stroke episode can lead to the emergence of cardiovascular diseases later in life is less clear. Given the risk of lethality, it is not possible to expose humans to an experimental heat stroke episode to perform hypothesis-driven research to definitively answer this question. This is where preclinical models of heat stroke can serve as a helpful tool to help answer these questions.

A study using a preclinical model of EHS in mice tested whether a single episode of EHS would result in myocardial abnormalities up to 14 days after the episode (776). It is worth highlighting that at this time point (e.g., 14 days), mice have clinically and behaviorally recovered from the heat stroke insult. Male and female C57bl/6J mice exercised for 1.5–3.0 h in a 37.5°C/40% RH until displaying central nervous system dysfunction characterized by loss of consciousness. Mice reached peak deep body temperatures of 42.2 ± 0.3°C. Interestingly, despite being smaller, females ran ∼40% longer, reaching ∼51% greater heat load in this model (804). Metabolic biomarkers of changes in cardiac metabolites that could indicate long-term consequences were studied. At 9–14 days, the myocardium of female mice developed marked elevations in free fatty acids, ceramides and diacylglycerols. Glycolytic and tricarboxylic acid cycle metabolites revealed bottlenecks in substrate flow, with build-up of intermediate metabolites consistent with oxidative stress and damage. The results suggest that the myocardium of female mice is vulnerable to a slowly emerging metabolic disorder following EHS that may harbinger long-term cardiovascular complications. An absence of similar findings in males was attributed to lower heat stress and shorter exercise duration. Differences in body size per se do not seem to explain sex differences observed in preclinical EHS (804). Studies exploring the role of sex hormones in the response to EHS are warranted.

6.3.4. Epigenetic alterations induced by heat stroke.

The field of epigenetics (i.e., how behavioral and/or environmental stimuli can alter the way genes work) and heat adaptation/maladaptation is a relatively new area of research. The first interest in epigenetics changes and heat stress was triggered by observations that repeated heat acclimatization following 18 days of acclimatization decay was profoundly faster than the period required to achieve initial acclimatization (805), suggestive of a “memory” phenomenon. This hypothesis was elegantly confirmed by preclinical experiments indicating that heat acclimation was indeed associated with posttranslational modification of histones in gene networks associated with cytoprotection (806, 807). These pioneering observations demonstrate that epigenetics alterations are associated with heat adaptation, with recent evidence suggesting that the long-term health consequences of heat stroke may involve epigenetic reprograming (808, 809). Epigenetic changes alter gene expression and change the ability of an organism to respond to future stimuli. At 30 days after a single episode of EHS, very specific alterations (e.g., hypo- or hypermethylation) were observed within the promoter regions of genes involved with the immune response. To determine whether these changes were related to phenotypic changes in the immune system, the whole blood, collected at 30 days post-EHS, was challenged with lipopolysaccharide (LPS) to measure cytokine secretion; monocytes were also challenged with heat shock to quantify mRNA expression. Whole blood collected from EHS mice showed markedly attenuated inflammatory responses to LPS challenge. Furthermore, monocyte mRNA from EHS mice showed altered responses to heat shock challenge. These results are consistent with EHS leading to a specific DNA methylation pattern in monocytes and altered immune and heat shock response 30 days after EHS. These preclinical data in mice support the hypothesis that EHS exposure can induce long-term physiological changes that may be linked to altered epigenetic profiles. Whether CHS results in similar epigenetic responses remains to be addressed. Future studies must address whether these preclinical observations hold true in humans.

7. SUMMARY AND PERSPECTIVES

This review has emphasized that the magnitude by which deep body (deep body) temperature rises during heat stress is related to the interaction between biophysical properties and processes of autonomic temperature regulation. From a biophysical point of view, metabolic heat production is the main determinant of deep body temperature. For a given absolute rate of heat production, the resultant change in deep body temperature is inversely related to mass (the body’s heat sink) and body composition (the body’s specific heat capacity), although the latter becomes apparent only when comparing individuals of vastly different adiposity levels. Increases in skin blood flow and sweat rate elevate the rate of skin surface heat loss in proportion to the imposed heat load in an effort to restore heat balance. The effectiveness of these thermoeffector responses in maintaining heat balance is limited to work intensities and ambient conditions in which the capacity for whole body heat loss does not exceed heat production. Numerous intrinsic/biological factors, diseases, and injuries interfere with the autonomic regulation of heat loss responses by disrupting afferent or efferent neural pathways or altering thermoeffector end-organ function.

With impairments in the control of skin blood flow and sweat rate, excessive heat strain may occur, placing the individual at greater risk of heat-related health complications (14). The possibility of such outcomes is likely to rise as the Earth’s average surface temperature (810), as well as the incidence, severity, and frequency of heat waves (6), is projected to increase in the coming decades. Therefore, a continued research effort to further our understanding of human temperature regulation during heat stress is needed to minimize the projected consequences of global warming for human health. Important unanswered questions include the following:

• What are the neural circuitries involved in the neural integration of thermosensory signaling in humans?

• To what extent does impaired thermosensory signaling impact temperature regulation in various populations?

• To what extent does impaired neural integration of thermosensory signaling contribute to altered temperature regulation in various populations?

• How does thermolytic capacity and limits of heat exposure vary in different populations? What, if any, lifestyle changes (e.g., diet, supplements, exercise) can augment thermolytic capacity in these individuals?

• When exposed to heat stress, what strategies beyond air conditioning can be employed to mitigate heat strain in the most vulnerable populations?

Answering these (and other) questions has important implications for heat management guidance and illness prevention in public health, occupational, and athletics sectors.

GRANTS

The presented work that was conducted by the authors was funded by the National Institutes of Health Grants GM068865 (to C.G.C.), R56AG069005 (to C.G.C.), R01AG096005 (to C.G.C.), R01HL061388 (to C.G.C.), Department of Defense W81XWH-15-1-0647 (to C.G.C.), Natural Sciences and Engineering Research Council of Canada (to M.N.C. and D.G.), and the Fonds de Recherche du Québec–Santé (to D.G.).

DISCLOSURES

No conflicts of interest, financial or otherwise, are declared by the authors.

AUTHOR CONTRIBUTIONS

M.N.C., D.G., and O.L. prepared figures; M.N.C., D.G., O.L., and C.G.C. drafted manuscript; M.N.C., D.G., O.L., and C.G.C. edited and revised manuscript; M.N.C., D.G., O.L., and C.G.C. approved final version of manuscript.