Internal and skin temperatures increase during exercise and/or exposure to elevated environmental temperatures. Excessive increases in body temperature can impair physical performance and may lead to heat-related illnesses and death. Evaporation of sweat as well as cutaneous vasodilation are necessary components for appropriate thermoregulation in humans. As skin and internal (i.e., primarily hypothalamic) temperatures increase, sweating from eccrine glands occurs after the release of acetylcholine, as well as co-released peptides, from sympathetic cholinergic nerves that innervate the sweat gland. In addition, hormonal influences can alter sweating via central mechanisms, as well as have a direct affect on the sweat gland. Upon neural stimulation, the gland secretes a fluid that is similar in composition to plasma but without plasma proteins into the duct portion of the sweat apparatus. As this fluid moves through the duct, it is modified by reabsorption of most of the sodium and chloride ions, as well as some water, resulting in a hypotonic fluid’s being secreted to the surface of the skin.
In addition to thermal and hormonal modulators of sweating, a number of studies have assessed the importance of non-thermal neural modulators of sweating (Fig. 1). During exposure to hyperthermic conditions, changes in sweating typically parallel changes in cutaneous vascular conductance, which led to the hypothesis that sweating and active cutaneous vasodilation are not independently controlled. However, we found that it is possible for sweating responses to occur in an opposite direction relative to active cutaneous vasodilation (3), and others have shown that cutaneous vasodilation and sweating responses are not always linked. These findings suggests that these responses can be independently controlled by efferent neural signals, and provides a rationale to study non-thermoregulatory control of sweating independent of non-thermoregulatory control of skin blood flow in humans.
The intent of this review is to summarize the primary non-thermal neural modulators of sweat rate; namely factors associated with exercise, baroreflex loading status, and dehydration. Because the control of sweating is different in glabrous skin (e.g., palms of hands, soles of feet, lips, etc.) relative to nonglabrous skin, and nonglabrous skin is the primary region for thermoregulatory sweating, this review will focus solely on studies involving the effects of non-thermal modifiers of sweating in nonglabrous skin of humans.
van Beaumont and Bullard (12) showed that, during short-term exercise in warm nonsweating subjects, sweating occurred almost immediately with the onset of exercise. Because the duration of exercise was insufficient to elevate internal temperature, the authors concluded that non-thermal factors associated with exercise contributed to the elevation in sweating. Similar conclusions were made from a study in which subjects cycled on an ergometer for longer periods, during which workload was changed in a sinusoidal manner (15). They found that sweating responses followed the sinusoidal pattern of exercise, not skin or internal temperatures, during the acute phase of the change in exercise load. Together, these and other studies suggest that non-thermal factors related to exercise are capable of modulating sweating responses. These factors may include stimulation of receptors in the exercising muscle that are sensitive to metabolic, mechanical, and thermal stimuli, as well as factors related to central control of motor unit recruitment (i.e., central command).
To gain insight into the effects of exercise on neural control of sweating, Vissing et al. (13) measured cutaneous nerve activity (i.e., postganglionic skin sympathetic nerve activity (SSNA)) during isometric handgrip exercise in normothermic and mildly heated subjects. Immediately after the exercise bout blood flow to the exercising limb was occluded. Limb occlusion was performed after exercise to remove the potential influence of central command and muscle mechanoreceptor stimulation in governing SSNA while trapping metabolites, and thus stimulating metaboreceptors, in the previously contracted muscle. These data indicated that central command, but not muscle metaboreceptor stimulation, caused the observed elevation in SSNA during isometric exercise, because the increase in SSNA during exercise returned to pre-exercise levels during postexercise ischemia (13). This hypothesis was confirmed in a follow-up study in which SSNA increased during attempted isometric handgrip exercise (via partial neuromuscular blockade), when neither muscle metaboreceptors nor mechanoreceptors were appreciably stimulated, but central command was augmented (13). Both findings provide strong evidence that central command is capable of modulating SSNA. However, these studies should be viewed with the understanding that SSNA measurements are not necessarily indicative of just sweating because the SSNA signal may contain sudomotor, cutaneous vasoconstrictor, cutaneous vasodilator, and/or pilomotor neural signals.
Perhaps due to the variety of end organ responses governed by SSNA, increases in SSNA during isometric exercise do not consistently lead to similar changes in efferent responses such as sweating. Thus, observations that SSNA can be modulated by central command do not necessarily confirm the hypothesis that central command is also capable of modulating sweating, nor do they refute the hypothesis that muscle afferents are capable of modulating sweating.
To investigate these questions, we measured sweating during isometric handgrip exercise and subsequent postexercise ischemia under a variety of thermal conditions (3,7,9), including measurements of sweating in normothermia from sites in which acetylcholinesterase was locally inhibited via microdialysis administration of neostigmine (9). We observed that in both normothermic and heat-stressed conditions sweat rate from nonglabrous skin of the nonexercising forearm (3,7,9) and the chest (9) increased during isometric exercise (Fig. 2A). In contrast to the aforementioned SSNA responses, sweat rate from these areas remained elevated during postexercise ischemia in both thermal conditions. Ischemia alone (i.e., not preceded by exercise) did not change sweat rate regardless of the thermal state. Key to the interpretation of these data is the return of sweat rate to pre-exercise levels upon the release of ischemia under normothermic and moderately heated conditions (see Fig. 2).
To investigate the contribution of central command in modulating the sweating response during handgrip exercise, we repeated the aforementioned protocol during partial neuromuscular blockade (notice the force curve in Fig. 2B). The subjects were unable to maintain the desired force, but continued to attempt to exercise throughout the 2-min period. Appropriate increases in heart rate and blood pressure were observed during this time. Importantly, sweat rate continued to increase even when force production was close to 0 kg. During postexercise ischemia, sweat rate and blood pressure returned to pre-exercise levels, indicating the absence of muscle metaboreceptor stimulation, which is in contrast to observations during the control condition (Fig. 2A). These preliminary data provide evidence that sweating can be modulated by factors associated with central command.
Blood pressure increases during isometric exercise and remains elevated during postexercise ischemia. This, coupled with the possibility that baroreceptors may modulate sweating responses (see baroreceptor section below), led us to test the hypothesis that increases in blood pressure during postexercise ischemia contribute to the elevation in forearm sweating observed in this condition (9). This objective was accomplished by restoring blood pressure during the period of postexercise ischemia to pre-exercise levels via systemic infusion of sodium nitroprusside (Fig. 2C). However, sweat rate remained elevated during the ischemic period both with and without the reduction in blood pressure. Thus, the elevation in forearm sweat rate during postexercise ischemia are unaffected by baroreceptor loading associated with increases in blood pressure during the ischemic period.
Taken together, these findings provide strong evidence that central command and stimulation of metabolic sensitive receptors in the exercising muscle contribute to the sweating response during isometric exercise. However, the contribution of stimulation of muscle mechanorceptors in modulating sweating responses during exercise remains an open question.
In humans it has been proposed that there are three primary baroreceptor populations: the cardiopulmonary baroreceptors (or so-called “low-pressure” baroreceptors), the aortic baroreceptors, and the carotid baroreceptors. The latter two are commonly called arterial baroreceptors because they reside within the arterial circulation. These baroreceptors are important for the regulation of blood pressure primarily through neural control of heart rate and vascular resistance. Less clear is whether baroreceptors are capable of modulating sweating in humans. Sweating will lead to a reduction in blood volume if fluid is not replaced, and reductions in blood volume may reduce blood pressure. If blood pressure were reduced due to excessive decreases in blood volume (e.g., prolonged sweating or hemorrhage), theoretically it would be beneficial for baroreceptors to attenuate sweating in an attempt to preserve blood volume.
Early conclusions regarding baroreflex modulation of sweating originated from SSNA recordings (1). Bini et al. (1) reported that in heated individuals SSNA bursts occurred in volleys time-locked to the cardiac cycle, and suggested that sudomotor activity might be under baroreceptor control. The investigators emphasized that cardiac rhythmicity of SSNA was most clearly observed when the sudomotor system was strongly activated by pronounced heating. However, because SSNA did not consistently change during maneuvers that caused larger changes in blood pressure than those that occur spontaneously during quiet rest, Bini et al. (1) also proposed that cardiac rhythmicity of the SSNA signal in heated individuals may be attributable to a common central mechanism independent of baroreflex pathways (possibly thermoregulatory in nature). Nevertheless, due to the aforementioned mixed nature of integrated SSNA, these data must be viewed with the understanding that the SSNA signal recorded by Bini et al. (1) may not have been exclusively sudomotor in nature.
To investigate the effects of baroreceptor unloading on sweat rate, Mack et al. (8) measured chest sweating from subjects exercising in a moderately warm environment during the application of 40 mm Hg lower-body negative pressure (LBNP). They reported that baroreceptor unloading delayed the onset of sweating. In a similar study, Solack et al. (10) measured forearm sweat rate during application of LBNP in subjects after internal temperature was elevated via passive heated (i.e., water-perfused suit). They reported that sweat rate did not decrease with the onset of LBNP, but the slope of the increase in sweat rate relative to the increase in internal temperature was reduced in many subjects. However, because this slope did not recover in most subjects when LBNP was discontinued, the authors concluded that sweat rate may not be modulated by baroreceptors. Vissing et al. (13) also investigated this question, but they suggested that the reduction in SSNA and skin electrodermal activity (as an indicator of sweating activity) during lower-body negative pressure was attributable to skin cooling associated with the application of the negative pressure and not to baroreceptor unloading. Thus, it may be that reductions in sweating observed by others during lower-body negative pressure were due to slight reductions in skin temperature (i.e., thermal factors) and not necessarily due to baroreceptor unloading.
In addressing this question, Dodt et al. (4) passively heated individuals using a warming lamp until sweating was noticeable, and then applied either −5 and −10 mm Hg lower-body negative pressure or 30° head-up tilt to unload baroreceptors. They reported that forearm SSNA (posterior cutaneous nerve or radial nerve) and skin electrodermal activity decreased during baroreceptor unloading and then returned to preperturbation levels upon cessation of baroreceptor unloading. Because reductions in both SSNA and an index of sweat rate (i.e., skin electrodermal activity) were observed during head-up tilt, which would not cause cooling, these investigators concluded that baroreceptor unloading could modulate SSNA and sweating in moderately warm subjects.
Given these apparently contradictory findings, we (14) measured the effects of acute baroreceptor unloading and loading, as well as sustained baroreceptor unloading, in normothermic and heat-stressed humans. Pharmacological agents were used to change blood pressure, thereby perturbing the baroreceptors without the potentially confounding influences of cooling or emotional factors that may be associated with other methods of baroreceptors loading/unloading. Throughout the protocol SSNA was measured from the peroneal nerve whereas sweat rate was measured within the field of afferent innervation of the recorded nerve. For the bolus infusion protocol, mean blood pressure was reduced via administration of nitroprusside and then elevated ∼60 sec later via bolus infusion of phenylephrine. For the steady-state protocol, mean blood pressure was reduced over a period of 8–10 min via steady-state infusion of nitroprusside. In spite of these acute and sustained changes in blood pressure, neither SSNA nor sweat rate were significantly affected (Fig. 3).
One possible explanation for the discrepancy between our findings (14) and the findings of others (4,8,10) may be related to differences in the baroreceptor populations unloaded. Both lower-body negative pressure and 30° head-up tilt likely cause greater unloading of cardiopulmonary baroreceptors (i.e., lower-pressure baroreceptors) relative to unloading of the arterial baroreceptors (i.e., carotid and aortic baroreceptors). In contrast, pharmacologically induced decreases in blood pressure may cause greater unloading of the arterial baroreceptors relative to the cardiopulmonary baroreceptors. Although unsubstantiated, it may be that cardiopulmonary baroreceptor unloading is necessary to alter SSNA and sweat rate, whereas arterial baroreceptor unloading will not affect these variables. In support of this hypothesis, prior work suggests that the carotid baroreceptors may not have an efferent limb governing skin blood flow (2) or SSNA. Thus, it is possible that the aortic and carotid baroreceptors do not have an efferent limb governing sweat rate. Nevertheless, this explanation alone is insufficient to account for all of the differences in responses between the cited (4,8,10,13,14) and other studies. Thus, whether baroreceptors are capable of modulating sweat rate, or the degree to which this occurs, remains an unanswered question.
HYPOHYDRATION AND HYPEROSMOLALITY
Prior studies have demonstrated that sweating responses are impaired when subjects are hypohydrated and/or have elevated plasma osmolality. Exercise, particularly in hot environmental conditions, as well as prolonged passive heat stress will lead to dehydration if fluid intake is inadequate. Sweating is the primary mechanism of this heat-induced dehydration. Because sweat is hypotonic, plasma osmolality increases and plasma volume decreases during heat-stress–induced dehydration. Given the precise regulation of plasma osmolality, coupled with the less precise although still tightly regulated plasma volume control, it is not surprising that both hydration status and plasma osmolality modify sweating responses.
Very small increases in plasma osmolality evoke physiological feedback responses, such as vasopressin secretion and increased thirst, to regulate plasma osmolality. Relatively large increases in plasma osmolality have been reported after prolonged exercise in the heat. A number of studies have investigated the effects of elevated plasma osmolality on sweating responses. Two of these studies, which represent the general findings found in the literature, are discussed below. Fortney et al. (6) conducted a study in which subjects were normovolemic but hyperosmotic (increased plasma osmolality from 283 to 294 mosmol·kg−1). This condition was accomplished via fluid restriction and mild exercise in hot ambient conditions, followed by infusion of 3% saline to restore plasma volume to pre-exercise levels but maintain elevated plasma osmolality. The subjects then exercised in a warm room (30°C; 40% relative humidity) for 30 min. During the hyperosmotic exercise bout, the internal temperature threshold for the onset of sweating was elevated by approximately 0.6°C relative to the responses during exercise under iso-osmotic conditions. However, the slope of the relationship between the elevation in sweat rate and the elevation in internal temperature was not affected by increased plasma osmolality. Similar results were reported by Takamata et al. (11) during passive heating (via lower leg immersion in warm water) in which plasma osmolality was elevated via infusion 3% saline. They also demonstrated that elevated plasma osmolality increased the internal temperature threshold for the onset of sweating (Fig. 4). Taken together, increased plasma osmolality independent of plasma volume impairs sweating responses primarily through elevating the internal temperature threshold for the onset of sweating.
Increased extracellular and cerebral spinal fluid osmolality promotes thirst, and the acute effects of drinking fluid may improve sweating responses in dehydrated subjects. To examine this question, Takamata et al. (11) had heat-stressed subjects drink ∼320 ml of 38°C water under hyperosmotic (3% saline infusion; plasma osmolality: 297 mosm·kg−1 H20) and iso-osmotic (0.9% saline infusion; plasma osmolality: 287 mosm·kg−1 H20) conditions. As previously reported, sweating responses were attenuated under the hyperosmotic conditions. However, within minutes of drinking, significant increases in sweat rate were observed (see Fig. 5). This rapid elevation in sweating occurred without changes in plasma osmolality or plasma volume. In contrast, under iso-osmotic conditions drinking warm water had no effect on sweating. It should be emphasized that although sweating responses were similar between conditions before drinking, to achieve a sweat rate of ∼0.7 mg·min−1·cm−2 under the hyperosmotic condition, the average esophageal temperature was ∼0.5°C higher (see Fig. 5). From these findings the investigators suggest that afferents in the oropharyngeal region are capable of contributing to the modulation of sweat rate under hyperosmotic conditions. The mechanism(s) by which drinking rapidly increases sweating remains to be identified.
Studies have shown that independent of osmotic status, hydration status is capable of altering sweating responses. For example, iso-osmotic hypovolemia induced by diuretics, in which blood volume was decreased by 8.7% without altering plasma osmolality, reduced both whole-body sweat rate during exercise as well as the slope of the relationship between the elevation in sweat rate and the elevation in internal temperature (5). Inhibition of sweating during hypovolemia may defend central venous pressure during a heat stress, perhaps through the aforementioned baroreflex mechanisms. In contrast to hypovolemic studies, expansion of plasma volume by 7.9% without altering plasma osmolality did not have a significant effect on sweating during passive heating (5).
Appropriate sweating responses are critical for thermoregulation in humans. In addition to the widely recognized thermal controllers of sweating (i.e., internal and skin temperatures), a number of non-thermal factors contribute to the overall sweating response. For example, during static exercise sweat rate is elevated by central command, by muscle metaboreceptor stimulation, and possibly by factors associated with muscle mechanoreceptor stimulation. Less understood is the potential contribution of these factors in modulating sweat rate during dynamic exercise (15). The effects of baroreceptor unloading on sweating also require further investigation. Differences in findings related to baroreceptor modulation of sweating may be attributable to the variables measured as an index of sweating (i.e., skin electrodermal activity or evaporative water loss) as well as the baroreceptor population perturbed. Finally, evidence exists to indicate that both elevated plasma osmolality and iso-osmotic hypovolemia impair sweating responses, whereas impaired sweating due to elevated plasma osmolality can be reversed by drinking, perhaps via stimulation of afferents in the oropharyngeal region.
The authors want to express their appreciation to John Johnson, Ph.D., Thad Wilson, Ph.D., and Kimberly Williams, R.N. for their review and comments regarding the manuscript. The authors also want to apologize to investigators working in this field of study whose work should be cited in this review but was not due to limitations on the number of citations allowed. Support for the authors was provided by the National Institutes of Health, National Heart, Lung, and Blood Institute (HL61388 and HL67422, C.G.C.), the National Aeronautics and Space Administration (NAG91033, C.G.C.), a Grant-in-Aid for the Encouragement of Young Scientists from the Japanese Society for the Promotion of Science (12780024, M.S.), and a Grant-in-Aid for Scientific Research from the Ministry of Education, Science, Sports and Culture (11780017, N.K.).
1. Bini, G., K. E. Hagbarth, and B. G. Wallin. Cardiac rhythmicity of skin sympathetic activity recorded from peripheral nerves in man. J. Auton. Nerv. Syst. 4: 17–24, 1981.
2. Crandall, C. G., J. M. Johnson, W. A. Kosiba, and D. L. Kellogg,Jr. Baroreceptor control of the cutaneous active vasodilator system. J. Appl. Physiol. 81: 2192–2198, 1996.
3. Crandall, C. G., D. P. Stephens, and J. M. Johnson. Muscle metaboreceptor activation reduces cutaneous active vasodilator activity during isometric exercise. Med. Sci. Sports Exerc. 30: 490–496, 1998.
4. Dodt, C., T. Gunnarsson, M. Elam, T. Karlsson, and B. G. Wallin. Central blood volume influences sympathetic sudomotor nerve traffic in warm humans
. Acta. Physiol. Scand. 155: 41–51, 1995.
5. Fortney, S. M., E. R. Nadel, C. B. Wenger, and J. R. Bove. Effect of blood volume on sweating rate and body fluids in exercising humans
. J. Appl. Physiol. 51: 1594–1600, 1981.
6. Fortney, S. M., C. B. Wenger, J. R. Bove, and E. R. Nadel. Effect of hyperosmolality on control of blood flow and sweating. J. Appl. Physiol. 57: 1688–1695, 1984.
7. Kondo, N., H. Tominaga, M. Shibasaki, K. Aoki, S. Koga, and T. Nishiyasu. Modulation of the thermoregulatory sweating response to mild hyperthermia during activation of the muscle metaboreflex
. J. Physiol. 515: 591–598, 1999.
8. Mack, G. W., D. Cordero, and J. Peters. Baroreceptor modulation of active cutaneous vasodilation during dynamic exercise in humans
. J. Appl. Physiol. 90: 1464–1473, 2001.
9. Shibasaki, M., N. Kondo, and C. G. Crandall. Evidence for metaboreceptor stimulation of sweating in normothermic and heat-stressed humans
. J. Physiol. (Lond). 534: 605–611, 2001.
10. Solack, S. D., G. L. Brengelmann, and P. R. Freund. Sweat rate
vs forearm blood flow during lower body negative pressure. J. Appl. Physiol. 58: 1546–1552, 1985.
11. Takamata, A., G. W. Mack, C. M. Gillen, A. C. Jozsi, and E. R. Nadel. Osmoregulatory modulation of thermal sweating in humans
: reflex effects of drinking. Am. J. Physiol. 268: R414–R422, 1995.
12. Van Beaumont, W., and R. W. Bullard. Sweating: its rapid responses to muscular work. Science. 141: 643–646, 1963.
13. Vissing, S. F. Differential activation of sympathetic discharge to skin and skeletal muscle in humans
. Acta Physiol. Scand. 161 (suppl639): 1–32, 1997.
14. Wilson, T. E., J. Cui, and C. G. Crandall. Absence of baroreflex
modulation of skin sympathetic nerve activity and sweat rate
during whole-body heating in humans
. J. Physiol. (Lond). 536: 615–623, 2001.
15. Yamazaki, F., R. Sone, and H. Ikegami. Responses of sweating and body temperature to sinusoidal exercise. J. Appl. Physiol. 76: 2541–2545, 1994.