Secondary Logo

Impaired Thermoregulatory Function during Dynamic Exercise in Multiple Sclerosis


Medicine & Science in Sports & Exercise: March 2019 - Volume 51 - Issue 3 - p 395–404
doi: 10.1249/MSS.0000000000001821

Introduction Impairments in sudomotor function during passive whole-body heating have been reported in multiple sclerosis (MS), a demyelinating disease of the CNS that disrupts autonomic function. However, the capability of the thermoregulatory system to control body temperature during exercise has never been assessed in MS. Thus, the aim of the present study was to test the hypothesis that thermoregulatory function is impaired in MS patients compared with healthy controls (CON) exercising at similar rates of metabolic heat production.

Methods Sweating and skin blood flow responses were compared between 12 individuals diagnosed with relapsing-remitting MS (9 females, 3 males) and 12 sex-, age-, mass-, and BSA-matched CON during a single bout of cycling exercise (rate of metabolic heat production: ∼4.5 W·kg−1) for 60 min in a climate-controlled room (25°C, 30% RH).

Results Individuals with MS exhibited an attenuated increase in cumulative whole-body sweat loss after 30 min (MS, 72 ± 51 g; CON, 104 ± 37 g; P = 0.04) and 60 min (MS, 209 ± 94 g; CON, 285 ± 62 g; P = 0.02), as well as lower sweating thermosensitivity (MS, 0.49 ± 0.26 mg·cm−2·min−1·°C−1; CON, 0.86 ± 0.30 mg·cm−2·min−1·°C−1; P = 0.049). Despite evidence for thermoregulatory dysfunction, there were no differences between MS and CON in esophageal or rectal temperatures at 30- or 60-min time points (P > 0.05). Cutaneous vasculature responses were also not different in MS compared with CON (P > 0.05).

Conclusion Taken together, MS blunts sweating responses during exercise while cutaneous vasculature responses are preserved. Altered mechanisms of body temperature regulation in persons with MS may lead to temporary worsening of disease symptoms and limit exercise tolerance under more thermally challenging conditions.

1Applied Physiology and Wellness, Southern Methodist University, Dallas, TX;

2Department of Health Sciences, Boston University, Boston, MA;

3Thermal Ergonomics Laboratory, Faculty of Health Sciences, The University of Sydney, Lidcombe, NSW, AUSTRALIA;

4Neurology and Neurotherapeutics, University of Texas Southwestern Medical Center, Dallas, TX; and

5Charles Perkins Centre, The University of Sydney, Camperdown, NSW, AUSTRALIA

Address for correspondence: Scott L. Davis, Ph.D., Department of Applied Physiology and Wellness, Southern Methodist University, 3101 University Blvd., Suite 163, Dallas, TX 75205; E-mail:

Submitted for publication May 2018.

Accepted for publication September 2018.

Multiple sclerosis (MS) is the most disabling neurological disorder of young adults (1). The pathophysiology of MS is principally described as a progressive deterioration and loss of functional capabilities as a result of myelin destruction and axonal loss within the CNS (2,3). Dysfunction along central axons can lead to a myriad of symptoms (e.g., fatigue, gait difficulties, numbness/tingling, visual dysfunction, depression, cognitive changes, etc.) and has been shown to cause downstream autonomic dysfunction (4–7). Among the autonomic processes thought to be disturbed by MS, there are growing indications of thermoregulatory abnormalities (8,9) as a result of the disease. However, most of the studies mentioned used subjective “sweat tests” elicited by an axon reflex, with no concern for body temperature regulation or central integration of thermal information. Adding complexity, it is well documented that a majority of MS patients are subject to a reversible temporary worsening of their symptoms with increases in core temperature, known as Uhthoff’s phenomenon (10). It remains unknown if individuals with MS are capable of sufficiently activating thermoeffector countermeasures enough to prevent excessive elevations in core temperature during thermal stress and, thus, prevent temperature-related symptom worsening.

In healthy humans, an increase in body temperature causes an increase in cutaneous blood flow and the secretion of sweat from eccrine sweat glands to augment radiative, convective, and evaporative heat loss from the skin to the surrounding environment, a process known as the thermoregulatory reflex. Recently, our laboratory used a passive heat stress protocol to examine thermoregulatory control of sweating and cutaneous vascular conductance (CVC, an index of skin blood flow) in MS (11). This study demonstrated that individuals with relapsing-remitting MS had attenuated increases in sweat rate, but not CVC in response to passive increases in core temperature. Although the aforementioned findings clearly indicate a compromised autonomic thermoregulatory reflex in MS, encapsulation of the body by the perfusion suit used in the study does not allow for any meaningful disparity in evaporative or dry heat exchange from most of the body, thus limiting our ability to assess the practical effect of these thermoregulatory deficits on the prevailing heat strain with MS.

It remains unknown if the attenuated sweat rates observed in our previous study translate to an impaired regulation of body temperature (i.e., thermoregulatory function) in a nonencapsulated, practical setting, which allows heat dissipation from the body to the environment to be freely modified by differences in sudomotor and/or vasomotor responses. Therefore, the aim of this study was to test the hypothesis that individuals with MS will 1) exhibit heightened increases in T core during exercise of a fixed rate of heat production compared with age-, height-, mass-, and sex-matched healthy controls (CON), and 2) these elevations in T core will likely be attributed to reduced potential for evaporative heat loss owing to attenuated sweating during exercise compared with CON. In addition, although sudomotor dysfunction is expected, we hypothesize that indices of skin blood flow control would be similar during exercise between the groups.

Back to Top | Article Outline



Individuals with definite relapsing-remitting MS (n = 12 [9 females, 3 males], expanded disability status scale, range 0–6) and age-, sex-, height-, and weight-matched CON (n = 12; 9 females, 3 males) voluntarily participated in this study. From pilot data of 6 (3 MS, 3 CON) subjects, a power calculation (G*power version with β set to 0.2 and α at 0.05, a minimum sample size of 16 subjects (8 MS, 8 CON) were required based on group mean whole-body sweat loss (WBSL) values (MS, 160 ± 106 g; CON, 281 ± 66 g; effect size: d = 1.37). In an effort to account for the heterogeneity of disease status in MS, additional subjects were recruited for the study.

All individuals with MS were otherwise healthy but clinically diagnosed and currently being treated by neurologists specializing in MS at the MS Clinic at the University of Texas Southwestern Medical Center. This study focused on relapsing-remitting MS because it is the most common disease course with ∼85% of individuals with MS initially diagnosed with this form of the disease. This form of MS is characterized by acute attacks (exacerbations) followed by periods of partial or complete recovery (remissions). All individuals with MS were on disease-modifying therapies. According to the 2015 consensus paper by the Multiple Sclerosis Coalition on the use of disease modifying therapies, FDA-approved disease-modifying treatment (DMT) is recommended and should be continued indefinitely. Individuals with MS on symptom modifying therapies that are known to affect the CNS and/or thermoregulatory responses (i.e., antidepressants, psychostimulants, anticonvulsants, antispasmatics, and anticholinergics) were excluded from the study. Individuals with MS remained on their DMT as prescribed by their neurologists but were asked to refrain from taking any additional supplements and/or over-the-counter medications before testing. Testing on individuals with MS taking interferon as their DMT were performed 3–5 d after injection to minimize the effect of flulike symptoms associated with the injection. These flulike symptoms resolve within 24–48 h after each injection. All individuals with MS were at least 6 months removed from their most recent relapse and reported experiencing no worsening in their disease symptomology.

All subjects were nonsmokers and had no history of metabolic, pulmonary, or cardiovascular disease. Subjects refrained from caffeine, alcohol, and vigorous exercise 24 h before the trial. All trials took place between 7:00 and 11:00 AM to avoid potential effects of circadian rhythm and outside of summer months in an effort to avoid any differences associated with acclimation status. All experimental procedures and protocols conformed to the Declaration of Helsinki and were approved by the Institutional Review Board at Southern Methodist University. All subjects provided written consent before volunteering to participate in the study.

Back to Top | Article Outline

Instrumentation and Measurements


Rectal temperature (T rec) was measured using a sterile thermocouple probe (Mon-a-Therm 400TM; Covidien, Mansfield, MA) inserted a minimum of 12 cm past the anal sphincter in all 24 subjects. Esophageal temperature (T eso) was measured by placing a sterile thermocouple probe (Mon-a-Therm 400TM, Covidien) inserted to the full length of the probe (40 cm) through the subject’s nostril into the esophagus. T eso measurements were only collected in seven matched pairs (MS vs CON) because of the inability for some individuals with MS to tolerate the thermocouple placement. Skin temperatures were collected at eight points (forehead, acromion, scapula, triceps, posterior hand, abdomen, quadriceps, and calf) using wireless temperature sensing buttons with a ±0.5°C accuracy and a resolution of 0.0625°C (iButton DS1922L; Embedded Data Systems, Lawrenceburg, KY). Temperature-sensing buttons were attached using surgical tape (Transpore; 3M, London, Ontario, Canada), and thermal data were sampled every 5 s before being exported to a personal computer to be subsequently collected and analyzed. Mean skin temperature (T sk) was estimated using a six-site weighted average as described elsewhere (12).

Back to Top | Article Outline

Sweat data

Body mass was measured in triplicate at 15-min intervals during the experimental protocol using a precision scale with an accuracy of ±2 g (Sartorius Combics 2, Goettingen, Germany). WBSL was determined as the change in body mass from baseline minus respiratory vapor exchange and metabolic mass losses (13). Local sweat rate (LSR) was collected continuously throughout the exercise protocol at the upper back and forearm using capacitance hygrometry (Vaisala HMT333, Helsinki, Finland) via the ventilated capsule method (14). In addition, sweat gland recruitment was quantified using iodine-stained resume 100% cotton paper (Southworth 24 lb Resume paper, Agawam, MA) printed with a 2 × 2-cm grid placed on the skin for ∼30 s. After the trial, this paper was immediately scanned (EPSON Perfection V6000, Owa, Suwa, Nagano, Japan) and quantified by hand counting each stained marking and validating it using a free online imaging software provided by the National Institutes of Health (ImageJ; NIH, Bethesda, MD) described elsewhere (15).

Back to Top | Article Outline

Metabolic data

Estimates for metabolic energy expenditure were determined via indirect calorimetry using a metabolic cart (Parvomedics TrueOne 2400, Sandy, UT). Subjects were instrumented with a mouthpiece and nose clip, while expired gases were continuously collected throughout exercise and analyzed in 1-min increments.

External work rate (W) was regulated and measured using an electronically braked semirecumbent cycle ergometer (Lode Corival, AEI Technologies, Naperville, IL). Estimations of the rate of heat production based on body mass or body surface area were expressed by dividing the difference between M and external work rate (MW = H prod) in watts by kilograms of body mass (W·kg−1) or body surface area (W·m−2) (16), respectively. A more detailed description of this technique is described below.

Back to Top | Article Outline

Hemodynamic data

Blood pressure (BP) was collected every 5 min via automated brachial artery auscultation (SunTech; Medical Instruments, Raleigh, NC). Heart rate (HR) was continuously monitored using a standard lead II surface ECG (Solar 8000i; General Electric, New York, NY) interfaced with a cardiotachometer (CWE Inc., Ardmore, PA). Skin blood flow, indexed as laser Doppler flux (LDF), was collected on the upper back and forearm using integrated laser Doppler probes (PF413; Perimed, Ardmore, PA) connected to a laser Doppler flowmeter (PF5010, Perimed). Probes were fitted inside thermostatic probe holders (PF450, Perimed) connected to a local heating device (PF5020, Perimed) used in this study to elicit thermally induced increases in cutaneous vasodilation.

Back to Top | Article Outline


Exercise intensity

The selection of exercise intensity was based on two criteria: 1) the workload selected should be sufficient to elicit significant increases in T core, while being modest enough for most individuals with MS to complete the entire 60-min trial; 2) the exercise intensity was chosen to elicit a similar rate of absolute metabolic heat production (H prod) between an MS individual and their CON match, as described in previous studies (17,18). In short, an absolute external workload of 70 W was initially assigned to all individuals to elicit ∼400 W of absolute metabolic energy expenditure and ∼330 W of metabolic heat production (assuming ∼17% efficiency for semirecumbent cycling based on pilot data) for comparisons of WBSL. Because MS individuals were tightly matched to CON of similar characteristics, this fixed rate of metabolic heat production was also a similar relative metabolic heat production on a watts per kilogram and watts per square meter basis. If individuals deviated from this H prod value (chosen a priori), workloads were adjusted accordingly in an effort to consistently maintain H prod at ∼330 W.

Back to Top | Article Outline

Experimental protocol

After completion of consent and medical history documents, each subject provided a urine sample, which was immediately analyzed for urine-specific gravity using a refractometer (Atago Pocket Refractometer, Kobe, Hyogo-ken, Japan), to ensure preexercise euhydration and that preexercise hydration status was similar between groups. A urine-specific gravity cutoff value of 1.020 was enforced as values below this threshold have been suggested to indicate euhydration (19). After confirmation of euhydration, subjects then changed into standard athletic clothing (running shorts, socks, and shoes) and were then instrumented with the aforementioned measurements outlined above. After instrumentation, baseline body mass measurements were taken while fully instrumented. Next, subjects were seated on a semirecumbent cycle ergometer (Lode Corival, Groningen, The Netherlands) in a climate-controlled room (Cantrol Environmental Systems Ltd., Markham, ON, Canada) set to 25°C 35% RH and asked to rest comfortably for 3 min of baseline measurements. Then subjects began cycling at a cadence of 60 to 80 rpm at an external workload of 70 W of resistance. At 15-min intervals throughout the exercise protocol, subjects stopped cycling and stepped on the adjacent scale for triplicate measures of body mass (fully instrumented). Once stable body mass measurements were collected (no longer than 2 min), subjects resumed cycling. After the final weigh-in at 60 min, subjects returned to their seat on the cycle ergometer for the local heating period, lasting 30 min. Neither food nor fluid were ingested during the experimental protocol, and air velocities remained minimal (less than 0.02 m·s−1) throughout the duration of the study.

Back to Top | Article Outline

Data Analysis

All data were continuously acquired at a sampling rate of 100 Hz on a 16-channel data acquisition system (Biopac, Santa Barbara, CA). Mean values from all measured variables were obtained from the final minute of each 15-min period during exercise. The rate of metabolic energy expenditure (M) was calculated using steady-state values for oxygen consumption (V˙O2) in liters per minute and RER during exercise using the following equation (16):

where e c is the caloric equivalent per liter of oxygen for the oxidation of carbohydrates (21.13 kJ·L−1), and e f is the caloric equivalent representing the oxidation of fats per liter of oxygen (19.62 kJ·L−1).

Thermosensitivity was calculated as the slope of the increase in effector organ (sweating and skin blood flow) response relative to increases in T eso. In the calculation of thermosensitivity, any data that had appeared to reach a plateau were excluded to avoid additional leverage on the regression line. Onset temperature threshold for sweating was calculated as the intercept of the thermosensitivity slope with resting sweat rate. Onset time threshold for sweating was calculated in a similar fashion, from the relationship of the increase in sweating relative to changes in time.

Skin blood flow is reported as CVC and calculated by dividing mean LDF values (arbitrary units [au]) by HR-weighted (20) mean arterial pressure calculated from brachial artery auscultation. CVC data were also normalized to maximal vasodilation obtained during the final minute of local heating at 44°C for 30 min and expressed as percentage of maximum CVC (%CVCmax).

Back to Top | Article Outline

Statistical Analysis

All values are presented as mean ± SD. Mixed-models ANOVA statistical analyses were not performed because of unanticipated differences in sample size between groups (see Exercise Tolerance section). Therefore, unpaired two-tailed t-tests were used to compare LSR, WBSL, CVC, T rec, and T eso at the 30- and 60-min interval of the exercise protocol. Unpaired t-tests were performed on baseline characteristics and thermal sensitivity. All analyses were conducted using IBM SPSS version 23 (IBM, Armonk, NY) and plotted via GraphPad Prism 7 (GraphPad Software Inc., La Jolla, CA). Statistical significance was accepted at P < 0.05. In instances where differences between the groups were statistically significant, a Cohen’s d was calculated to determine effect size.

Back to Top | Article Outline


Subject characteristics

By design, no differences were observed in age (MS, 40 ± 9 yr; CON, 36 ± 13 yr; P = 0.47), body mass (MS, 78 ± 14 kg; CON, 76 ± 18 kg; P = 0.73), height (MS, 175 ± 11 cm; CON, 173 ± 10 cm; P = 0.74), and BSA (MS, 1.9 ± 0.2 m2; CON, 1.9 ± 0.3 m2; P = 0.71) between MS and CON groups. In response to preparticipation screening questions, MS and CON groups reported similar activity levels (MS, 232 ± 168 min·wk−1; CON, 233 ± 151 min·wk−1; P = 0.98). In addition, groups had similar resting HR (MS, 76 ± 17 bpm; CON, 73 ± 10 bpm; P = 0.52), systolic BP (MS, 126 ± 25 mm Hg; CON, 116 ± 10 mm Hg; P = 0.08), diastolic BP (MS, 80 ± 10 mm Hg; CON, 81 ± 10 mm Hg; P = 0.90), and mean arterial pressures (MS, 96 ± 10 mm Hg; CON, 93 + 9 mm Hg; P = 0.43). DMT used by individuals with MS included the following: Avonex (interferon beta-1a), n = 2; Gilenya (fingolimod), n = 1; Copaxone (glatiramer acetate), n = 3; Tecfidera (dimethyl fumarate), n = 1; Tysabri (natalizumab), n = 3; no DMT reported, n = 2.

Back to Top | Article Outline

Exercise tolerance

The exercise intensity selected for this study was well tolerated in all 12 CON; however, only 10 persons with MS were able to complete the entire 60-min exercise bout. Two individuals with MS were unable to complete the exercise protocol and terminated exercise as a result of symptom worsening. One person with MS experienced overwhelming fatigue, whereas the second lost motor control in one lower limb. Despite exercise termination, hemodynamic and thermoregulatory responses from both of individuals were similar to those of the rest of the MS group. As such, final exercise values for these two individuals with MS were compared with their matched CON at the same time point MS subjects ended exercise (e.g., if MS subject completed 32 min of exercise, data were compared with values up to minute 32 of exercise of their matched CON subject).

Back to Top | Article Outline

Heat production

By design, absolute rates of H prod (MS, 314 ± 65 W; CON, 309 ± 49 W; P = 0.85), rates of H prod relative to body mass (MS, 4.2 ± 0.7 W·kg−1; CON, 4.5 ± 1.1 W·kg−1; P = 0.40), and rates of H prod relative to BSA (MS, 162 ± 22 W·m−2; CON, 170 ± 25 W·m−2; P = 0.41) were all similar between groups. In addition, rate of external work (MS, 70 ± 1 W; CON, 74 ± 7 W; P = 0.06) and gross mechanical efficiency (MS, 18.7% ± 3.95%; CON, 18.4% ± 1.5%; P = 0.85) were not different between groups.

Back to Top | Article Outline


Baseline T eso was not different between MS and CON groups (MS, 37.1°C ± 0.4°C vs CON, 37.1°C ± 0.2°C; P = 0.89). Likewise, differences in T eso did not reach statistical significance at 30 min (n = 7 vs 7; MS, 37.7°C ± 0.4°C; CON, 37.6°C ± 0.4°C; P = 0.56) and 60 min (n = 5 vs 5; MS, 37.8°C ± 0.2°C; CON, 37.6°C ± 0.4°C; P = 0.40) of exercise. Therefore, when represented as change from baseline, esophageal temperature was not different between groups at the 30-min (n = 7 vs 7, P = 0.08) and 60-min (n = 5 vs 5, P = 0.16) time points (Fig. 1A). Individual time course changes in T eso across the exercise protocol are presented for all CON (Fig. 1B) and all individuals with MS (Fig. 1C).



Similar to T eso, T rec at baseline (MS, 37.2°C ± 0.4°C vs CON, 37.3°C ± 0.2°C; P = 0.49), 30 min (n = 12 vs 12; MS, 37.7°C ± 0.3°C; CON, 37.7°C ± 0.2°C; P = 0.59), and 60 min (n = 10 vs 10; MS, 38.0°C ± 0.2°C; CON, 37.9°C ± 0.3°C; P = 0.33) were not different between groups. As a result, the MS group had statistically similar increases in T rec compared with CON at both 30 min (n = 12 vs 12, P = 0.07) and 60 min (n = 10 vs 10, P = 0.19; Fig. 2A). Individual time course changes in T rec across the exercise protocol are presented for all CON (Fig. 2B) and all individuals with MS (Fig. 2C).



Baseline T sk was similar between groups (MS, 31.14°C ± 0.7°C vs CON, 31.63°C ± 1.1°C; P = 0.15). In addition, T sk was not different at both 30-min (n = 12 vs 12; MS, 31.5°C ± 0.8°C vs CON, 31.9°C ± 0.9°C; P = 0.19) and 60-min (n = 10 vs 10; MS, 31.8°C ± 0.8°C vs CON, 31.8°C ± 1.2°C; P = 0.85) time points. Therefore, when reported as change from baseline, T sk was not different at both 30 min (n = 12 vs 12; MS, 0.40°C ± 0.4°C vs CON, 0.25°C ± 0.7°C; P = 0.74) and 60 min (n = 10 vs 10: MS, 0.76°C ± 0.9°C vs CON, 0.03°C ± 0.8°C; P = 0.15).

Back to Top | Article Outline

Sweating responses

Whole-body sweating responses between groups at the cessation of exercise are presented in Figure 3. WBSL was significantly lower at the 30-min (P = 0.03; effect size: d = 0.76) as well as the 60-min (P = 0.02; effect size: d = 1.03) time points, indicating a blunted cumulative sweating response in MS as exercise duration increased. Within groups, ΔLSR was not different between sites (dorsal forearm and upper back) at 30 min (MS, 0.30 ± 0.1 vs 0.39 ± 0.3 mg·cm−2·min−1, P = 0.32; CON, 0.39 ± 0.3 vs 0.32 vs 0.1 mg·cm−2·min−1, P = 0.47) and 60 min (MS, 0.44 ± 0.2 vs 0.40 ± 0.2 mg·cm−2·min−1, P = 0.69; CON, 0.44 ± 0.1 vs 0.47 ± 0.3 mg·cm−2·min−1, P = 0.79). In addition, ΔLSR was similar between groups at the dorsal forearm at the 30-min (MS, 0.29 ± 0.1 mg·cm−2·min−1; CON, 0.32 ± 0.1 mg·cm−2·min−1; P = 0.39) and 60-min (MS, 0.37 ± 0.1 mg·cm−2·min−1; CON, 0.43 ± 0.1 mg·cm−2·min−1; P = 0.25) time points, as well as the upper back at 30 min (MS, 0.39 ± 0.3 mg·cm−2·min−1; CON, 0.39 ± 0.3 mg·cm−2·min−1; P = 0.92) and 60 min (MS, 0.44 ± 0.2 mg·cm−2·min−1; CON, 0.47 ± 0.3 mg·cm−2·min−1; P = 0.69). Therefore, values from these sites were averaged together in all subsequent LSR analysis and figures. At 30 min (n = 12 vs 12; MS, 0.26 ± 0.17 mg·cm−2·min−1; CON, 0.27 ± 0.15 mg·cm−2·min−1; P = 0.80) as well as at 60 min (n = 10 vs 10; MS, 0.33 ± 0.19 mg·cm−2·min−1; CON, 0.35 ± 0.14 mg·cm−2·min−1; P = 0.97) of exercise, there were no differences in the steady-state ΔLSR between groups (Fig. 4). Individual time course changes in ΔLSR across the exercise protocol are presented for all CON (Fig. 4B) and all individuals with MS (Fig. 4C).





In addition, total sweat gland recruitment at the 60-min time point was similar between groups (n = 10 vs 10; MS, 84.9 ± 51.2 glands per square centimeter; CON, 67.1 ± 29.5 glands per square centimeter; P = 0.38).

Back to Top | Article Outline

Cutaneous vasculature responses

Baseline LDF (n = 12 vs 12; MS, 51 ± 14 au; CON, 61 ± 24 au; P = 0.20), CVC (n = 12 vs 12; MS, 0.52 ± 0.14 au·mm Hg−1; CON, 0.67 ± 0.30 au·mm Hg−1; P = 0.13), and CVC relative to maximum (CVCmax; n = 10 vs 10; MS, 18% ± 6%; CON, 21% ± 9%; P = 0.26) were similar between groups. Comparisons were between sites of LDF collection to determine whether there were any regional differences in cutaneous vasodilation in response to exercise. Within groups, there were no differences in ΔCVCmax between the dorsal forearm and the upper back at 30 min (n = 12 vs 12; MS, 29.6% ± 14% vs 33.2% ± 17%, P = 0.47; CON, 38.1% ± 14% vs 34.0% ± 8%, P = 0.84) and 60 min (n = 10 vs 10; MS, 41.0% ± 23% vs 31.4% ± 17%, P = 0.63; CON, 36.7% ± 8% vs 37.4% ± 14%, P = 0.89). Likewise, there were no differences in ΔCVCmax between groups at the dorsal forearm at 30 min (n = 12 vs 12; MS, 29.6% ± 14%; CON, 38.1% ± 14%; P = 0.21) and 60 min (n = 10 vs 10; MS, 41.0% ± 23%; CON, 37.4% ± 14%; P = 0.87), or at the upper back at 30 min (n = 12 vs 12; MS, 33.2% ± 17%; 34.0% ± 8%; P = 0.85) and 60 min (n = 10 vs 10; MS, 31.4% ± 17%; 37.4% ± 14%; P = 0.87) of exercise. As there were no differences in ΔCVCmax between the two sites (forearm and upper back), all subsequent measures of skin blood flow are presented as the average between the two sites for each group. There were no differences in ΔCVCmax between groups at the 30-min (n = 12 vs 12; MS, 12.2% ± 8.4%; CON, 13.6% ± 13.0%; P = 0.71) and 60-min (n = 10 vs 10; MS, 17.8% ± 12.4%; CON, 15.0% ± 10.4%; P = 0.44) point of exercise (Fig. 5A). Individual time course changes in ΔCVCmax across the exercise protocol are presented for all CON (Fig. 5B) and all individuals with MS (Fig. 5C).



Groups had similar increases (change from baseline) in LDF (n = 10 vs 10; MS, 205 ± 155 au; CON, 216 ± 69 au; P = 0.81) and CVC (n = 10 vs 10; MS, 2.27 ± 1.61 au·mm Hg−1; CON, 2.53 ± 0.84 au·mm Hg−1; P = 0.63) in response to local heating for ∼30 min to elicit a maximal vasodilatory response.

Back to Top | Article Outline

Thermosensitivity and sweating onset

The slope of the relationship between ΔT eso and mean ΔLSR, often called thermosensitivity, was significantly blunted (MS, 0.49 ± 0.26; CON, 0.86 ± 0.30; P = 0.049; d = 1.33) in MS compared with CON (Fig. 6A). Representative individual tracings of ΔLSR responses as a function of ΔT eso during exercise from one individual with MS and a matched healthy control are presented in Figure 6B. By contrast, the slope of the relationship between ΔT eso and mean ΔCVCmax was not different (P = 0.22) in MS compared with CON (Fig. 6C). Representative individual tracings of ΔCVCmax responses as a function of ΔT eso during exercise from one individual with MS and a matched healthy control are presented in Figure 6D.



The onset time for sweating was not different between groups (MS, 10 ± 7 min; CON, 7 ± 4 min; P = 0.31). In addition to the timing of sweat onset, threshold changes in T eso (n = 12 vs 12; MS, 0.29°C ± 0.24°C; CON, 0.22°C ± 0.15°C; P = 0.40) at which a detectable increase in LSR occurred was also not different between groups. Because of the necessity for BP measurements to accurately calculate CVC, the timing analysis and the analysis of onset thresholds for increases in CVC were not possible.

Back to Top | Article Outline


The primary finding of the present study is that individuals with MS had a lower WBSL during exercise due to a blunted thermosensitivity (i.e., smaller rise in local sweat rate as a function of exercise-induced changes in T eso) compared with CON. Despite this, differences in internal temperature (T rec and T eso) between groups did not reach statistical significance. In addition, in support of our previous findings (11), physiological control and function of the cutaneous vasculature is preserved in this clinical population.

Upon exercise onset, there is a rapid, if not immediate, increase in metabolism to convert stored chemical energy into mechanical work. Because gross mechanical efficiency in humans is only ∼20% efficient, the net result of this increase in metabolism is an increase in heat production, body heat storage, and a rise in core temperature. As this heat is dispersed throughout the body, the body’s core temperature (i.e., tissues and blood) will inevitably rise; however, the magnitude of this rise is partly dependent on thermoregulatory function. With intact afferent signaling and thermosensory integration, there will be a signal proportional to the rise in body temperature (core temperature and skin temperature) out of the CNS to increase sweat secretion and cutaneous vasodilation, facilitating evaporative and dry heat loss from the body, respectively, and thus preventing harmful increases in core temperature. It is this complex autonomic coupling of core temperature (T eso) and ΔLSR, that appears to be dampened by MS, as evidenced by the significantly attenuated relationship between T eso and LSR (i.e., thermosensitivity). Similarly, individuals with MS had significantly blunted WBSL during the exercise bout relative to CON, despite having to reconcile similar absolute heat loads. Although the precise mechanisms by which these abnormalities occur remain unclear, clearly a disease disrupting conduction within regions of the CNS involved in the thermoregulatory reflex arc would have a significant effect on these thermoregulatory processes. The stripping of myelin sheaths surrounding neural axons in the CNS is a hallmark of MS pathophysiology (21,22). This disruption in signaling can have slowing (22–24) and weakening (22) effects on neural conduction that could potentially be a contributor in the reduced thermosensitivity and attenuated WBSL responses seen in this study. It is likely that this aforementioned demyelination process results in increased axonal exposure, thus allowing more current to escape from the axon before reaching the axon terminal. In such case, if the current available is unable to meet the current required for signal transmission (i.e., reduced “safety factor”), there may become instances of inadequate signaling or even conduction block (21,22). Given that individuals with MS were able to achieve similar steady-state LSR as CON, it is likely that these neural signaling deficits simply result in a weaker sweating response however adequate to permit heat dissipation at modest heat loads (environment and/or activity). The effect of this demyelination is certainly not limited to the effector responses observed here, but could also be contributed to: disruptions in thermal sensory processing (25), disruptions in processing within the CNS, blunted efferent signaling to effector organs through the spinal cord, or some combination therein. More research is warranted to undercover the exact mechanisms responsible for these deficits.

Although we observed differences in WBSL at multiple time points during exercise in this study, measures of LSR did not reach statistical significance at any point. This is somewhat perplexing given that subjects were matched for rate of absolute metabolic heat production, BSA, and therefore rates of relative heat production (W·m−2 of BSA). However, the changes in LSR in this context only provide a snapshot of sweat rates at the 30- and 60-min time points and perhaps do not adequately reflect the cumulative sweating responses that are essential to achieve heat balance. Therefore, it is possible these traditional sites for collection of LSR were unaffected by the disease. Similarly, because of the local nature of this measurement, these findings do not rule out that there may be impairments at other locations on the body. Because of the heterogeneities associated with MS, perhaps LSR should be obtained at numerous sites in this population going forward.

Counter to our original hypothesis, changes in T core, represented as T eso or T rec, were not significantly different between groups in response to the exercise bout selected for this study. Therefore, it appears the environment and exercise intensities chosen for this study were well within the limits of thermal compensability for subjects in both groups. This is somewhat surprising given the evidence of blunted sweating responses in this study and elsewhere (8,9,11,26,27). As indicated by our data, MS subjects produced similar amounts of heat as CON while dissipating less. This dissociation of heat dissipation and heat production must result in elevations in temperature, perhaps in sites that were not directly measured. Although it is beyond the scope of this study to speculate exactly where and how this heat is being transferred, perhaps the storage of heat in muscle (or other tissues) could play a pivotal role in the onset of Uhthoff’s phenomenon and thermal perception in the MS population. Clearly, more research is warranted on this issue.

Supporting previous findings from our laboratory (11), it appears the cutaneous vasculature in persons with MS responds similarly versus CON to an increase in T core during exercise in a nonencapsulated environment. This raises questions regarding how the disease could selectively affect sweating responses while the control of the cutaneous vasculature remains intact. It is widely accepted that the primary site of integration of thermoregulatory information is the preoptic area of the hypothalamus (28–31). There is also growing evidence that this information is relayed from the hypothalamus to the medullary raphe regions before activating the intermediolateral nucleus of the spinal cord (29–31). Upon leaving the spinal cord, signals branch along nerves in the peripheral nervous system, ultimately arriving at the thermoregulatory end organs. Although the aforementioned pathway is the culmination of many studies and detailed analysis, the majority of this work has been conducted in rats or other species rather than humans. With this in mind, perhaps lesions as a result of MS are more prevalent in CNS pathways for sweating, and these pathways are yet to be elucidated. As an alternative explanation, it is possible the exercise intensities chosen were not demanding enough to elicit discernable differences in the CVC responses to exercise. However, as previously mentioned, the intensities selected in the current study were chosen to elicit the greatest change in T core while remaining low enough that this clinical population could complete the trials. Therefore, given that moderate environmental conditions (25°C) were used in the present study, future research should urgently examine potential thermoregulatory impairments with MS under warmer conditions. This research would be particularly relevant to clinical reports of Uhthoff’s phenomenon, which dramatically peak in the warmer summer months of the year.

Back to Top | Article Outline


Although our findings are consistent with CNS impairments, it remains possible that our results were influenced by DMT prescribed to individuals with MS. Patients were carefully screened for additional medications (i.e., antidepressants, psychostimulants, anticonvulsants, antispasmatics, and anticholinergics), and precautions were taken to remove any clear outliers that may influence group data. With this in mind, immediate prescription of DMT upon diagnosis is the current standard of care; thus, a critical understanding of individuals with MS currently on these medications is warranted.

The exercise intensity chosen was both of low intensity for CON individuals while being quite challenging for some individuals with MS (2 of 12 individuals had to stop the exercise trial ∼30 min early). Due to the conservative workload chosen, any differences might be more apparent and robust with higher workloads/heat loads. However, it is important to examine heat loss responses at ecologically valid intensities in which the majority of the MS population is capable of performing in their daily lives.

As mentioned previously, LSR and LDF/CVC values were collected over small skin areas (<3 cm2). Although this was consistent across all subjects (i.e., CON and MS), it is unclear how the lesions caused by MS could affect the efferent signal going to these specific sites, for each person. With a clinical condition that elicits seemingly random lesions throughout the CNS, no two people are affected in the same magnitude and/or locations of lesions, and therefore our measurements could be biased by the local nature of these measurements. However, as shown in Figures 3A and 5A, our localized measurements of LSR and CVC are similar between MS and CON; therefore, we may not be capturing the dysfunction in areas not being measured. If this is the case, this would add to the severity of thermoregulatory dysfunction among individuals with MS observed in this study.

In addition, sweat gland density measurements were taken from different areas of the forearm than LSR sites. As such, sweat gland output was unable to be calculated at the two locations. Therefore, it remains possible that there are differences in sweat gland output between the groups at one or both of sites. Future studies should examine potential regional differences in sweat gland recruitment and subsequent sweat output in individual with MS.

Back to Top | Article Outline

Clinical significance

On the basis of these findings, individuals with MS could face a substantial obstacle when it comes to maintaining an active lifestyle. Although there is much evidence in support of the benefits of exercise in this population (32–34), it appears even exercise intensities that generate modest amounts of heat are capable of eliciting the rapid onset of Uhthoff’s phenomenon in some. This is practically significant because 40%–80% of the MS population (35,36) can experience this acute worsening of their symptoms as a result of modest elevations in T core (37) or T sk (38). Therefore, although the benefits of exercise in the MS population remain absolutely undisputed, the obtainment of these beneficial adaptations may come at a risk for those seeking to live an active lifestyle, especially as their disease course progresses or as warm summer months approach. On the basis of these findings, it is imperative that individuals with MS are aware of these thermoregulatory limitations and have a ready and accessible cooling strategy in place before engaging in even modest exercise intensities. In addition, although promising studies have determined that exercise tolerance can be improved with cooling strategies (39) in MS, more research is warranted to identify the most impactful means to accomplish these goals.

Back to Top | Article Outline


Individuals with MS are at a greater risk for thermoregulatory dysfunction and thus more vulnerable to exaggerated elevations in body temperature (internal and/or skin) than disease free individuals. Although our results indicate that individuals with MS are able to achieve a steady-state T core and LSR in the exercise intensity and environmental conditions chosen for this study, WBSL values as well as sweating thermosensitivity were significantly blunted in those with MS compared with CON. Taken together, it appears individuals with MS can achieve a thermal steady state to exercise intensities of a modest heat load; however, there is evidence that this ability may reach its capacity before the healthy population. Given the prevalence of Uhthoff’s phenomenon among this population, it is imperative that more research is done to define the limits of compensability in this population, as well as to identify ways to mitigate the risks of excessive heating during physical activity in MS.

The considerable time and effort of the subjects are greatly appreciated.

This study was supported by National Heart, Lung, and Blood Institute grant R15-HL-117224 (S. L. Davis); National Multiple Sclerosis Society grants RG4043A1/1 and RG4696A3/2 (S. L. Davis); MS Research Australia Incubator grant 14-009 (O. Jay and S. L. Davis); MS Research Australia Postgraduate Fellowship 15-087 (G. K. Chaseling); and the Kuzell Institute (S. L. Davis and O. Jay).

No conflicts of interest, financial or otherwise, are declared by the authors. The results of this study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation. The results and conclusions of the study do not constitute endorsement by the American College of Sports Medicine.

Back to Top | Article Outline


1. Multiple Sclerosis: Hope through Research. National Institute of Neurological Disorders and Stroke [Internet]. [cited 2018 Aug 29]. Available from:
2. Mcdonald WI. Pathophysiology in multiple sclerosis. Brain. 1974;97(1):179–96.
3. Smith KJ, McDonald WI. The pathophysiology of multiple sclerosis: the mechanisms underlying the production of symptoms and the natural history of the disease. Philos Trans R Soc Lond B Biol Sci. 1999;354(1390):1649–73.
4. Adamec I, Habek M. Autonomic dysfunction in multiple sclerosis. Clin Neurol Neurosurg. 2013;115(1 Suppl):S73–8.
5. Flachenecker P, Wolf A, Krauser M, Hartung HP, Reiners K. Cardiovascular autonomic dysfunction in multiple sclerosis: correlation with orthostatic intolerance. J Neurol. 1999;246(7):578–86.
6. McDougall AJ, McLeod JG. Autonomic nervous system function in multiple sclerosis. J Neurol Sci. 2003;215(1–2):79–85.
7. Sterman AB, Coyle PK, Panasci DJ, Grimson R. Disseminated abnormalities of cardiovascular autonomic functions in multiple sclerosis. Neurology. 1985;35(11):1665–5.
8. Saari A, Tolonen U, Pääkkö E, et al. Sweating impairment in patients with multiple sclerosis. Acta Neurol Scand. 2009;120(5):358–63.
9. Noronha MJ, Vas CJ, Aziz H. Autonomic dysfunction (sweating responses) in multiple sclerosis. J Neurol Neurosurg Psychiatry. 1968;31(1):19–22.
10. Tsolaki F, Tagarakis GI, Karangelis D, et al. Uthoff phenomenon—a rare manifestation of a rare disease. J Paediatr Child Health. 2011;47(6):396–6.
11. Allen DR, Huang M, Parupia IM, Dubelko AR, Frohman EM, Davis SL. Impaired sweating responses to a passive whole body heat stress in individuals with multiple sclerosis. J Neurophysiol. 2017;118(1):7–14.
12. Taylor WF, Johnson JM, Kosiba WA, Kwan CM. Cutaneous vascular responses to isometric handgrip exercise. J Appl Physiol. 1989;66(4):1586–92.
13. Mitchell JW, Nadel ER, Stolwijk JA. Respiratory weight losses during exercise. J Appl Physiol. 1972;32(4):474–6.
14. Bullard RW. Continuous recording of sweating rate by resistance hygrometry. J Appl Physiol. 1962;17:735–7.
15. Gagnon D, Ganio MS, Lucas RA, Pearson J, Crandall CG, Kenny GP. Modified iodine-paper technique for the standardized determination of sweat gland activation. J Appl Physiol. 2012;112(8):1419–25.
16. Cramer MN, Jay O. Selecting the correct exercise intensity for unbiased comparisons of thermoregulatory responses between groups of different mass and surface area. J Appl Physiol. 2014;116(9):1123–32.
17. Jay O, Bain AR, Deren TM, Sacheli M, Cramer MN. Large differences in peak oxygen uptake do not independently alter changes in core temperature and sweating during exercise. Am J Physiol Regul Integr Comp Physiol. 2011;301(3):R832–41.
18. Cramer MN, Jay O. Explained variance in the thermoregulatory responses to exercise: the independent roles of biophysical and fitness/fatness-related factors. J Appl Physiol. 2015;119(9):982–9.
19. Kenefick RW, Cheuvront SN. Hydration for recreational sport and physical activity. Nutr Rev. 2012;70(2 Suppl):S137–42.
20. Moran D, Epstein Y, Keren G, Laor A, Sherez J, Shapiro Y. Calculation of mean arterial pressure during exercise as a function of heart rate. Appl Human Sci. 1995;14(6):293–5.
21. Schauf CL, Davis FA. Impulse conduction in multiple sclerosis: a theoretical basis for modification by temperature and pharmacological agents. J Neurol Neurosurg Psychiatry. 1974;37(2):152–61.
22. Smith KJ. Sodium channels and multiple sclerosis: roles in symptom production, damage and therapy. Brain Pathol. 2007;17(2):230–42.
23. Davis SL, Frohman TC, Crandall CG, et al. Modeling Uhthoff’s phenomenon in MS patients with internuclear ophthalmoparesis. Neurology. 2008;70(13 Pt 2):1098–106.
24. Frohman TC, Davis SL, Frohman EM. Modeling the mechanisms of Uhthoff’s phenomenon in MS patients with internuclear ophthalmoparesis. Ann N Y Acad Sci. 2011;1233(1):313–9.
25. Filingeri D, Chaseling G, Hoang P, Barnett M, Davis SL, Jay O. Afferent thermosensory function in relapsing-remitting multiple sclerosis following exercise-induced increases in body temperature. Exp Physiol. 2017;102(8):887–93.
26. Davis SL, Wilson TE, Vener JM, Crandall CG, Petajan JH, White AT. Pilocarpine-induced sweat gland function in individuals with multiple sclerosis. J Appl Physiol. 2005;98(5):1740–4.
27. Vas CJ. Sexual impotence and some autonomic disturbances in men with multiple sclerosis. Acta Neurol Scand. 1969;45(2):166–82.
28. Benzinger TH. On physical heat regulation and the sense of temperature in man. Proc Natl Acad Sci. 1959;45(4):645–59.
29. Blessing W, McAllen R, McKinley M. Control of the cutaneous circulation by the central nervous system. Compr Physiol. 2016;6(3):1161–97.
30. Shibasaki M, Wilson TE, Crandall CG. Neural control and mechanisms of eccrine sweating during heat stress and exercise. J Appl Physiol. 2006;100(5):1692–701.
31. Smith CJ, Johnson JM. Responses to hyperthermia. Optimizing heat dissipation by convection and evaporation: neural control of skin blood flow and sweating in humans. Auton Neurosci. 2016;196:25–36.
32. Huang M, Jay O, Davis SL. Autonomic dysfunction in multiple sclerosis: implications for exercise. Auton Neurosci. 2015;188:82–5.
33. Petajan JH, White AT. Recommendations for physical activity in patients with multiple sclerosis. Sports Med. 1999;27(3):179–91.
34. Petajan JH, Gappmaier E, White AT, Spencer MK, Mino L, Hicks RW. Impact of aerobic training on fitness and quality of life in multiple sclerosis. Ann Neurol. 1996;39(4):432–41.
35. Frohman TC, Davis SL, Beh S, Greenberg BM, Remington G, Frohman EM. Uhthoff’s phenomena in MS—clinical features and pathophysiology. Nat Rev Neurol. 2013;9(9):535–40.
36. Frohman TC, Davis SL, Frohman EM. Modeling the mechanisms of Uhthoff’s phenomenon in MS patients with internuclear ophthalmoparesis. Ann N Y Acad Sci. 2011;1233(1):313–9.
37. Rasminsky M. The effects of temperature on conduction in demyelinated single nerve fibers. Arch Neurol. 1973;28(5):287–92.
38. Poh PYS, Adams AN, Huang M, et al. Increased postural sway in persons with multiple sclerosis during short-term exposure to warm ambient temperatures. Gait Posture. 2017;53:230–5.
39. Chaseling GK, Filingeri D, Barnett M, Hoang P, Davis SL, Jay O. Cold water ingestion improves exercise tolerance of heat-sensitive people with MS. Med Sci Sports Exerc. 2018;50(4):643–8.


© 2019 American College of Sports Medicine