We assessed whether comparisons of thermoregulatory responses between groups unmatched for body mass and surface area (BSA) should be performed using a metabolic heat production (Ḣprod) in Watts or Watts per kilogram for changes in rectal temperature (ΔTre), and an evaporative heat balance requirement (Ereq) in Watts or Watts per square meter for local sweat rates (LSR). Two groups with vastly different mass and BSA [large (LG): 91.5 ± 6.8 kg, 2.12 ± 0.09 m2, n = 8; small (SM): 67.6 ± 5.6 kg, 1.80 ± 0.09 m2, n = 8; P < 0.001], but matched for heat acclimation status, sex, age, and with the same onset threshold esophageal temperatures (LG: +0.37 ± 0.12°C; SM: +0.41 ± 0.17°C; P = 0.364) and thermosensitivities (LG: 1.02 ± 0.54, SM: 1.00 ± 0.38 mg·cm−2·min−1·°C−1; P = 0.918) for sweating, cycled for 60 min in 25°C at different levels of Ḣprod (500 W, 600 W, 6.5 W/kg, 9.0 W/kg) and Ereq (340 W, 400 W, 165 W/m2, 190 W/m2). ΔTre was different between groups at a Ḣprod of 500 W (LG: 0.52 ± 0.15°C, SM: 0.92 ± 0.24°C; P < 0.001) and 600 W (LG: 0.78 ± 0.19°C, SM: 1.14 ± 0.24°C; P = 0.007), but similar at 6.5 W/kg (LG: 0.79 ± 0.21°C, SM: 0.85 ± 0.14°C; P = 0.433) and 9.0 W/kg (LG: 1.02 ± 0.22°C, SM: 1.14 ± 0.24°C; P = 0.303). Furthermore, ΔTre was the same at 9.0 W/kg in a 35°C environment (LG: 1.12 ± 0.30°C, SM: 1.14 ± 0.25°C) as at 25°C (P > 0.230). End-exercise LSR was different at Ereq of 400 W (LG: 0.41 ± 0.18, SM: 0.57 ± 0.13 mg·cm−2·min−1; P = 0.043) with a trend toward higher LSR in SM at 340 W (LG: 0.28 ± 0.06, SM: 0.37 ± 0.15 mg·cm−2·min−1; P = 0.057), but similar at 165 W/m2 (LG: 0.28 ± 0.06, SM: 0.28 ± 0.12 mg·cm−2·min−1; P = 0.988) and 190 W/m2 (LG: 0.41 ± 0.18, SM: 0.37 ± 0.15 mg·cm−2·min−1; P = 0.902). In conclusion, when comparing groups unmatched for mass and BSA, future experiments can avoid systematic differences in ΔTre and LSR by using a fixed Ḣprod in Watts per kilogram and Ereq in Watts per square meter, respectively.
- core temperature
- local sweat rate
- body mass
- body surface area
studies of human thermoregulation often employ a between-groups experimental design to isolate the independent effect of a particular physiological factor, e.g., age (5, 24, 50), sex (18, 25), aerobic fitness (19, 39), disease (29, 30), injury (21), on the core temperature and sudomotor responses to exercise.1 The exercise intensity prescribed to facilitate these comparisons is fundamentally important, since the introduction of any inherent bias to an experimental design may lead researchers to incorrectly attribute different changes in core temperature and/or sweating between groups to the physiological factor under examination.
Since the seminal work of Saltin and Hermansen in 1966 (44), many exercise and thermal physiologists have interpreted their findings to mean that a fixed relative exercise intensity [a percentage of the maximum rate of oxygen uptake (%V̇o2max)] should be administered to compare thermoregulatory responses between independent groups due to the prevailing notion that V̇o2max profoundly influences the change in core temperature and sweating during exercise (13, 19, 20, 39). However, we recently reported that two groups matched for body mass and body surface area (BSA), but differing greatly in V̇o2max, exhibit almost identical changes in core temperature and whole body sweat loss (WBSL) during exercise at the same absolute rate of metabolic heat production (Ḣprod) (540 W) in a physiologically compensable environment, despite large differences in relative intensity (58 vs. 40% of V̇o2max) (27). It is now also clear that protocols utilizing %V̇o2max can lead to systematically different changes in core temperature and sweating between groups that may otherwise respond similarly from a physiological perspective, due to differences in Ḣprod and the evaporation required for heat balance (Ereq) (16, 27). However, since the participants in our previous study were matched for body mass (27), it is still unknown whether an absolute Ḣprod [in Watts (W)] should be used to prescribe exercise intensity for between-group experimental designs, or if Ḣprod should be normalized for body mass (W/kg), if groups are unmatched for this physical trait. The practical importance of this question is emphasized by the fact that matching groups for body mass may, in some cases, be impossible for researchers investigating the consequences of potential thermoregulatory dysfunction in special populations, such as multiple sclerosis patients (10), the obese (23, 33), spinal cord injury victims (21, 49), sympathectomy patients (8), and skin-graft recipients (36, 47).
From a biophysical perspective, changes in core temperature are determined by the cumulative imbalance between Ḣprod and net heat loss to the environment (i.e., body heat storage), body mass (i.e., internal heat sink), and body composition (i.e., specific heat capacity of body tissue). Previous studies have shown that large variations in body mass lead to diverse core temperature responses during exercise at the same absolute work rate or Ḣprod (15, 22, 35, 43). It therefore stands to reason that normalizing Ḣprod for body mass should lead to similar changes in core temperature between groups of dissimilar body mass, unless heat loss is altered as a function of the physiological parameter under investigation (e.g., age, sex, etc.).
Using direct calorimetry and, therefore, under conditions permitting full evaporation, Gagnon et al. (16) recently demonstrated that whole body sweat rate (WBSR) in grams per minute is determined by the absolute rate of Ereq in W, irrespective of %V̇o2max, core temperature, BSA, and body mass. However, local sweat rate (LSR) is typically measured in milligrams per square centimeter per minute over a fixed surface area with either a ventilated capsule (18, 25, 50) or absorbent patch (3, 7, 11, 48); therefore, a higher mean LSR would be expected at a fixed absolute Ereq (and therefore WBSR) in individuals with a lower BSA, because the same absolute amount of sweat would have to be secreted over a smaller area. The prescription of an exercise intensity that elicits the same absolute Ereq may, therefore, lead to a systematically different LSR between independent groups unmatched for BSA, yet an intensity eliciting the same Ereq per unit BSA (in W/m2) may remove this inherent bias (11, 18).
The purpose of this study was to derive the optimal methods for comparing changes in rectal temperature (ΔTre) and LSR between groups unmatched for body mass and BSA so that any inherent bias due to biophysical factors is removed. To this end, we compared responses between groups vastly different in body mass (∼90 vs. ∼65 kg) and BSA (∼2.10 vs. 1.80 m2), but matched for age, sex, and heat acclimation status, and with identical operational parameters for sudomotor control (i.e., onset threshold and thermosensitivity). Values for ΔTre were compared using fixed levels of 1) absolute Ḣprod (in W), and 2) Ḣprod per unit total body mass (in W/kg). Values for mean LSR were compared using fixed levels of 1) absolute Ereq (in W), and 2) Ereq per unit BSA (in W/m2). It was hypothesized that 1) Ḣprod in W would yield a greater ΔTre in the small body mass group due to a greater Watts per kilogram (W/kg), but Ḣprod in W/kg would lead to a similar ΔTre between groups, despite differences in body mass, and 2) Ereq in W would yield similar WBSR between large and small BSA groups, but mean LSR would be greater in the small BSA group due to a greater Ereq in Watts per square meter (W/m2); however, Ereq in W/m2 would lead to similar mean LSR between groups, despite differences in BSA.
Approval of the experimental protocol was obtained from the University of Ottawa Health Sciences and Science Research Ethics Board (file no. H12-11-05). All procedures conformed to the principles set forth in the Declaration of Helsinki. Volunteers were fully informed of the experimental protocol and potential risks before providing written, informed consent. Also, a Physical Activity Readiness Questionnaire and an American Heart Association/American College of Sports Medicine Health/Fitness Facility Pre-participation Screening Questionnaire were completed before participation.
Using a power calculation (G*power version 3.1.5) with conventional β- (0.1) and α-values (0.05), a minimum sample size of 12 participants (6 per group) was required based on a mean ΔTre of 0.35°C and a standard deviation of 0.15°C following 60 min of exercise at a fixed Ḣprod of 500 W between independent groups with a 17.7-kg difference in body mass (15). Sixteen men of large (LG; n = 8) or small (SM; n = 8) body mass and BSA volunteered for this study. Groups were matched for age, but not aerobic fitness, to ensure differences in %V̇o2max at each Watt per kilogram, and thereby isolate whether systematic differences in ΔTre are avoided by prescribing a fixed W/kg between groups unmatched for body mass. All participants were nonsmokers, reported no history of cardiovascular, respiratory, neurological, or metabolic disease, and were not taking any medications at the time of participation.
Each participant visited the laboratory for a preliminary session that included an explanation of the experimental protocol, anthropometric measurements, and an exercise test. Height was measured using a wall-mounted stadiometer (HR-200, Tanita, Arlington Heights, IL), and body mass was measured using a digital scale (BWB-800, Tanita, Arlington Heights, IL). These values were used subsequently to estimate BSA (12). Body composition was measured via dual-energy X-ray absorptiometry (GE-LUNAR Prodigy module, GE Medical Systems, Madison, WI).
The exercise test was performed in a climate-controlled room set to 22°C on a semirecumbent cycle ergometer (Lode Corival, Groningen, the Netherlands) in two phases. The first phase was performed to determine the relationship between external work rate and steady-state rate of oxygen consumption (V̇o2) (and thus Ḣprod) for each participant over the full range of Ḣprod targeted in the experimental trials. This procedure permitted greater accuracy in achieving each target Ḣprod from the onset of exercise (see appendix for step-by-step instructions for prescribing exercise intensity to achieve target Ḣprod). Participants completed four 5-min submaximal stages, which began at 80 W (SM) or 100 W (LG) and increased by 20 W/stage. Expired gases were analyzed throughout exercise via indirect calorimetry using a metabolic cart (Vmax Encore, CareFusion, Yorba Linda, CA). Following a 10-min rest period, the second phase of exercise included an incremental exercise test to exhaustion to determine V̇o2max. This protocol commenced at an external work rate of 80 W and increased by 20 W/min until volitional exhaustion, in accordance with guidelines from the Canadian Society for Exercise Physiology (9).
Before experimentation, each participant performed 7 consecutive days of low-intensity cycling at 35°C and 35% relative humidity (RH) for 90 min/day to improve exercise tolerance and to minimize potential variance in the operational parameters of sudomotor activity (i.e., onset threshold and thermosensitivity) that could possibly explain differences in LSR between groups (45).
Experimental trials were separated by 2–3 days and were performed in a randomized, counterbalanced order at the same time of day to eliminate any systematic differences between groups due to circadian variation. Participants were asked to abstain from alcohol and caffeine, avoid strenuous exercise in the 12 h before each experimental session, and consume a light meal and 500 ml of water ∼2 h before arriving at the laboratory. On arrival, each participant provided a urine sample, which was immediately analyzed for urine specific gravity (USG) to ensure preexercise hydration status was similar between groups. A USG cutoff value of 1.025 was enforced, as values below this threshold have been suggested to indicate normal hydration (28). Participants then inserted the rectal thermocouple, and, while wearing only a standardized pair of cotton running shorts, an initial body mass measurement was taken to determine the rate of Ḣprod for each Watt per kilogram trial. Next, the participants put on a pair of cotton socks and running shoes and sat on the ergometer while they were instrumented. Following 30 min of baseline data collection while seated on the ergometer, participants then performed 60 min of semirecumbent cycling in one of four experimental conditions: three trials in neutral ambient conditions (25.1 ± 0.5°C, 36.8 ± 12.7% RH, and 1.2 ± 0.1 m/s air velocity) at exercise intensities eliciting 500 W, 6.5 W/kg, or 9.0 W/kg of Ḣprod, and one trial in the heat (34.7 ± 1.7°C, 34.1 ± 8.7% RH, and 1.1 ± 0.3 m/s air velocity) at 9.0 W/kg. This latter trial was performed to determine whether similar ΔTre would be observed within each group in different, but compensable, ambient conditions that remained within the “prescriptive zone” (34). Two LG subjects could not complete the protocol in the heat and were, therefore, not included. Air flow was provided by three 46-cm mechanical fans stacked vertically and positioned 1.25 m in front of the ergometer. By virtue of the targeted differences in body mass between LG and SM groups, comparisons of ΔTre at Ḣprod of 600 W were also possible from data collected in the 6.5 W/kg and 9.0 W/kg trials in the LG and SM groups, respectively. Due to differences in BSA between LG and SM, LSR comparisons were possible at Ereq of 165 and 190 W/m2. Specifically, exercise at 500 W in LG and 6.5 W/kg in SM corresponded to an Ereq of 165 W/m2, while exercise at 6.5 W/kg for LG and 500 W for SM corresponded to an Ereq of 190 W/m2, in both groups. Cycling cadence was maintained at 80 revolutions/min in all trials. Core temperature, skin temperature (Tsk), and LSR on the upper back and forearm were measured continuously. Body mass measurements were taken in triplicate while clothed and fully instrumented immediately before exercise (i.e., as a baseline for WBSL estimations) and every 15 min throughout exercise, which required a 2-min break from cycling.
Core temperatures were measured using general-purpose pediatric thermocouple probes (Mon-a-therm, Mallinckrodt Medical, St. Louis, MO). Tre was measured at a depth of 12 cm beyond the anal sphincter. Esophageal temperature (Tes) was measured at a maximum depth of ∼40 cm (37) for the first 15 min of exercise to determine the sudomotor onset threshold and thermosensitivity (see below). Both Tre and Tes are expressed as changes from baseline (i.e., ΔTre and ΔTes). Skin temperature was measured at eight sites with thermistors integrated into 2.5-cm2 heat flux sensors (Concept Engineering, Old Saybrook, CT). These sensors were affixed to the skin using double-sided adhesive disks (3M Health Care, Neuss, Germany) and surgical tape (Transpore, 3M, London, ON, Canada). Mean Tsk was calculated using weighting coefficients according to ISO 9886 (26): forehead, 0.07; shoulder, 0.07; triceps, 0.07; chest, 0.175; scapula, 0.175, hand, 0.05; thigh, 0.19; and calf, 0.20. Values for Tsk are reported as an average over the 60 min of exercise. Core temperature and Tsk were recorded with a data acquisition system (NI cDAQ-9172, National Instruments, Austin, TX) and LabView software (version 7.0, National Instruments, Austin, TX), sampled at 0.2 Hz.
Estimations of WBSR were made from changes in body mass every 15 min. Body mass was measured in triplicate using a platform scale accurate to the nearest ±2 g (Combics 2, Sartorius, Mississauga, ON, Canada) and corrected for metabolic mass loss and vapor losses from the respiratory tract (38). Values for WBSR are reported for each 15-min time period in grams per minute (g/min). Cumulative WBSL for the 60-min exercise period is also reported in grams.
LSR was measured using ventilated capsules (4.1 cm2) placed on the forearm ∼5 cm distal to the antecubital fossa and the upper back ∼5 cm above the scapular spine over the trapezium, and secured with adhesive (Collodion HV, Mavidon Medical, Lake Worth, FL) and surgical tape. The flow of anhydrous air through each capsule was regulated at 1.80 l/min (FMA-A2307, Omega Engineering, Stamford, CT). The vapor concentration of effluent air was measured at 0.2 Hz using factory-calibrated capacitance hygrometers (HMT333, Vaisala, Vantaa, Finland). LSR is reported as the product of the vapor concentration and the flow rate, normalized to the skin surface area covered by the capsule, and expressed in milligrams per square centimeter per minute (mg·cm−2·min−1). The sudomotor onset threshold and thermosensitivity were determined via segmented regression using 1-min averages for the mean LSR response and ΔTes (6).
Heat balance parameters.
Heat balance parameters were estimated via partitional calorimetry and are presented as the mean value within each experimental condition. All heat exchange parameters were calculated in W/m2, but are presented in W, W/m2, or W/kg, where appropriate. Due to a minimal clothing ensemble, dry insulation and evaporative resistance of clothing were considered negligible.
The rate of metabolic energy expenditure (M) was estimated as: (1) where RER is the respiratory exchange ratio, and ec and ef represent the energy equivalent of carbohydrate (21.13 kJ) and fat (19.69 kJ), respectively, per liter of O2 consumed (l/min). Ḣprod was determined as the difference between M and the external work rate (W): (2) Heat loss via radiation (R) was calculated as: (3) where Ta denotes ambient temperature (°C), and hr is the radiant heat transfer coefficient: (4) where ϵ is the emissivity of the skin (0.95), σ is the Stefan-Boltzmann constant (5.67·10−8 W·m2·K−4); BSAr/BSA is the effective radiant surface area (ND), equal to 0.70 (31); and Tr is the mean radiant temperature, assumed to be equivalent to Ta (°C). Convective heat exchange from the skin, C, was calculated as: (5) where hc is the convective heat transfer coefficient for a seated individual facing an air velocity (v) between 0.2 and 4.0 m/s (41): (6) Respiratory heat losses through evaporation (Eres) and convection (Cres) were determined by: (7) where Pa is the ambient vapor pressure (kPa). The Ereq was calculated as: (8)
Mean participant characteristics were compared using independent-samples t-tests. Data for ΔTre and LSR were analyzed as 5-min averages ending at 0, 15, 30, 45, and 60 min of exercise. For each Ḣprod, a two-way mixed-model ANOVA with the repeated factor of time (five levels: baseline, 15, 30, 45, 60 min) and the nonrepeated factor of body size (two levels: LG and SM) were performed to compare ΔTre, WBSR, and LSR with a Bonferroni correction for multiple comparisons (i.e., at each time point). Independent-samples t-tests were used for single comparisons of heat balance parameters, %V̇o2max, 60-min ΔTre, Tsk, and cumulative WBSL, as well as sudomotor onset threshold and thermosensitivity. All statistical analyses were performed with GraphPad Prism (version 6.0, GraphPad Software, La Jolla, CA). All data are expressed as means ± SD. P values ≤ 0.05 were considered statistically significant.
Mean participant characteristics are presented in Table 1. No differences in age (P = 1.000) existed between groups. Body mass (P < 0.001), height (P = 0.017), BSA (P < 0.001), and body fat percentage (P < 0.001) were greater in LG, while relative V̇o2max (expressed in ml·kg−1·min−1) was higher in SM (P = 0.019). Preexperimental USG was similar between groups in each trial, with mean values of 1.019 ± 0.006 and 1.015 ± 0.008 in LG and SM, respectively.
In the fixed absolute Ḣprod trials of 500 and 600 W (Fig. 1), end-exercise ΔTre after 60 min was significantly greater in SM at both 500 W (LG: 0.52 ± 0.15°C, SM: 0.92 ± 0.24°C; P < 0.001) and 600 W (LG: 0.78 ± 0.19°C, SM: 1.14 ± 0.24°C; P = 0.007). Differences in ΔTre were observed between groups from 30 min of exercise onwards in both trials. Due to differences in body mass, the corresponding W/kg was greater in SM at both 500 W (P < 0.001) and 600 W (P < 0.001). Furthermore, the relative exercise intensity (%V̇o2max) was higher in SM at 500 W (P = 0.038) and tended to be higher at 600 W (P = 0.053).
In contrast, when comparing SM and LG groups at the same fixed Ḣprod per unit mass trials of 6.5 and 9.0 W/kg (Fig. 1), end-exercise ΔTre after 60 min was similar between groups at both 6.5 W/kg (SM: 0.85 ± 0.14°C, LG: 0.79 ± 0.21°C; P = 0.433) and 9.0 W/kg (SM: 1.14 ± 0.24°C, LG: 1.02 ± 0.22°C; P = 0.303). Furthermore, no differences in ΔTre were observed between SM and LG at any time at 6.5 W/kg (P = 0.129) or 9.0 W/kg (P = 0.635). While there were no differences in ΔTre, the corresponding absolute Ḣprod in W were higher in LG due to their greater mass at both 6.5 W/kg (P < 0.001) and 9.0 W/kg (P < 0.001). The %V̇o2max was also greater in LG at 6.5 W/kg (P = 0.019) and 9.0 W/kg (P = 0.002).
When exercise at 9.0 W/kg was repeated in a hotter environment (35°C), a similar ΔTre was observed over time relative to a neutral environment (25°C) within both the LG (P = 0.398) and SM (P = 0.646) groups (Fig. 2).
WBSR and LSR.
Absolute Ereq was ∼340 W for both LG and SM at a fixed Ḣprod of 500 W (P = 0.330), and absolute Ereq was ∼400 W for both LG and SM at a fixed Ḣprod of 600 W (P = 0.453). In parallel, similar WBSR values were observed between groups in both trials (Fig. 3), resulting in almost identical cumulative WBSL values at an Ereq of 340 W (LG: 383 ± 108 g, SM: 380 ± 52 g; P = 0.956) and 400 W (LG: 473 ± 156 g, SM: 493 ± 65 g; P = 0.734). At 6.5 and 9.0 W/kg, WBSR was greater in the LG group in parallel to a higher absolute Ereq in the LG group in both conditions (Fig. 3), leading to greater cumulative WBSL in the LG compared with the SM group at 6.5 W/kg (LG: 473 ± 156 g, SM: 298 ± 35 g; P = 0.008) and 9.0 W/kg (LG: 774 ± 210 g, SM: 493 ± 65 g; P = 0.003). Although Ḣprod and ΔTre were similar between groups, absolute Ereq was higher in the heat for both LG (35°C: 792 ± 83 W, 25°C: 575 ± 73 W; P < 0.001) and SM groups (35°C: 585 ± 60 W, 25°C: 391 ± 44 W; P < 0.001). Accordingly, cumulative WBSL was greater in the heat for LG (35°C: 1,067 ± 218 g, 25°C: 774 ± 210 g; P < 0.001) and SM (35°C: 817 ± 159 g, 25°C: 493 ± 65 g; P < 0.001) groups.
Despite similar WBSR and WBSL values, LSR was greater in the SM group when absolute Ereq was 340 W (P = 0.007) and 400 W (P = 0.032) (Fig. 4). These greater LSR values in the SM group at the same absolute Ereq corresponded with a greater Ereq in W/m2 in the SM group in both cases (Fig. 4). In contrast, when comparing LG and SM groups at the same Ereq values in W/m2, no differences in LSR were evident throughout exercise at an Ereq of 165 or 190 W/m2, despite very different absolute Ereq values in W in both conditions (Fig. 4). After 60 min of exercise, LSR was similar between groups at an Ereq of 165 W/m2 (LG: 0.28 ± 0.06 mg·cm−2·min−1, SM: 0.28 ± 0.12 mg·cm−2·min−1; P = 0.988) and 190 W/m2 (LG: 0.41 ± 0.18 mg·cm−2·min−1, SM: 0.37 ± 0.15 mg·cm−2·min−1; P = 0.902).
Onset threshold and thermosensitivity.
The mean LSR response relative to ΔTes is shown in Fig. 5, and the onset threshold ΔTes and thermosensitivity of the mean LSR response are presented in Table 2. In support of our aim to ensure that no differences in the physiological control parameters for sudomotor activity existed between the SM and LG group, neither the onset threshold ΔTes (P ≥ 0.360) nor thermosensitivity (P ≥ 0.351) of the mean LSR response was different between groups during any of the experimental trials.
Values for Tsk were similar between groups at Ḣprod of 500 W (LG: 31.31 ± 0.52°C, SM: 31.30 ± 0.48°C; P = 0.965), 600 W (LG: 31.39 ± 0.45°C, SM: 31.54 ± 0.40°C; P = 0.493), and 6.5 W/kg (LG: 31.39 ± 0.45°C, SM: 31.02 ± 0.50°C; P = 0.139). A higher Tsk was observed in LG at 9.0 W/kg in 25°C (LG: 32.10 ± 0.40°C, SM: 31.54 ± 0.40°C; P = 0.016). Similar Tsk were also observed at each Ereq of 165 W/m2 (LG: 31.31 ± 0.52°C, SM: 31.02 ± 0.50°C; P = 0.275) and 190 W/m2 (LG: 31.39 ± 0.45°C, SM: 31.30 ± 0.48°C; P = 0.692).
The present study clearly demonstrates that a large difference in body mass systematically alters ΔTre during exercise at a fixed Ḣprod (in W; Fig. 1) between independent groups that are otherwise matched for age, sex, heat acclimation status, and physiologically identical in terms of their control parameters for sudomotor activity (i.e., onset threshold and thermosensitivity; Fig. 5, Table 2). However, when an exercise intensity eliciting a fixed Ḣprod per unit mass is prescribed (W/kg; Figs. 1 and 2), the systematic difference in ΔTre is eliminated, despite differences in body mass, absolute Ḣprod (in W), and relative exercise intensity (%V̇o2max). The present study also demonstrates that, despite an almost identical WBSR (in g/min) between groups differing greatly in BSA during exercise at a fixed absolute Ereq (in W; Fig. 3), as would be expected given the recent findings of Gagnon et al. (16), LSR measured with a ventilated sweat capsule (in mg·cm−2·min−1) is systematically greater in the group with a smaller BSA (Fig. 4). However, when an exercise intensity eliciting a fixed Ereq per unit surface area (in W/m2) is prescribed, changes in LSR throughout 60 min of exercise are the same (Fig. 4), despite different BSA and absolute Ereq in W. These findings demonstrate that future studies aiming to isolate the independent influence of a particular physiological factor (e.g., age, sex, injury, autonomic diseases) on thermoregulatory responses by comparing ΔTre and LSR between experimental and control groups unmatched for body mass and BSA should use a fixed Ḣprod in W/kg for ΔTre comparisons and a fixed Ereq in W/m2 for LSR comparisons. It follows that, if different ΔTre or LSR responses are subsequently observed, they can be confidently attributed to the physiological factor under examination and are not a consequence of an inherent bias arising from the prescription of exercise intensity, such as with the %V̇o2max approach (27).
From a biophysical perspective, different changes in core temperature will arise from differences in heat storage (cumulative differences between Ḣprod and heat dissipation throughout exercise), body composition, or body mass. In the present study, the greater ΔTre observed in the SM group at the same absolute rates of Ḣprod (Fig. 1) is directly explained by the influence of body mass per se and not by any differences in heat dissipation or body composition. First, while factors such as age (32), sex (17), and heat acclimation status (42) are known to alter thermoeffector responses, sudomotor control, and heat dissipation, all of these factors were controlled in the present study. Second, at a Ḣprod of both 500 and 600 W, no differences in Tsk and, therefore, dry heat loss were evident between groups, resulting in a similar absolute Ereq and, therefore, the same WBSR (Fig. 3) and presumably evaporation. Although a high body fat percentage may alter core temperature changes due to a lower average specific heat capacity of adipose tissue (1), a nearly twofold difference (11.9 vs. 22.2%) in body fat percentage does not alter ΔTre in mass-matched participants exercising at the same absolute Ḣprod (27). As such, it is unlikely that the difference in body fat percentage between LG and SM (Table 1) contributed to the observed difference in ΔTre. While it may be possible that much larger differences in body fat percentage alter changes in core temperature, the independent influence of high vs. low adiposity (i.e., while controlling for Ḣprod and body mass) has not yet been evaluated and merits further investigation.
By prescribing the same Ḣprod in W/kg, the influence of body mass is effectively normalized, resulting in similar ΔTre between two groups, despite a 23.9-kg difference in body mass (Fig. 1). A retrospective assessment of data from previous studies examining core temperature responses over a range of relative intensities (i.e., %V̇o2max) in groups unmatched for aerobic fitness and body mass also supports the use of the W/kg method for eliminating systematic differences in ΔTre. For example, aerobically trained individuals exercising at 50% V̇o2max demonstrated a similar rate of Ḣprod (∼9.0 W/kg) and ΔTes (∼0.8°C) as aerobically untrained individuals exercising at 70% V̇o2max, despite an 8.2-kg difference in body mass between groups (13). Similarly, a closer look at the data of Mora-Rodriguez et al. (39) reveals a ΔTre of ∼0.6°C in trained and untrained groups of dissimilar mass (10-kg difference) cycling at 40% and 60% V̇o2max, respectively, which actually corresponded to a Ḣprod of ∼8.2 W/kg in both groups. As noted by Jay et al. (27), it follows that different changes in core temperature attributed to some physiological effect [e.g., age (24, 50), aerobic fitness (19, 39, 44), burn injury (36)] may be explained simply by differences in W/kg of as little as 1.8 W/kg (Fig. 1). Therefore, a reevaluation of some of these potential physiological alterations to heat balance may be warranted. To further demonstrate the validity of the W/kg approach, an additional trial was performed at 9.0 W/kg in a hotter environment (35°C) but within the classical prescriptive zone (34). For both the LG and SM groups, ΔTre was the same compared with 25°C (Fig. 2), with a compensatory rise in sweating and evaporative heat loss in association with the higher Ereq (34, 40).
The present data provide further evidence that a %V̇o2max approach is not appropriate for comparing changes in core temperature between individuals and groups of different V̇o2max (27). The LG group had a lower V̇o2max than the SM group (Table 1), and while exercise at 500 and 600 W resulted in a higher %V̇o2max and a greater ΔTre in SM in both cases (Fig. 1), exercise at 6.5 and 9.0 W/kg resulted in a significantly greater %V̇o2max in LG, but no differences in ΔTre (Fig. 1). Furthermore, although it may be argued that there was a slightly greater end-exercise ΔTre in the SM group at 6.5 and 9.0 W/kg, %V̇o2max was in fact lower in the SM group, which, according to conventional wisdom, should have led to a lower, not a higher, change in core temperature. Nevertheless, two points regarding the prescription of %V̇o2max should be noted. First, the prescription of %V̇o2max may be used without concern in a within-subjects (repeated measures) experimental design to compare changes in core temperature, provided that the rate of Ḣprod is not altered between conditions. Second, it is possible that, despite differences in V̇o2max and body mass between groups, combinations of these factors may yield a similar Ḣprod in W/kg, and, therefore, core temperature changes during exercise at a fixed %V̇o2max. However, by maintaining a fixed Ḣprod in W/kg, irrespective of relative exercise intensity, the present approach ensures an unbiased comparison at all combinations of V̇o2max and body mass. This approach may be especially useful in studies comparing core temperature responses during weight-bearing exercise (e.g., walking and running), during which Ḣprod varies with body mass, and a high interindividual variability in movement economy at a given speed is often observed. Future studies should evaluate the present approach for between-groups comparisons during treadmill exercise.
Gagnon et al. (16) recently demonstrated that absolute Ereq (in W) is the principal determinant of WBSR (in g/min), irrespective of %V̇o2max. Accordingly, WBSR was similar between the SM and LG groups at an Ereq of 340 W (Ḣprod: 500 W) and 400 W (Ḣprod: 600 W), despite greater %V̇o2max in the SM group, while differences in body mass led to greater absolute Ereq and WBSR in the LG group at 6.5 and 9.0 W/kg (Fig. 3). However, at an absolute Ereq of 340 and 400 W, greater mean LSR values (in mg·cm−2·min−1) were observed in the SM group (Fig. 4), demonstrating that the conclusions of Gagnon et al. (16) do not necessarily hold for measurements of local sudomotor activity in individuals of different morphological characteristics. Although it has been suggested that LSR is determined by the absolute external work rate (46), there were no differences in work rate between groups at either fixed absolute Ereq value (Fig. 4). Therefore, the differences in LSR between groups at the same absolute Ereq are attributed to the influence of body size alone. Considering that LSR is measured across a fixed surface area, it is most logical that this influence is related to BSA; that is, at a given absolute Ereq, the same absolute rate of sweat production (in g/min) must be secreted over a smaller surface area in the SM group; therefore, the mean rate of sweating per unit area (in mg·cm−2·min−1) should be greater with a smaller BSA. For the purpose of comparing LSR responses between groups unmatched for BSA, this systematic difference in LSR due to differences in BSA at a fixed absolute Ereq can theoretically be removed by prescribing the same Ereq per unit of BSA (in W/m2). In the present study, this notion is strongly supported by the similar mean LSR values during exercise at Ereq values of 165 and 190 W/m2 (Fig. 4), despite differences in absolute Ereq and BSA. By removing this systematic difference in LSR due to differences in BSA between groups, researchers can isolate the independent influence of physiological factors on local sudomotor activity, since any difference will be due to the factor under investigation, as opposed to inherent bias associated with the exercise intensity prescribed.
Although previous research has clearly highlighted the importance of changes in core temperature and Tsk for sudomotor control (4), the present study emphasizes the large influence of biophysical factors on ΔTre, WBSR, and LSR among individuals of different morphological characteristics (Table 1) but identical functional parameters for the physiological control of sudomotor activity (Fig. 5, Table 2). In participants unmatched for mass and BSA, separate experimental approaches are necessary to isolate the influence of other factors that are different between participants on ΔTre, WBSR, and LSR. For example, WBSR should be compared between groups using a fixed absolute Ereq in W (16), whereas a fixed Ḣprod in W/kg is most appropriate for comparing core temperature changes between groups. The latter, however, would not be valid for simultaneous comparisons of WBSR between groups of dissimilar mass, because absolute Ereq (in W) would be different. Similarly, LSR can only be compared using a fixed Ereq in W/m2, so, if groups are of dissimilar mass, WBSR could not be independently compared, whereas changes in core temperature could only be compared if groups had similar BSA-to-mass ratios, since a fixed W/m2 would simultaneously yield the same W/kg between groups.
Finally, the present findings may only be applicable in compensable conditions. In an uncompensable environment [i.e., Ereq exceeds the maximum potential for evaporation (Emax)] differences in BSA-to-mass ratio will raise Ereq (in W/m2) in larger individuals for a given Ḣprod in W/kg, while Emax is unchanged. The greater difference between Ereq and Emax in larger individuals should theoretically result in a higher rate of heat storage; however, this remains to be experimentally proven.
In conclusion, to prevent the introduction of systematic bias to an experimental design related to differences in Ḣprod and body morphology, the present data suggest that exercise should be prescribed to elicit the same Ḣprod in W/kg to compare changes in core temperature and the same Ereq in W/m2 to compare LSR responses. These approaches may be particularly useful for researchers investigating thermoregulatory responses between healthy/control and special populations that may potentially demonstrate impaired heat dissipation secondary to alterations in thermoeffector function, such as diseases that lead to autonomic dysfunction (e.g., multiple sclerosis) or injuries that dennervate sweat glands (e.g., spinal cord injury).
This research was supported by a Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant (no. 386143-2010, O. Jay). M. N. Cramer is supported by a University of Ottawa Excellence Scholarship and a NSERC Postgraduate Scholarship (PGS-D).
No conflicts of interest, financial or otherwise, are declared by the author(s).
Author contributions: M.N.C. and O.J. conception and design of research; M.N.C. performed experiments; M.N.C. analyzed data; M.N.C. and O.J. interpreted results of experiments; M.N.C. prepared figures; M.N.C. and O.J. drafted manuscript; M.N.C. and O.J. edited and revised manuscript; M.N.C. and O.J. approved final version of manuscript.
This study was performed by M. N. Cramer in partial fulfillment of the degree Doctor of Philosophy from the University of Ottawa. The authors thank the volunteers for participation, Zuzana Novak for assistance with data collection, Dr. Éric Doucet for use of the DXA scanner, and Isabelle LaForest for performing DXA scans.
Prescribing Exercise Intensity to Elicit a Fixed Ḣprod
During a preexperimental visit, height and body mass must first be measured if prescribing Ḣprod in W/kg. BSA can be estimated using equation of DuBois and DuBois (12).
Before testing, identify the target absolute Ḣprod (in W) to be used. For example, if a fixed Ḣprod of 7.0 W/kg is required and the individual is 75 kg, the target absolute Ḣprod is 7.0 × 75 = 525 W.
The exercise intensity required to elicit each target absolute Ḣprod may be estimated from the relationship between the V̇o2 and external work rate. To establish this relationship, have each participant perform a submaximal incremental exercise test that includes a range of work rates that will incorporate the experimental target absolute Ḣprod. The work rates in this test may be estimated based on pilot testing, previous research, or, in the case of cycling, assumed gross efficiency values. For example, if Ḣprod values of 400 and 600 W will be targeted, assuming a gross efficiency of 17% (14), work rates of ∼80 and ∼125 W, respectively, would be expected. Therefore, during the preliminary test, the initial work rate may be set to 80 W and increased by 20 W/stage for four stages (i.e., up to 140 W) to include all estimated target work rates. The duration of each stage should be sufficient to attain steady-state V̇o2 values (i.e., 3–5 min). Metabolic data (i.e., V̇o2 and RER) should be collected throughout this test.
Take the final 1-min (i.e., steady-state) V̇o2 value of each stage and, using conventional equations (i.e., Eq. 1 in methods), calculate M and then subtract W to obtain Ḣprod for each stage. As the Ḣprod-work rate relationship is linear at submaximal intensities (2), the work rate required to elicit each target absolute Ḣprod may be estimated using the equation of a straight line (y = mx + b). It is also important to note the corresponding V̇o2 value for each required work rate.
During experimentation, set the initial work rate as that predicted to elicit the target absolute Ḣprod. The actual Ḣprod should be verified using real-time V̇o2 measurements, with slight work rate adjustments potentially necessary to ensure a constant Ḣprod throughout exercise. To this end, it is crucial that V̇o2 is monitored closely.
Prescribing Exercise Intensity to Elicit a Fixed Ereq
Since Ereq is primarily determined by Ḣprod (see Eq. 8 in methods), prescribing work rates that elicit a fixed Ḣprod in W or Ḣprod in W/m2 should result in fixed Ereq in W or Ereq in W/m2, respectively, provided that the experimental environmental conditions (ambient temperature, air velocity) are constant. To calculate the actual Ereq, dry and respiratory heat exchange must be calculated using mean Tsk, air velocity, and ambient temperature measurements (see Eqs. 3–7 in methods).
- Copyright © 2014 the American Physiological Society