Advertisement

European Journal of Applied Physiology

, Volume 114, Issue 11, pp 2399–2410 | Cite as

Do greater rates of body heat storage precede the accelerated reduction of self-paced exercise intensity in the heat?

  • Nicholas M. Ravanelli
  • Matthew N. Cramer
  • Yannick Molgat-Seon
  • Anthony N. Carlsen
  • Ollie Jay
Original Article

Abstract

Aim

To reevaluate the previous hypothesis that greater reductions in self-paced exercise intensity in the heat are mediated by early differences in the rate of body heat storage (S).

Methods

Eight trained volunteers cycled in 19 °C/1.8 kPa (COOL), 25 °C/1.2 kPa (NORM), and 34 °C/1.6 kPa (HOT), while maintaining an RPE of 16. Potential differences in S following the onset of exercise were assessed by comparing rates of esophageal temperature change (ΔT est); and estimated S values using a traditional two-compartment thermometric model (S therm) of changes in rectal (T re) and skin (T sk) temperature, and partitional calorimetry (S cal).

Results

After 15 min of exercise, workload decreased more in HOT vs. COOL (P = 0.03), resulting in a shorter time (HOT: 40.7 ± 14.9 min; COOL: 53.5 ± 18.7 min; P = 0.04) to 70 % of initial workload. However, there were no preceding differences in ΔT est between conditions (P = 0.18). S therm values were different between HOT and COOL during the first 5 min of exercise (P < 0.05), primarily due to negative S therm values (−32 ± 15 kJ min−1) in COOL, which according to partitional calorimetric measurements, required improbably high (~56 kJ min−1) rates of evaporation when no sweating on the back and thigh was observed until after 7.6 ± 1.5 min and 4.8 ± 1.7 min of exercise, respectively. S cal values in the first 5 min of exercise confirmed S was actually positive in COOL (+21 ± 8 kJ min−1) and not negative. Different S therm values following the onset of exercise at different environmental temperatures are simply due to transient differences in the rate of change in T sk.

Conclusion

Reductions in self-paced exercise intensity in the heat are not mediated by early differences in S following the onset of exercise.

Keywords

Core temperature Exercise performance Partitional calorimetry Rating of perceived exertion Thermoregulation 

Abbreviations

Ar/AD

Effective radiant surface area (ND)

BF

Biceps femoris

AD

Body surface area (m2)

C

Convective heat exchange (W m−2)

Cres

Convective heat exchange via respiration (W m−2)

Esk

Evaporative heat loss from the skin (W m−2)

Eres

Evaporative heat loss via respiration (W m−2)

EC

Caloric equivalent of carbohydrates (kJ LO−2)

EF

Caloric equivalent of lipids (kJ LO−2)

EMG

Electromyography (% of initial EMG)

ε

Area-weighted emissivity of the skin (ND)

fcl

Area-weighted clothing factor (ND)

GM

Gluteus maximus

h

Combined heat transfer coefficient (W m−2 °C−1)

hc

Convective heat transfer coefficient (W m−2 °C−1)

hr

Radiative heat transfer coefficient (W m−2 °C−1)

Hprod

Rate of metabolic heat production (W m−2)

iEMG

Integrated electromyography (% of initial iEMG)

LG

Lateral gastrocnemius

LSRback

Local sweat rate on the lower back (mg cm−2 min−1)

LSRthigh

Local sweat rate on the thigh (mg cm−2 min−1)

M

Metabolic energy expenditure (W m−2)

PAR-Q

Physical Activity Readiness Questionnaire

Pa

Ambient water vapor pressure (kPa)

R

Radiant Heat exchange (W m−2)

Rcl

Thermal resistance of clothing (m2 °C W−1)

RER

Respiratory exchange ratio (ND)

RPE

Rating of perceived exertion

S

Rate of body heat storage (kJ min−1)

Scal

Rate of body heat storage measured with partitional calorimetry (kJ min−1)

Stherm

Rate of body heat storage estimated with a traditional 2-compartment thermometry model (kJ min−1)

σ

Stefan-Boltzmann constant (5.67 × 10−8W m−2 °C−4)

Ta

Ambient air temperature (°C)

ΔTb

Change in mean body temperature (°C)

Tcl

Temperature of the clothing (°C)

Tes

Esophageal temperature (°C)

ΔTest

Rate of oesophageal temperature change (°C min−1)

To

Operative temperature (°C)

Tr

Mean radiant temperature (°C)

Tre

Rectal temperature (°C)

Tsk

Mean skin temperature (°C)

VO2peak

Peak rate of oxygen consumption (L min−1)

VO2

Rate of oxygen consumption (L min−1)

v

Air velocity (m s−1)

VL

Vastus lateralis

W

Rate of external work (W m−2)

Wpeak

Peak rate of external work (W)

Introduction

It has been proposed that catastrophic levels of hyperthermia may be avoided during self-paced exercise in the heat via an anticipatory (i.e. feed forward) reduction of power output, motor unit recruitment (and therefore metabolic heat production), which are subconsciously mediated by the rate of body heat storage (S) during the first 4 min of exercise (Tucker et al. 2006). This interpretation was subsequently challenged using a theoretical heat balance assessment that demonstrated that the observed S values following the onset of exercise were potentially incorrect (Jay and Kenny 2009). It was further proposed that the underlying reason for these inaccurate S values was the erroneous nature of the two-compartment thermometric model employed that uses minute-by-minute changes in rectal temperature (T re) and mean skin temperature (T sk) to estimate S (Stolwijk and Hardy 1966; Horstman and Horvath 1972; Vallerand 1992; Snellen 2000; Jay et al. 2007). Specifically, early changes in T sk, which represent the body “shell”, seem to exaggerate any potential differences in S during the early stages of exercise between hot and cold environments; and changes in T re, which represent the body “core”, do not seem to reflect any changes in S following the start of exercise due to its well documented time lag (Jay and Kenny 2009). If this is the case, then the original conclusion—that differences in S between hot and cool conditions only occur during the first 4 min of exercise, with no differences observed thereafter (when changes in self-regulated power output then occur)—may be greatly dependent upon the “core” to “shell” weightings used to estimate S.

In the absence of whole-body direct calorimetry, a different thermometric approach is required to re-evaluate whether S plays a potential role in a feed-forward regulatory model of self-paced exercise. Firstly, any early changes in T sk can likely be discounted since they do not contribute to the description of changes in heat storage during the first 10 min of exercise (Jay et al. 2007). Secondly, a temperature probe placed in the esophagus (T es) should provide a more responsive indicator of early changes in S than T re. Since the tissue temperature changes occurring in the body as a consequence of the large rates of initial heat storage will primarily be centralized around the main source of exercise-induced thermogenesis—the active musculature (Jay et al. 2007)—temperature of out-flowing blood from active muscle and returning to the heart should therefore closely resemble T es, which has been shown to approach aortic blood temperature with only a minor lag of 80–160 s during high rates of changing blood temperature (Cooper and Kenyon 1957; Shiraki et al. 1986). Therefore, while replacing T re with T es as a representative of the body “core” in the conventional two-compartment thermometry model probably does not yield precise absolute values for S during the early stages of exercise, it stands to reason that different rates of change of T esT est) will occur in parallel to proportional physiologically significant differences in S with a minimal time delay.

The primary purpose of the present study was to reevaluate the potential role of early changes in S in an anticipatory feed-forward regulatory model of self-paced exercise (Tucker et al. 2006). Previous procedures were replicated and S was assessed using a range of methods in subjects self-regulating their cycling power output at a fixed rating of perceived exertion (RPE) (Borg 1982) in cool, neutral, and warm environments. Specifically, we sought to determine if subjects only show different ΔT est during the early stages of exercise, indicative of different S, with similar ΔT est occurring thereafter irrespective of environmental conditions due to a reduction in power output that is greatest in the hottest environment. A secondary purpose of the present study was to assess whether previous conclusions related to the role of S were dependent upon the weighting coefficients of the “core” and “shell” components of the two-compartment thermometry model used to estimate S, and to determine how the S values previously reported in cooler environments using thermometry during the early stages of exercise compare to S values measured using partitional calorimetry. It was hypothesized that: (1) ΔT est would be similar between cool, neutral, and hot conditions throughout self-regulated exercise at a fixed RPE; (2) different “core” and “shell” weighting coefficients employed in the two-compartment thermometry model alter the conclusion of different S values between conditions during the early stages of exercise; and (3) S values measured with partitional calorimetry during the early stages of self-regulated exercise in cooler environments are positive and not negative.

Methods

Participants

A power calculation using esophageal temperature data from pilot studies was performed using the calculated effect size of 1.2, α = 0.05, and β = 0.2 which determined that eight participants were required to demonstrate a difference in ΔT est. Therefore, 8 endurance-trained volunteers (seven males, one female) participated in this study (VO2peak 59.3 ± 6.3 ml kg−1 min−1; W peak 355 ± 49 W, 5.0 ± 0.6 W kg−1; Age 26.0 ± 5.5 years; Mass 72.0 ± 7.3 kg; Height 1.88 ± 0.12 m2). Prior to experimentation, all participants provided written informed consent, and completed a Physical Activity Readiness Questionnaire (PAR-Q) as well as an American Heart Association/American College of Sports Medicine Health/Fitness Facility Pre-participation Screening Questionnaire. The experimental protocol was approved by the University of Ottawa Research Ethics Committee, and conformed to the guidelines set forth in the Declaration of Helsinki.

Preliminary trials

During the first visit to the laboratory, each participant performed an incremental exercise test to exhaustion on an upright cycle ergometer to determine their peak rate of oxygen consumption (VO2peak) and peak external workload (W peak): Following a 10 min warmup the incremental protocol began at an external workload of 80 W, which was then increased by 20 W min−1 until volitional exhaustion. This test was conducted in accordance with the guidelines of the Canadian Society for Exercise Physiology (CSEP 1996). The second visit involved familiarization with the fixed RPE protocol: In the familiarization trial, participants commenced cycling at a workload of 100 W, which was subsequently increased by 10 W every 5 min until an RPE of 16 (corresponding to a rating between ‘hard’ and ‘very hard’ on the Borg scale) was reached. The participants were then instructed to adjust their workload to maintain this sensation for an additional 10 min. This approach allowed the participants to become familiar with the Borg scale and the sensation associated with the target RPE. Thus for each participant the initial workload for each experimental trial was established as the average workload during the final 10 min of the familiarization trial.

Experimental protocol

In separate sessions, experimental trials were performed in hot (HOT: 33.7 ± 0.8 °C, 1.6 ± 0.5 kPa), thermo neutral (NORM: 25.2 ± 0.7 °C, 1.2 ± 0.5 kPa), and cool (COOL: 19.3 ± 1.4 °C, 1.8 ± 0.5 kPa) conditions, with trial order counterbalanced across participants. Levels of humidity were chosen to elicit a similar absolute skin-air vapor pressure gradient at each ambient air temperature, and to thereby produce a similar drive for evaporation. In the 24 h prior to each experimental trial, participants were asked to avoid alcohol, caffeine, and strenuous exercise. Additionally, they were asked to consume a light meal and 500 ml of water ~2 h prior to arrival at the laboratory. Upon arrival, the participants voided their bladders providing a urine sample to ensure hydration and changed into a standardized clothing ensemble consisting of cotton shorts, cotton socks, and cycling shoes. Trials would be postponed and rescheduled if urine specific gravity measured using a clinical refractometer (Reichert TS 400, Depew, NY, USA) was greater than 1.020. Following a 45 min instrumentation period, a 10 min warmup with the cycle ergometer set at 125 W was performed. Following this, the participants entered the laboratory and rested quietly in a seated position on the ergometer while instrumentation was completed. A 20 min baseline data collection period preceded exercise, during which time a body mass measurement was taken 5 min before exercise with a platform scale accurate to the nearest ± 2 g (Combics 2; Sartorius, Mississauga, ON, Canada) with sensor cables secured to an adjacent table. Immediately prior to the start of exercise participants were reminded to maintain a RPE of 16 by adjusting the workload. The ergometer was set to a cadence-independent mode, so that a preferred cadence could be maintained for any self-selected workload. The participants then began cycling, facing a mechanized fan providing an air velocity of 1.7 ± 0.4 m s−1. Participants were blinded to physiological and performance feedback throughout all trials. Exercise continued until workload declined to <70 % of the initial 5 min average. A final body mass measurement was taken immediately upon the completion of exercise, prior to the subject drying off, and with sensor cables secured to an adjacent table at exactly the same point as during the pre-exercise body mass measurement.

Measurements

Ambient conditions: Dry-bulb temperature and absolute humidity were measured every 5 s using a dew point mirror (473 RH Systems, Albuquerque, NM, USA). Air velocity was measured with a hot wire anemometer (HHF42, Omega Engineering, Stamford, CT, USA).

Indirect calorimetry: Participants were equipped with a mouthpiece and nose clip, and were instructed to breathe normally at all times. Expired gases were analyzed breath-by-breath throughout exercise using a metabolic cart (Vmax Encore, CareFusion, Yorba Linda, CA, USA).

Thermometry: Core temperature was measured continuously in the rectum and the esophagus using general purpose pediatric T-type (copper/constantan) thermocouple probes (Mon-a-therm®, Mallinckrodt Medical, St. Louis, MO, USA). Rectal temperature (T re) was measured at a depth of 12 cm beyond the anal sphincter. Esophageal temperature (T es) was measured at a maximum depth of 40 cm, estimated to be at the level of the left ventricle (Mekjavic and Rempel 1990). Skin temperature was measured at four sites on the left side of the body using T-type thermocouples integrated into heat flux sensors (Concept Engineering, Old Saybrook, CT, USA), which were secured to the skin with surgical tape (Transpore®, 3M, London, ON, Canada). Mean skin temperature (T sk) was estimated using a weighted average of four sites in accordance with Ramanathan (1964): chest 30 %, triceps 30 %, thigh 20 %, and calf 20 %. Thermometric measurements were sampled every 5 s (NI cDAQ-91722 module, National Instruments, Austin, TX, USA) and displayed in real-time on a desktop computer using customized LabView software (v7.0, National Instruments, Austin, TX, USA).

Local sweat rates were measured on the lower back (LSRback) and thigh (LSRthigh) on the right side of the body using the ventilated capsule technique. These sites were chosen as they demonstrate the relatively early onset times for thermoregulatory sweating (Cotter and Taylor 2005). Anhydrous air was supplied to each 4.1 cm2 capsule at a constant flow rate of 1.80 L min−1 (Omega FMA-A2307, Omega Engineering, Stamford, CT, USA). Capsules were secured using skin glue (Collodion USP MD0002, Mavidon, Lake Worth, FL, USA) and surgical tape. The temperature and humidity of effluent air were measured every 5 s using factory-calibrated capacitance hygrometers (HMT333, Vaisala, Vantaa, Finland). Local sweat rate was then calculated as the product of absolute humidity and flow rate, and expressed relative to the area under the capsule in milligrams per square centimeter per minute (mg cm−2 min−1).

Electromyography (EMG): Surface EMG data were collected from the right vastus lateralis (VL), gluteus maximus (GM), biceps femoris (BF), and lateral gastrocnemius (LG) using bipolar, pre-amplified surface electrodes connected to an external amplifier system (Delsys Bagnoli-8, Natick, MA, USA). The recording sites were prepared and the electrodes were attached using double-sided adhesive strips, and secured using surgical tape. A grounding electrode (Kendall Q-Trace 5400, Covidien Inc., Mansfield, MA, USA) was placed over the participant’s right anterior superior iliac spine.

Unfiltered raw EMG was digitally sampled at 1 kHz (National Instruments PCI-6052E, Austin, TX, USA) for the first 10 s of each minute during cycling and EMG values for all four muscles were calculated from a time window corresponding to 10 cycles within each 10 s time block irrespective of cadence. The magnitude of the integrated EMG (iEMG) was calculated for each muscle by numerically integrating the rectified EMG signal within the 10-cycle window on each trial. To quantify muscle output for the whole leg (LEG) in a more holistic manner, LEG weighted average iEMG was calculated from the normalized iEMG and expressed as a percentage of initial LEG. The weightings used were: VL, 41 %; GM, 28 %; BF, 10 %; LG, 21 % (Ericson 1986). To normalize between testing sessions, iEMG was expressed as a value per unit of work (W) by dividing iEMG by the cycle ergometer power output observed at minute 5 in each session (see above). Initial iEMG was then calculated as the mean of iEMG at minutes 4, 5, and 6. For minutes 10, 15 and 20, iEMG values were calculated as the mean iEMG values from the preceding 5 min (i.e., 6–10, 11–15, 16–20), and expressed as a percentage of initial iEMG.

Calculations

Rate of heat storage (partitional calorimetry): Calorimetric estimates of the rates of heat storage (S cal) were calculated for NORM and COOL conditions according to the conceptual heat balance equation:
$$S_{\text{cal}} = \, M \, {-} \, W \, {-} \, \left( {C \, + \, R \, + E_{\text{sk}} + \, C_{\text{res}} + \, E_{\text{res}} } \right) \, \left( {{\text{W m}}^{ - 2} } \right)$$
Metabolic rate (M) was calculated as:
$$M \, = \, V{\text{O}}_{2} \frac{{\left( {\left( {\frac{{{\text{RER}} {-} 0.7}}{0.3}} \right)e_{\text{c}} } \right) + \left( {\left( {\frac{{1.0 {-} {\text{RER}}}}{0.3}} \right)e_{\text{f}} } \right)}}{{\left( {60} \right)(A_{\text{D}} )}} \; \times \;1000 \, \left( {{\text{W m}}^{ - 2} } \right)$$
where: VO2 is the rate of oxygen consumption (L min−1); RER is the non-dimensional respiratory exchange ratio; e c and e f are the energetic equivalents of carbohydrate (21.13 kJ L−1 O2) and fat (19.62 kJ L−1 O2), respectively; A D is body surface area in m2, estimated using the equation of DuBois and Dubois (1916). External workload (W) was regulated using an electromagnetically-braked upright cycle ergometer (Ergo Race, Kettler, Ense-Parsit, Germany). The rate of metabolic heat production (H prod) was calculated as the difference between metabolic rate and external workload (in W m−2).
Combined respiratory heat exchange (H res) via convection (C res) and evaporation (E res) was calculated as:
$$C_{\text{res}} + \, E_{\text{res}} = \, \left[ {0.0014\; \times \;\left( {H_{\text{prod}} } \right)\; \times \;\left( {34 \, {-} \, T_{\text{a}} } \right)\left] { \, + \, } \right[0.0173\;\left( {H_{\text{prod}} } \right)\; \times \;\left( {5.87 \, {-} \, P_{\text{a}} } \right)} \right]\;\left( {{\text{W m}}^{ - 2} } \right)$$
where: P a is the ambient vapor pressure in kilopascals (kPa).
Heat exchange via convection (C) and radiation (R) was calculated as:
$$C \, + R \, = \frac{{(T_{\text{sk}} {-}T_{\text{o}} )}}{{R_{\text{cl}} + \frac{1}{{f_{\text{cl}} \cdot h}}}}\;\left( {{\text{W m}}^{ - 2} } \right)$$
where: R cl is the thermal resistance of clothing (m2 °C W−1), but was assumed to be negligible in the present study due to a minimal clothing ensemble; f cl is the non-dimensional area-weighted clothing factor (assumed to be 1.0 with minimal clothing); h is the combined heat transfer coefficient (W m−2 °C−1), which is the sum of the convective (h c) and radiant (h r) heat transfer coefficients (see below).
Operative temperature (T o) was estimated as:
$$T_{\text{o}} = \frac{{(h_{\text{c}} T_{\text{a}} ) + (h_{\text{r}} T_{\text{r}} )}}{{h_{\text{r}} + h_{\text{c}} }} \;\left( {^\circ {\text{C}}} \right)$$
where: T a is ambient temperature (°C) and T r is the radiant temperature (°C), assumed to be equal to T a.
Values for h c and h r were calculated as:
$$\begin{aligned} &h_{\text{c}} = { 8}. 3v^{0. 6} \left[ {{\text{W}}\;{\text{m}}^{ - 2} \;^\circ {\text{C}}^{ - 1} } \right] \\ & h_{\text{r}} = \, 4\varepsilon \sigma \frac{{A_{\text{r}} }}{{A_{\text{D}} }}\; \left[273.15 \, + \, (\frac{{T_{\text{cl}} + T_{\text{r}} }}{2})\right]^{3} \left( {{\text{W}}\;{\text{m}}^{ - 2} \;^\circ {\text{C}}^{ - 1} } \right) \\ \end{aligned}$$
where: v is air velocity (m s−1); ε is the area-weighted emissivity of the skin, taken as 0.95; σ is the Stefan-Boltzmann constant (5.67 × 10−8W m−2 °C−4); A r/A D is the effective radiant surface area (ND), estimated to be 0.7 for a seated individual (Fanger 1970); and T cl is the temperature of the clothing (°C), assumed to be equal to T sk.

Evaporative heat loss was assumed to be negligible in NORM and COOL conditions until the onset of sweating. No estimate of evaporative heat loss and therefore S cal was made in HOT since sweating was evident prior to exercise.

Although estimated in W m−2, all heat balance parameters were converted to kJ min−1 (1 W = 0.06 kJ min−1) to compare the present results to those of previous studies.

Rate of heat storage (thermometry): Minute-by-minute changes in mean body temperature (ΔT b) were estimated using four different two-compartment thermometric models: T re/T sk (A) 0.66/0.34, (B) 0.79/0.21, (C) 0.90/0.10, and (D) 1.00/0.00. Thermometric estimates of heat storage (S therm) were calculated as the product of body mass, an estimated average-specific heat capacity of 3.47 kJ kg−1 °C−1 for body tissue (Gephart and DuBois 1915), and ΔT b. Values of S therm using T re/T sk weighting coefficients of 0.79/0.21 were compared directly to S cal during the first 5 min of exercise in COOL and NORM.

Statistical analysis

All data are reported as the mean ± standard deviation (SD) for each variable. Mean trial duration was analyzed using a one-way analysis of variance (ANOVA) employing the independent variable of ambient temperature (three levels: HOT, NORM, COOL). Since trial duration varied between subjects, analyzes were performed for the first 20 min of exercise, which was the minimum trial duration for all participants. A two-way repeated measures ANOVA, with the independent factors of ambient temperature and time (four levels: Baseline, 10, 15, 20 min) were used to analyze the dependent variables of percentage change in workload, ΔT est, ΔT skt, H prod, S threm, LSRback, LSRthigh, individual muscle iEMG, and whole-leg iEMG. S cal values in each of the first 5 min of exercise were compared between NORM and COOL using a two-way ANOVA. A Greenhouse-Geisser correction was applied if assumptions of sphericity were not met. When a significant main effect was detected, individual differences were compared using the Holm–Bonferonni method. The probability of a Type I error was set at the 0.05 level. All statistical analyzes were performed with SPSS 19.0 (SPSS Inc., Chicago, IL, USA).

Results

Trial duration

Mean trial duration was 53.5 ± 18.7, 54.5 ± 18.5, and 40.7 ± 14.9 min for COOL, NORM, and HOT, respectively (Fig. 1a). Mean trial duration was significantly shorter in the HOT compared to the NORM and COOL condition (P < 0.05).
Fig. 1

Mean trial duration (a) and self-selected power output expressed as a percentage of the absolute workload at 5 min of exercise (b) for HOT, NORM, and COOL.*Significantly higher than HOT (P < 0.05). #Significantly lower in HOT compared to COOL (P < 0.05)

External workload

Changes in workload expressed as a percentage of the initial workload (i.e., at 5 min) are depicted in Fig. 1b. Initial workloads were not different between the COOL, NORM, and HOT (219 ± 27, 224 ± 35, and 219 ± 35 W, respectively; P > 0.05). However, the reduction in workload was significantly greater in HOT compared to COOL after 15 min (P < 0.05; Fig. 1b).

Rates of change of esophageal and skin temperatures

The rate of change of esophageal temperature (ΔT est) is shown in Fig. 2. ΔT est changed significantly with exercise time (P < 0.001); however, there was no difference between HOT, NORM and COOL (P = 0.18) (Fig. 2).
Fig. 2

Rate of change in T es during the first 20 min of exercise

The rates of change of rectal temperature (ΔT ret) and mean skin temperature (ΔT skt) are shown in Fig. 3a, b. No differences were observed between conditions for ΔT ret during the first 20 min of exercise. ΔT skt declined rapidly following the onset of exercise, with the highest rates of skin cooling observed at 2 min in all conditions. While ΔT skt changed significantly with exercise time (P < 0.001), there was also a significant interaction between time and environmental temperature (P < 0.001) showing that the change in ΔT skt with exercise time was most negative and greatest in magnitude in the COOL (Fig. 3b).
Fig. 3

Rate of change in T re (a) and T sk (b) during the first 20 min of exercise. *Significantly greater in HOT compared to NORM and COOL (P < 0.05)

Differences in absolute T sk were evident between conditions (P < 0.001), with the highest values observed in HOT (33.89 ± 0.20 °C) and the lowest values in COOL (28.30 ± 0.57 °C) during the first 20 min.

Rate of heat storage estimated using thermometry (S therm)

Compared to HOT, S therm in NORM and COOL was negative and significantly lower during the first 4 min of exercise using model A (core/shell: 0.66/0.34) and B (core/shell: 0.79/0.21), and between 2–4 min of exercise using model C (core/shell: 0.90/0.10). In models A, B, and C, S therm reached a nadir after 2 min (Fig. 4a–c). However, as the proportion of the “shell” compartment became progressively smaller (from 0.34 to 0.10), the estimated values of S therm became less negative in the COOL and NORM conditions and became increasingly similar to the HOT condition. When the “shell” compartment was completely eliminated in model D (core/shell: 1.00/0.00) (Fig. 4b), no difference in S therm was observed between conditions.
Fig. 4

Thermometric estimates of the rate of heat storage, using four different T re/T sk weighting coefficients for T b: a 0.66/0.34, b 0.79/0.21, c 0.90/0.10 and d 1.00/0.00. *Significantly greater in HOT compared to NORM and COOL (P < 0.05)

Rate of heat storage estimated in COOL and NORM conditions using partitional calorimetry (S cal)

S cal in the HOT condition could not be estimated using partitional calorimetry because sweating had already begun and minute-by-minute rates of evaporation could not be measured (see “Methods”). The average of the first 5 min of exercise heat balance parameters for NORM and COOL condition are detailed in Fig. 5. While H prod increased significantly with exercise time (P < 0.001), this change was not altered by environmental temperature (P = 0.98). In both COOL and NORM, S cal was positive (i.e. heat energy was stored in the body), and increased during the first 5 min of exercise while the first 5-minute average was negative and significantly lower for S therm (P < 0.001). While the rate of evaporation could not be measured, when compared to the values estimated in the COOL and NORM using thermometry (S therm), rates of evaporation needed for these S therm values to be correct would have been between 30 and 56 kJ min−1, respectively.
Fig. 5

Heat balance parameters averaged over the first 5 min of exersice for COOL and NORM conditions. H prod, rate of metabolic heat production (i.e. M−W); C + R + H res, rates of convective, radiant, and respiratory heat exchange, respectively; S cal, rate of heat storage estimated via partitional calorimetry; S therm, core/shell weighting of 0.79/0.21, respectively. *Significantly lower in S therm compared to S cal (P < 0.001)

Local sweat rates

Significantly greater LSRback was observed in the HOT compared to the COOL during the first 20 min (P < 0.05), while LSRthigh was significantly greater in the HOT compared to the COOL during the first 11 min (P < 0.05). The onset of sweating at LSRback and LSRthigh in COOL and NORM occurred after 7.6 ± 1.5 and 4.8 ± 1.7 min of exercise, respectively. Sweating on both the back and thigh had already initiated prior to the onset of exercise in the HOT condition.

Integrated EMG (iEMG)

For iEMG values, a main effect of time was found for all four muscles (Fig. 6; P < 0.05). There were no other main effects or interactions. Pre-planned comparisons showed that for BF and GM, iEMG was significantly lower than initial iEMG at 10, 15, and 20 min (P < 0.05). For VL, iEMG was significantly lower than initial iEMG at the 15 and 20 min (P < 0.05), and for LG, iEMG was only significantly lower than initial iEMG at 20 min (P < 0.05).
Fig. 6

iEMG during the first 20 min of exercise. VL vastus lateralis, GM gluteus medius, BF biceps femoris; LG lateral gastrocnemius. *Significantly different from initial for all conditions (P < 0.05)

LEG weighted average iEMG data are presented in Fig. 6. Similar to the individual muscles, for LEG there was a main effect of time on iEMG (P = 0.001). Pre-planned comparisons showed that for LEG, iEMG was significantly lower at all time points compared to initial iEMG (P < 0.05). There was no main effect for temperature (P = 0.387) and no interaction between the variables (P = 0.497).

Discussion

In the present study an ambient temperature-dependent reduction in power output occurred after 15–20 min of self-paced fixed-RPE exercise, a result similar to that reported by Tucker et al. (2006). However, in contrast to their conclusions, we did not observe any difference in ΔT est between COOL, NORM and HOT conditions in advance of these changes in power output, or thereafter, suggesting a similar S in all conditions. The negative S estimated by Tucker et al. (2006) during the first 4 min of exercise using the traditional two-compartment thermometry model in the COOL and NORM conditions, were also obtained with the S therm  thermometry model in the present study, but it is clear that these negative values are the result of a disproportionate influence of negative rates of skin temperature change during the first 5 min of exercise arising from an increased self-generated air velocity across the legs (Nishi and Gagge 1970). Taken together our findings do not support the hypothesis that S during the early minutes of exercise mediates subsequent changes in self-selected exercise intensity.

It has been previously argued that T es may not provide an accurate absolute estimation of changes in body heat content when incorporated within a two-compartment thermometry model (Jay et al. 2007). However, while an accurate absolute S value can probably not be derived, considering the responsiveness of this core temperature measurement to changes in arterial blood temperature, it stands to reason that physiologically significant differences in the rate of body heat storage will be paralleled by different rates of change of T es. Theoretically, differences in S will be best detected by measuring rates of temperature change at a site that is well perfused with blood, and serves as a major conduit for heat transfer from the working muscles to the skin. Despite its close proximity to the working muscles, the pelvic region has a high tissue density and relatively low blood perfusion, therefore the rate of change in rectal temperature is unsuitable to assess rapid changes in S such as those seen during the early stages of exercise (Blight 1957; Jay and Kenny 2009). In contrast, T es is measured at the level of the right atrium/aorta (Mekjavic and Rempel 1990); therefore, rapid changes in blood temperature due to changes in heat storage are adequately reflected in the change in T es (Cooper and Kenyon 1957; Molnar and Read 1974; Shiraki et al. 1986). Indeed, heat production within the muscle begins almost immediately at the onset of dynamic exercise, followed by heat transfer to the blood (with a rise in blood temperature) and a rise in T es (González-Alonso et al. 2000). Therefore, based on the time course for heat exchange between muscle and blood, as well as the regional differences between the esophagus and rectum, it would be expected that only a slight delay between the rise in H prod and ΔT est should be observed, while ΔT ret would be delayed. Accordingly, H prod reached its peak between 2 and 4 min into exercise; meanwhile T es and T re reached their peak after 4 and 8 min of exercise, respectively (Fig. 2, 3a).

Large differences in S therm between conditions during the first 5 min of exercise are clearly the result of a disproportionate influence of early and marked differences in the rate of skin temperature change (ΔT skt) (Fig. 3b). In both COOL and NORM conditions, the greatest (negative) ΔT skt occurred after 2 min, coinciding with the most negative S therm values when T sk was included in the thermometry model (Fig. 4a–c). Moreover, the thermometry models with the greatest weighting of T sk yielded the most negative S therm values during the first 5 min of exercise. Notably, when T sk was eliminated from the thermometry model, no differences in S therm were observed between COOL, NORM and HOT conditions (Fig. 4).

While the rate of metabolic heat production rises rapidly at the start of exercise, the increase in the rate of total heat dissipation lags considerably, resulting in heat imbalance and a positive S. Based on this well-established pattern of thermogenesis and thermolysis, Jay and Kenny (2009) argued that the negative S values estimated using the traditional two-compartment thermometric model in cold environments would require rates of heat dissipation that were impossible based on the characteristics of the environment (i.e., ambient temperature, humidity, air velocity) and the lag in thermoeffector onset (i.e., thermoregulatory sweating). Indeed, given the H prod, C+R, and H res values measured in the present study, for the S therm values estimated in the COOL and NORM conditions after 2 min of exercise to be correct, rates of whole-body sweating (and subsequent evaporation) equivalent to 1.39 and 1.31 L h−1, respectively, would have been required when in fact local sweat rate on the back and thigh had not changed from baseline pre-exercise levels. Assuming little evaporation from the skin was occurring, S cal in COOL and NORM conditions actually increased rapidly in a positive direction during the first 3 min of exercise, as would be expected (Jay and Kenny 2009), suggesting early positive whole-body heat storage—which is also supported by the positive change in T es by ~0.4 °C from rest to 5 min. In the absence of any means to measure dynamic rates of whole-body evaporation, and since sweating had already begun prior to exercise, S cal could not be determined in the HOT condition. Nevertheless, since H prod (Fig. 5) and ΔT es/dt (Fig. 2) were similar during the first 5 min of exercise it is likely that S cal would have also been similar in all three environmental conditions. As such, higher evaporation in HOT during this time would have likely been counterbalanced by smaller dry heat losses via convection and radiation.

A unique feature of the fixed-RPE protocol is the ability to isolate the physiological factor(s) that affects RPE-mediated changes in work rate, and therefore performance (Tucker 2004). As in previous studies, exercise duration (i.e., the time to reach 70 % of initial work rate) was shorter in HOT compared to NORM and COOL (Fig. 1a) with similar variability as previously reported (Tucker et al. 2006; Crewe et al. 2008) due to an accelerated decline in self-selected work rate (Fig. 1b), suggesting some factor related to a high ambient temperature influenced the selection of work rate for a RPE of 16. The present findings do not support any influence of S on the decline in exercise performance; however, this does not discount the possibility of other thermally-mediated effects. While a reduction in force production has been shown following hyperthermia (Nybo and Nielsen 2001), absolute core temperature responses were not different between conditions. However, large differences in skin temperature were observed. A high T sk or thermal sensation alters RPE and performance (Schlader et al. 2011a, b, c; Flouris 2011), and therefore may have been the primary stimulus for the reduction in work rate during the early minutes of exercise in the present study. Yet, since a high T sk was present before the onset of exercise, one might expect that the initial self-selected workload would have been lower in the HOT condition due to a higher T sk. Although this was not the case, a high T sk coupled with a rising core temperature may have precipitated the subsequent reduction in self-selected work rate (Ely et al. 2009). Future studies should investigate the independent role of T sk on exercise performance through acute mean or local T sk manipulations as recent work has suggested afferent feedback from skin thermoreceptors could mediate differences in self-paced work rate (Schlader et al. 2011a).

Another important unresolved issue regarding exercise performance is whether the accelerated decline in work rate during the HOT condition occurred in an anticipatory fashion. That is, it may be that neuromuscular activation (and therefore exercise intensity) was regulated to prevent deleterious homeostatic disturbances. However, it is unclear whether this was the case in the present study. Exercise performance in the HOT condition was impaired (Fig. 1), with a more accelerated decline in self-selected work rate resulting in a shorter time to reach 70 % of the initial work rate (the condition for terminating exercise). An anticipatory reduction in exercise intensity is argued to be mediated through a subconsciously altered pattern of neuromuscular activation (Tucker et al. 2006). In the present study, iEMG declined similarly over the first 20 min of exercise in HOT, NORM, and COOL conditions (Fig. 6), yet the reduction in work rate was significantly greater in the HOT condition when expressed as a percentage of the initial work rate (Fig. 1). It is possible that greater reductions in work rate in the HOT condition, while significant relative to NORM and COOL condition, in terms of the percentage decline from the initial work rate, were not large enough in absolute terms to yield significantly different iEMG signals between conditions. These iEMG results are similar to those previously presented by Tucker et al. (2006), even though the methods for normalizing iEMG varied. In fact, our normalization method (averaging over the first 5 min of dynamic exercise) led to a decrease in variability compared to normalizing to MVC, a static contraction. Importantly, however, in the present study the observed decrease in iEMG did not seem to occur in an anticipatory manner but rather in response to changes in core temperature.

The main limitation of the present study was the inability to measure S cal in the HOT condition. The determination of S cal depends on the ability to measure (1) the rate of metabolic heat production and (2) the estimation of all avenues of heat exchange. This latter condition is complicated by the inability to measure minute-by-minute changes in evaporative heat loss from the skin (E sk) that would have required a direct calorimeter or continuous body mass measurements while accounting for dripping sweat (if necessary). No sweating was evident until 3 and 5 min of exercise in NORM and COOL, respectively; therefore, there was no need to account for E sk in the estimation of S cal in those conditions (see “Methods”). However, sweating was apparent in the HOT condition during the baseline period, preventing an accurate determination of E sk and S cal. Nevertheless, since the rates of change of both T es and T re were similar in all conditions, it seems reasonable that a higher E sk counterbalanced the negligible dry heat loss in the HOT condition to elicit similar S cal values as the COOL and NORM conditions. Another minor limitation of the present study was that the ambient temperature of the COOL condition was slightly higher (~19 °C) than the COOL condition (~15 °C) in the main comparator study (Tucker et al. 2006). However, air temperature was still sufficiently low to recreate the performance differences previously observed between environmental conditions, and the two-compartment thermometric estimates of heat storage were similarly negative in our COOL condition to those previously reported (Tucker et al. 2006; Crewe et al. 2008).

In conclusion, despite a greater decline in self-paced exercise intensity in the HOT condition after ~15 min of exercise as previously reported (Tucker et al. 2006), ΔT est was similar in COOL, NORM and HOT conditions suggesting no early differences in the rate of body heat storage, thus apparently ruling this factor out as a potential mediator of the subsequent changes in power output. The accelerated decline in power output in the HOT condition occurred in parallel with a lower neuromuscular activation but in the absence of any evidence of neuromuscular fatigue, which may indicate an anticipatory regulation of exercise intensity based on factors unrelated to S, such as a high mean skin temperature.

Notes

Acknowledgments

The authors thank the volunteers for their participation and effort. This research was supported by a Discovery Grant from the Natural Sciences and Engineering Research Council (NSERC) of Canada (#386143-2010, held by O. Jay). Funding for the metabolic cart and dew point sensor were provided by a Canadian Foundation for Innovation (CFI) Infrastructure grant (co-applicant: O. Jay). N.M. Ravanelli was supported by a University of Ottawa Faculty of Health Sciences Research Bursary. M.N. Cramer is supported by a Postgraduate Scholarship (doctoral) from NSERC and an Excellence Scholarship from the University of Ottawa. Y. Molgat-Seon was supported by a University of Ottawa Master’s Scholarship. A.N. Carlsen is supported by NSERC and the University of Ottawa, Faculty of Health Sciences.

Conflict of interest

The authors report no competing interests, financial or otherwise.

References

  1. Blight J (1957) The relationship between the temperature in the rectum and of the blood in the bicarotid trunk of the calf during exposure to heat stress. J Physiol 136:393–403PubMedCentralGoogle Scholar
  2. Borg GA (1982) Psychophysical bases of perceived exertion. Med Sci Sports Exerc 14:377–381PubMedGoogle Scholar
  3. Cooper KE, Kenyon JR (1957) A comparison of temperatures measured in the rectum, œsophagus, and on the surface of the aorta during hypothermia in man. Br J Surg 44:616–619PubMedCrossRefGoogle Scholar
  4. Cotter JD, Taylor NA (2005) The distribution of cutaneous sudomotor and alliesthesial thermosensitivity in mildly heat-stressed humans: an open-loop approach. J Physiol 565:335–345PubMedCrossRefPubMedCentralGoogle Scholar
  5. Crewe H, Tucker R, Noakes T (2008) The rate of increase in rating of perceived exertion predicts the duration of exercise to fatigue at a fixed power output in different environmental conditions. Euro J Appl Physiol 103:569–577CrossRefGoogle Scholar
  6. CSEP (1996) Canadian Society for Exercise Physiology: certified fitness appraiser resource manual. CSEP, OttawaGoogle Scholar
  7. DuBois D, Dubois E (1916) A formula to estimate surface area if height and weight are known. Arch Intern Med 17:863CrossRefGoogle Scholar
  8. Ely BR, Ely MR, Cheuvront SN, Kenefick RW, DeGroot DW, Montain SJ (2009) Evidence against a 40°C core temperature threshold for fatigue in humans. J Appl Physiol 107:1519–1525PubMedCrossRefGoogle Scholar
  9. Ericson MO (1986) On the biomechanics of cycling. A study of joint and muscle load during exercise on the bicycle ergometer. Scand J Rehabil Med Suppl 16:1–43Google Scholar
  10. Fanger PO (1970) Thermal comfort. Danish Technical Press, CopenhagenGoogle Scholar
  11. Flouris AD (2011) Functional architecture of behavioural thermoregulation. Eur J Appl Physiol 111:1–8PubMedCrossRefGoogle Scholar
  12. Gephart FC, DuBois EF (1915) Fourth paper the determination of the basal metabolism of normal men and the effect of food. Arch Intern Med (Chic) 15:835–867CrossRefGoogle Scholar
  13. González-Alonso J, Mora-Rodríguez R, Coyle EF (2000) Stroke volume during exercise: interaction of environment and hydration. Am J Physiol Heart Circ Physiol 278:H321–H330PubMedGoogle Scholar
  14. Horstman DH, Horvath SM (1972) Cardiovascular and temperature regulatory changes during progressive dehydration and euhydration. J Appl Physiol 33:446–450PubMedGoogle Scholar
  15. Jay O, Kenny GP (2009) Current evidence does not support an anticipatory regulation of exercise intensity mediated by rate of body heat storage. J Appl Physiol 107:630–631PubMedCrossRefGoogle Scholar
  16. Jay O, Gariépy LM, Reardon FD, Webb P, Ducharme MB, Ramsay T, Kenny GP (2007) A three-compartment thermometry model for the improved estimation of changes in body heat content. Am J Physiol Regul Integr Comp Physiol 292:R167–R175PubMedCrossRefGoogle Scholar
  17. Mekjavic IB, Rempel ME (1990) Determination of esophageal probe insertion length based on standing and sitting height. J Appl Physiol 69:376–379PubMedGoogle Scholar
  18. Molnar GW, Read RC (1974) Studies during open-heart surgery on the special characteristics of rectal temperature. J Appl Physiol 36:333–336PubMedGoogle Scholar
  19. Nishi Y, Gagge AP (1970) Direct evaluation of convective heat transfer coefficient by naphthalene sublimation. J Appl Physiol 29:830–838PubMedGoogle Scholar
  20. Nybo L, Nielsen B (2001) Hyperthermia and central fatigue during prolonged exercise in humans. J Appl Physiol 91:1055–1060PubMedGoogle Scholar
  21. Ramanathan NL (1964) A new weighting system for mean surface temperature of the human body. J Appl Physiol 19:531–533PubMedGoogle Scholar
  22. Schlader ZJ, Simmons SE, Stannard SR, Mündel T (2011a) The independent roles of temperature and thermal perception in the control of human thermoregulatory behavior. Physiol Behav 103:217–224PubMedCrossRefGoogle Scholar
  23. Schlader ZJ, Simmons SE, Stannard SR, Mündel T (2011b) Skin temperature as a thermal controller of exercise intensity. Eur J Appl Physiol 111:1631–1639PubMedCrossRefGoogle Scholar
  24. Schlader ZJ, Stannard SR, Mündel T (2011c) Evidence for thermoregulatory behavior during self-paced exercise in the heat. J Therm Biol 36:390–396CrossRefGoogle Scholar
  25. Shiraki K, Konda N, Sagawa S (1986) Esophageal and tympanic temperature responses to core blood temperature changes during hyperthermia. J Appl Physiol 61:98–102PubMedGoogle Scholar
  26. Snellen JW (2000) An improved estimation of mean body temperature using combined direct calorimetry and thermometry. Eur J Appl Physiol 82:188–196PubMedCrossRefGoogle Scholar
  27. Stolwijk JA, Hardy JD (1966) Partitional calorimetric studies of responses of man to thermal transients. J Appl Physiol 21:967–977PubMedGoogle Scholar
  28. Tucker R, Rauch L, Harley YXR, Noakes TD (2004) Impaired exercise performance in the heat is associated with an anticipatory reduction in skeletal muscle recruitment. Pflugers Arch Eur J Physiol 448:422–430CrossRefGoogle Scholar
  29. Tucker R, Marle T, Lambert EV, Noakes TD (2006) The rate of heat storage mediates an anticipatory reduction in exercise intensity during cycling at a fixed rating of perceived exertion. J Physiol 574:905–915PubMedCrossRefPubMedCentralGoogle Scholar
  30. Vallerand A (1992) How should body heat storage be determined in humans: by thermometry or calorimetry? Eur J Appl Physiol O 65:286–294CrossRefGoogle Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2014

Authors and Affiliations

  • Nicholas M. Ravanelli
    • 1
  • Matthew N. Cramer
    • 1
  • Yannick Molgat-Seon
    • 1
    • 2
  • Anthony N. Carlsen
    • 3
  • Ollie Jay
    • 1
    • 4
  1. 1.Thermal Ergonomics Laboratory, School of Human KineticsUniversity of OttawaOttawaCanada
  2. 2.Health and Integrative Physiology LaboratorySchool of Kinesiology, University of British ColumbiaVancouverCanada
  3. 3.Sensorimotor Neuroscience Laboratory, School of Human KineticsUniversity of OttawaOttawaCanada
  4. 4.Exercise and Sports Science, Faculty of Health SciencesUniversity of SydneyLidcombeAustralia

Personalised recommendations