Does attenuated skin blood flow lower sweat rate and the critical environmental limit for heat balance during severe heat exposure?

Abstract

New Findings

• What is the central question of this study?

  Does attenuated skin blood flow diminish sweating and reduce the critical environmental limit for heat balance, which indicates maximal heat loss potential, during severe heat stress?

• What is the main finding and its importance?

  Isosmotic hypovolaemia attenuated skin blood flow by ∼20% but did not result in different sweating rates, mean skin temperatures or critical environmental limits for heat balance compared with control and volume-infusion treatments, suggesting that the lower levels of skin blood flow commonly observed in aged and diseased populations may not diminish maximal whole-body heat dissipation.

Attenuated skin blood flow (SkBF) is often assumed to impair core temperature (T_c) regulation. Profound pharmacologically induced reductions in SkBF (∼85%) lead to impaired sweating, but whether the smaller attenuations in SkBF (∼20%) more often associated with ageing and certain diseases lead to decrements in sweating and maximal heat loss potential is unknown. Seven healthy men (28 ± 4 years old) completed a 30 min equilibration period at 41°C and a vapour pressure (P_a) of 2.57 kPa followed by incremental steps in P_a of 0.17 kPa every 6 min to 5.95 kPa. Differences in heat loss potential were assessed by identifying the critical vapour pressure (P_crit) at which an upward inflection in T_c occurred. The following three separate treatments elicited changes in plasma volume to achieve three distinct levels of SkBF: control (CON); diuretic-induced isosmotic dehydration to lower SkBF (DEH); and continuous saline infusion to maintain SkBF (SAL). The T_c, mean skin temperature (T_sk), heart rate, mean laser-Doppler flux (forearm and thigh; LDF_mean), mean local sweat rate (forearm and thigh; LSR_mean) and metabolic rate were measured. In DEH, a 14.2 ± 5.7% lower plasma volume resulted in a ∼20% lower LDF_mean in perfusion units (PU) (DEH, 139 ± 23 PU; CON, 176 ± 22 PU; and SAL, 186 ± 22 PU; P = 0.034). However, LSR_mean and whole-body sweat losses were unaffected by treatment throughout (P > 0.482). The P_crit for T_c was similar between treatments (CON, 5.05 ± 0.30 kPa; DEH, 4.93 ± 0.16 kPa; and SAL, 5.12 ± 0.10 kPa; P = 0.166). Furthermore, no differences were observed in the skin-air temperature gradient, metabolic rate or changes in T_c (P > 0.197). In conclusion, a ∼20% reduction in SkBF alters neither sweat rate nor the upper limit for heat loss from the skin during non-encapsulated passive heat stress.

Introduction

In response to rising core and/or skin temperatures, reflex and locally mediated mechanisms initiate cutaneous vasodilatation, leading to greater skin blood flow (SkBF). As internal heat flux to the skin surface is governed primarily by the core-to-skin temperature gradient and SkBF, it is reasonable to suggest that physiological conditions impairing cutaneous vasodilatation and/or SkBF can diminish the rate of whole-body heat loss from the skin to the external environment (Johnson & Proppe, 1996; Charkoudian, 2010).

Fundamentally, heat exchange between the skin and the environment is driven by the physical properties of the skin–air interface; specifically, dry heat transfer is governed by the skin–air temperature gradient, whereas evaporative heat loss is dependent on the skin–air vapour pressure gradient and the area of skin sweat coverage. In fixed ambient conditions, attenuated SkBF could thereby influence heat exchange with the environment through changes in mean skin temperature (T_sk) and/or diminished sweat production. Exposure to very high ambient temperatures favours dry heat gain, but higher T_sk secondary to greater SkBF would theoretically assist core temperature regulation by reducing the negative skin–air temperature gradient. With respect to a possible relationship between SkBF and sweating, local ischaemia following arterial occlusion abolishes reflex sweating in warm and hot environments (Randall et al. 1948; Elizondo, 1973), and noradrenaline-induced reductions in SkBF attenuate the rate of local sweat production by as much as 50% for a given change in core temperature (ΔT_c) during passive heating in a water-perfused suit (Wingo et al. 2010). It must be noted, however, that the ∼85–100% reductions in SkBF achieved with noradrenaline infusion and local ischaemia are well above the ∼20% reduction in SkBF typically reported with ageing (Minson et al. 1998), disease (Carberry et al. 1992; Cui et al. 2005; Green et al. 2006; Wick et al. 2006; Sokolnicki et al. 2009) or dehydration (Nadel et al. 1980; Montain & Coyle, 1992a,b). Therefore, to the best of our knowledge, whether reductions in SkBF of a more physiologically relevant magnitude diminish sweating, and thereby maximal evaporative heat loss, have not been previously examined.

Many previous studies have examined cutaneous vasodilatory and SkBF responses during exercise with submaximal thermolytic requirements (i.e. compensable conditions) or during passive heating in encapsulated conditions, such as a water-perfused garment, which yields an extremely high T_sk under the suit while local skin temperature at the SkBF measurement site is held constant. However, in order to assess the upper limit for heat loss from the skin surface, a non-encapsulated approach that identifies the transition from compensability to uncompensability must be adopted, such as an incremental humidity protocol consisting of short-duration steps of increasing ambient vapour pressure at a constant air temperature and metabolic rate. Above a critical ambient vapour pressure (P_crit), an upward inflection in core temperature is observed, indicating that the thermolytic requirements (a combination of metabolic heat production and any dry heat gain from the environment) have exceeded the maximal rate of heat loss that can be attained physiologically via evaporation. Values for P_crit can then be compared between different conditions in which SkBF levels are manipulated to assess whether skin surface heat loss is meaningfully altered. Historically, this approach has been used to define critical environmental limits between populations (Kamon et al. 1978; Kenney & Zeman, 2002; Dougherty et al. 2010), at various air velocities and ambient temperatures (Belding & Kamon, 1973; Kamon & Avellini, 1979; Ravanelli et al. 2015) and with different clothing ensembles (Belding & Kamon, 1973; Kamon et al. 1978; Kenney et al. 1993), and may be especially useful to assess whether physiological differences in SkBF are sufficient to alter maximal heat loss potential in non-encapsulated heat stress conditions.

The present investigation sought to determine whether physiologically relevant reductions in SkBF attenuate sweating rates and whole-body heat loss to the extent that the P_crit for core temperature is ultimately lowered. The following hypotheses were made: (i) diminished SkBF would result in lower sweat rates and a lower P_crit for core temperature inflection; and (ii) the maintenance of a higher SkBF would preserve sweat rates and the P_crit for core temperature inflection.

Methods

Subjects

Seven healthy young men completed the study (age, 28 ± 4 years; body mass, 82.3 ± 17.2 kg; height, 1.76 ± 0.07 m; body surface area, 1.98 ± 0.23 m²). Based on previously reported data (Ravanelli et al. 2015) and using conventional β (0.20) and α (0.05) values, a sample size of seven subjects was estimated to be sufficient to detect a significant difference in the P_crit for core temperature (G^*power version 3.1.9.2, Universität Düsseldorf, Düsseldorf, Germany). Subjects were non-smokers; reported no history of cardiovascular, metabolic, respiratory or neurological disease; were not taking any medications; and were not acclimatized to the heat. The protocol was explained to each subject before obtaining written informed consent, and a complete medical history was obtained prior to testing. The experimental protocol was approved by the University of Ottawa Health Sciences and Science Research Ethics Board and the Institutional Review Boards of the University of Texas Southwestern Medical Center and Texas Health Presbyterian Hospital Dallas. All procedures conformed to the standards set forth by the Declaration of Helsinki.

Instrumentation and measurements

To assess SkBF, cutaneous red blood cell flux was measured using laser-Doppler velocimetry (Moor Instruments Ltd, Axminster, UK) on the dorsal forearm and the mid-point of the anterior thigh. Skin blood flow data are reported as the mean of both sites as absolute laser-Doppler flux in arbitrary perfusion units (LDF_mean), as a percentage of maximal laser-Doppler flux (%LDF_max), as cutaneous vascular conductance (CVC_mean), which was calculated as the quotient of LDF_mean and mean arterial pressure (MAP), and as the percentage of the maximal cutaneous vascular conductance (%CVC_max). Additionally, forearm blood flow (FBF) was measured via venous occlusion plethysmography (Whitney, 1953) with an indium–gallium strain gauge (Hokanson, Bellevue, WA, USA) every 12 min during the first 72 min of the protocol and every 6 min thereafter. Forearm vascular conductance (FVC) was subsequently calculated as the quotient of FBF and MAP. Maximal laser-Doppler flux was assessed by heating each measurement site for ∼45 min by setting the local heater to 44°C. Arterial blood pressures were measured with an electrosphygmomanometer (Tango; SunTech Medical Instruments, Inc., Raleigh, NC, USA), with MAP values calculated as (pulse pressure/3) + diastolic blood pressure. Heart rate was taken from an ECG (Solar 8000M; GE Medical Systems, Madison, WI, USA). Venous blood samples (4 ml) were collected from a forearm vein in lithium heparin tubes (BD Vacutainer, Franklin Lakes, NJ, USA) and analysed in triplicate for haemoglobin concentration (Hemoximeter, OSM3; Radiometer, Copenhagen, Denmark), haematocrit and plasma osmolality (Micro-Osmometer, Model 3MO plus; Advanced Instruments Inc., Norwood, MA, USA). Changes in blood, cell and plasma volumes were estimated with the equations of Dill & Costill (1974).

Oesophageal temperature was measured with a general-purpose thermocouple probe (Mon-a-therm; Mallinckrodt Medical, St Louis, MO, USA) inserted to a maximal depth of 40 cm, which is a level estimated to be adjacent to the left ventricle and aorta (Mekjavic & Rempel, 1990). Skin temperatures were measured on the chest, shoulder, abdomen, lower back, thigh and calf using thermocouples secured to the skin with soft cloth medical tape (Medipore, 3M), and T_sk was calculated as a weighted average of these six sites (Taylor et al. 1989). Oesophageal and skin temperatures were measured with thermocouple readers (Sable Systems International, Las Vegas, NV, USA). Gastrointestinal temperature was also measured in six subjects using a telemetric pill (HQ Inc., Palmetto, FL, USA) ingested ∼4 h prior to baseline data collection.

Local sweat rates were measured using two 4.1 cm² ventilated capsules placed adjacent to each laser-Doppler probe. Nitrogen gas was supplied to each capsule at a flow rate of 300 ml min^–1, and the effluent air was analysed for vapour concentration using capacitance hygrometry (Vaisala, Vantaa, Finland). Local sweat rate for each site was calculated as the product of the vapour concentration and flow rate relative to the skin surface area covered by a sweat capsule (in milligrams per square centimetre per minute). Local sweat rate (LSR_mean) values are reported as the average of the two sites.

Expired gases were analysed with a metabolic cart (TrueOne 2400; Parvo Medics, Sandy, UT, USA). The rate of oxygen uptake () and the respiratory exchange ratio (RER) were used to determine metabolic rate (M) via indirect calorimetry, as follows:

Where e_c and e_f are the energetic equivalents for carbohydrate [21.12 kJ (l O₂)^–1] and fat [19.61 kJ (l O₂)^–1], respectively. Dry heat exchange was calculated as the sum of convection (C) and radiation (R) using the following equations, which assumed negligible clothing insulation:

The value of the convective heat transfer coefficient (h_c) was taken as 3.1 W m^–2 K^–1 because the air velocity inside the thermal chamber was measured to be <0.2 m s^–1 (Mitchell, 1974). Ambient temperature (T_a) and mean radiant temperature (T_r) were assumed to be equal. Ambient temperatures were measured in degrees Celsius, but converted to degrees kelvin to calculate dry heat transfer coefficients. The radiant heat transfer coefficient (h_r) was calculated as follows:

Here, ε is the emissivity of the skin, assumed to be 0.95 (no dimensions); σ is the Stefan–Boltzmann constant (5.67  ×10⁻⁸ W m⁻² K⁻⁴); and A_rA_D⁻¹ is the effective radiant surface area (non-dimensional), which was taken as 0.7 (Guibert & Taylor, 1952; Kerslake, 1972).

Experimental protocol

In a randomized, counterbalanced order separated by at least 3 days, subjects performed three experimental trials, as follows: in control conditions (CON), in which no treatment was administered; a diuretic-induced isosmotic dehydration trial (DEH) with the goal of lowering SkBF relative to CON; and a continuous saline infusion trial (SAL) with the goal of elevating SkBF relative to CON. Prior to experimentation, subjects were instructed to avoid cold, allergy and anti-inflammatory medicines for 36 h, alcohol consumption and exercise for 24 h and caffeine intake for 12 h. They were also asked to consume plenty of water the night before and the morning of each visit to ensure adequate hydration and to eat breakfast before leaving for the laboratory.

Subjects arrived at the laboratory between 08.00 and 09.00 h. Nude body mass, standing height and urine specific gravity (USG) were first recorded, with euhydration accepted at a USG ≤ 1.025 (Kenefick & Cheuvront, 2012). A forearm venous catheter was then inserted, and after 30 min of supine rest an initial blood sample was drawn. In DEH, 40–80 mg of furosemide was then administered orally, after which the subjects rested for ∼3.5 h to allow the drug to take effect (Ikegawa et al. 2011). Furosemide was selected to induce hypovolaemia without any change in plasma osmolality, which can independently affect sweating responses (Fortney et al. 1984). In CON and SAL, subjects rested for an equivalent duration to ensure that heating commenced at a similar time of day during each visit. Drinking water was provided ad libitum during the initial rest period in CON and SAL only, but a small volume of fluid was ingested in DEH to facilitate insertion of the oesophageal temperature probe during instrumentation. Urine output was also recorded during this period. Nude body mass was measured again at the end of the rest period.

Following instrumentation, subjects entered the chamber. The experimental protocol began with a 30 min baseline equilibration period at 41°C, 33% relative humidity (RH; 2.57 kPa), during which steady-state sweat rates and core temperature were attained. Afterwards, an incremental humidity protocol began, with P_a increased by 0.17 kPa every 6 min from 2.57 to 5.95 kPa (maximum of 20 stages or 120 min). Throughout baseline and the incremental humidity protocol, T_a was controlled at 41.0 ± 0.1°C. This T_a was selected to elicit skin temperatures similar to those achieved in encapsulated protocols (e.g. water-perfused suit studies) and induce high sweating rates that would facilitate examination of a link between SkBF and LSR. The pattern of change in P_a over time was nearly identical between trials, resulting in steps of 0.17 ± 0.01 kPa per 6 min stage. Termination criteria were as follows: (i) completion of the 20th stage; (ii) reaching a 1.0°C change (Δ) in T_c from baseline; (iii) hypotension (systolic blood pressure <80 mmHg); or (iv) voluntary withdrawal because of excessive discomfort. Trials were conducted in an upright seated position (n = 4) on a steel-framed mesh chair (height, 83 cm) or in a supine position (n = 3) on a hospital bed. Experimentation was conducted initially in the upright seated posture, but because of episodes of hypotension in some DEH trials, testing was completed in a supine posture by the remaining three subjects to minimize the possibility of syncope. Each subject conducted his trials in the same posture. During SAL, 0.9% saline was infused continuously at a rate of 0.05 ml kg⁻¹ min⁻¹ throughout the incremental humidity trial (total, 479 ± 105 ml). Core and skin temperatures, local sweat rates, skin blood flow and heart rate were measured throughout. Blood pressure was measured at the end of each stage; expired gases were sampled at baseline and every four stages. Upon completion of the incremental humidity protocol, a final blood sample was drawn and the thermal chamber was cooled. Local heating of the laser-Doppler measurement sites was performed for ∼45 min to induce maximal LDF values. A final nude body mass measurement was then performed.

Data analysis

Continually measured variables were sampled at 25 Hz (MP150; Biopac Systems, Inc., Santa Barbara, CA, USA). For each 6 min stage, data were averaged over the final 2 min. In accordance with recent work, these values were subsequently used to define P_crit for core temperature and heart rate using segmental regression analysis (Ravanelli et al. 2015). This latter statistical approach is based on the biphasic core temperature response during an incremental humidity protocol; the first phase is a baseline phase, during which the slope of the relationship between core temperature and P_a is approximately zero, and an ‘inflected’ phase, during which core temperature rises abruptly and linearly from its baseline level as the heat loss requirement has exceeded heat loss potential owing to increasing P_a. The segmental regression analysis identifies the P_a value at the intersection between these two slopes, which is then accepted as P_crit (see Fig. 1). The core temperature P_crit was determined from oesophageal temperature in six subjects and from gastrointestinal temperature in one subject.

Figure 1. Representative core temperature data from the incremental humidity protocol, demonstrating the determination of critical ambient vapour pressure (P_crit)

After 30 min baseline at 41°C with 2.57 kPa (30% relative humidity), ambient vapour pressure was elevated by 0.17 kPa (∼2% relative humidity) every 6 min while air temperature remained fixed. Core temperature stayed relatively stable despite the rising ambient vapour pressure. At P_crit, core temperature inflects and rises rapidly. The value of P_crit is identified as the intersection of the two slopes (slope 1, pre-inflection; and slope 2, postinflection), determined via segmental regression analysis.

Statistical analysis

Data are presented as means ± SD. A one-way repeated-measured ANOVA, using the independent factor of treatment (three levels: CON, DEH and SAL), was used to compare USG, fluid intake and urine output during rest, whole-body sweat loss, P_crit values and metabolic rate. A two-way repeated-measures ANOVA, with the independent factors of treatment (three levels: CON, DEH and SAL) and time, was also used to compare changes in blood and plasma volumes and plasma osmolality (four levels: initial, baseline, at 60 min and at the end of the protocol); LDF_mean, %LDF_max, CVC_mean, %CVC_max, LSR_mean, T_sk, ΔT_c, T_sk–T_a and dry heat exchange (15 levels: baseline and every 6 min stage to ∼5.0 kPa or 84 min of the protocol); as well as FBF and FVC (eight levels: baseline and every two stages to ∼4.8 kPa or 78 min of the protocol). A Greenhouse–Geisser correction was applied if the assumption of sphericity had been violated. In the event of a significant time-by-treatment interaction, post hoc analysis was performed using Tukey's range test for multiple comparisons. All remaining statistical analyses were performed with Prism 6 for Windows (GraphPad Software, La Jolla, CA, USA). The α value was set at the 0.05 level.

Results

Initial rest period

Initial USG values upon arrival were not different between trials (CON, 1.014 ± 0.007; DEH, 1.012 ± 0.007; and SAL, 1.013 ± 0.006; P = 0.633). Ad libitum fluid intake in CON (613 ± 239 ml) and SAL (614 ± 184 ml) was greater than in DEH (253 ± 109 ml; P = 0.002). Diuretic ingestion resulted in greater urine output in DEH (1997 ± 633 ml) compared with CON (566 ± 361 ml) and SAL (878 ± 507 ml; P < 0.001). Consequently, blood volume declined to a greater extent during this time in DEH compared with CON and SAL because of a contraction of plasma volume but not cell volume (P < 0.001; Table 1). Fluid intake, urine output, changes in blood, cell and plasma volumes and changes in body mass were not different between CON and SAL during the initial rest period (P > 0.168).

Table 1. Changes in blood volumes, plasma osmolality and body mass throughout experimentation

Conditions	Paramter	Rest	Baseline	60  min	End
CON	Δ Blood volume (%)	—	−1.7 ± 1.3a	−2.3 ± 1.8ab	−6.1 ± 2.9ab
	Δ Cell volume (%)	—	0.5 ± 2.2	−0.2 ± 2.0	−0.7 ± 3.4
	Δ Plasma volume (%)	—	−3.4 ± 2.0a	−4.1 ± 3.0ab	−10.4 ± 3.4ab
	Plasma osmolality (mosmol kg^−1)	296 ± 5	293 ± 5	295 ± 3	297 ± 4
	Δ Body mass (kg)	—	−0.3 ± 0.5a	—	−1.1 ± 0.3
	Δ Body mass (%)	—	−0.4 ± 0.6a	—	−1.4 ± 0.4
DEH	Δ Blood volume (%)	—	−8.1 ± 2.5ac	−9.0 ± 3.5ac	−11.8 ± 2.8ac
	Δ Cell volume (%)	—	−0.5 ± 2.1	−1.3 ± 1.5	−1.8 ± 2.1
	Δ Plasma volume (%)	—	−14.2 ± 5.7ac	−15.3 ± 6.8ac	−19.7 ± 5.1ac
	Plasma osmolality (mosmol kg^−1)	293 ± 5	293 ± 4	296 ± 7	298  ± 5
	Δ Body mass (kg)	—	−2.0 ± 0.6ac	—	−0.9 ± 0.4
	Δ Body mass (%)	—	−2.4 ± 0.7ac	—	−1.1 ± 0.6
SAL	Δ Blood volume (%)	—	−0.7 ± 2.6c	1.2 ± 3.4bc	−1.6 ± 3.1bc
	Δ Cell volume (%)	—	0.6 ± 1.9	0.8 ± 2.2	−0.3 ± 1.7
	Δ Plasma volume (%)	—	−1.6 ± 5.0c	1.5 ± 6.8bc	−2.7 ± 5.3bc
	Plasma osmolality (mosmol kg^−1)	294 ± 4	290 ± 2	294 ± 5	296 ± 5
	Δ Body mass (kg)	—	−0.4 ± 0.4c	—	−1.1 ± 0.3
	Δ Body mass (%)	—	−0.5 ± 0.5c	—	−1.3 ± 0.4

• Data are means ± SD for seven subjects. Abbreviations: CON, control conditions without treatment; DEH, diuretic-induced dehydration; and SAL, continuous saline infusion. Note that changes (Δ) in blood, cell and plasma volumes are expressed as percentage changes from rest; changes in body mass are expressed as absolute and percentage changes from rest to baseline (i.e. from arrival until the start of the incremental humidity protocol, ‘Baseline’) and from baseline to the end of the incremental humidity protocol (‘End’). Superscript letters a, b and c indicate differences between CON and DEH, between CON and SAL and between DEH and SAL, respectively (P < 0.05).

Incremental humidity protocol

The greater reduction in blood and plasma volumes during DEH persisted throughout the incremental humidity protocol (P < 0.001), but no differences in plasma osmolality were observed between treatments (P = 0.546; Table 1). This greater hypovolaemia in DEH mitigated the increase in laser-Doppler SkBF during the protocol (Fig. 2). Significant main effects of treatment were found for LDF_mean (P = 0.034), %LDF_max (P = 0.003) and %CVC_max (P = 0.016), while CVC_mean tended to show an effect of treatment (P = 0.056). Treatment-related effects resulted in an ∼20% difference in each laser-Doppler index of SkBF between DEH and CON, as well as between DEH and SAL (Fig. 2), while no differences were observed between CON and SAL treatments. A similar pattern of change in all indices of SkBF were observed between forearm and thigh sites. No effect of treatment was evident for FBF and FVC (P ≥ 0.769).

Figure 2. Laser-Doppler-derived skin blood flow responses in control conditions without treatment (CON), after diuretic-induced dehydration (DEH) and with continuous saline infusion (SAL)

The following four indices of skin blood flow are presented: mean absolute laser-Doppler flux (LDF_mean); mean laser-Doppler flux expressed as a percentage of maximal values (%LDF_max); mean cutaneous vascular conductance (CVC_mean); and mean cutaneous vascular conductance expressed as a percentage of maximal values (%CVC_max). Analyses were performed for seven subjects up to ∼5 kPa (84 min), after which early termination criteria were met at different times. A significant main effect of treatment was observed for LDF_mean, %LDF_max and %CVC_max (P < 0.05). A trend towards an effect of treatment on CVC_mean was also observed.

Despite the attenuation of SkBF, no treatment effect was observed for LSR_mean (P = 0.279; Fig. 3), while changes in body mass, indicating whole-body sweat loss, were not different between conditions (P = 0.482; Table 1). Likewise, no effects of treatment on T_sk (Fig. 3), the T_sk–T_a thermal gradient, and the calculated rate of dry heat exchange (Fig. 4) were found between trials (P > 0.193). Moreover, metabolic rate was not different between treatments, averaging 119 ± 13, 115 ± 15 and 109 ± 18 W in CON, DEH and SAL, respectively (P = 0.121). Consequently, P_crit for core temperature was not different across treatments (P = 0.166; Fig. 5). A similar rate of increase in ΔT_c was also observed postinflection between trials (CON, 0.020 ± 0.003°C min⁻¹; DEH, 0.019 ± 0.004°C min⁻¹; and SAL, 0.018 ± 0.002°C min⁻¹; P = 0.353).

Figure 3. Local sweating and skin temperature responses throughout the incremental humidity protocol in control conditions without treatment (CON), after diuretic-induced dehydration (DEH) and with continuous saline infusion (SAL)

Abbreviations: LSR_mean, mean local sweat rate (two sites: forearm and thigh); P_a, ambient vapour pressure; and T_sk, mean skin temperature. Analysis was performed up to ∼5 kPa (84 min) in seven subjects. Beyond this point, early termination criteria were met at different times. No treatment effect was observed for LSR_mean and T_sk.

Figure 4. The skin-to-air temperature gradient (T_sk–T_a) and the corresponding rate of dry heat exchange throughout the incremental humidity protocol in control conditions without treatment (CON), after diuretic-induced dehydration (DEH) and with continuous saline infusion (SAL)

Abbreviations: P_a, ambient vapour pressure; T_a, ambient temperature; and T_sk, mean skin temperature. Data are for seven subjects. Note that negative values for dry heat exchange indicate dry heat gain. No effect of treatment was observed for T_sk–T_a and dry heat exchange (P > 0.193).

Figure 5. Critical ambient vapour pressures (P_crit) for core temperature in control conditions without treatment (CON), after diuretic-induced dehydration (DEH) and with continuous saline infusion (SAL) for seven subjects

No effect of treatment on P_crit for core temperature was observed (P = 0.166).

Baseline MAP was not different between treatments (CON, 87 ± 7 mmHg; DEH, 84 ± 7 mmHg; and SAL, 84 ± 10 mmHg; P = 0.740), and no differences in ΔMAP were observed during the protocol (Fig. 6; P = 0.189). Volume depletion caused a significantly higher heart rate throughout DEH compared with CON and SAL (P = 0.009; Fig. 6), amounting to a consistent difference of 8 ± 2 beats min⁻¹ in CON versus DEH and 11 ± 2 beats min⁻¹ in SAL versus DEH.

Figure 6. Mean arterial pressure (MAP) and heart rate responses during the incremental humidity protocol in control conditions without treatment (CON), after diuretic-induced dehydration (DEH) and with continuous saline infusion (SAL)

Abbreviations: ΔMAP, change in mean arterial pressure from baseline; and P_a, ambient vapour pressure. Analyses were performed for seven subjects up to ∼5 kPa (84 min), after which early termination criteria were met at different times. A significant main effect of treatment was observed for heart rate but not ΔMAP.

Discussion

In the present study, the influence of physiologically relevant differences in SkBF on sweating and skin surface heat loss potential were assessed by using an incremental humidity protocol that enabled us to identify the P_crit above which T_c could no longer be compensated; that is, above P_crit, rates of metabolic heat production and dry heat gained from a 41°C environment could no longer be balanced by the total rate of evaporative heat loss from the skin. Using an isosmotic volume-depletion model, a ∼20% reduction in SkBF had no effect on sweating (Fig. 4) and, ultimately, the P_crit for core temperature (Fig. 6) compared with control and volume-infusion treatments, suggesting that sudomotor output and the upper limit of heat dissipation from the skin is not sufficiently compromised by an attenuated SkBF within the parameters presently used.

Changes in body mass (−2.4%) and plasma volume (−14.0%) after oral furosemide administration but prior to heat exposure were consistent with those reported previously (Ikegawa et al. 2011; Romero et al. 2011). It is particularly noteworthy that the subsequent reduction in SkBF (Fig. 3) was similar to that observed with exercise- (Montain & Coyle, 1992a) and diuretic-induced (Ikegawa et al. 2011) hypohydration compared with euhydration. Despite the lower SkBF in DEH (Fig. 3), the values for LSR (Fig. 4), whole-body sweat rate and P_crit (Fig. 6) were similar relative to CON and SAL, suggesting that maximal heat dissipation was not affected by lower SkBF. At a constant metabolic rate and ambient temperature, a difference in heat loss potential (i.e. P_crit) secondary to altered SkBF would theoretically arise from changes in the physical properties of the skin surface (T_sk, maximal skin wettedness and/or evaporative efficiency) that alter heat exchange with the environment. Specifically, if attenuated SkBF caused a lower T_sk, a wider skin–air temperature gradient would follow, and the rate of dry heat gain (convection and radiation) would rise. Likewise, if SkBF compromised sweat production and/or efficiency, the rate of evaporative heat loss would be reduced. In the present study, given that ambient temperature was constant (41°C) and T_sk was not different between treatments, no differences in the skin–air temperature gradient and the rate of dry heat exchange were observed (Fig. 5). Furthermore, it stands to reason that evaporative heat loss was similar between conditions given the absence of any differences in local and whole-body sweating responses, T_sk and the P_crit for core temperature (Figs 4-6). Although a functional relationship between SkBF and sweating has been reported previously (Wingo et al. 2010), the supraphysiological reduction in local SkBF induced pharmacologically (noradrenaline infusion) in the study by Wingo et al. (2010) cannot be compared with the whole-body observations reported in the present study. Importantly, we show that physiologically relevant reductions in SkBF, such as those that occur with ageing (Kenney et al. 1997; Minson et al. 1998) and disease (Kenney & Kamon, 1984; Green et al. 2006), are unlikely to affect whole-body heat loss potential.

Although this is the first study to investigate a direct link between altered SkBF and the potential for skin surface heat loss, previous data suggest that high levels of SkBF may not be advantageous for core temperature regulation. For example, a 15% reduction in FVC did not alter local sweat rates, T_sk or ΔT_c in hypoxic versus normoxic conditions (Miyagawa et al. 2011) or following a 2.6% reduction in body mass with diuretic-induced hypohydration (Ikegawa et al. 2011) during exercise. Additionally, chronic heart failure patients with a 20% lower CVC demonstrated similar ΔT_c and T_sk (sweat rate was not measured) compared with healthy control subjects during non-encapsulated passive heat stress (38°C, 50% RH; Green et al. 2006). Likewise, a 40% reduction in FBF in older versus younger individuals resulted in similar ΔT_c, T_sk and sweat rates during exercise in the heat despite similarities in the rate of metabolic heat production, body size and aerobic fitness between groups (Kenney, 1988). Although this has not been a consistent finding in aged versus younger groups (Kenney et al. 1997), it is important to note that elderly individuals also demonstrate attenuated sudomotor responses and rates of evaporative heat loss, which perhaps contribute to greater ΔT_c independently of any parallel reductions in SkBF (Anderson & Kenney, 1987; Larose et al. 2013). Collectively, current and previous data suggest that core temperature in elderly individuals is regulated in a similar manner despite a ≥15% lower SkBF, provided that sweating and evaporative heat loss are not physiologically limited. It follows that populations demonstrating concurrent vasodilatory and sudomotor impairments, such as the elderly (Minson et al. 1998), type 2 diabetics (Wick et al. 2006; Sokolnicki et al. 2009), skin graft recipients (Davis et al. 2007, 2009) and spinal cord injury victims (Freund et al. 1984), may be at greater risk of thermal injury during severe heat stress because of problems associated with sudomotor function and/or evaporation, and perhaps not because of impairments of skin blood flow per se.

In many previous studies, physiological regulation of cutaneous vasodilatation and SkBF has been examined during passive heating to a fixed elevation in core temperature using a water-perfused suit (e.g. Freund et al. 1984; Minson et al. 1998; Wick et al. 2006; Davis et al. 2007, 2010), which controls skin temperature under the suit at a set level (∼38°C) and simultaneously impedes evaporative heat loss over most of the skin surface. Given the mostly encapsulated nature of water-perfusion suit studies, the influence of SkBF on sweating and whole-body heat loss cannot be established with this experimental approach. Likewise, studies conducted during exercise (Montain & Coyle, 1992a,b) or non-encapsulated passive heat stress (Green et al. 2006) with submaximal heat loss requirements would not necessarily observe any effect on core temperature because any reductions in skin surface heat loss attributable to lower SkBF are likely to be physiologically compensated by increased sweat production and evaporation. In contrast, the incremental humidity protocol used in the present study permits the evaluation of whole-body heat loss potential in non-encapsulated conditions of high environmental heat stress. The basis for this approach lies in examining how various factors or interventions affect the heat loss side of the conceptual heat balance equation. Given that the rate of heat storage is equal to the difference between rates of heat production and loss, P_crit represents the threshold ambient vapour pressure above which thermal balance (i.e. a rate of heat storage equal to zero) cannot be maintained. Therefore, a shift in P_crit at a fixed metabolic heat production and ambient temperature, attributed to the factor or intervention in question, directly indicates a change in maximal whole-body skin surface heat loss. The present experiment represents a novel use of this classic protocol (Belding & Kamon, 1973).

Although the present iteration of the incremental humidity protocol used similar stepwise increases in P_a (∼0.17 kPa per stage) and stage durations (6 min) as in previous studies (Belding & Kamon, 1973; Kenney et al. 1993; Ravanelli et al. 2015), it could be argued that the present approach might lack the sensitivity required to detect differences in P_crit secondary to a reduction in SkBF. One way to demonstrate the sensitivity of the present protocol is through the effect of stepwise increases in P_a on the skin wettedness required for heat balance (Gagge, 1937). Theoretically, skin wettedness reflects the fraction of the skin surface that must be covered with sweat to achieve a certain rate of evaporative heat loss, while maximal skin wettedness refers to the value of skin wettedness at the threshold for compensability (i.e. P_crit). Presently, stepwise increases in P_a of 0.17 kPa corresponded to average elevations in skin wettedness of 0.05 ± 0.02 (range, 0.02–0.10). Given that interventions (e.g. heat acclimatization) alter maximal skin wettedness by ∼0.15, a possible shift in the maximal skin wettedness by <0.05 (and thus P_crit) by an attenuated SkBF response would have been trivial. Furthermore, a lack of any effect of lowered SkBF on P_crit is supported by a similar rate of change in core temperature above P_crit. If a true shift in P_crit to a lower ambient vapour pressure attributable to lowered SkBF was not detected using the present analysis, a higher rate of change in core temperature would have been expected above P_crit in the DEH trial; however, this was not the case.

Continuous or rapid saline infusion has been an effective approach to elevate SkBF above non-infusion levels, resulting in a higher FBF during exercise (Nose et al. 1990). The lack of any effect of saline infusion (0.05 ml kg^–1 min^–1) on indices of SkBF, LSR_mean, T_sk and, ultimately, ΔT_c (Figs 2 and 5) in the present study is perhaps explained by the relatively modest infusion rate compared with previous studies. Although a higher infusion rate could have shifted P_crit to a higher value by affecting SkBF and sweating, this possibility requires further investigation.

Estimates of SkBF are typically expressed in a variety of ways, including absolute LDF or FBF, CVC or FVC values that are normalized to arterial blood pressure, or as a percentage of maximal LDF or CVC. While indices of SkBF expressed relative to blood pressure or a percentage of maximal flux indicate aspects of vasomotor control, it must be recognized that heat exchange between cutaneous blood and skin tissue will be dependent on the absolute level of blood flow through cutaneous vessels, as well as the blood-to-skin thermal gradient, not absolute CVC or CVC normalized to maximal values. Absolute SkBF was represented herein by the LDF_mean value, which was consistently lower in DEH, while %LDF_max, CVC_mean and %CVC_max were also included given the conventional use of these parameters (Fig. 2). Regardless of how it was expressed, interpretation of the SkBF data was similar. Absolute SkBF was also estimated using venous occlusion plethysmography to determine FBF in the present study; however, unlike previous investigations using a volume-depletion experimental model (Nadel et al. 1980; Fortney et al. 1983), no treatment effect was detected. Previously, Johnson et al. (1984) demonstrated good agreement between LDF and FBF values measured during passive heating. However, in that study the experimental conditions were very different from those of the present investigation; specifically, LDF and FBF measurements would have been performed on an arm exposed to the ambient air, with skin much cooler than skin under the heating garment (Johnson et al. 1984). As such, LDF and FBF values would have represented reflex drive for vasodilatation only. Although the severity of the heat stress imposed in the present study would have been likely to induce reflex and locally mediated cutaneous vasodilatation, there is no conspicuous reason for the present discrepancy between laser-Doppler and plethysmographic measurements, unless alterations in cutaneous blood flow were balanced by changes in blood flow through underlying tissues. Nevertheless, the consistent attenuation of LDF_mean observed among all subjects during DEH in the present study (Fig. 2) suggests that blood flow in cutaneous vascular beds was consistently altered as expected.

Limitations and future research

Owing to instances of hypotension following volume depletion, body posture was adjusted from an upright seated position to a supine posture (note that seven total subjects were tested, with four seated in the upright position and three supine). Although postural differences with heat stress influence baroreflex control of SkBF (Lind et al. 1968), it should be noted that: (i) all analyses were performed within subject, with the position of a given subject consistent between all three trials; (ii) perturbations that unload baroreceptors do not appear to alter thermoregulatory sweat rate (Wilson et al. 2005; Kenny et al. 2010; Lynn et al. 2012); (iii) comparisons within posture still revealed no effect of treatment on P_crit for core temperature, metabolic rate, whole-body sweat rate, LSR_mean and T_sk; and (iv) the slightly higher P_crit for core temperature in the upright seated position (CON, 5.2 ± 0.1 kPa; DEH, 5.0 ± 0.2 kPa; and SAL, 5.2 ± 0.1 kPa; n = 4) compared with the supine posture (CON, 4.9 ± 0.4 kPa; DEH, 4.9 ± 0.1 kPa; and SAL, 5.1 ± 0.1 kPa; n = 3) can be explained by the greater effective skin area from which heat loss could occur while seated.

The applicability of the present findings is limited to healthy young male subjects at one high ambient temperature (∼41°C), in resting conditions and with minimal clothing. Future studies should examine the implications of lower SkBF on physiological compensability to environmental heat stress in special populations with well-known impairments of cutaneous vasodilatation, between sexes, at different ambient temperatures, with various clothing ensembles and even during exercise at different metabolic rates.

Conclusion

The present study examined whether physiologically relevant reductions in SkBF alter sweating and subsequently lower the potential for skin surface heat loss during non-encapsulated passive whole-body heat stress. Despite a ∼20% lower SkBF achieved using a volume-depletion model, sweating and heat loss potential were not altered compared with control and continuous saline infusion treatments.

Additional information

Competing interests

None declared.

Author contributions

All experiments were conducted at the Institute for Exercise and Environmental Medicine, Texas Health Presbyterian Hospital of Dallas. M.N.C., D.G., C.G.C. and O.J. conceived and designed experiments; M.N.C. and D.G. conducted experiments; M.N.C. analysed data; M.N.C., D.G., C.G.C. and O.J. interpreted data; M.N.C. drafted the manuscript; and M.N.C., D.G., C.G.C. and O.J. edited, revised and approved the final draft of the manuscript. All authors agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. All persons designated as authors qualify for authorship, and all those who qualify for authorship are listed.

Funding

This research was supported by National Institutes of Health R01GM068865 (C.G.C.), by Department of Defense W81XWH-12-1-0152 (C.G.C.) and by a Discovery Grant (#386143-2010) from the Natural Sciences and Engineering Research Council (O.J.). M.N.C. was supported by an Excellence Scholarship and a Student Mobility Bursary from the Faculty of Graduate and Postdoctoral Studies at University of Ottawa, as well as an Ontario Graduate Scholarship (OGS). D.G. was supported by a Natural Sciences and Engineering Research Council of Canada Postdoctoral Fellowship.

Acknowledgements

This study was performed by M.N.C. in part fulfilment of the degree Doctor of Philosophy from the University of Ottawa. The authors are grateful for the participation of all volunteers; to Amy Adams, MS, Naomi Kennedy, RN, and Drs Paula Poh, Steven Romero and Zachary Schlader for assistance with data collection; and to Dr Robert Kenefick for his valuable input in the development of the current protocol.