Fan cooling after cardiovascular drift does not reverse decrements in maximal oxygen uptake during heat stress

ABSTRACT

Cardiovascular (CV) drift, the progressive increase in heart rate (HR) and decrease in stroke volume (SV) during constant rate, moderate intensity exercise, is related to reduced maximal oxygen uptake (V̇O_2max) during heat stress. Once it has already occurred, it is unknown whether the detrimental effects of CV drift on V̇O_2max can be reversed. This study tested the hypothesis that fan cooling after CV drift has occurred attenuates decrements in V̇O_2max associated with CV drift. Eight men completed a control graded exercise test (GXT) in 22°C to measure V̇O_2max. Then on separate, counterbalanced occasions, they completed one 15-min (15MIN) and two 45-min bouts (45NF and 45FAN) of cycling in 35°C, 40% RH at 60% V̇O_2max, each immediately followed by a GXT to measure V̇O_2max. For one of the 45-min trials (45FAN), fan airflow (4.5 m/s) was directed at participants beginning ~5 min before the GXT and continuing throughout the remainder of exercise. The purpose of the separate 15- and 45-min trials was to measure V̇O_2max during the same time interval that CV drift occurred. HR increased (13.8% and 11.4%) and SV decreased (14.4% and 14.1%) for 45NF and 45FAN, respectively; trials were not different (all P > 0.05). Despite a decrease in mean skin temperature of ~1°C with fan use, V̇O_2max decreased similarly between conditions (17% vs. 15% for 45NF and 45FAN, P = 0.54). Fan cooling after CV drift was insufficient to reverse the negative consequences of CV drift on V̇O_2max after prolonged exercise in a hot environment.

Abbreviations: 15MIN: 15-min trial; 45FAN: 45-min, fan trial; 45NF: 45-min, no fan trial; ANOVA: Analysis of variance; CV: Cardiovascular; GXT: Graded exercise test; HR: Heart rate; SV: Stroke volume; T̅_b: Mean body temperature; T_re: Rectal temperature; T̅_sk: Mean skin temperature; V̇O_2max: Maximal oxygen uptake

Introduction

Heat acclimated endurance athletes who remain euhydrated in a hot ambient environment experience minimal cardiovascular and thermal strain during prolonged exercise with fan airflow [1]. However, for the vast majority of other individuals – including endurance athletes without fan airflow [2] – sustained, constant rate, moderate intensity physical activity in hot conditions is accompanied by elevated cardiovascular strain, which can be indexed by cardiovascular drift (CV drift), a progressive increase in heart rate (HR) and decrease in stroke volume (SV) beginning after about 10–15 min [3]. CV drift in the heat – in both highly trained and less trained people – is associated with reduced work capacity reflected by decreased maximal oxygen uptake (V̇O_2max) [4,5].

The mechanism(s) responsible for reduced V̇O_2max after CV drift is unclear, but decreased SV – either from tachycardia [6] or peripheral displacement of central blood volume in conjunction with cutaneous vasodilation [7] – associated with CV drift during submaximal exercise that is sustained during subsequent maximal exercise has been implicated [2,4,8,9]. Another mechanism reputed to reduce V̇O_2max in the heat – that typically occurs in conjunction with CV drift – is elevated skin temperature (T_sk) relative to core temperature (T_c) [10–12]. Greater skin blood flow is required for heat loss when the T_c – T_sk gradient is narrow [10,12,13], and elevated T_sk results in elevated skin blood flow driven by direct effects of T_sk on cutaneous blood vessels and reflex thermoregulatory heat loss responses [14]. Under circumstances of sufficient motivation at high exercise intensities, muscle blood flow is preserved at the expense of skin blood flow [12,15]; but under conditions of extensive blood flow directed to the skin – like a narrowed T_c – T_sk gradient – skeletal muscle blood flow and oxygen delivery are diminished despite preservation of leg vascular conductance [16], which can diminish aerobic capacity [12].

Regardless of the mechanism explaining reduced V̇O_2max associated with CV drift, mitigation of CV drift has been shown to also mitigate reductions in V̇O_2max in the heat. For example, fan airflow (4.5 m/s) directed at participants for 30 min during submaximal exercise in 35°C resulted in T̅_sk and T_re just prior to measurement of V̇O_2max that were 0.8 °C and 0.4°C, respectively, lower than a no fan condition; the magnitude of CV drift and the magnitude of decline in V̇O_2max were one fourth and one third, respectively, that of the trial without fan airflow [2]. Other studies manipulating CV drift using exercise intensity [17], hydration [18], and ambient temperature [19] have shown similar results. All these studies have in common that CV drift was manipulated during exercise. The extent to which the effects of CV drift on V̇O_2max can be reversed, after CV drift has already occurred, is unknown. The reversibility of detrimental effects on V̇O_2max has important implications for strategies to address CV drift-related work capacity limitations in hot environments because V̇O_2max sets the upper limit of the rate of aerobic metabolism. As such, activities that require a low percentage of V̇O_2max are sustained more easily than those that require a high percentage of V̇O_2max [20]. Therefore, even though most workers and many athletes will not push themselves to maximum during a work bout, maintaining as high a V̇O_2max as possible permits performance of work at a more sustainable perceived exertion – i.e. lower %V̇O_2max because as V̇O_2max declines any given absolute level of work then represents a higher %V̇O_2max – which better sustains physical performance. In contrast, if V̇O_2max diminishes over time during work in the heat, any given level of work will be perceived as more strenuous (i.e. higher %V̇O_2max), which can reduce performance [5].

Is applying a countermeasure after CV drift has already occurred as effective in preserving V̇O_2max as taking steps to attenuate CV drift while it occurs? Research by Shaffrath et al. [21] supports the possibility of effectively reversing CV drift (and its accompanying negative consequences) by demonstrating that fan airflow administered for 5 min after 70 min of cycling at 60% V̇O_2max without fan airflow decreased HR ~11 beats/min. However, data collection only involved 1 participant, occurred in temperate conditions (24.2 ± 0.8°C, 62.2% RH), and did not involve measurement of SV or V̇O_2max.

In light of the proposed mechanisms explaining reduced V̇O_2max associated with CV drift described above, it is reasonable to posit that application of fan airflow for a brief period after CV drift has occurred in the heat could 1) increase the evaporative capacity of the environment [22–25]; 2) lower skin temperature; 3) reduce skin blood flow [10,12,13]; and thereby 4) mobilize peripherally displaced blood to partially or fully restore SV and thereby maximal cardiac output (Q̇). Accordingly, the purpose of this study was to test the hypothesis that a period of brief fan cooling, just prior to and during assessment of V̇O_2max after prolonged submaximal exercise in the heat during which CV drift has occurred, alleviates the decrease in V̇O_2max associated with CV drift.

Methods

Research design

A repeated measures design was used in which all participants were tested under all conditions. After an initial trial to determine V̇O_2max (CONTROL), 3 experimental trials were completed on separate days, and counterbalanced treatment orders were randomly assigned. Experimental trials consisted of one 15-min bout of cycling at 60% V̇O_2max (15MIN) and two 45-min bouts of cycling at 60% V̇O_2max (45NF and 45FAN). Each submaximal exercise bout was followed immediately by a graded exercise test (GXT) to measure V̇O_2max. Separate 15- and 45-min trials were necessary to measure V̇O_2max over the same time interval (between 15 and 45 min of exercise) that CV drift occurred since V̇O_2max could not be measured at both 15 min and 45 min within the same trial. For one of the 45-min trials, fan airflow (~4.5 m/s) was directed at participants beginning at 38 min and continuing throughout the remainder of submaximal exercise (until 45 min) and the subsequent GXT (45FAN). The air velocity chosen was based on previous research incorporating fan airflow 1) during exercise that attenuated CV drift [2,21], reductions in V̇O_2max [2], and time to exhaustion [23], and 2) near the end of exercise after CV drift had already occurred that effectively reduced HR [21]. Each participant was tested around the same time of day to minimize the influence of circadian rhythm on HR and T_c. The initial control GXT trial to determine V̇O_2max took place in a temperate environment (~22°C, 40% RH) while the 3 experimental (15- and 45-min) trials took place in a hot environment (35°C, 40% RH). All trials were separated by at least one day but not more than 1 week.

Participants

Eight healthy, active men (mean ± SD; age = 23 ± 5 y, height = 177.9 ± 3.9 cm, weight = 77.3 ± 5.3 kg, percent body fat estimated from 3-site skinfolds = 13.9 ± 3.1%) volunteered to participate. This sample size was sufficient to detect a moderate effect [d = 0.5, as described in Potvin and Schutz [26]] among the measures of V̇O_2max (CONTROL, 15MIN, 45NF, 45FAN; 1-way ANOVA) and for an interaction effect for CV drift (2-way ANOVA with factors for treatment and time), assuming α = 0.05 and power ≈ 0.8 [26,27]. Participants were recreationally active, operationally defined as performing aerobic exercise 3 or more days per week for 30 min or more each session, and they were free of any symptoms of or known cardiovascular, metabolic, or respiratory disease as determined by health history questionnaire. Each participant provided written informed consent prior to participating and all study procedures were approved by the Institutional Review Board of the University of Alabama.

Procedures

Control V̇O_2max and practice session

For the first trial, participants were instructed to avoid eating 2 hours before arrival, to maintain adequate hydration, and avoid consuming alcohol, caffeine, and non-prescription drugs the day before and the day of testing. Upon arrival, they completed a 24-h history form to verify adherence to pre-test instructions, medical and physical activity questionnaires, and they provided written informed consent for study procedures. Next, height and weight were measured and body fat estimated from the sum of 3 skinfolds [28]. Then they completed a 10-min warm-up on a cycle ergometer (Velotron Dynafit Pro, SRAM, Spearfish, SD) at a self-selected intensity in an environmental chamber set to ~22°C, 40% RH. After the warm-up, the GXT began at 200 W and increased 25 W every 2 min until participants reached volitional fatigue or until they were unable to maintain 40 rpm. Approximately 2–3 min after the end of the GXT, a blood sample was obtained from a forearm vein and used to measure blood lactate concentration (Lactate Plus, Nova Biomedical, Waltham, MA). HR (Vantage XL, Polar, Bethpage, NY) and V̇O₂ (TrueOne 2400, ParvoMedics, Inc., Salt Lake City, UT) were measured continuously throughout the GXT, and rating of perceived exertion (RPE) using the Borg 6–20 scale [29] was obtained at the end of each stage. V̇O_2max was taken as the highest 2 consecutive 30-s values averaged together.

Approximately 20 min after the initial GXT, a follow-up verification procedure was performed to ensure a plateau in V̇O₂ had been attained. Participants mounted the cycle ergometer and cycled for as long as possible at: 1) the same maximal power output achieved on the initial GXT (if less than 1 min was completed during the final stage of the initial GXT), or 2) a power output 25 W higher (if greater than 1 min was completed during the final stage of the initial GXT). A plateau in V̇O₂ was based on Taylor et al. [30] and defined as less than half (135 mL·min⁻¹) of the expected increase (270 mL·min⁻¹) in V̇O₂ based on the American College of Sports Medicine metabolic equation for leg ergometry [31]. Based on this criterion, all participants reached a plateau either during the initial GXT or during the follow-up procedure. After the follow-up verification procedure, participants rested for ~20 min, after which they cycled for an additional 20 min while the power output corresponding to 60% V̇O_2max was verified and the CO₂-rebreathing procedure used to estimate cardiac output (Q̇) [32] in the experimental trials was practiced.

Experimental trials

For the first experimental trial, participants followed the same instructions as the first trial, and adherence to pre-test instructions was verified using a 24-h history form. Hydration status was verified using a refractometer to measure urine specific gravity (USG). USG values < 1.020 were considered euhydrated [33]. Nude body mass was measured before the trial using a digital scale (BWB-800, Tanita, Arlington Heights, IL). For the 15-min trial, participants strapped on a HR monitor and inserted a rectal temperature probe (RET-1, Physitemp Instruments, Clifton, NJ) ~10 cm past the anal sphincter for measurement of rectal temperature (T_re). They rode the cycle ergometer for 15 min at a power output corresponding to 60% V̇O_2max, followed immediately (with no cessation of exercise) by a GXT to measure V̇O_2max. The GXT began by initially increasing power output 25 W, with 25-W increases every 2 min thereafter. Participants cycled until volitional fatigue or until they were unable to maintain 40 rpm. Like the control GXT, approximately 2–3 min after test termination, a blood sample was obtained and analyzed for blood lactate concentration. After being de-instrumented, participants towel dried and measured nude body mass again.

For the 45-min trials, procedures were similar but additional instrumentation was used. After measuring nude body mass and inserting the rectal temperature probe, a flexible venous catheter was inserted in a superficial forearm vein and kept patent with 0.9% normal saline. Thermocouples (Type T, Omega Engineering, Stamford, CT) were taped to the lateral calf, lateral thigh, lower back, lower abdomen, upper back, and chest for measurement of mean skin temperature (T̅_sk; TC-1000 Thermocouple Meter, Sable Systems, Las Vegas, NV) using the weighted average of the 6 sites. Mean body temperature (T̅_b) was calculated using the following equation [34]:

$${\overline{\mathrm{T}}}_{\mathrm{b}}=0.8\cdot {\mathrm{T}}_{\mathrm{r}\mathrm{e}}+0.2\cdot {\overline{\mathrm{T}}}_{\mathrm{s}\mathrm{k}}.$$

Next, participants entered an environmental chamber and sat upright on the cycle ergometer. A laser-Doppler flowmetry probe (VMS-LDF2, Moor Instruments, Wilmington, DE) was taped to the left posterior forearm for measurement of skin blood flow. After instrumentation, resting measures were taken for all variables (including blood pressure via auscultation of the brachial artery), a 2-mL blood sample was obtained, and then participants began cycling at a power output corresponding to 60% V̇O_2max. During exercise, between 8 and 15 min and again between 38 and 45 min, blood pressure, HR, V̇O₂ and V̇CO₂, RPE, and 2–3 trials of CO₂-rebreathing (for measurement of Q̇) were performed, in that order. Values presented for Q̇ represent the mean of the trials for a given time point. After 45 min, participants immediately began a GXT to measure V̇O_2max in the same manner as the 15-min trials. Within 3 min after the end of the GXT, a final 2-mL blood sample was obtained. Participants cooled down for ~5 min, and after being de-instrumented, they exited the chamber, toweled off, and measured nude body mass. V̇O₂ values measured at maximum in the experimental trials are referred to as V̇O_2peak because the procedure used to ensure a plateau in the control trial could not be performed. A V̇O_2peak value was considered valid if: 1) a plateau was reached based on the criterion used for the control test, or 2) a HR within 5 beats/min of the maximum reached during the control test was achieved.

The two 45-min trials were identical to one another except that in 45FAN, 2 floor fans (model HASF-99, Holmes Products Corporation, Milford, MA) on stands (~1.25 m off the floor) placed ~1 m away from participants and directed at their anterior torso were turned on a high setting (eliciting ~4.5 m/s air velocity) at ~38 min of exercise. The fans remained on throughout the rest of the submaximal exercise period as well as during the subsequent GXT.

Systolic and diastolic blood pressures were used to calculate mean arterial pressure (MAP) using the following equation [35]:

$$\mathrm{M}\mathrm{A}\mathrm{P}=\mathrm{D}\mathrm{B}\mathrm{P}+\left({\mathrm{S}}_{\mathrm{t}}\cdot \mathrm{P}\mathrm{P}\right),$$

where DBP = diastolic blood pressure, PP = pulse pressure, and S_t = the fraction of systole from the heart cycle, calculated as [35]:

$${\mathrm{S}}_{\mathrm{t}}=0.01{\mathrm{e}}^{[4.14-(40.74/\mathrm{H}\mathrm{R})]},$$

where HR = heart rate. MAP was divided by Q̇ to calculate systemic vascular resistance. Blood samples were collected into Vacutainers containing EDTA and were used to measure hemoglobin in duplicate using a HemoPoint H2 photometer (Stanbio, Boerne, TX) and to measure hematocrit in triplicate using the microhematocrit method. These values were then used to calculate plasma volume change using the equation by Dill and Costill [36].

Data analysis

T_re, T̅_sk, T̅_b, and skin blood flow data were recorded continuously and averaged every minute. HR data were recorded every 5 s and averaged every minute. Statistical analyses were performed using IBM SPSS Statistics v. 24 for Windows (IBM Corp., Armonk, NY). Data are reported as mean ± SD. A one-way repeated measures ANOVA was used to test the significance of mean differences in V̇O_2max among the treatment conditions (CONTROL, 15MIN, 45NF, and 45FAN). For variables during the experimental trials, such as cardiovascular, thermoregulatory, and metabolic measures, two-way repeated measures ANOVAs [treatment (45NF vs. 45FAN) × time (15 vs. 45 min)] were used to test the significance of mean differences. In the event of a significant omnibus test, paired samples t-tests using an adjustment to control family-wise α was used [37]. Outcome measures at maximum for the 45NF and 45FAN treatments – such as plasma volume change from rest, T̅_sk, and T̅_b – were analyzed using paired samples t-tests. All hypothesis tests used an α level of 0.05.

Results

Indicators of hydration

Body mass before exercise was not different between CONTROL [77.3 ± 5.3 kg)] and the experimental trials [15MIN = 76.8 ± 5.0 kg, 45NF = 76.6 ± 5.2 kg, and 45FAN = 76.7 ± 4.8 kg; P = 0.13]. USG also was not different among treatments (15MIN = 1.011 ± 0.005, 45NF = 1.012 ± 0.005, and 45FAN = 1.008 ± 0.005; P = 0.30) which, combined with the pre-exercise body mass data, suggests participants began all trials in a similar hydration state, as intended. Body mass decreased from pre- to post-exercise, and, as expected, the change was smaller in 15MIN [−0.82 ± 0.31%] compared to 45NF [−1.82 ± 0.27%; P < 0.001] and 45FAN [−1.91 ± 0.33%; P < 0.001], which were not different from one another (P = 0.48). Despite greater mass lost in the longer exercise trials, whole-body sweat rate, adjusted for blood drawn and respiratory water loss, was not different among experimental conditions [1500 ± 553 mL·hr⁻¹, 1616 ± 198 mL·hr⁻¹, and 1667 ± 236 mL·hr⁻¹ for 15MIN, 45NF, and 45FAN, respectively; P = 0.52].

Responses during submaximal exercise

Table 1 and Figure 1 show that a substantial CV drift occurred but the magnitude of drift was not different between conditions. Additionally, the increase in HR was proportional to the decrease in SV to the extent that Q̇ was essentially maintained over time, along with V̇O₂ and blood lactate. MAP was 2.9% higher in 45FAN compared to 45NF across time points (main effect of treatment, P = 0.02), which was most likely attributable to a higher (but non-significant difference across treatments, P = 0.86 for interaction) systemic vascular resistance at 45 min (main effect time, P = 0.01). Plasma volume decreased from rest to a similar extent in each condition (P = 0.85 for interaction). The increases in T_re, T̅_b, and RPE over time (all P < 0.05; Table 1) were not different between conditions (all P > 0.05). Figure 2 shows the effect of fan cooling in lowering T̅_sk ~1°C by 45 min, which was sustained during maximal exercise. Individuals varied in terms of the T̅_sk response to the fan airflow, with the decrements ranging from ~ 0.2–2.3°C by 45 min and ~ 0.5–2.3°C by the end of maximal exercise. The core-to-skin thermal gradient was higher at 45 min compared to 15 min in both conditions, and as a result of fan airflow the gradient was larger in 45FAN (Table 1). Cooling the skin reduced SkBF from a change from rest of ~350% (Table 1) to a change from rest of ~300%, but this difference (~15%) was not significant (P = 0.22).

Table 1.

Responses during submaximal exercise (mean ± SD).

	45NF		45FAN
	15-min	45-min	15-min	45-min
V̇O2 (L·min−1)	2.2 ± 0.2	2.3 ± 0.2	2.3 ± 0.2	2.2 ± 0.2
V̇O2 (% control V̇O2max)	59.9 ± 3.0	60.7 ± 3.2	61.2 ± 5.7	59.8 ± 4.1
Q̇ (L·min−1)	18.0 ± 1.1	17.5 ± 1.2	18.0 ± 1.4	17.1 ± 1.1
HR (beats·min−1)*	155 ± 16	176 ± 12	155 ± 17	173 ± 18
SV (mL)*	117.2 ± 13.6	99.7 ± 7.2	117.1 ± 16.0	100.6 ± 15.3
MAP (mm Hg)†	104.7 ± 7.9	109.6 ± 6.2	109.2 ± 5.5	111.2 ± 6.8
SVR (dyn·cm·s−5)*	468.3 ± 57.1	504.0 ± 45.0	490.0 ± 53.7	522.1 ± 55.1
Blood lactate (mmol·L−1)	3.3 ± 1.0	3.3 ± 1.3	3.5 ± 1.6	3.7 ± 1.7
ΔSkBF from rest (%)	361 ± 331	363 ± 270	273 ± 195	348 ± 249
ΔPV from rest (%)	−9.2 ± 2.1	−11.4 ± 5.9	−9.0 ± 2.4	−10.9 ± 5.7
Tre (°C)*	37.4 ± 0.1	38.3 ± 0.2	37.5 ± 0.2	38.4 ± 0.4
T̅b (°C)*	37.0 ± 0.2	37.7 ± 0.2	37.1 ± 0.2	37.6 ± 0.4
Tre – T̅sk (°C)	2.2 ± 0.8	2.9 ± 0.7§**	2.0 ± 0.8	4.0 ± 1.2§
RPE*	12 ± 2	15 ± 1	12 ± 1	16 ± 2

45NF = 45-min trial with no fan cooling; 45FAN = 45-min trial with fan cooling at the end of exercise; V̇O₂ = oxygen uptake; V̇O_2max = maximal oxygen uptake; Q̇ = cardiac output; HR = heart rate; SV = stroke volume; MAP = mean arterial pressure; SVR = systemic vascular resistance; ΔSkBF = change in skin blood flow; ΔPV = change in plasma volume; T_re = rectal temperature; T̅_b = mean body temperature; T_re – T̅_sk = core-to-skin thermal gradient; RPE = rating of perceived exertion. SkBF data are based on n = 6. For all other data, n = 8. Results of repeated measures ANOVAs are noted as: *P < 0.05 for main effect of time; ^†P < 0.05 for main effect of treatment; ^§P < 0.05 comparing 15- and 45-min values within the same condition; **P < 0.05 vs. the value at the same time point in FAN.

Figure 1.

Mean ± SD change in heart rate (HR) and stroke volume (SV) between minutes 15 and 45 during submaximal cycling at 60% V̇O_2max in 35°C with or without fan cooling starting at 38 min of exercise (n = 8). Means were compared using a 2-way repeated measures ANOVA. *P < 0.05 for main effect of time for each variable.

Figure 2.

Average ± SD mean skin temperature responses during 45 min of cycling at 60% V̇O_2max and at maximum after a graded exercise test in 35°C with or without fan cooling starting at 38 min of exercise (n = 8). Means were compared using a 2-way repeated measures ANOVA. *P = 0.001 vs. the 15-min value within the Fan treatment; ^†P = 0.004 vs. Fan treatment.

Responses during maximal exercise

Participants exerted a maximal effort during the GXTs at the end of the experimental trials as evidenced by no differences in maximal HR and RPE (HR, P = 0.07; RPE, P = 0.30; Table 2). Nonetheless, the durations of the GXTs were shorter [36% and 28% for 45NF (P < 0.001) and 45FAN (P < 0.001), respectively, compared to 15MIN] and the power outputs achieved were lower [20% and 18% for 45NF compared to CONTROL (P < 0.001) and 15MIN (P < 0.001), respectively, and 14% and 13% for 45FAN compared to CONTROL (P < 0.001) and 15MIN (P < 0.001), respectively] at maximum compared to CONTROL and 15MIN, which probably explains the lower blood lactate levels and RER values (both P < 0.05) during 45NF and 45FAN (Table 2). Given the longer duration of submaximal exercise during 45NF and 45FAN, it is not surprising that T_re at maximum was higher in 45NF and 45FAN compared to 15MIN (both P = 0.002), but experimental treatments were not different from one another (P = 0.85). Despite maintaining reduced T̅_sk and a wider T_re – T̅_sk thermal gradient during maximal exercise in 45FAN, V̇O_2peak still decreased by 15% compared to 15MIN. Furthermore, despite GXTs that were on average ~ 18% longer (but quite variable) in 45FAN compared to 45NF (P = 0.26), the 15% decrease in V̇O_2peak from 15MIN in 45FAN was no different (P = 0.54) than the 17% reduction in V̇O_2peak from 15MIN in 45NF (Figure 3).

Table 2.

Responses during maximal exercise (mean ± SD).

	Trial
Variable	Control	15-min	45NF	45FAN
HR (beats·min−1)	194 ± 9	196 ± 5	199 ± 6	196 ± 6
Blood lactate (mmol·L−1)	9.9 ± 2.5	9.0 ± 3.0	6.0 ± 1.8*†	6.7 ± 2.7*†
RER	1.10 ± 0.04	1.01 ± 0.03*	0.93 ± 0.04*†	0.94 ± 0.04*†
RPE	19 ± 1	20 ± 1	20 ± 1	20 ± 0
Power output (W)	288 ± 40	280 ± 37	230 ± 42*†	246 ± 42*†
Test duration (min)	10.8 ± 1.9	10.1 ± 1.0	6.4 ± 1.7*†	7.3 ± 1.6*†
ΔSkBF from rest (%)	–	–	448 ± 356	325 ± 207
ΔPV from rest (%)	–	–	−12.5 ± 5.2	−13.0 ± 5.5
Tre (°C)	–	38.0 ± 0.3	38.5 ± 0.2†	38.6 ± 0.4†
T̅b (°C)	–	–	37.9 ± 0.2	37.7 ± 0.4
Tre – T̅sk (°C)	–	–	2.9 ± 0.6**	4.2 ± 1.1

15-min = 15-min trial; 45NF = 45-min trial with no fan cooling; 45FAN = 45-min trial with fan cooling at the end of exercise; HR = heart rate; RER = respiratory exchange ratio; RPE = rating of perceived exertion; ΔSkBF = change in skin blood flow; ΔPV = change in plasma volume; T_re = rectal temperature; T̅_b = mean body temperature; T_re – T̅_sk = core-to-skin thermal gradient. SkBF data are based on n = 6. For all other data, n = 8. Results of one-way repeated measures ANOVAs with post hoc tests are noted as: *P < 0.05 vs. Control; ^†P < 0.05 vs. 15-min trial. Results of paired samples t-test are noted as: **P < 0.05 vs. 45FAN.

Figure 3.

Vertical scattergram showing mean (horizontal lines) V̇O_2max data for each condition as well as individual responses (n = 8). Means were compared using a 1-way repeated measures ANOVA. *P < 0.001 vs. Control. ^†P < 0.001 vs. 15-min trial.

Discussion

The purpose of this study was to determine if a period of brief fan cooling – after CV drift had already occurred – was sufficient to attenuate the negative effects of CV drift on V̇O_2max in a hot environment. The primary finding was that fan cooling at the level utilized in this study (4.5 m/s), administered after CV drift had occurred, was insufficient to mitigate the negative effects of CV drift on V̇O_2max after prolonged submaximal exercise in the heat, despite permitting longer GXT durations in 6 of 8 subjects and higher power outputs in 4 of 8 subjects relative to the condition with no fan. A given magnitude of CV drift resulted in a proportional decrease in V̇O_2max.

The magnitude of CV drift and corresponding reductions in V̇O_2max in 45FAN and 45NF were comparable to our previous studies utilizing similar conditions [2,4,5,17,19]. In our prior study involving fan cooling during the period of CV drift between 15 and 45 min the magnitude of CV drift was blunted – similar to Shaffrath et al. [21] – and the corresponding decrease in V̇O_2max was smaller [2], relative to this study.

In terms of comparison to other studies investigating maximal exercise capacity after manipulating the combination of core and skin temperatures, our findings are similar to Trangmar et al. [11]. They showed that combined internal and skin hyperthermia (T_c – T̅_sk gradient ≈ 2.4°C) resulted in a reduction in V̇O_2max of about 8%. However, they also showed that V̇O_2max was not compromised under a condition with a larger T_c – T̅_sk gradient, unlike in the present study. Trangmar et al. [11] utilized incremental exercise, not prolonged exercise with CV drift, so methodological differences may explain the discrepant results.

Comparison with other studies using a period of brief cooling is difficult because limited studies are available investigating such cooling strategies. Studies involving cooling during exercise – so called per-cooling – have generally used cooling packs, cooling vests, cold fluid/ice slurry ingestion, facial wind or water spray, and application of menthol. Generally, cooling was initiated early in the exercise bout in these studies, not for a short duration near the end as was the case in the current study [38]. Among the studies that applied cooling later in the exercise protocol, one study used a period of brief wind cooling (at km 3–12) during a 15-km time trial (after a 15-min bout of submaximal cycling) and showed an ~4% improvement in performance in 28°C, 80% RH compared to a similar condition without wind cooling [39]. T̅_sk was approximately 1.5 °C lower in the wind condition relative to the control condition [39]. Our decrease in T̅_sk was smaller, but GXT duration and power output at maximum (i.e. markers of performance) were on average 18% longer and 7% higher, respectively, although responses were quite variable. In another study, menthol was sprayed on the shirt of participants after completing 10 km of a 16.1-km cycling time trial in 33.5°C; menthol improved perception but did not affect performance or body core and mean skin temperatures [40].

Our findings are in contrast to those of Shaffrath and Adams [21],who showed an ~10 bpm decrease in HR with fan airflow at 4.3 m/s administered during the final 5 min of a 75-min bout at 60% V̇O_2max. In the present study, HR was not different between conditions at 45 min (Table 1), and it did not change between 38 and 45 min after the fans were turned on in 45FAN. A more mild ambient air temperature condition (~24°C) in the Shaffrath and Adams [21],study compared to the hot air temperature (35°C) used in the current study may explain the conflicting results. Given that hyperthermia (elevated core and skin temperatures) has been shown to limit aerobic exercise capacity in the heat [11,12,15,41], the general mechanism by which per-cooling is reputed to improve performance is by increasing heat storage capacity [38] and blunting the rise in T_c [42]. However, a reduced T_c is not requisite for improved performance in the heat [42] or reversal of CV drift under temperate conditions [21].

The mechanisms responsible for CV drift have been thoroughly covered elsewhere [5,7,43,44] and a lengthy discussion is beyond the scope of this paper. However, briefly, the increased skin blood flow resulting from the heat stress would be expected to increase cutaneous venous volume and thereby reduce central venous pressure [7]. Under such conditions, the increase in HR could be triggered by a baroreceptor-mediated reflex to maintain Q̇ and blood pressure [7]. Others have eloquently shown that the decrease in SV could be attributable to reduced ventricular filling time associated with tachycardia [43,44] instead of peripheral displacement of blood volume associated with elevated skin blood flow. It is plausible that mechanisms asserted by both explanations may have interacted to elicit CV drift in the current study [5].

The precise mechanism(s) by which V̇O_2max was reduced in the 45-min trials in the present study cannot be determined from the data collected. Maximum HR was not different between 15MIN and the 45-min trials, which suggests cardiovascular capacity had been attained and leg fatigue was improbable as a limiting factor. Likewise, maximal arteriovenous oxygen difference would not be expected to be reduced at maximum under these conditions [16]. Therefore, reduced maximal SV likely explains the reductions in V̇O_2max observed. While SV likely increased during maximal exercise from levels observed during the submaximal exercise period, we suspect the peak level achieved was less than that achieved during the control and 15MIN trials. SV could have been lower at maximum because of reductions in central blood volume, central venous pressure, and end-diastolic volume in conjunction with elevated skin blood flow [45].

In light of this proposed mechanism for reduced SV, it is unclear why V̇O_2max decreased similarly in the 45-min trials since skin blood flow was expected to be higher at the end of submaximal exercise during 45NF vs. 45FAN. The level of skin blood flow is partly dependent on the core-to-skin thermal gradient [10,12,13], and elevated T̅_sk results in greater skin blood flow at a given T_c [10]. At the end of exercise in the 45-min trials, just before commencement of the graded exercise test to elicit V̇O_2peak, the thermal gradient for 45NF was 2.9°C on average, which was sustained throughout maximal exercise. The fan airflow in 45FAN, even though applied relatively briefly, was sufficient to increase the evaporative capacity of the environment [24,25] and thereby lower T̅_sk, which was sustained throughout the maximal exercise portion of the bout (Figure 2). This resulted in an average thermal gradient of 4.0°C, which was sustained during maximal exercise. We speculate the difference in the thermal gradients between conditions was not sufficient to mobilize enough peripherally displaced blood from the skin to restore SV to control max or 15MIN max levels, hence similar decrements in V̇O_2max between 45-min trials. Additionally, while skin blood flow is dependent on the T_c – T̅_sk gradient, it is also driven by absolute T_c [46]; given that absolute T_c at maximum was not different between 45FAN and 45NF (Table 2), the thermoregulatory driver of skin blood flow may have been comparable under both conditions.

Interestingly, in our previous study utilizing fan cooling between 15 and 45 min [2], T̅_sk at 45 min was higher than in the present study (35.2°C vs. 34.4°C), and T_re was lower (38.2°C vs. 38.4°C), so the T_c – T̅_sk thermal gradient was larger in the current study (4°C vs. 3°C). Despite this larger thermal gradient (and presumably lower skin blood flow), V̇O_2max decreased proportionally over twice as much in the present study (15% vs. 5.7%). Fan cooling during exercise in the aforementioned study [2] blunted the magnitude of CV drift (4% increase in HR and 3% decrease in SV) and thereby blunted the magnitude of decrement in V̇O_2max. These data support the hypothesis that independent of the combination of skin and core temperatures and the thermal gradient, a given magnitude of CV drift typically results in a proportional decrease in V̇O_2max.

Given that V̇O_2max and HR at maximum were unaffected by fan cooling, it is unclear why some subjects were able to sustain the maximal GXTs for longer and thereby achieve higher maximal power outputs. We speculate the drier and cooler skin concomitant with fan airflow may have permitted these individuals to temporarily perceive the exercise as more tolerable during the GXT [47,48], so they could perform more anaerobic work, even though at the point of maximum their perceived exertion was the same as without fan airflow.

A possible limitation of the study was that participants were in various stages of acclimation. Trials took place in the southeastern United States throughout the year (data were collected over many months because of interruptions in collection due to other ongoing projects). The repeated trials in the heat could have had an acclimation effect. Counterbalancing of trials, however, should have avoided any systematic effects of acclimation on outcome measures. Another limitation was that skin blood flow and SV were not measured at maximum, so we cannot be certain of the extent to which fan airflow may have affected these variables during maximal exercise.

Brief administration of fan airflow at the level and for the duration used in this study did not mitigate the negative consequences of CV drift on V̇O_2max after prolonged submaximal exercise in a hot environment during which CV drift had already occurred. These data have implications regarding countermeasures used to maintain aerobic reserves. Cooling modalities initiated earlier during exercise and thereby applied for a longer duration with the goal of minimizing CV drift – or if applied only briefly, resulting in a greater T_c – T_sk thermal gradient – appear necessary to attenuate decreases in V̇O_2max associated with CV drift during heat stress. Future studies are therefore warranted to explore more robust, acute countermeasures to better develop CV drift mitigation strategies for situations in which prolonged administration of countermeasures is not feasible. Moreover, this study adds further evidence that the magnitude of CV drift is proportional to the decrease in V̇O_2max, and suggests that the decrease in SV associated with CV drift may cause the reduction in V̇O_2max.

Funding Statement

This work was supported by the University of Alabama [College of Education].

Acknowledgments

The authors thank the volunteers for participating in the study. The efforts of Robert Herron and Jason Casey are also appreciated for helping with data collection. This work was supported by a grant from the College of Education at the University of Alabama.

Author contribution statement

JW conceived and designed the research, collected and analyzed data, and drafted the manuscript. JN, CK, and SC collected data and edited the manuscript. JN and SC reduced data. SC drafted some parts of the manuscript. All authors read and approved the manuscript.

Disclosure statement

No potential conflict of interest was reported by the authors.