Effect of acute ambient temperature exposure on cardio-pulmonary and respiratory kinetics in men

Abstract

Aim: The goal of the study was to compare the kinetic responses of heart rate (HR) and pulmonary (V̇O₂pulm) and muscular (V̇O₂musc) oxygen uptake during dynamic leg exercise across different acute ambient temperature conditions in a climatic chamber.

Methods: Thirteen physically healthy, active, male volunteers demonstrated pseudorandom binary sequence (PRBS) work rate (WR) changes between 30 and 80 W at 15 °C, 25 °C and 35 °C, respectively. HR was measured beat-to-beat using an echocardiogram and V̇O₂pulm by breath-by-breath gas exchange; V̇O₂musc estimations were assessed by applying a circulatory model and cross-correlation functions.

Results: No significant differences were observed across the various temperature conditions in each case for HR, V̇O₂pulm or V̇O₂musc kinetics (p > 0.05). Baroreflex regulation based on HR kinetics does not seem to be influenced between ambient temperatures of 15 °C and 35 °C during dynamic exercise.

Conclusions: The results imply that ambient temperatures of 15 °C, 25 °C and 35 °C have no effect on HR, V̇O₂pulm or V̇O₂musc kinetics during dynamic moderate exercise. The applied approach may be of interest for assessments of the cardio-pulmonary and respiratory health statuses of individuals working or performing sports in extreme temperature environments. Furthermore, differentiation between systemic (e.g. cardio-dynamic: HR) and specific (e.g. exercising tissues: V̇O₂musc) determinants of the relevant physiological systems may improve the evaluation of an individual’s health status.

Keywords:

• Thermoregulation
• exercise
• gas exchange
• circulatory modelling
• PRBS
• time-series analysis
• ambient temperature

Introduction

At different ambient temperature conditions (e.g. warm or hot environments), the rapid delivery and extraction of oxygen (O₂) to and within the exercising skeletal musculature are essential for endurance sports (e.g. skating, bicycling, skiing and running) and specific occupations (e.g. firefighters and Special Forces). The quality of the regulatory processes of oxygen delivery and extraction can be described via oxygen uptake (V̇O₂) kinetics. It is well-known that improved exercise capacity (V̇O₂max) is accompanied by faster V̇O₂ kinetics, which is an indicator of exercise performance [1,2].

Alterations in thermoregulatory responses – e.g. based on ambient temperature conditions – are associated with changes in cardiac output (Q̇) redistribution, which impacts the skin, visceral and skeletal muscle blood flow and mean arterial blood pressure [3]. The thermoregulatory interplay becomes more complex with the combination of thermal and dynamic exercise stress. In this case, the redistribution of Q̇ to the skin for heat reduction and to the exercising skeletal muscles for the maintenance of aerobic metabolism can create a conflict in the Q̇.

Owing to the impacts of redistribution on the cardiovascular system during dynamic exercise in the heat, V̇O₂ kinetics could be slowed because O₂ delivery may be limited at the muscle site [4,5], improved by higher muscle temperature [6,7], or possibly unchanged as a net effect of counterbalancing the positive and negative impacts as aforementioned.

Only limited information is available on V̇O₂ kinetics across different acute ambient temperature exposures during dynamic exercise. If ambient temperature affects muscle temperature, then aerobic metabolism could be influenced. For instance, higher temperatures induce faster reactions of the aerobic enzymes, resulting in faster muscle V̇O₂ kinetics. Based on previous experiments, muscle or core hyperthermia at moderate [8] or intense exercise intensity [9] did not influence V̇O₂ kinetics. The reason for this lack of information is caused by two issues: first, the appropriate equipment (e.g. climatic chambers) is rare and hardly accessible. Second, applied cooling or heating devices (e.g. jackets, suits) only establish a locally restricted temperature environment on or within body segments [8–12]. These locally restricted temperature environments can lead to altered interactions of centrally (e.g. the hypothalamic area) and peripherally (e.g. leg, thorax skin) affected thermoreceptor responses in contrast to all-encompassing ambient conditions. For instance, a real integrated temperature environment should consider both peripheral (skin) and central (hypothalamus) receptors for thermoregulatory temperature control. Therefore, it is important to strive for this environment to establish real ambient conditions, which can be achieved by applying a climatic chamber.

From a sophisticated point of view, V̇O₂ kinetics must be subdivided into the measurement at the mouth (pulmonary oxygen uptake [V̇O₂pulm]) and the oxygen consumption across the exercising skeletal muscles (muscular oxygen uptake [V̇O₂musc]). This differentiation is important during work rate (WR) transitions because V̇O₂pulm is – in contrast to V̇O₂musc – non-linearly time-delayed and distorted because of venous return and Q̇ dynamics [13,14]. Therefore, appropriate models are required to account for these deformations and for the proper estimation of V̇O₂musc kinetics, which is the parameter of interest [13,15–20]. The differentiation between the pulmonary and the muscle levels is of general relevance because hyperthermia-induced alterations in Q̇ – based on heart rate (HR) responses – result in the redistribution of Q̇ and can lead to altered venous muscle-to-lung transit times, impacting V̇O₂pulm kinetics.

In this regard, HR is an indicator of baroreflex sensitivity by means of blood pressure regulation and can subsequently impact Q̇, as well. There is evidence that the baroreflex response is not impaired by passive heat strain [3], but there is no information available to suggest that HR kinetics may be impacted as a result of different ambient temperature conditions during dynamic exercise.

Typically, V̇O₂ is non-invasively measured at the mouth by the application of repeated step responses in WR and is analysed with multiple exponential mathematical equations and fitting procedures [15,21]. In this regard, multi-exponential approaches cannot be applied for fast changes in WR by definition [22–25]. Alternatively, the application of pseudorandom binary sequence (PRBS) WR protocols with a time-series analysis (e.g. auto- correlation function [ACF] and cross-correlation function [CCF]) allows a practical evaluation of the parameters of interest in the transient WR phases in a single test run [13,26,27].

The novelty of the presented study is the combined application of the sophisticated analysis of V̇O₂musc and V̇O₂pulm kinetics with a climatic chamber. Here, we are interested in the acute effects of different ambient temperature conditions on cardio-pulmonary and respiratory parameters during moderate dynamic leg exercise.

We intended to investigate whether the application of a PRBS WR protocol combined with a time-series analysis and the utilisation of a climatic chamber provided comparable results with previous studies that applied locally restricted temperature environments. The following hypothesis was tested: HR, V̇O₂pulm and V̇O₂musc kinetics are not different across ambient temperatures of 15 °C, 25 °C and 35 °C.

Methods

Subjects

Thirteen physically healthy, active, male volunteers (age: 36 ± 10 years; height: 180 ± 6 cm; body mass: 82 ± 10 kg; body mass index: 25 ± 3 kg/m² [mean ± standard deviation]) participated in the study and gave their written informed consent, which was in accordance with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. Informed consent was approved by the Ethics Committee of the German Sport University Cologne.

Figure 1. Work rate protocol for exercise testing for the three different ambient temperatures. Gray horizontal bars indicate time periods for mean value calculations during rest (30 s), at low and high constant phases (30 s), during the PRBS phases (300 s) and at the end of the WR steps of 100, 125 and 150 W (30 s each). PRBS: pseudorandom binary sequence.

Exercise protocol and climatic chamber

Each subject performed three identical exercise tests, each at different ambient temperatures (15 °C, 25 °C and 35 °C) on a semi-recumbent (seat back at 45° to ground level) cycle ergometer (ergometrics er900L, Schiller AG, Baar, Switzerland) with a cadence of 60 rpm. The exercise protocol includes two parts (Figure 1). In the first part (time: 0–1000 s), PRBS WR changes between 30 and 80 W were applied for the kinetics analysis with constant low and high phases before (30 W; 200 s) and after (80 W; 200 s) the PRBS.

In the second part (time >1000 s), a step increase in WR (25 W·min⁻¹) was applied until the subjects reached their symptom-limited maximums. For each participant, at least 24- to 48-h recoveries were prescribed between the tests.

A climatic chamber (NEMA NCZ3036N(S), WTD 61, Manching, Germany) was used to control the ambient conditions (see Table 1).

Table 1. Ambient conditions in the climatic chamber during the exercise tests (mean ± SD).

N = 13	15 °C	25 °C	35 °C	Sig
Temperature (°C)	15.1 ± 0.3	24.5 ± 1.3	35.4 ± 0.3	a, b, c
Pressure (mbar)	997 ± 7	968 ± 8	973 ± 8	b, c, d
Relative humidity (%)	34 ± 2	33 ± 2	33 ± 5	n.s.

SD: standard deviation; Sig: significant difference.

n.s.: Not significantly different, p > 0.05.

^a Significant difference between 15 °C and 25 °C, p < 0.01 (2-tailed).

^b Significant difference between 25 °C and 35 °C, p < 0.01 (2-tailed).

^c Significant difference between 35 °C and 15 °C, p < 0.01 (2-tailed).

^d Significant difference between 15 °C and 25 °C, p < 0.05 (2-tailed).

Test procedure

Before each exercise test, the ambient temperature and the relative humidity in the climatic chamber were verified to be within the expected range.

The subjects were prepared outside the climatic chamber at a room temperature of ∼20 °C. The instrumentation of the subjects was then performed in the climatic chamber followed by a resting phase for acute acclimatisation and baseline measurements. Next, the exercise test was performed either at 15 °C, 25 °C or 35 °C in a randomised order (Figure 2). In the climatic chamber, all subjects wore only shorts, socks and sports shoes.

Figure 2. Test procedure time line. See text for details.

Heart rate, oxygen uptake and core body temperature measurements

HR was measured beat-to-beat using an ECG (Schiller AT-104 PC EKG, Schiller Medizintechnik GmbH, Feldkirchen, Germany), and V̇O₂pulm was measured breath-by-breath (ZAN 680, ZAN Messgeräte GmbH, Oberthulba, Germany) by applying the algorithms of Beaver et al. [28] for true alveolar gas exchange.

To determine core body temperature (CBT), a heat flux sensor labelled “Double Sensor” (DS) was used. The DS was fixed above each subject’s left eye, below the hairline on a nearly flat surface/area (regio frontalis). A description of the sensor design, precision, accuracy and validity is given elsewhere [29–32]. During the 15 °C temperature condition, the DS was insulated on its exterior with a textile pad for CBT correction because the sensor is physically disturbed by cold environmental conditions below 20 °C.

Stroke volume and gas exchange threshold estimations

Stroke volume (SV) was estimated for the constant phases before and after the PRBS intervals for 30 and 80 W by applying Equation (1(1) ), which was taken and adapted from Geigy Scientific Tables [33,34]. (1)

The gas exchange threshold was calculated by applying the V-slope method [35] across the different temperature conditions.

Data processing

All measured breath-by-breath, beat-to-beat and CBT data were time-aligned and interpolated into 1-s intervals for further analysis.

Kinetics analysis

The kinetics analyses of HR and V̇O₂pulm and the estimation of V̇O₂musc were assessed according to the approach of Hoffmann et al. [13]. The CCF of the WR and the parameter of interest (e.g. HR, V̇O₂pulm or V̇O₂musc) were applied to estimate the maximum (CCF_max) of the CCF and the related lag (CCF_lag) as surrogates of the kinetics responses. Higher CCF_max-values indicate faster system responses, and higher CCF_lag-values indicate delayed system responses.

Statistical analysis

All data were analysed for normal distribution (ND) by applying a Kolmogorov–Smirnov test with Lilliefors significance correction.

For comparisons of peak responses, the gas exchange threshold, exercise time, model parameters and the CCF_max- and CCF_lag-values, a repeated-measures analysis of variance (ANOVA) with Temperature as factor was applied if ND was given. Otherwise, the Friedman test with post hoc pairwise Wilcoxon signed-rank test comparisons with alpha error accumulation correction was utilised.

In addition, for the influence of ambient temperature on thermoregulatory parameters, a two-way repeated-measures ANVOA was performed for HR and CBT with Temperature and Work rate as factors during low (30 W), high (80 W) and dynamic WR changes (PRBS1&2; 53.3 W) and also during the following WR steps (100, 125 and 150 W) as long as all 13 participants were able to continue exercising on the leg ergometer.

For the dynamic kinetics analysis, a two-way ANOVA with Temperature and Lag as factors was performed for HR, V̇O₂pulm and V̇O₂musc at specific lag times (−20 to 120 s; stepwise: 10 s) to account for comparisons between the CCF time course and temperature.

In case of significant factors or interactions, post hoc pairwise comparisons (Bonferroni) were performed.

Relationships between the parameters were detected using the Pearson product–moment correlation coefficient in the case of ND. Otherwise, the Spearman rank correlation coefficient was applied. Statistical significance was set to 0.05 at the alpha level.

Results

Peak responses

Table 2 summarises the peak responses of power output, V̇O₂, HR, CBT and the values of the gas exchange threshold and exercise time.

Table 2. Peak responses (mean ± SD) during exercise for 15 °C, 25 °C and 35 °C ambient temperatures.

N = 13	15 °C	25 °C	35 °C	Sig
Peak power output (W)	237 ± 35	244 ± 31	231 ± 29	a
Peak power output (W·kg^−1)	2.9 ± 0.6	3.0 ± 0.4	2.8 ± 0.4	n.s.
Peak oxygen uptake (L·min^−1)	3.8 ± 0.8	3.8 ± 0.8	3.4 ± 0.7	a
Peak oxygen uptake (mL·min^−1·kg^−1)	48 ± 12	47 ± 9	42 ± 9	a
Peak heart rate (min^−1)	161 ± 16	167 ± 9	164 ± 10	n.s.
Peak core body temperature (°C)	39.2 ± 0.6	39.0 ± 0.4	37.9 ± 0.3	b, c
Gas exchange threshold (L·min^−1)	1.6 ± 0.4	1.7 ± 0.5	1.6 ± 0.4	n.s.
Exercise time (s)	1389 ± 88	1404 ± 74	1367 ± 69	a

SD: standard deviation; Sig: significant difference.

n.s.: not significantly different, p > 0.05.

^a significant difference between 25 °C and 35 °C, p < 0.05 (2-tailed).

^b significant difference between 15 °C and 35 °C, p < 0.001 (2-tailed).

^c significant difference between 25 °C and 35 °C, p < 0.001 (2-tailed).

Basic thermoregulatory responses

Figure 3(a) illustrates the absolute values of HR as a function of WR as grouped mean data for the different ambient temperature conditions. According to the two-way repeated-measures ANOVA with factors Work rate and Temperature, the factors Work rate (p < 0.001) and Temperature (p < 0.01) were significant, but no significant interactions (p > 0.05) could be observed. For Temperature, pairwise post hoc comparisons (Bonferroni) revealed that the conditions 15 °C and 35 °C (p < 0.01) were significantly different. For Work rate, the HR values were significantly different for all pairwise comparisons (Bonferroni; p < 0.001).

Figure 3. (a,b) Absolute values of HR (mean ± SE, 3a) and CBT (mean ± SE, 3b) as function of WR across the ambient temperatures (15 °C, 25 °C and 35 °C). CBT: core body temperature; HR: heart rate; SE: standard error of the mean; WR: work rate.

Figure 3(b) illustrates the absolute values of CBT as function of WR as grouped mean data for the different ambient temperature conditions. According to the two-way repeated-measures ANOVA with factors Work rate and Temperature, the factors Work rate (p < 0.001) and Temperature (p < 0.01) were significant, but no significant interactions (p > 0.05) could be observed. For Temperature, pairwise post hoc comparisons (Bonferroni) revealed that the conditions 15 °C and 35 °C (p < 0.01) and 25 °C and 35 °C (p < 0.05) were significantly different. For Work rate, the CBT values were significantly different for all pairwise comparisons (Bonferroni; at least p < 0.05) without a comparison of 30 W and PRBS1 (53.3 W; p > 0.05).

To establish comparable conditions for further analysis during exercise, we calculated the relative changes (Δ; based on resting values) as averages of HR and CBT for 30 W, PRBS1&2 (53.3 W), 80, 100, 125 and 150 W as functions of WR (Figure 4(a,b)). The relative changes for HR and CBT and for the ΔHR–WR and ΔCBT–WR relationships show only a significant influence of Work rate (p < 0.001). Neither Temperature nor the interactions illustrate significant effects (p > 0.05) expressed as relative changes.

Figure 4. (a–c) Relative changes of HR (ΔHR, mean ± SE, 4a), CBT (ΔCBT; mean ± SE, 4b) and PSI (mean ± SE, 4c) based on the resting values as function of work rate across the ambient temperatures. CBT: core body temperature; HR: heart rate; PSI: physiological strain index; SE: standard error of the mean.

The physiological strain index (PSI) [36] is a common measure to evaluate heat stress. Accordingly, we calculated the PSI as the PSI–WR relationship (Figure 4(c)) for each of the three temperature conditions. According to a two-way repeated-measures ANOVA with Work rate and Temperature as factors, Work rate was significant (p < 0.001), but Temperature and the interactions were not significant (p > 0.05). For Work rate, the PSI values were significantly different for all pairwise comparisons (Bonferroni; at least p < 0.01).

Kinetics analysis – time-series analysis

A two-way repeated-measures ANOVA revealed a significant effect of factor Lag (p < 0.001), but not factor Temperature or the interactions (p > 0.05) in each case for HR, V̇O₂pulm and V̇O₂musc kinetics (Figure 5). Post hoc comparisons for Lag are not displayed because potential differences between lags are natural and were not of interest in the current investigation. Nevertheless, HR, V̇O₂pulm and V̇O₂musc differ in their relative CCF courses, demonstrating different kinetics over the entire responses.

Figure 5. (a–c) Means (±SE) of the cross-correlation functions (CCF) of WR with HR (5a), pulmonary (V̇O₂pulm 5 b) and muscular (V̇O₂musc, 5c) oxygen uptake, respectively, for the different ambient temperatures. The grey triangle displays the ACF of WR. The upper right enhanced graphs display the mean values (± SE) at specific lag times (−20, −10, 0, 10, 20, 30, 40, 50, 60, 70, 80, 90, 100, 110 and 120 s) for a comparison applying a two-way ANOVA with factors Lag and Posture. Significant differences for factor Lag are not displayed graphically to avoid overloading the figure. See text for further explanation. ACF: auto-correlation function; ANOVA: analysis of variance; HR: heart rate; SE: standard error of the mean; WR: work rate.

The maxima of the CCF courses and their corresponding lags allow visualisation of the kinetics responses of HR, V̇O₂pulm and V̇O₂musc in a compressed manner as grouped mean data (Figure 6). This allows an easy and direct comparison of the kinetics responses across the temperature conditions. No significant differences in CCF_max or in CCF_lag were found for HR, V̇O₂pulm and V̇O₂musc across the temperature conditions (see also Table 3). The CCF_max-values of HR and V̇O₂musc are located close to or on the descending part of the ACF (WR). This result presents indicative – though insufficient – information that these parameters may be in line with the assumption of first-order system properties. In contrast, the CCF_max-values of V̇O₂pulm, located at lags between 30 and 40 s, indicate that this parameter is substantially inconsistent with first-order system properties.

Figure 6. Mean (±SE) of the CCF [maxima and lags (CCF_max, CCF_lag)] of work rate with HR, pulmonary (V̇O₂pulm) and muscular (V̇O₂musc) oxygen uptake, respectively, for 15 °C, 25 °C and 35 °C ambient temperatures. The grey line illustrates the auto-correlation function of WR (in part). Data points that are close to or on this line represent therefore possible assumptions of a linear first-order response of the physiological system or parameter behind the computational rational. CCF: cross-correlation function; HR: heart rate; SE: standard error of the mean.

Table 3. Means and SD of model parameter estimates and time-series analysis between HR, muscular (V̇O₂musc) and pulmonary (V̇O₂pulm) oxygen uptake for 15 °C, 25 °C and 35 °C ambient temperatures.

					HR		V̇O2pulm		V̇O2musc
N = 13		Vv (mL)	Q̇rem (mL·min^−1)	V̇O2rem (L·min^−1)	CCFmax (a.u.)	CCFlag (s)	CCFmax (a.u.)	CCFlag (s)	CCFmax (a.u.)	CCFlag (s)
Temperature	Sig	*	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.	n.s.
15 °C	Mean	3446	3580	0.28	0.440	10	0.331	32	0.432	13
	SD	584	1606	0.06	0.116	3	0.056	7	0.093	4
25 °C	Mean	3319	3831	0.27	0.420	11	0.339	36	0.446	13
	SD	556	1347	0.06	0.102	4	0.045	4	0.091	4
35 °C	Mean	2962	4404	0.27	0.417	10	0.325	32	0.409	12
	SD	597	1740	0.06	0.118	3	0.054	6	0.081	3

HR: heart rate; Q̇_rem and V̇O₂rem: perfusion and oxygen uptake for the remainder of the body; SD: standard deviation; Sig: significant difference; V_v: venous blood volume between the exercising muscles and the lungs.

n.s.: not significantly different, p > 0.05.

* Tendency for factor Temperature, 0.10 > p > 0.05.

The model parameter estimates, the CCF_max- and the CCF_lag-values for HR, V̇O₂pulm and V̇O₂musc are summarised in Table 3. For differences in venous blood volume (V_v), a trend was observed (0.10 > p > 0.05). With increasing ambient temperature, V_v decreases (15 °C–25 °C: −3.6%; 15 °C–35 °C: −14.0%). At the same time, the perfusion of non-exercising tissues (Q̇_rem) increases (15 °C–25 °C: +7.0%; 15 °C–35 °C: +23.0%).

Kinetics analysis – first-order system properties

To verify the CCF courses of HR, V̇O₂pulm and V̇O₂musc for the first-order system properties (Figure 7), we conducted the following steps:

1. Averaging of the three temperature conditions to obtain a single CCF course for each parameter (HR, V̇O₂pulm and V̇O₂musc).

2. Computation of the error bars as 95% confidence intervals for all data points for each parameter.

3. Determination of the CCF_max of the averaged single CCF course to estimate the corresponding time constant (τ) for each parameter (HR, V̇O₂pulm and V̇O₂musc).

4. Using the estimated τ, we simulated CCF courses equivalent to the first-order system properties in the respective figures for a one-to-one comparison.

5. Calculation of the residuals between the averaged single CCF course from the estimated (V̇O₂musc) and measured (HR, V̇O₂pulm) data and the simulated CCF course of a first-order system.

Generally, we assume that a first-order system responds without a time delay (TD). The comparisons between predefined and estimated system responses are illustrated in Figure 7(a)–(d). However, V̇O₂pulm is delayed and distorted because of cardio dynamic influences and venous return, so a TD or an equivalent approach is obviously necessary for an analysis of the kinetics responses. Therefore, we shifted the simulated CCF course (τ = 43.8 s) and the ACF (WR) in Figure 7(d) to the right side (+20 s; lag-correction) to account for the delayed response and specifically for an improved fitting procedure as applied for the phase II analysis of V̇O₂pulm.

Figure 7. (a–d) Averages (±95% confidence intervals) of the CCF of WR with HR (7a), muscular (V̇O₂musc, 7 b) and pulmonary oxygen uptake without (V̇O₂pulm, 7c) and with lag-correction (+20 s, 7d). The grey dotted lines illustrate the simulated CCF course of first-order system derived from the CCF_max of the corresponding parameter in each panel. The grey triangle illustrates the ACF of WR. The lower panels show the residuals between the measured or estimated parameters and the simulated CCF course with τ as the first-order system in each case. ACF: auto-correlation function; CCF: cross-correlation function; HR: heart rate; WR: work rate.

Next, we calculated the magnitude of the aberrations between the first-order system responses and a particular parameter from the residuals in Figure 7 as the sum of deviation squares (SDS; see Figure 8).

Figure 8. SDS of HR, muscular (V̇O₂musc) and pulmonary oxygen uptake kinetics [V̇O₂pulm, without and with lag-correction (+20 s)]. The deviance is calculated between simulated CCF courses derived from the corresponding CCF_max value and the measured or estimated parameter (from lags −20 to 180 s, see Figure 7). Smaller values indicate better agreements with first-order system properties. CCF: cross-correlation function; HR: heart rate; SDS: sum of deviation squares.

The SDS values are high for V̇O₂pulm both with and without lag-correction. In contrast, HR shows minor deviations, but V̇O₂musc shows the best agreement with the first-order system properties.

Significant correlations were found between the CCF_max-values of V̇O₂pulm and V̇O₂musc at 15 °C (r = 0.732, p < 0.01), 25 °C (r = 0.688, p < 0.01) and 35 °C (r = 0.744, p < 0.01). For the relationship between the absolute peak V̇O₂ and the CCF_max of HR, a significant correlation was found at 35 °C (r = 0.708, p < 0.01), but not at 15 °C (r = 0.208, p > 0.05) or 25 °C (r = 0.512, p > 0.05). A significant correlation was found between the absolute peak V̇O₂ and the CCF_max of V̇O₂musc at 35 °C (r = 0.712, p < 0.01), but not at 15 °C (r = 0.160, p > 0.05) or 25 °C (r = 0.093, p > 0.05). The relative peak power output showed a significant correlation with the CCF_max of V̇O₂musc at 35 °C (r = 0.666, p < 0.01), but not at 15 °C (r = 0.145, p > 0.05) or 25 °C (r = 0.049, p > 0.05). Between the peak CBT and the relative peak power output, a correlation was observed at 35 °C (r = 0.681, p < 0.05), but not at 15 °C (r = 0.142, p > 0.05) or 25 °C (r = 0.388, p > 0.05).

Discussion

To the best of our knowledge, this study is the first to investigate the impact of different acute ambient temperature exposures on cardio-pulmonary and respiratory kinetics during moderate leg exercise using a climatic chamber. In particular, the discrimination between the pulmonary and the muscle responses in V̇O₂ kinetics provides sophisticated insights into the thermoregulatory interactions between the involved physiological systems. The most significant findings can be summarised as the following:

1. During dynamic moderate exercise, in each case, no significant differences in HR, V̇O₂pulm or V̇O₂musc kinetics could be identified across the different temperature exposures in a climatic chamber.

2. Baroreflex regulation based on HR kinetics does not appear to be influenced by different temperature conditions.

3. HR and especially V̇O₂pulm kinetics show significant deviations from the first-order system properties.

4. V̇O₂musc kinetics show the smallest deviations from the first-order system properties.

5. Virtual V_v shows a tendency for an effect of temperature.

6. Absolute CBT is reduced at 35 °C in contrast to 15 °C and 25 °C during moderate exercise.

Peak values

The estimated peak values must be reviewed with caution because the subjects were indeed stressed to the submaximal intensity domain but were not ultimately exhausted. This was due to aborted procedures because of reaching symptom-limited criteria (e.g. predicted maximal HR) or subjects’ request for recovery. Therefore, we chose not to compare our absolute peak data with the literature.

Thermoregulatory responses

HR showed significantly higher values at 35 °C in contrast to the 15 °C ambient temperature condition. This is due to the hyperthermia-induced tachycardia, which is basically described in the literature [37,38]. However, between 25 °C and 35 °C, there were no significant differences in HR between the two conditions, illustrating that thermoregulatory responses by means of increased HR become apparent with higher increasing ambient temperatures. With increasing ambient temperatures, CBT also increases due to the accumulation of heat from the exercising tissues. To prevent heat strain, Q̇ is redistributed from the exercising body compartments and the central circulation to non-exercising tissues (e.g. cutaneous vascular beds) to reduce CBT. This results in hyperthermia-induced tachycardia to sustain a sufficient Q̇ during exercise.

In this context, CBT showed significantly lower absolute values at 35 °C compared with 15 °C and 25 °C, respectively. This may be a compensatory effect based on blood flow redistribution towards the external vascular beds, resulting in cooling of the body and a decreased CBT in the higher ambient temperature condition during dynamic moderate exercise.

To establish comparable conditions of HR and CBT, we calculated the relative changes of these parameters as functions of WR. Here, ΔHR and ΔCBT showed no significant differences relative to ambient temperature effects. Additionally, PSI – an estimate for the physiological strain index based on HR and CBT measurements – revealed no systematic differences between the three ambient temperature conditions. In summary, the thermoregulatory responses showed similar influences at the three different ambient temperatures on PSI at moderate exercise intensities up to 150 W. Additionally, there are indications that heat stress has no effect on motor and cognitive functions [39], so motor control can be assumed to be equal during dynamic exercise across the different temperature conditions.

Kinetics analysis

The kinetics analysis, which included CCF_max, CCF_lag and the entire CCF course, revealed no significant differences between the ambient temperature conditions for HR, V̇O₂pulm or for V̇O₂musc in each case. However, it can be recognised that the relative CCF courses of HR, V̇O₂pulm and V̇O₂musc are different. HR and V̇O₂musc seem to be relatively similar, but they are indeed different.

The reduced kinetics responses as reflected by CCF_max- and CCF_lag-values must be seen as simplifications that are accompanied by losses in the kinetics information. Then, the CCF_max- and the CCF_lag-values have only descriptive characteristics, without sufficient significance for the entire kinetics responses. Therefore, it is more valuable to evaluate the entire CCF course as demonstrated in the following.

It was hypothesised that HR kinetics would not be altered at different ambient temperatures, but absolute HR values would be different in the time domain based on Q̇ redistributions. So, absolute HR values and HR kinetics do not seem to be coupled and can therefore be altered independently. It is known that heat stress does not affect the baroreflex regulation of HR in non-exercise conditions [40–42]. Yamazaki and Sone [43] showed that baroreflex sensitivity increased during cooling, but no alterations were observed relative to heating. Assuming that baroreflex regulation would not be altered with the combination of heat and dynamic exercise, HR control would not be influenced either and then HR kinetics would be similar. We have shown this in the present study, but only in the temperature range of 15 °C–35 °C. Therefore, we hypothesise that HR kinetics can serve as a proxy for baroreflex sensitivity in normothermia and in hyperthermia, especially in the range of 15 °C–35 °C.

For V̇O₂pulm, we also hypothesised that there is no Temperature effect between the conditions, which was derived from the literature. The kinetics analysis by Nybo et al. [9] demonstrated that marked hyperthermia does not alter V̇O₂ on-kinetics. Koga et al. [8] showed that increased muscle temperature had no effect on V̇O₂-kinetics, both in moderate and in heavy exercise domains. Alternatively, it could be argued that the redistribution of Q̇ and an increased HR at higher ambient temperatures resulted in faster or slower V̇O₂pulm kinetics. The literature and our data suggest that this is not the case, but these alterations may exist. However, their impacts on the V̇O₂pulm measurement could be too small to identify.

One significant novelty of the present study in the field of kinetics analyses is the non-invasive discrimination between V̇O₂pulm and V̇O₂musc kinetics using a climatic chamber. Here, we were able to “remove” the dynamic distortive effects of Q̇ and venous blood transit times on V̇O₂pulm to obtain a clear and non-sluggish V̇O₂musc response. The results imply that V̇O₂musc kinetics showed no significant differences across the acute temperature conditions. This assumption is based on the fact that the exercising leg muscles are equally demanded (e.g. the same recruitment patterns of muscle fibers), resulting in similar metabolic rates and also in the same kinetics responses. It can also be assumed that muscle temperature is not influenced, but if it is affected, there are no significant alterations in V̇O₂musc kinetics. Additionally, we ensured that all participants were tested within 2 weeks and that no training procedures were performed, which could speed V̇O₂ kinetics. This is supported by the gas exchange threshold (GET) values, which are not significantly different, implying that the aerobic fitness level was not influenced or altered in each subject.

In the literature, it is currently unclear whether V̇O₂musc kinetics behave as a pure first-order system with or without a TD as we will show below. Di Prampero and Margaria [44] demonstrated – consistent with our data – a mono-exponential time course of oxygen consumption (gastrocnemius muscle) during exercise and in recovery without a TD in anaesthetised dogs. In this regard, Behnke et al. [45] illustrated an increased muscle V̇O₂ without a noticeable TD in rats. Mahler [46] described a good approximation of first-order system properties for muscle oxygen consumption in frog sartorius muscles. In contrast, Grassi et al. [19] demonstrated a TD in V̇O₂ during leg exercise in humans. Additionally, Grassi et al. [47,48] showed sluggish V̇O₂ on-kinetics in isolated in situ canine muscles, illustrating deviations from mono-exponential system properties.

In sum, Figures 7 and 8 show that V̇O₂musc kinetics show the best approximation of first-order system properties. For HR, greater deviations have been estimated, followed by V̇O₂pulm, which shows major aberrations from a first-order system.

General discussion

Stroke volume estimation

As demonstrated by Drescher [49], variations in Q̇ (i.e. based on SV estimations) show negligible effects on the assessment of V̇O₂musc kinetics but have effects on V_v and muscular arteriovenous oxygen content difference estimations. Because we applied this procedure in all subjects and across all three ambient temperature conditions, the tests are therefore standardised and should be comparable among each other in the present study. However, one issue must be considered. Although the relative SV estimations among the temperature conditions are reasonable, the absolute SV values and, therefore, the absolute V_v data may not be adequate because SV is used for the estimation of V_v.

Venous blood volume (V_v) and the perfusion of the non-exercising compartments (Q̇_rem)

With elevated ambient temperatures, a tendency towards decreased V_v was observed (15 °C–25 °C: −3.6%; 15 °C–35 °C: −14.0%). At the same time, Q̇_rem showed an increase that was not significant. Both circumstances can be attributed to thermoregulatory adaptations. With increasing ambient temperature, CBT also increases over time. This has an impact on the redistribution of Q̇ to more external body segments, transferring heat from the core to remote body areas and increasing Q̇_rem. Simultaneously, the overall V_v is redistributed to cell interstitials owing to the initial mechanisms of sweating. Additionally, sweating causes fluid loss, which reduces the overall V_v. It must be noted that V_v represents the amount of venous blood volume between the exercising muscles and the lungs only. Therefore, it is a virtual value that cannot be measured practically and serves as more of a vague estimation to illustrate trends or drifts.

Comparisons with women

There is evidence that the thermoregulatory responses of women to combined heat and exercise strain are comparable with those of men [50–53]. However, in women, luteal and pre- and postmenopausal phases must be taken into account because CBT thresholds are altered during these periods [54] and could have an effect on cardio-pulmonary and respiratory responses. Therefore, in these phases, the kinetics responses of HR, V̇O₂pulm and V̇O₂musc may be different in women compared with those of men.

Conclusions

The results imply that (a) ambient temperature (range: 15 °C–35 °C) has no effect on HR, V̇O₂pulm or V̇O₂musc kinetics at moderate exercise intensities (30–80 W) during dynamic exercise, (b) baroreflex regulation based on HR kinetics does not seem to be influenced by different temperature conditions, (c) V̇O₂musc kinetics show good agreement with first-order system properties in contrast to HR and V̇O₂pulm kinetics, which show greater deviations and (d) the application of the non-invasive method presented here may allow a rough estimation of V_v alterations based on trends.

For practical application, the applied approach may be of interest for assessments of the cardio-pulmonary and respiratory health statuses of individuals working or performing sports in extreme temperature environments. Furthermore, differentiation between systemic (e.g. cardio-dynamic: HR) and peripheral (e.g. exercising tissues: V̇O₂musc) determinants of the involved physiological systems may improve the evaluation of an individual’s health status. This could be beneficial for improved cardio-pulmonary and respiratory fitness via therapeutic interventions for participants suffering from chronic diseases (e.g. diabetes mellitus) or after injuries and surgeries because potential prolonged bed rest phases could result in common fitness losses.

Furthermore, for exercise or work in warm and hot environments, the cardiodynamic and metabolic responses at the muscular level do not seem to be influenced by ambient temperature at moderate exercise intensity if hydration status can be controlled with appropriate water intake and if sweating is not negatively influenced by clothing or equipment.

Acknowledgements

The authors thank Stephan Klatt for support during the exercise tests and Michael Voigt for data preparation. The authors would also like to thank all the participating subjects from the military services and employees of the German Federal Armed Forces.

Disclosure statement

The authors declare that they have no conflict of interest.

Additional information

Funding

The study was supported by the German Space Agency (DLR – Deutsches Zentrum für Luft- und Raumfahrt [grant No. FKZ 50WB1426]), Center for Health and Integrative Physiology in Space (CHIPS, German Sport University Cologne) and German Federal Armed Forces.