The time course of adaptations in thermoneutral maximal oxygen consumption following heat acclimation

Heat acclimation (HA) describes a systematic process, whereby serial exposures to artificially heated environments, often in combination with exercise, can lead to rapid adaptations that enhance the capacity to thermoregulate in the heat (Gibson et al. 2015; Taylor and Cotter 2006). These adaptations improve heat tolerance (Sawka et al. 2011), characterised by increased sudomotor function (Fox et al. 1963), reduced heart rate (HR) (Senay et al. 1976), hypervolemia (Nielsen et al. 1993), increased cardiac output (Lorenzo et al. 2010) and a reduced core body temperature (Tc) for a given heat exposure or exercise intensity (Nielsen et al. 1993). This combination of physiological adaptations can enhance endurance performance in hot (Racinais et al. 2015) and thermoneutral environments (Lorenzo et al. 2010).

The maximal rate of oxygen uptake (\(\dot{V}\)O_2max) is an important determinant of endurance performance, explaining ~ 20–60% of the variation in performances of different mode and distance (Coyle et al. 1988; Schabort et al. 2000; Jacobs et al. 2011). \(\dot{V}\)O_2max is chiefly limited by central cardiovascular factors, such as O₂ transport (Bassett and Howley 2000). Heat acclimation can improve \(\dot{V}\)O_2max, and is thought to occur owing to heat-induced cardiovascular changes (Périard et al. 2016). Indeed, there have been historical observations of a relationship between \(\dot{V}\)O_2max and heat tolerance (Shvartz et al. 1978; Pandolf 1979; Havenith and van Middendorp 1990). However, evidence of the efficacy of HA on \(\dot{V}\)O_2max is equivocal, with a number of studies demonstrating 4–13% changes (Nadel et al. 1974; Shvartz et al. 1977; Sawka et al. 1985; Pivarnik et al. 1987; Lorenzo et al. 2010; James et al. 2017) and others reporting no change or a reduction following a range of HA protocols (Houmard et al. 1990; Gore et al. 1997; Chen et al. 2013; Karlsen et al. 2015; Keiser et al. 2015; Neal et al. 2016a, b; Rendell et al. 2017; Sotiridis et al. 2018).

There are various explanations for discrepancies in the aforementioned findings. First, not all studies have evaluated adaptations in \(\dot{V}\)O_2max in thermoneutral conditions, thus limiting the understanding of a transfer between heat-induced adaptation and aerobic capacity in temperate environments, which is a topic of ongoing debate (Nybo and Lundby 2016). This could be related to limited \(\dot{V}\)O_2max trainability (Bouchard et al. 2011) or inter-individual and inter-system differences in adaptation reported after 11 days of isothermal acclimation (Corbett et al. 2018). It is therefore possible that \(\dot{V}\)O_2max increases more substantially among HA-responsive individuals compared to their less responsive counterparts (Minson and Cotter 2016). Daanen et al. (2011) also demonstrated that the acute stress imposed by HA acted, initially, to suppress the adaptation in Tc, before adaptation beyond baseline in the days following acclimation. Thus, a period of recovery (from heat, exercise or both) might be necessary to fully realise adaptation in multi-system physiological measurements, such as \(\dot{V}\)O_2max. This reasoning is consistent with the tenets of general adaptation syndrome (Selye 1950), which has been incorporated into exercise training guidelines (ACSM 2009). However, there has been no study to evaluate the detailed time course of change in thermoneutral \(\dot{V}\)O_2max after HA. Therefore, it is possible that previous studies have not: (1) monitored the inter-individual time course of responses of \(\dot{V}\)O_2max to HA across successive days and (2) provided adequate time for adaptation to the combined thermal and exercise training stimuli.

Based on this reasoning, we investigated the effects of a 10-day HA programme on the time course of changes in thermoneutral \(\dot{V}\)O_2max, during and up to 10 days post-HA against a control training group. We hypothesised that increases in \(\dot{V}\)O_2max would occur in the HA group in the days following the intervention but presumed variability between individuals in the time course of this response in the post-HA 10-day period.

We are the first to evaluate the time course of changes in thermoneutral \(\dot{V}\)O_2max adaptation during and after a HA programme. Changes in \(\dot{V}\)O_2max were elicited in a non-linear manner during the 20-day (combined HA and post-HA) period, with acute reductions apparent at HA days 5 and 10 among both groups. Delayed increase in thermoneutral \(\dot{V}\)O_2max occurred in the post-HA (recovery) period. Specifically, by post-HA day 4, there were pairwise differences between the HA and control groups, where mean changes in \(\dot{V}\)O_2max peaked (4.9% increase from baseline) for the HA group. The control group’s \(\dot{V}\)O_2max fluctuated between − 4.5 and 0.3% approximating the 3.0% error of the test. There were notable inter-individual differences in the time course of adaptation among the HA group, yet only 1 of the 12 participants did not respond to the HA programme, demonstrating their highest value prior to the post-HA period. These adaptations were supported by larger plasma volume expansion in the HA group, which was maintained across the study period.

There are a number of studies demonstrating HA-induced increase in thermoneutral \(\dot{V}\)O_2max, ranging between 4 and 13% (Nadel et al. 1974; Shvartz et al. 1977; Sawka et al. 1985; Pivarnik et al. 1987; Lorenzo et al. 2010; James et al. 2017). Similarly, we report a 4.9% increase on post-HA day 4 compared to baseline, which is substantial, given the limited potential for adaptation in \(\dot{V}\)O_2max (Bouchard et al. 2011) and the trained status of the participants. Furthermore, this value was increased to a mean of 7.1% if the largest individual changes in \(\dot{V}\)O_2max observed across the study are accounted for. These data suggest that the changes reported previously are likely to have underestimated the full adaptation potential, owing to inter-individual variability in this measure. Our findings highlight that the timing of the post-acclimation testing is crucial to the outcome, despite most studies not precisely reporting this. Based on the current data, we recommend that post-acclimation \(\dot{V}\)O_2max testing is not performed within the post-HA 48 h and that a 96-h post-test is likely to produce the highest value for most individuals. Of note, the one participant’s \(\dot{V}\)O_2max that did not peak in the post-HA period was highest at baseline and remained lower for the entire 20-day period. This denotes an important individual feature of training adaptations and questions rigid training models. As reported elsewhere, it is also possible that some individuals experience a delayed response (absence of decay) to heat acclimation (Daanen et al. 2011) of up to 26 days (Weller et al. 2007), which would not have been captured in our 10-day post-HA window and is one explanation for the participant’s response in the current study. This descriptive analysis provides insight into the dynamics of heat-induced adaptation in \(\dot{V}\)O_2max between individuals and could lead to imprecise reporting in studies of this type, depending on the selected post-testing period.

As noted previously (Armstrong and Maresh 1991; Nielsen et al. 1993), there were some characteristic signs of early adaptation demonstrated across the sub-maximal HA sessions, with Tc, HR and Ts declining across the first 3 days in the HA group. These changes indicate an improved tolerance to the exercise and thermal stimuli. However, the time course of \(\dot{V}\)O_2max adaptation to HA was not as immediate and, therefore, did not coincide with early adaptations in Tc, HR and Ts. This is important to recognise because HA-induced changes in \(\dot{V}\)O_2max are typically ascribed to central cardiovascular adaptations that are responsible for O₂ transport. For example, \(\dot{V}\)O_2max is chiefly determined by the delivery of oxygenated blood to the working musculature and relies on factors, such as stroke volume and muscle capillarisation, to support this (Bassett and Howley 2000). The lower HR responses to the HA programme reported here and elsewhere (Nielsen et al. 1993; Lorenzo et al. 2010), and increased plasma volume, are consistent with an improved cardiac efficiency, which is commonly observed following HA (Périard et al. 2016). These changes also infer an increase in stroke volume and cardiac output (Senay et al. 1976; Nielsen et al. 1997). Thus, the temporal differences in \(\dot{V}\)O_2max adaptation, highlighted by acute decline in \(\dot{V}\)O_2max, must be attributed to other factors in the O₂ transport pathway, perhaps indicating an acute impairment and perfusion capacity of muscle microvasculature function during the HA programme. There have been reports of increased vascular conductance after 10–14 days of HA in the peripheral cutaneous microvasculature (Lorenzo and Minson 2010) and others have reported changes in O₂ pulse at high exercise intensities, which were ascribed to increase in arterial–venous O₂ differences (Chen et al. 2013). Together, these findings indicate that HA induces an increase in peripheral blood flow that could lead to enhanced perfusion of blood in the working muscles. However, there has been no study of the time course of this adaptation to monitor early-phase changes in the microvasculature of the skeletal muscle. Further work is required to elucidate this.

\(\dot{V}\)O_2max relies on multiple physiological factors; therefore, a less predictable time course of adaptation might have been anticipated. General adaptation syndrome describes a process of systematic overload to impose controlled ‘stress’, leading to changes in that system, such that subsequent exposures to the same stress are more tolerable (Selye 1950). Indeed, the discordant temporal adaptations between different physiological processes observed here are supported by Selye’s (1938) concept of ‘adaptation energy’, whereby simultaneous adaptations to given stimuli are constrained by finite energy resources, which was originally demonstrated by exposing rats to cold temperatures and exercise. Similar adaptation kinetics have been demonstrated in the exercise training literature, where 1–3 weeks of overreaching in cyclists acutely reduced \(\dot{V}\)O_2max prior to rebound improvements in the recovery period (Jeukendrup et al. 1992; Aubry et al. 2014). These changes have not been fully explained but attributed to a general training fatigue or psychological factors that ensue during heavy training periods. This theory has also been previously applied to the process of HA, where Daanen et al. (2011) reported lower resting and exercising Tc—a principal feature of adaptation to heat exposure—to manifest most prominently during the post-HA period. This had been reported in other investigations (Pandolf et al. 1977) and was explained by the severity of the imposed thermal stimulus. It is plausible that the same reasoning applies to the current data, since the participants trained for the duration of the study, in addition to the thermal load and serial \(\dot{V}\)O_2max testing. However, the absence of significant \(\dot{V}\)O_2max changes over time in the control group or differences in training characteristics compared to the HA group suggests that the thermal stimulus was the responsible factor. Alternatively, it is possible that kinetics of \(\dot{V}\)O_2max adaptation are different to other measures but has not yet been considered with the necessary scrutiny.

The suggestion that \(\dot{V}\)O_2max adaptation occurs in the post-HA period is at odds with the characteristic rapid decay reported in other physiological measures, such as Tc and HR (Garrett et al. 2009, 2011; Daanen et al. 2018). Indeed, it is the concern over rapid post-HA decay that has most likely prompted testing in the following 72-h period (i.e., Sotiridis et al. 2018). Of note, some studies have reported descriptive increase in \(\dot{V}\)O_2max of up to 3.5% and significant increase in end test power of 8.5% (Sotiridis et al. 2018), yet re-tested across 3 consecutive days, immediately following HA. Based on the current data, \(\dot{V}\)O_2max does not appear to enter a period of sustained ‘decay’ for the first 4 days, and perhaps up to 10 days, after HA. Future studies should focus on manipulating the intensity of the stimulus and extend the period of post-HA monitoring to establish the full time course of decay in \(\dot{V}\)O_2max. In this regard, the chosen HA protocol followed a so-called ‘fixed-intensity’ model, which is not thought to provide an optimal stimulus for adaptation to the heat, as Tc is uncontrolled by the investigator (Daanen et al. 2018). Indeed, others have suggested that acclimation > 38.5 °C Tc is necessary for adaptation (Fox et al. 1963). However, recent studies have questioned this recommendation, reporting no relationship between thermal load (time spent > 38.5 °C) and changes in Tc or HR (Corbett et al. 2018). Others have used fixed-intensity HA programmes to induce changes in \(\dot{V}\)O_2max (Lorenzo et al. 2010) and more recent studies have demonstrated similar heat adaptations between isothermal or fixed-intensity approaches (Gibson et al. 2015). There are some potential advantages to adopting a fixed-intensity protocol. For example, the absolute power output and relative intensity sustained in a fixed-intensity protocol is higher than reported in isothermal models (Gibson et al. 2015), which might be of greater importance when a gross outcome (\(\dot{V}\)O_2max) is targeted, requiring a mixture of thermal and exercise stimuli. The Tc and HR responses were also significantly higher than the control group across all sessions, with mean Tc values beginning at a mean of 39 °C and finishing marginally above 38.2 °C on day 10. Furthermore, this model can be practically simpler to run with groups of athletes, which is often necessary for applied practitioners. Therefore, the approach we have adopted appears to have provided sufficient stimulus for adaptation in \(\dot{V}\)O_2max and the delayed increase could, theoretically, suit the needs of individuals who cannot acclimate for the entire pre-competition training camp, owing to travel arrangements or the details of their pre-event taper.

There are some limitations to the current study, such as the types of athletes used, who were all cyclists and well trained. While we do not anticipate that delayed \(\dot{V}\)O_2max adaptation will be specific to the mode of exercise, it is possible that less trained participants would demonstrate a different time course of adaptation to training stimuli and that the lower aerobic capacity might prevent them from fully engaging in the intensive acclimation process. Furthermore, all of the cyclists underwent an acclimation intervention, rather than acclimatisation, where there is typically less control over the ambient conditions (i.e., heat, humidity and wind speed). Thus, we cannot be certain that the same effects would be observed in a natural environment.

The time course of \(\dot{V}\)O_2max adaptation is non-uniform, mimicking a typical supercompensation response. These findings have implications for researchers and athletes wishing to evaluate adaptations in \(\dot{V}\)O_2max after performing HA programmes. We advise that post-testing is not performed in the following 48 h and that 96 h provides the peak adaptation for most individuals. Whilst the current findings are of fundamental scientific interest, there are also many practical advantages of using heat acclimation to enhance thermoneutral aerobic capacity. This includes the ability to induce a variety of physiological adaptations (reduced Tc and HR, increased sweat response), alongside changes in \(\dot{V}\)O_2max, in response to a training programme.