Pre-cooling for endurance exercise performance in the heat: a systematic review

Different pre-cooling interventions are proposed to act via different mechanisms to reduce core body temperature and thus cool the body prior to exercise.

Repeated measures crossover studies and randomized controlled trials comparing a pre-cooling method(s) to control or no intervention in healthy adults were considered for inclusion. The pre-cooling method could be any that cooled a participant prior to commencing an endurance exercise protocol or event. A measure of aerobic endurance was required to be one of the outcome measures in each study. The ambient environmental temperature during the performance trials had to be at or above the human thermoneutral zone of environmental temperatures (≥ 28°C) [32, 33].

Unpublished studies, case series studies, non-peer-reviewed publications, studies not involving humans, reviews, letters, opinion articles, articles and abstracts not in English were excluded. Studies that included participants with pathological conditions known to increase susceptibility to heat strain, such as spinal cord injury [34], were also excluded, as were studies that attempted to cool participants during exercise, those that used intermittent or team-based sport exercise protocols or protocols that primarily stressed the anaerobic energy pathway. Unpublished research was not sought. Although this may potentially lead to publication bias [35], it was deemed impractical to identify all unpublished work on pre-cooling and endurance exercise performance from all authors and institutions around the world.

The following databases were searched in May 2012 (week 4): MEDLINE (Ovid Web, 1948 to 2012 and Medline In-process and Other Non-Indexed Citations), EMBASE (1974 to 2012), CINAHL (1981 to 2012), Web of Science (1899 to 2012) and SPORTDiscus. Key terms used in the search strategy and results of the search are shown in Table 1. Reference lists and lists of citing articles were searched to ensure that no relevant studies had been missed by the search strategy. No additional papers were identified.

All titles and abstracts were downloaded into EndNote X4 (Thomson Reuters, Philadelphia, PA, USA) giving a set of 9922 citations. The set was crossreferenced and any duplicates were deleted, leaving a total of 4454 citations. Each title and abstract were evaluated for potential inclusion by two independent reviewers (PRJ and CB) using a checklist developed from the inclusion/exclusion criteria outlined above. If insufficient information was contained in the title and abstract to make a decision on a study, it was retained until the full text could be obtained for evaluation. Any disagreements regarding studies were resolved by a consensus meeting between the two reviewers, and a third reviewer (DM) was available if necessary.

Quality assessment was performed using the Physiotherapy Evidence Database (PEDro) Scale, which is a valid measure of the methodological quality of clinical trials [36]. Each study is rated according to ten separate criteria on the PEDro scale that assess a study's internal validity and statistical reporting, then totaled to give a score out of 10. An additional criterion, 'sample size calculation', was included in the quality assessment as the authors felt it to be an important component of study methodology. This criterion did not contribute to the PEDro score. Two reviewers (PRJ and CB) applied the PEDro scale to each included study independently, and any scoring discrepancies were resolved through a consensus meeting, with a third reviewer (DM) available if necessary. Studies were considered high quality if PEDro scores were greater than 6, and low quality if 6 and below.

Means and standard deviations for all continuous data were extracted and effect sizes (Cohen's d) (with 95% CIs) calculated to allow comparison between the results of each study. Data were pooled using RevMan for Mac version 5.1.2 (The Nordic Cochrane Centre, Copenhagen, Denmark). If inadequate data were available from original studies to complete effect size calculations, attempts were made via email to contact the study's corresponding author for the required data. The presence of publication bias was determined by evaluating funnel plot asymmetry graphically [37, 38].

Definitions for 'levels of evidence' were guided by recommendations made by van Tulder et al. [39] and are as given below:

'Strong evidence' = consistent findings among multiple high quality randomized controlled trials.

'Moderate evidence' = consistent findings among multiple low quality randomized controlled trials and/or non-randomized controlled trials, or one high quality randomized controlled trial.

'Limited evidence' = findings from one low quality randomized or non-randomized controlled trial.

'Conflicting evidence' = inconsistent findings among multiple trials.

Of the 13 studies included in the review, 8 studies attained a PEDro score of 6/10 [16‐19, 26, 27, 40, 41], 4 attained a score of 5/10 [24, 25, 42, 43], and 1 study received a score of 4/10 (Table 3) [25]. Sample size calculations were not performed by any of the reviewed studies.

Corresponding authors of two additional studies eligible for review were contacted via email to request additional data necessary for inclusion in the review [44, 45]. The required data had not been supplied at the time of going to press. A symmetrical funnel plot indicated the absence of publication bias [37, 38].

Each included study used a repeated measures crossover design. However, methodological quality was varied with PEDro scores ranging from 4/10 to 6/10, indicating no high quality randomized controlled trials evaluating the effectiveness of pre-cooling to improve endurance exercise performance in the heat. Some studies did not randomize participant allocation, possibly introducing allocation bias [24, 25, 28, 42, 43]. All except four studies [18, 19, 25, 43] used participants who were moderately to well trained (Table 4) in sports with high endurance components (cycling, triathlon and distance running), and within that only cycling and running exercise protocols were used, limiting the applicability of the findings to the broader, less well trained population. Lack of participant, investigator and outcome assessor blinding was consistent across all studies, likely due to practical difficulties. Consequently, some results could have been unintentionally biased, either by observer bias, such as encouraging participants in the pre-cooled group, or a placebo effect.

Participant numbers in each study were low, ranging from 6 [27] to 20 [26], limiting the validity of conclusions that can be drawn from the results. None of the reviewed studies performed sample size calculations, and therefore certain data trends could not be substantiated due to inadequate statistical power. There was a high level of methodological heterogeneity between studies, including: exercise performance protocol, pre-cooling duration, exercise duration and outcome measure, making comparison of studies and recommendations for enhancing sporting performance difficult. This was further compounded by the absence of comparisons between the three main individual pre-cooling maneuvers (cold water immersion, cooling garment and ice slurry ingestion) in all but one study [19]. Therefore, the relative efficacy and practicality of one pre-cooling method to another could not be made. In one study, subjects exercised at 60% VO₂max for 60 minutes followed by an effort at 80% VO₂max to volitional fatigue [43]. However, mean performance time ± standard error were only reported for the short effort at 80% VO₂max to fatigue. This is likely to have inflated the effect size compared to other studies.

Moderate evidence currently exists to support the use of cold water immersion as a pre-cooling intervention to improve endurance exercise performance in the heat. Three studies showed a significant performance improvement in the pre-cooled compared to control condition [19, 40, 43], with the remaining three studies showing a positive trend to improved performance [28, 41, 42]. In each of the immersion studies there was a significant reduction in core temperature compared to control at some point during the exercise protocol. Additionally, the rate of heat storage was greater in three of the four studies that reported this variable [19, 28, 42], conferring a greater margin for metabolic load during exercise in the pre-cooling condition. Gonzalez-Alonso et al. [40] reported that rate of heat storage was equal between both conditions. However, as the pre-cooled group commenced exercise with a core temperature 1.5°C lower than the control condition, their total heat storage capacity was greater. Although not conclusive evidence of a precise mechanism, it seems that pre-cooling using cold water immersion could possibly improve performance by reducing core temperature prior to exercise, or blunting the rate of rise in core temperature during exercise, increasing heat storage capacity and enabling athletes to perform at a greater relative intensity or for a greater duration [29].

Despite a more rapid reduction in core temperature with water immersion compared with traditional cold air exposure [46], the required length of pre-cooling remains significant (30 to 60 minutes) [29, 30]. Marino and Booth [21], in one of the first studies investigating the potential use of pre-cooling via cold water immersion prior to endurance exercise, reduced core temperature by gradually reducing the temperature of the immersion bath over a 60-minute period. This was to avoid the potentially detrimental cold stress responses that had previously been seen with cold air exposure, such as shivering [29]. Such a regimented technique, which also precludes a concomitant warm-up, is limited in its practicality in an elite sports setting immediately prior to athletic competition, in addition to other logistical issues such as expense, transportation of equipment, and access to such a large volume of water and electricity in the field.

Limited evidence currently exists to support the use of ice slurry as a pre-cooling intervention to improve endurance exercise performance in the heat. One study [18] showed a significant performance improvement in the ice slurry ingestion pre-cooled compared to the control condition and the three remaining studies showed a positive trend to improved performance [16, 17, 19]. Each study reported that core temperature was significantly lower in the pre-cooling condition than control after the cooling intervention and prior to the start of the exercise task, increasing heat storage capacity. Alternatively, the participants' lower core body temperatures prior to exercise may have enabled them to select a faster pacing strategy by influencing central regulation of exercise intensity [47].

Two studies [18, 19] reported that the pre-cooled group exhibited a significantly higher core temperature at exhaustion. The authors suggest that this could be due to the generation of higher metabolic heat loads as a result of either a direct cooling effect on the brain, or an effect on core temperature afferent nerves [48], altering perception of effort and increasing time to exhaustion. Increased core temperature above normal tolerable limits is an important safety consideration and may be detrimental to athlete health, increasing the risk of heat-related illness, and is something that requires attention in future studies.

Ice slurry ingestion offers a number of practical benefits over cold water immersion, as it is not subject to the same logistical restrictions. The ice slurry can be produced using a commercially available machine or simply freezing and part-thawing sports drinks prior to the event, and transporting them in a cool box. This is particularly useful at events where there is no provision for electrical equipment, or where transportation is an issue. Pre-cooling athletes in this way is quick and simple. The amount of ice slurry required to achieve effective cooling is low and similar in volume to pre-exercise fluid hydration protocols, ranging from 6.8 g/kg [16] to 14 g/kg [17] of body mass. In each reviewed study, the volume of ice slurry was administered over a 30-minute period at a standardized rate that ranged from 5 [18, 19] to 15 minutes [17]. Although not yet investigated, there is the potential that ice slurry ingestion could enable athletes to warm-up during cooling, making it much more time efficient than cold water immersion. In addition to providing a greater cooling effect than cold water alone [23], a much smaller volume is required to produce this response, reducing the potential for detrimental effects that the ingestion of large volumes of fluid may have. As well as cooling athletes, the ice slurry can be used to hydrate athletes too so that combined fluid and slurry ingestion is not necessary.

None of the studies showed a significant improvement of wearing a cooling garment on subsequent exercise performance [24, 26, 27]. This likely resulted from the lack of effect on core body temperature. In two studies [26, 27], despite the pre-cooling groups having significantly lower skin temperatures while wearing the cooling garment, core temperature was not significantly lower at any time point during either pre-cooling or subsequent exercise. Arngrïmsson et al. [24] reported significantly lower rectal temperatures in the pre-cooling group for the last 18 minutes of the warm-up and first 3.2 km of the running exercise task compared to the control group. However, this effect was not strong enough to have caused a significant improvement in performance and may have resulted from the high rectal temperatures, and therefore reduced heat storage capacity, at the start of the performance task in both the cooling garment (38.0°C) and control group (38.2°C) compared to all other studies that reported rectal temperature at the onset of the exercise task [18, 19, 25, 27, 28, 41‐43].

Kay et al. [28] suggested a reduction in core body temperature achieved through lowering skin temperature, effecting heat loss from core to skin. This is the mechanism by which cooling garments are believed to act to cool athletes prior to exercise. However, cooling in Kay et al.'s [28] study was achieved via whole-body cold water immersion, which likely provided a greater cooling stimulus than cooling garments, especially at peripheral areas of the body. This could explain why cooling garments were found to have little effect on core body temperature in the present study. One study reported than the application of a cooling garment reduced skin blood flow across the body by stimulating vasoconstriction, preventing efficient heat transfer between the skin and the cooling garment. Core body temperature of subjects remained unaltered, likely from the redistribution of blood to the core [44]. If the hypothesis that a critical core temperature limits exercise performance in the heat is correct, then, by failing to reduce core body temperature cooling garments were unable to improve endurance exercise performance.

Some studies combined more than one pre-cooling intervention to cool participants prior to the exercise component of the trial. Two studies immersed subjects in cold water, followed by a period wearing a cooling garment [17, 27]. Quod et al. [27] reported a significant decrease in core body temperature prior to exercise, likely as a result of an 'afterdrop' effect [49]; that is, a continued fall in core temperature after the initial hypothermic exposure, as opposed to any further cooling effect of the garment. Indeed, the same study reported that wearing the cooling garment alone failed to reduce core temperature compared to control. Exercise performance was significantly better than control and cooling garment conditions. Conversely, Ross et al. [17] reported that, despite a significantly lower core body temperature after cooling and throughout warm-up compared to controls, there was no improvement in performance. The authors suggest that the larger cooling response in the combined condition may have led the athletes to select poorer pacing strategies. An alternative explanation could be that the cold water immersion protocol used may have been too abrupt compared to that used in other studies [21], and may therefore have elicited a cold stress response that was detrimental to performance, similar to that reported for cold air exposure [29].

A combination of cold air and a cooling garment, with or without thigh cooling, showed trends to improved performance in both conditions compared to control in one study [25]. Both pre-cooling groups had a lower core temperature after pre-cooling, and power output was significantly greater compared to controls during the 15-minute performance trial. There was no difference in power output between the two cooling conditions. It is difficult to determine whether the cooling garment conferred any additional benefits than have been shown to be conferred by cold air cooling alone [12, 13, 50]. In practice, cold air cooling has a number of logistical limitations including equipment transport and cost, the significant time required to adequately cool athletes, and a noted cold stress response that can impair exercise performance [29].

There is a high level of heterogeneity in study design examining the effectiveness of pre-cooling strategies, and optimal cooling protocols have yet to be established. Variables such as cooling duration and time between pre-cooling and commencing exercise are likely to exert considerable influence on study outcomes and require greater attention. Once repeatable pre-cooling protocols have been identified for each individual modality, then more reliable comparisons of effectiveness can be made between modalities. One study directly compared ice slurry ingestion to cold water immersion and found it to be similarly effective at improving performance (Figure 6) [19]. As potentially the cheaper, more practical strategy, this result is encouraging and warrants further investigation of ice slurry ingestion in the field. Additionally, following Quod et al.'s study [27], it would be instructive to determine whether the combination of a cooling garment following cold water immersion confers any additional benefit compared to immersion only.

Hydration strategies employed, and reporting of these strategies was inconsistent (Table 5). Water ingestion, especially cool water, may lower core body temperature via a similar mechanism as ice slurry ingestion. Potentially, if control participants were permitted to drink cool water either before or throughout the exercise trial this may confound the effectiveness of the pre-cooling strategy. However, this is a difficult variable to control for and depends on the comparison being made. For example, studies investigating ice slurry ingestion used water ingestion of an equal volume as the control condition to determine that any improvements in performance were a result of the pre-cooling effect of ice slurry ingestion as opposed to the ergogenic effect of adequate hydration [51]. Interestingly, Siegel et al. found a greater effect of ice slurry ingestion on performance when compared to controls drinking cool fluid (4°C) [18] than when compared to controls drinking warmer fluid (37°C) [19], which suggests that cool water ingestion may not blunt the effectiveness of pre-cooling as much as expected. However, more consistent hydration protocols will enable greater analysis of this relationship. Hasegawa et al. [43] reported that continuous cool water ingestion during exercise following cold water immersion significantly improved performance compared to cold water immersion alone and negated the rise in core body temperature towards the end of the performance protocol. The authors attributed this to increased evaporative sweat loss, sweat efficiency and decreased heat strain in the continuous water ingestion group. This finding suggests that the benefits of pre-cooling may be augmented by maintaining hydration during exercise. Ice slurry ingestion acts to pre-cool athletes and could also be used to maintain cooling and hydration during exercise. Therefore, a comparison of combined cold water immersion and water beverage with continuous ice slurry ingestion is warranted.

The level of fitness of participants varied across studies as did the consistency of reporting of fitness and experience of endurance exercise (Table 4). It is difficult, therefore, to determine whether more experienced or less experienced athletes would benefit more from pre-cooling. Furthermore, those who are less experienced are likely to be less accurate when anticipating a required pacing strategy to complete a given exercise trial [47]. By lowering core body temperature using exogenous means, participants may perceive their level of exertion to be lower than their body's thermal load should dictate, that is, a discrepancy between their perceived and actual homeostatic state, which could cause them to develop heat illness due to the masking of thermal strain. This is acknowledged in the two studies that reported participants to have an elevated core body temperature at volitional fatigue [18, 19]. Notably, these studies used untrained participants so it is possible that more experienced athletes may be better attuned to their physiological limits and hence less at risk of heat illness, but this is speculative and warrants further investigation before ice slurry ingestion can be recommended.

Given the lack of blinding of participants and researchers in the reviewed studies, a placebo effect cannot be excluded from having influenced results. Future studies should consider introducing a separate, placebo-controlled group, and participant and assessor blinding to improve methodological validity. The placebo-control could, for example, use menthol to provide a cooling sensation for participants without causing an actual change in temperature [52].

Each study included in this review was limited by low participant numbers. It was therefore difficult to determine whether certain reported trends or lack thereof were the result of the studies being underpowered. A priori power calculations should be performed to increase the statistical significance of any trends reported in the results. Study participants were predominantly male, therefore the findings of this review may not be applicable to females, especially because certain anthropometric and hormonal differences, including stage of the menstrual cycle [53] and body composition [54], can affect thermoregulation under heat stress. It should be acknowledged that the majority were performed with the intention of applying the findings to highly trained athletes, and that recruiting large numbers of such compliant volunteers is difficult. However, the inclusion of larger sample sizes and inclusion of similar proportions of female and male participants in future research will allow both improved external validity for broader populations and between sex comparisons to be made.

Laboratory studies grant assessors strict control of certain variables, such as the environmental conditions under which exercise is performed, which is necessary with preliminary studies to establish intervention efficacy and optimal protocols. However, future studies of pre-cooling should focus on real-world testing to determine whether the promising laboratory findings translate to tangible performance gains in the field. This is also important to evaluate the practicality of each method of pre-cooling during competition.

There was also a lack of safety or adverse event reporting. It remains unknown what effect increased heat storage capacity may have on other bodily systems other than those directly involved in thermoregulation. Therefore, until these can be elucidated, it would be prudent for future research to include consideration of athlete safety, as this will be of primary concern to coaches and athletes alike, given the physiologically stressful environment in which they will be competing.

Although consistently the most effective method of pre-cooling and enhancing endurance exercise performance in the heat, cold water immersion has limited practicality in sporting settings due to expense, transportation issues, difficulty accessing large volumes of water and time required to achieve a reduction in core body temperature. Ice slurry ingestion is a relatively cheap and much more practical alternative to whole body immersion and effectively lowers core body temperature, approaching the improvements in performance seen with immersion. Additionally, there is currently limited evidence from one study that indicates the effectiveness of ice slurry ingestion and cold water immersion are comparable [19]. However, safety concerns raised by two studies that reported a raised core body temperature at volitional fatigue need to be addressed before its use can be recommended for competing athletes. Cooling garments failed to improve endurance exercise performance and therefore are of limited use in this regard.