Pre-cooling for football training and competition in hot and humid conditions
Abstract
Pre-cooling studies report positive physiological and performance benefits in laboratory conditions, although research studies have not investigated these reported benefits in ecologically valid team-sport training and competition settings. Accordingly, this study investigated the effect of field-based pre-cooling strategies for professional football players during training and competition in the heat. Ten professional football players from an Australian A-League club performed two training sessions and competitive matches in hot ambient conditions (29±3oC, 78±8% relative humidity) with or without pre-cooling. The pre-cooling intervention involved 20-min of an ice-vest, cold towels and 350 mL ice-slushie drink. Training sessions (n=9) were randomised, and consisted of 2 × 10-min interval training, followed by 6 × 3-min of 5v5 small sided games. Competitions (n=7) involved official A-League matches during the 2009–10 season. Player movement characteristics, core temperature (gastrointestinal), skin temperature, nude mass, heart rate, capillary blood (glucose, K+, Na+, haematocrit), perceptual exertion and thermal stress measures were recorded. No significant differences (P>0.05) were present between conditions for any measure of physical performance, although moderate-large effects for a greater total and relative distance covered during training were present (d > 0.8). While mean skin temperature was reduced following cooling, core temperature was only lower until following the warm-up in training and was even less evident during matches (P>0.05; d < 0.6). However, a smaller change in mass (sweat loss) and reduced perceptual exertion and thermal stress were evident during training following cooling (d > 0.9), although again, to a much lesser extent in matches (d = 0.6). In conclusion, equivocal findings were present for the effects of pre-cooling for professional football players during competitive training and matches in the heat. However, performance and thermoregulatory response trends showed similarities to previous laboratory evidence. The field-based nature of the current study may highlight that the transfer of lab findings to field settings is difficult or the strength of the intervention is diminished by the settings.
Introduction
Exercise performed in high ambient temperatures and humidity can increase the thermoregulatory, physiological and perceptual loads endured by the athlete (Reilly, Drust, & Gregson, 2006). These increased stresses can reduce exercise performance, either due to the earlier onset of volitional exhaustion (Gonzalez-Alonso et al., 1999; Morris, Nevill, & Williams, 2000) or self-selected reductions in exercise intensity (Kay et al., 2001; Tucker et al., 2006). Accordingly, football training and competition performed in hot/humid conditions may result in exacerbated physiological and perceptual stress and reduced physical performances (Duffield, Coutts, & Quinn, 2009; Edwards & Clark, 2006; Reilly et al., 2006).
Methods to counter heat-induced stress include cooling of skin, core and muscle temperatures and blunting the augmented rise in endogenous thermal stress (Kay, Taaffe, & Marino, 1999; Quod et al., 2008). Pre-cooling has been reported to be beneficial for prolonged continuous (Duffield, Green, Castle, & Maxwell, 2010; Kay et al., 1999; Quod et al., 2008) and intermittent-sprint (Castle et al., 2006; Duffield & Marino, 2007) exercise in warm conditions; possibly via reduced thermoregulatory load, slowing the rise in core temperature (Hessemer, Langusch, Bruck, Bodeker, & Breidenbach, 1984; Price, Boyd, & Goosey-Tolfrey, 2009), and reducing perceptual stress (Duffield & Marino, 2007) during prolonged exercise in the heat. Previous research in laboratory conditions highlight benefits of cooling prior to exercise in the heat (Castle et al., 2006; Duffield et al., 2010; Duffield & Marino, 2007; Hessemer et al., 1984; Kay et al., 1999; Price et al., 2009; Quod et al., 2008); however, evidence for the transfer of these findings to the ecologically valid football environment is limited. Given these previously reported benefits (Duffield et al., 2010; Kay et al., 1999; Quod et al., 2008), pre-cooling may reduce physiological and perceptual loads to improve performance for football training and competition in hot environmental conditions.
Inconsistent findings in the literature regarding the ergogenic benefits of pre-cooling seem to relate to the mode and duration of the type of exercise (Cheung & Robinson, 2004; Duffield et al., 2010; Quod et al., 2008). Some, (Castle et al., 2006; Duffield & Marino, 2007) but not all (Cheung & Robinson, 2004; Drust, Cable, & Reilly, 2000), laboratory-based studies have demonstrated physiological and performance benefits of cooling before intermittent-sprint exercise. Moreover, other studies have investigated the efficacy of pre-cooling during simulated training or competition in team sport athletes (Duffield, Steinbacher, & Fairchild, 2009; Hornery, Farrow, Mujika, & Young, 2007; Price et al., 2009). A recent study showed increased distance covered during free-paced intermittent-sprint exercise in a simulated lacrosse training session (Duffield, Steinbacher et al., 2009). Moreover, Price et al. (2009) reported pre-cooling reduced core and skin temperatures during a simulated football match, although no measures of physical performance were provided. These studies suggest that the pre-cooling induced reductions in physiological and perceptual load for the same or greater exercise workload in the heat may be of benefit for football players. However, no study has yet assessed its efficacy as used by a sporting team during actual competition or training in the heat. In addition, there is some evidence that suggests difficulties in transferring laboratory findings into practice in the field (Hornery et al., 2007). As such, while several studies indicate potential benefits of pre-cooling, no research has investigated this potentially beneficial intervention in an ecologically valid environment, such as football training and competition in the heat.
Whole- or part-body immersion in cold water is assumed to be the most potent cooling stimulus (Quod et al., 2008); however, in non-emergency situations the logistical constraints involved with applying this intervention may be difficult. Pre-cooling many players during pre-game or training routines, along with transportation of equipment may make whole-body immersion difficult to use in practice. Further immersion techniques may not be optimal due to the reduction of local muscle temperatures and ensuing requirement for peak power (Duffield et al., 2010). However, previous research has indicated that there may be a dose-dependent response to cooling, in that the larger the volume of cooling the greater the potential benefits (Castle et al., 2006; Duffield & Marino, 2007), and that both internal and external methods of cooling may be valid (Ross et al., 2010). As such, the combination of a mixture of several smaller and portable cooling methods, of both internal and external application, may be a viable method for providing pre-cooling to athletes in training or competition. Accordingly, the current study aimed to investigate the effects of field-based pre-cooling on performance, and physiological and perceptual responses during both high-performance football training and competition in hot and humid conditions.
Methods
Participants
Ten male, professional football players (mean±standard deviation (s) age 23.1±3.2 years, body mass 78.4±6.3 kg, and height 178±7 cm) from the same Australian A-League football club were recruited as participants for the present study. From these ten players, data was collected during both training and competitive matches. However, due to incomplete participation in cross-over testing (injury, non-selection etc), data sets were n=9 for training and n=7 for matches. All participants were regular professional players in the elite national, Australian A-League competition, and trained 6 days a week, normally with multiple sessions a day. Following explanation of all procedures, all participants gave verbal and written informed consent to participate in the studies and ethics approval was granted by the Ethics in Human Research Committee of the University.
Overview
The study consisted of two parts, involving the use of field-based pre-cooling during both training sessions and competitive matches, respectively. For the training study, players (n=9) performed two, standardised sessions in a randomised, cross-over fashion, separated by 7 days, with only a pre-cooling intervention differing between sessions. Training sessions were conducted at the same time of day (09:00) on an outdoor football pitch in an environment of 30±2oC, 75±5% relative humidity (RH) and 28±2oC Wet Bulb Globe Temperature (WBGT). The participants were accustomed to the training session, including both conditioning and small-sided games (SSG), and were familiarised with all measures, procedures and protocols. The training session was set by the coaching panel and was repeated on two standardised occasions by the entire team. The sessions started with a 15-min warm up consisting of jogging, movement drills and sprints as set out in the Fédération Internationale de Football Association (FIFA) 11+ warm up (Dvorak & Junge, 2005). The training session included 2 × 10-min of standardised running intervals around a marked football pitch, separated by a 5-min recovery period and 6 × 3-min bouts of SSG. For the interval running, players were required to perform 30 s of running between two cones on opposite sides of the field, separated by 30 s of self-paced jogging recovery. The respective interval bouts involved 30 s work, 30 s recovery at intensities approximating 105% and 50% Maximal Aerobic Speed, respectively (based on pre-season testing). In total, participants performed 20 × 30-s hard efforts during the 20 min of interval work. Following the 2 × 10-min interval sessions, players performed 6 × 3-min bouts of 5v5 SSG separated by 1-min recoveries on a pitch size of 40 × 29 m. The same teams, pitch dimensions and rules were enforced for each respective session and were overseen by the Head and Assistant coaches. This session represents an actual and ecologically valid training session used by a professional football club and accordingly some constraints were present regarding access to players and regularity of measures. Finally, sessions were held the day following a standard recovery day to ensure similar physical activity patterns and players were required to complete food diaries documenting food and fluid intake the day of testing and replicate for the subsequent session. Fluid consumption during training was not set, but players drank from individual drink bottles at set times throughout the training session.
To further increase the ecological validity of the research design, the pre-cooling procedures were also applied before and during competitive matches. For match-based data collection, players were recruited in a semi-randomised fashion, depending on selection in the starting team, with all feasible attempts made to ensure a cross-over design during five different A-League games. Data was compared between an intervention and a control match respectively based on each player being selected in the same starting position as the previous match. All matches were 90-min games (plus injury time) and collected data was included in analysis if players were on the field for at least 60 min of the match (n=7). The A-League matches were conducted on the same FIFA accredited outdoor football pitch in an environment of 27±2oC, 80±10% RH and 26±2oC WBGT. All matches were played at a standardised time (20:00). Before each match, the participants completed a similar, but not standardised, warm-up under instructions from relevant senior coaching staff. Within reason for this professional team, players were asked to standardise all food and fluid ingested prior to the match in accordance with their normal pre-game routine. However, while pre-match hydration status was checked, due to logistical constraints of the testing environment, no measure of fluid consumed during the match was obtained.
Methodology
Pre-cooling. Thirty minutes prior to the start of the warm up, players performed the pre-cooling intervention, consisting of 20 min of field-based, mixed-method cooling procedures. While previous research suggests a dose-dependent volume effect of pre-cooling (Duffield & Marino, 2007), the ability to cool players prior to training/competition with whole body immersion is limited. Accordingly, we attempted to combine multiple smaller methods to increase the cooling intervention in a way that would provide minimal intrusion to these professional athletes. Therefore, following arrival at the ground and collection of resting measures, the players wore an ice-vest (Arctic Heat, Brisbane, Australia) that had been frozen overnight and was placed in direct contact with the skin. The players also placed a cold towel, that had been soaked in 5oC coldwater, over their head and neck. Further, in addition to the external cooling, players also consumed a standardised 350 mL ice-slushie that consisted of frozen and then partially thawed commercially available sports drink (Gatorade, Chicago, USA). While only an absolute volume (rather than relative) was provided, this volume was to ensure minimal gastro-intestinal discomfort based on prior pilot testing. Further, during the half-time break of the match, the ice-vest and cold towels were also applied for 5 min. Conversely, during the control condition, players were allowed to continue with their normal pre-training or match routine and consumed 350 mL of room temperature sports drink (Gatorade, Chicago, USA). Generally there is a period of 30 min during which players are allowed personal preparation time, and hence this time frame guided the duration of the intervention. Players were permitted to prepare themselves normally for training and competition and as such, the cooling interventions allowed players to undergo standard preparation. Accordingly, pre-cooling was applied until approximately 10 min prior to start of warm-up in training and 10 min prior to warm-up in the match.
Physical performance
During matches and training, players wore the same Global Positioning Satellite (GPS) unit fitted in a customised harness between the scapulae of the shoulders (SPI elite, GPSports Systems, Australia). GPS devices recorded distance and velocity every second (1 Hz). Following both matches and training sessions, data were downloaded and analysed using specific customised software (Team AMS; GPSports Systems, Australia). The GPS data for distance travelled and mean velocities collected during each bout of training (aerobic intervals and SSG's) and for each half (and total) of the competitive matches were categorised into three pre-defined zones (Coutts & Duffield, 2010); low-speed activity (speeds < 7.0 km · h−1), moderate-speed activity (speeds 7.0–14.4 km · h−1), and high-speed running (speeds>14.5 km · h−1). Relative distance covered (m · min−1), maximum speed attained and number of efforts above 23 km · h−1 were reported as physical performance indices for both training and competition sessions.
Physiological and perceptual measures
Participants swallowed a telemetric core temperature capsule 5 h prior to the start of each session to measure core temperature (Vital Sense, MiniMitter, NJ, USA). For training, core temperature was recorded at rest, following the intervention, warm-up, respective interval bouts and SSG. Additionally, a heart rate monitor (Team2, Polar Electo Oy, Finland) was worn during the intervention and training session. Skin temperature was measured from three sites including the sternum, forearm and calf to calculate mean skin temperature (Burnham, McKinley, & Vincent, 2006). The three respective sites were marked with a permanent marker and skin temperatures were recorded via a hand-held infra-red thermometer (Thermoscan 3000, Braun, USA) at rest, following cooling, after the 20 min of interval training and at the end of training. This method of measuring skin temperature was selected due to easier access to multiple players with minimal intrusion and has previously been reported to be accurate compared to skin thermometers (r = 0.92) and a reliable measure of skin temperature (intra-class correlation = 0.96) (Burnham et al., 2006). Further, prior to and following the training session, participants were weighed nude to calculate sweat loss and prior to the session provided a mid-stream sample of urine to measure Urine Specific Gravity (USG) (Digital Refractometer, ATAGO, Japan) as an indicator of hydration status. Fluid consumed during training was recorded via the measurement of individual drink bottles to calculate mass change and estimate sweat rate during the session. Finally, a sample of capillary blood was obtained from the fingertip prior to and following the training session. Capillary samples were analysed for metabolites (Glucose), potassium (K+), sodium (Na+), haematocrit (Hct) and haemoglobin (Hb) concentration (i-Stat, Abbot Point-of-Care, USA). A rating of perceptual exertion (RPE) and thermal stress (TSS) were recorded following the intervention, warm-up, each 10-min interval and the end of the SSG session. RPE was measured based on the Borg scale (CR-10; 1–10) (Borg, Ljunggren, & Ceci, 1985), while thermal comfort was recorded based on the Thermal Sensation Scale (TSS; 0–8) (Young, Sawka, Epstein, Decristofano, & Pandolf, 1987).
Given the difficulties in obtaining measures during a professional football match, all measures were recorded before and after each respective half. Core temperature was recorded at rest, and following the intervention, warm-up and each respective half of football. Further, prior to and following the match, participants were weighed nude to estimate sweat loss; although due to the difficulties in quantifying in-game fluid consumption, no measure of fluid consumption during the match was recorded. Prior to the game, players provided a mid-stream sample of urine to measure USG. Finally, RPE and TSS measures were recorded following the intervention, warm-up and each half of the match.
Statistical analyses
All data are reported as mean±s. A two-way (condition x time) repeated measures ANOVA was used to determine differences between the two conditions (cooling v control). Post-hoc paired t-test analyses were performed to determine the location of significant differences with Bonferroni corrections. Significance was set at P <0.05. Given the nature of data collection and inappropriateness of reliance on statistical significance; effect size (ES) data were calculated (Cohen's d) to determine the magnitude of effect of the cooling intervention on physical performance and physiology with an ES of < 0.2 classified as ‘trivial’, 0.2–0.4 as ‘small’, 0.4–0.7 as ‘moderate’ and > 0.8 as ‘large’ effects (Cohen, 1988).
Results
Training
No significant differences were evident between conditions for any physical performance measure, including total distance, high, moderate or low speed distance or the number of entries or peak of speeds reached above 23 km · h−1 (Table I; P =0.28–0.93). However, pre-cooling resulted in a moderate-large ES for a trend for increased total distance covered during the training session (Table I; d = 0.80; P =0.25). Additionally, a large effect for relative distance covered was also present in the cooling condition (145±6 v 140±8 m · min-1; d = 0.81; P =0.26). There were no differences between conditions in high-speed distance covered or efforts above 23 km · h−1 during the session and no differences in peak speed following pre-cooling (d < 0.32; P =0.97).
| Total Distance (m) | Relative Distance (m · min−1) | HSR (m) | MSA (m) | LSA (m) | Above 23 km · h−1 | Maximum speed (km · h−1) | |
|---|---|---|---|---|---|---|---|
| Total session | |||||||
| Pre-cooling | 6996±296# | 146±6# | 3152±167 | 2470±207# | 1372±103 | 2±2 | 24.5±2.3 |
| Control | 6789±423 | 141±8 | 3149±179 | 2283±364 | 1377±149 | 2±2 | 24.6±2.3 |
| Interval | |||||||
| Pre-cooling | 4595±160 | 229±7 | 2960±79 | 1500±146 | 135±63 | 1±1 | 22.3±2.0 |
| Control | 4486±242 | 224±9 | 2931±203 | 1393±183 | 162±68 | 1±1 | 22.0±2.0 |
| SSG | |||||||
| Pre-cooling | 2400±236 | 86±8 | 192±61 | 970±180 | 1238±57 | 1±1 | 23.0±2.0 |
| Control | 324±262 | 83±9 | 218±67 | 890±204 | 1216±119 | 0.5±0.5 | 22.9±1.8 |
- # represents a large effect size (d > 0.80) compared to control condition.
Pre-cooling prevented any increase in core temperature during pre-training preparation and resulted in a reduced core temperature following the warm-up (Figure 1; d > 1.00; P =0.01–0.07). However, the reduced core temperature following pre-cooling had dissipated by 10 min, with no differences between conditions for the remainder of the session (P >0.05). Mean skin temperature was reduced following pre-cooling until the 10th min, with no differences present at 20th minute of training (Figure 1; d > 1.00; P =0.01–0.09). Heart rate did not differ between conditions following either the interval (190±9 vs. 191±5 bpm) or SSG (182±6 vs. 186±9 bpm) sessions (d < 0.40; P =0.15–0.89) for cooling and control conditions, respectively. Estimated sweat loss (1.8±0.7 vs. 2.3±0.6 kg), representing a change of 2.4±0.8 vs. 2.9±0.6% body mass was not significantly different but indicated a large ES for a lower sweat loss in cooling compared to control, (d > 0.90; P =0.10). Pre-exercise USG did not differ between conditions (1.012±0.006 vs. 1.011±0.004 for cooling and control; P =0.33; d < 0.30). There were no significant differences between conditions (P =0.24–0.30; Table II) in Hct or Hb concentrations following the training session; although, a moderate ES for a smaller change was present in the pre-cooling condition for Hct (d = 0.71). No differences and small effect sizes were evident for other blood measures of glucose, Hb, Na+ and K+ (d < 0.3; P =0.12–0.59). While not significantly different, pre-cooling resulted in moderate-large ES for a reduced RPE following both the respective interval and SSG sessions (P =0.07–0.60; d > 0.80; Figure 2). Further, as expected, perceived thermal strain (TSS) was significantly reduced following the cooling intervention and remained lower throughout the training session (P =0.01–0.89; d > 0.90; Figure 2).
| Measure | Condition | Pre | Post |
|---|---|---|---|
| Glu (mmol · L−1) | Pre-cooling | 5.2±1.1 | 5.0±1.1 |
| Control | 5.2±1.3 | 6.1±1.5 | |
| Hb (g · L−1) | Pre-cooling | 149.2±11.0 | 163.8±16.9 |
| Control | 143.9±17.9 | 155.6±16.6 | |
| Hct (%) | Pre-cooling | 0.44±0.03 | 0.48±0.05^ |
| Control | 0.42±0.05 | 0.47±0.05 | |
| Na+ (mmol · L−1) | Pre-cooling | 140±1 | 142±2 |
| Control | 140±2 | 142±2 | |
| K+ (mmol · L−1) | Pre-cooling | 1.7±0.7 | 4.3±0.4 |
| Control | 5.1±0.6 | 4.4±0.3 |
- No significant differences (P > 0.05) were present between conditions for any measure. ^ represents a moderate effect (d = 0.60) for change compared to control.
Mean ± s A) Core temperature, and B) skin temperature during cooling and football training in the heat with and without cooling. # represents a large effect size (d > 0.80) compared to control condition. * represents significant difference (P < 0.05) compared to control condition.
Mean ± s A) Rating of perceived exertion (RPE), and B) Thermal Sensation (TSS) during football training in the heat with and without pre-cooling. # represents a large effect size (d > 0.80) compared to control condition. * represents significant difference (P < 0.05) compared to control condition.
Match
No significant differences were evident between conditions for any match physical performance measure, including relative total distance, relative high, moderate or low speed distance or the number of entries or peak of speeds reached above 23 km · h−1 (Table III; P =0.26–0.67). A moderate ES (d = 0.60; Table III) was evident indicating a trend for increased relative total distance following cooling, of which the predominant increase occurred in the second half (d = 0.77; P =0.27). While total high-speed running did not differ (d < 0.30; P =0.55) between conditions for the match, trends for increased moderate and low speed relative distances, predominantly via second half increases, were present with cooling (d = 0.80–0.95). Again, number of efforts above 23 km · h−1 or peak speeds were not significantly different between conditions, although both number of efforts and peak speeds tended to be higher (Table III; d = 1.5; P =0.15) in the control compared to cooling condition. Finally, time spent on field during matches did not differ and showed trivial ESs between respective conditions (83±17 vs. 80±15 min; P =0.68; d = 0.32).
| Relative Distance (m · min−1) | HSR (m · min−1) | MSA (m · min−1) | LSA (m · min−1) | Above 23 km · h−1 | Maximum speed (km · h−1) | |
|---|---|---|---|---|---|---|
| Total match | ||||||
| Pre-cooling | 108±10 | 19±4 | 44±7 | 45±3 | 11±3 | 27.7±1.0 |
| Control | 106±12 | 22±6 | 42±7 | 42±3 | 12±3 | 29.9±1.5 |
| 1 st half | ||||||
| Pre-cooling | 110±11 | 19±5 | 46±7 | 45±3 | 6±3 | 27.8±1.6^ |
| Control | 112±13 | 23±6 | 46±12 | 42±4 | 8±2 | 29.5±0.9 |
| 2 nd half | ||||||
| Pre-cooling | 105±10^ | 19±5 | 42±6^ | 45±3^ | 5±3 | 26.1±2.1 |
| Control | 101±15 | 21±5 | 39±7 | 42±6 | 4±3 | 26.7±2.9 |
- ^ represents a moderate effect (d = 0.60–0.80) compared to control condition.
Core temperature was not significantly different (P =0.09–0.80) between conditions at any time point during the match; however, large ESs (d = 1.10; Figure 3) were evident for trends towards lower core temperatures in the cooling condition immediately following cooling application both before and at half-time of the match. Further, the change in body mass was not significantly different between conditions, although a moderate ES (P =0.34; d = 0.50) was evident for a trend toward lowered sweat loss (1.5±0.7 vs. 1.8±0.6 kg) during the cooling condition. Finally, pre-game USG was not different between conditions for cooling or control, (1.008±0.003 vs. 1.007±0.003; P =0.49; d = 0.31). No significant differences (P =0.13–0.62) and small ESs (d < 0.40) were evident between conditions for RPE at all time points during competitive matches (Figure 3). However, pre-cooling reduced the subjective thermal strain of competitive matches, with large ESs evident for reduced TSS following the cooling intervention at the start of the match and again at half-time (P =0.08–0.15; d > 1.00; Figure 3).
Mean ± s A) Core temperature, B) Rating of perceived exertion (RPE), and C) Thermal Sensation (TSS) during competitive football matches in the heat with and without pre-cooling. # represents a large effect size (d > 0.80) compared to control condition. ^ represents a moderate effect (d = 0.60–0.80) compared to control condition.
Discussion
The current results showed that the performance, physiological and perceptual benefits of pre-cooling in the field were minimal, and less explicit than previous laboratory evidence (Duffield, Steinbacher et al., 2009; Price et al., 2009). Although some moderate-large effects were evident, no significant benefits of pre-cooling were observed. The limitations to collecting research data in field settings, particularly within professional teams, may dampen the strength of the findings due to reduced control of the subjects, setting, hydration status and access to players during matches. However, the similarities in observed trends to previous laboratory research, and the lack of a negative influence of this intervention may still suggest some application of pre-cooling interventions to professional football players.
While mostly beneficial for endurance exercise (Duffield et al., 2010; Kay et al., 1999; Quod et al., 2008), the evidence to support the use of pre-cooling for team-sport exercise is from non-specific, intermittent-sprint exercise protocols (Castle et al., 2006; Duffield, Steinbacher et al., 2009). In contrast however, studies that have used exercise protocols consisting of short-duration sprints, separated by controlled exercise intensities have not shown ergogenic effects of pre-cooling (Cheung & Robinson, 2004; Drust et al., 2000; Duffield, Dawson, Bishop, Fitzsimons, & Lawrence, 2003). These data suggest pre-cooling may be more beneficial for prolonged duration, sub-maximal exercise (Duffield et al., 2010; Kay et al., 1999; Quod et al., 2008). It could be argued that given many football training sessions consist of either controlled speed interval training, SSG's and/or match simulation drills (Bangsbo, Mohr, & Krustrup, 2006), there may be some difficulty in transferring laboratory findings to this context. Accordingly, the current findings indicate that pre-cooling prior to football training, including both interval and SSG based exercise, provides no significant performance or physiological benefits in hot conditions. That said, given the logistical constraints of such data collection, the large effects for small increases in distance covered observed throughout the conditioning session, alongside trends for suppression of core temperature during warm-ups, a smaller sweat loss and lower perception of thermal load, indicates pre-cooling might have some small practical benefits.
The current study showed minimal significant effects of pre-cooling for physical performance responses during competitive football matches. However, it must be acknowledged that match performance is a multi-factorial phenomena, incorporating a myriad of potential determinants of physical performance, not least of which include, the opposition, conditions, physiological state and tactical or technical role (Rampinini, Coutts, Castagna, Sassi, & Impellizzeri, 2007; Rampinini, Impellizzeri, Castagna, Coutts, & Wisløff, 2009). Moreover, within this context, recent evidence highlights large inter-match variance with regards to physical performance; with between match differences potentially ranging up to 20% (Gregson, Drust, Atkinson, & Salvo, 2010; Rampinini et al., 2007). Research in such settings is difficult due to limited access and control of players and procedures. Further, within game hydration is difficult to measure due to water delivery, consumption and restrictions on filming. Additionally, the limited number of players who are available from game to game reduces sample sizes and the statistical power of this study. That said, interventions are tested within laboratory settings for use in field-based practice, and as such should also be investigated in the environments in which they are used (Bishop, 2008). While acknowledging the importance of these factors in interpreting the physical performance during competitive matches, the standardisation of the ground, and similarity of conditions and players collected over multiple games allows some discussion of the effect of an external intervention on physical performance.
Equivocal findings were evident for physical performance during competitive matches. Performance benefits are supported by some (Price et al., 2009), but not all (Drust et al., 2000) simulated laboratory protocols that highlight improvements to exercise performance during the later stages of simulated protocols. The present observations for trends of increased relative distance covered in the second half are not as explicit as those observed during laboratory studies (Castle et al., 2006; Duffield & Marino, 2007). Current evidence derived from laboratory studies indicates that self-paced exercise in the heat results in an earlier reduction in exercise intensity compared to cooler environments (Tucker et al., 2006). Further, pre-cooling has been reported to result in the maintenance of higher exercise intensity in laboratory (Kay et al., 1999) and simulated field environments (Duffield & Marino, 2007). The present results do not strongly support these findings, although it is acknowledged that this study may be underpowered. The collective findings of laboratory studies (Duffield & Marino, 2007; Duffield, Steinbacher et al., 2009), combined with the current observations indicates that pre-cooling before and during football matches provides equivocal benefits.
Previous studies suggest pre-cooling can improve exercise performance due to reduced core temperature and thermal gradients (Cheung & Robinson, 2004; Hessemer et al., 1984; Price et al., 2009), lowered physiological load (Kay et al., 1999; Quod et al., 2008), or prevention of the down-regulation of muscle recruitment (Duffield et al., 2010) during exercise in the heat. The present pre-cooling intervention blunted the rise in core temperature during preparation for training, but only resulted in lower core temperatures until the end of the warm-up. Further, by 20 min of training, all skin and core temperature differences between conditions had dissipated, although trends for lowered sweat loss and changes in Hct were evident post-training. While not as evident in the present study, the aforementioned findings are commonly reported in laboratory-based pre-cooling research (Hessemer et al., 1984; Lee & Haymes, 1995). For example, the post-cooling reduction of core temperature in the current study was diminished by the end of the warm-up, while previous studies highlight reduced core temperatures for up to 30 min of exercise (Duffield et al., 2010; Lee & Haymes, 1995). Such equivocal findings in the present study may suggest small to minimal physiological effects resulting from the cooling stimulus. The cooling procedures used here utilised similar interventions as reported in previous studies (Duffield, Steinbacher et al., 2009; Hornery et al., 2007; Price et al., 2009; Ross et al., 2010). Further, most of these studies report improved performance following cooling interventions for single-persons under passive rest in normothermic, controlled laboratory settings (Castle et al., 2006; Duffield et al., 2010; Hessemer et al., 1984). However, it may be that in the crowded player change rooms, with movement and activities of players and time between cooling and start of session, the volume of cooling did not create a sufficient change in physiological state to benefit performance.
In the match-context core temperature was not reduced, and while this may result in the questioning of the strength of the intervention, practically few other options may be available. The physiological effects of pre-cooling seem to have not lasted much longer than the warm-up, and suggest that effective implementation requires the re-application of cooling, although this may be limited by coach/player access. Trends were evident at half-time for faster reduction in core temperature and perceived thermal stress, which may highlight that effective re-application may be beneficial. It is reported that the application of cooling procedures slows the rate of rise in core temperature, while reducing the perceived thermal stress during simulated matches (Price et al., 2009), field-based training (Duffield, Steinbacher et al., 2009), and laboratory studies (Duffield & Marino, 2007). As such, to observe trends for similar responses during real world competition represents an extension of laboratory evidence to a practical field-setting. However, since there was no evidence of negative influence on physical performance, this intervention may still be a worthwhile intervention for football players competing in the heat.
Reductions in exercise intensity in the heat are hypothesised to result due to an increase in endogenous thermal load (Gonzalez-Alonso et al., 1999), to regulate the physiological load against the anticipated requirements of the exercise bout (Tucker et al., 2006), or by the conscious perception of effort (RPE) (Marcora, Staiano, & Manning, 2009). While many previous pre-cooling studies highlight a physiological basis to explain performance improvement, few studies discount the perceptual effect (Drust et al., 2000; Duffield et al., 2010; Duffield, Steinbacher et al., 2009). Despite minor or equivocal cooling-induced physiological alterations, pre-cooling did improve the perceptual responses within these environmental conditions, which may also lead to altered pacing strategies adopted for an exercise session (Arngrimsson, Petitt, Stueck, Jorgensen, & Cureton, 2004; Tucker et al., 2006). Given the difficulties of effectively cooling a team prior to training, pre-cooling football players prior to training may assist in lowering the subjective thermal load of the session, although the resulting effect on exercise intensity of training in the heat remain speculative. However, perceptual benefits of pre-cooling should not be discounted, particularly given the noted trends for increased training distance covered.
Finally, the overall aim of this study was to assess the transfer of laboratory evidence to an actual training environment in elite football. Previous research suggests a dose-dependent cooling response (Duffield & Marino, 2007), and despite greater reductions in physiological parameters i.e. core temperature reductions > 0.3oC (Duffield et al., 2010; Quod et al., 2008), the use of immersion techniques in field-based team-sports provides logistical difficulties. As such, the present study attempted to maximise cooling, while minimising player disruption by using a mixed method cooling strategy. In the present study, these practical pre-cooling methods did not show the magnitude of performance improvements shown in previous studies (Duffield & Marino, 2007; Duffield, Steinbacher et al., 2009). Additionally, we did not observe a reduced core temperature following pre-cooling in the training environment, which is not uncommon in part-body cooling (Castle et al., 2006;Duffield et al., 2010; Kay et al., 1999; Quod et al., 2008), although this is not a necessity for performance enhancement (Kay et al., 1999; Quod et al., 2008). Consequently, it is more likely that activities during pre-cooling exposures and/or the exercise stimulus between these studies are the most likely explanations for the different results, resulting in difficulty replicating ergogenic findings of laboratory studies.
Conclusion
In conclusion, equivocal findings were present for the effects of pre-cooling (and re-cooling) professional football players training and competing in hot and humid climates. Although significant differences for altered performance or physiological responses were not evident, trends similar to previous laboratory findings were present. As such, the physical performance benefits of cooling for both football-specific training and competitive matches are not as explicit as laboratory studies. However, when trends for improved physical performance are aligned with trends for lowered physiological load, the use of pre-cooling procedures may provide some benefit for players training and competing in the heat. In particular, players who are unaccustomed to such conditions may find the use of such cooling interventions of additional use.
Acknowledgements
The research team would like to acknowledge the support of Football Federation Australia for their financial support of this project. We would also like to acknowledge the players, coaches and support staff of the NQF FC for their involvement with this study.
