Secondary Logo

Share this article on:

In-Play Cooling Interventions for Simulated Match-Play Tennis in Hot/Humid Conditions


Medicine & Science in Sports & Exercise: May 2017 - Volume 49 - Issue 5 - p 991–998
doi: 10.1249/MSS.0000000000001183

Purpose This study aimed to assess the efficacy of different in-play cooling strategies for mitigating heat strain during simulated tennis match-play activity in a hot/humid environment representing the most extreme conditions during the US Open (36°C, 50% relative humidity).

Methods On three occasions, nine males completed an intermittent treadmill protocol with an exercise intensity and activity profile simulating a four-set tennis match, with 90-s breaks between odd-numbered games and 120-s breaks between sets, according to International Tennis Federation rules. During breaks, 1) the currently used cooling strategy—an ice-filled damp towel around the neck and a cold-damp towel on the head and thighs (ICE); 2) wetting of arms, neck, face, and lower legs with a sponge in front of an electric fan (FANwet); or 3) no cooling (CON) were applied. Rectal (T re) and mean skin (T sk) temperature and HR were measured throughout. Thermal sensation and RPE were assessed during breaks. Trials were terminated upon reaching a T re ≥ 39.5°C or volitional exhaustion.

Results Seven, five, and one participant completed FANwet, ICE, and CON, respectively. By end set 1, ΔT re was lower in FANwet (0.92°C ± 0.15°C) compared with CON (1.09°C ± 0.09°C, P = 0.01), and by end set 2, ΔT re was lower (P < 0.001) in FANwet (1.55°C ± 0.23°C) and ICE (1.59°C ± 0.17°C) compared with CON (1.99°C ± 0.19°C). Mean RPE (FANwet = 13.9 ± 2.2, ICE = 13.6 ± 1.8, CON = 16.6 ± 1.8), HR (FANwet = 163 ± 21, ICE 164 ± 22, CON = 175 ± 19 bpm), T sk (FANwet = 36.56°C ± 0.69°C, ICE 36.12°C ± 0.44°C, CON = 37.21°C ± 0.42°C), and thermal sensation were lower in FANwet and ICE (P < 0.05) compared with CON by end set 2.

Conclusions The currently recommended ICE strategy successfully mitigates thermal strain during simulated tennis match play in hot/humid conditions. The FANwet intervention is an equally effective alternative that may be more practical in limited resource settings.

1Technical University of Munich, Institute of Ergonomics, Garching, GERMANY; 2Thermal Ergonomics Laboratory, Exercise and Sport Science, Faculty of Health Sciences, University of Sydney, NSW, AUSTRALIA; 3Medical Department, Koninklijke Nederlandse Lawn Tennis, Amersfoort, THE NETHERLANDS; and 4Athlete Health and Performance Research Centre, Aspetar Orthopaedic and Sports Medicine Hospital, Doha, QATAR

Address for correspondence: Ollie Jay, Ph.D., Thermal Ergonomics Laboratory, Exercise and Sports Science, University of Sydney, 75 East St., Lidcombe, NSW 2141, Australia; E-mail:

Submitted for publication September 2016.

Accepted for publication December 2016.

Among the tennis Grand Slam tournaments, the greatest number of heat-related health problems has occurred at the Australian and US Open tournaments. During the 2014 US Open alone, 12 male players had to abort play, at least four of them because of heat-related health issues. All incidents occurred at air temperatures ≥35°C with a relative humidity (RH) of ~50%. Under such conditions, the ability to dissipate heat to the surrounding environment is mostly dependent on evaporation, which is greatly limited. The resultant accumulation of heat inside the body can prove deleterious to performance and health if exercise is continued and countermeasures are not taken (6). Indeed, peak core temperatures of 38.9°C ± 0.3°C and 39.1°C ± 0.3°C have been reported during professional hard court tennis match play in males and females, respectively (4).

The principal in-play cooling intervention presently recommended by the International Tennis Federation (ITF) is the application of ice wrapped in a wet towel around the neck with damp, cold towels simultaneously placed on the head and thighs (“ice towels”) (10). Despite their widespread use, the efficacy of this cooling strategy has not yet been assessed. Although some neck-cooling devices have been demonstrated to improve running performance in hot and humid conditions (29), there is no evidence that the currently recommended “ice towel” intervention is an effective cooling strategy for mitigating thermal strain when applied for only the short duration of in-play breaks (90–120 s) permitted by the standard structure of tennis play.

The principal heat loss avenue with the application of “ice towels” is conduction. Heat is transferred from the skin through direct contact with the colder surface of the towel. The heat loss capacity of this intervention is therefore determined by the total skin surface area in contact with the towels, the temperature difference between the skin, and the towel surface and the latent heat of fusion of ice (334 J·g−1) (22). However, the efficacy of this cooling strategy may be limited because of the small surface area covered (neck, head, and thighs), which is consequently available for heat exchange. Further, temperature differences between the towels and the skin will likely become progressively smaller as the surface temperatures of the towels and skin equilibrate during application, whereas the evaporation of sweat may also be simultaneously impeded from covered skin surfaces (22).

Relative to the melting of ice, the evaporation of water from the skin surface liberates approximately seven times more heat per gram (2427 J·g−1) (30). A cooling strategy promoting evaporative heat loss may therefore present a more efficacious alternative than the “ice towel” method. Public health heat wave guidance has previously suggested that fan use does not enable body cooling. However, recent studies have demonstrated that fan use can be beneficial at peak air temperatures experienced at the US Open (35°C–36°C) irrespective of ambient humidity (15). Moreover, maximizing skin moisture coverage by wetting the skin with a damp sponge should ensure the maximum potential for heat loss via evaporation is realized.

The aim of the present study was to assess the efficacy of 1) the currently recommended in-play cooling strategy for professional tennis (“ice towels” [ICE]) and 2) a new cooling strategy that optimizes evaporative heat loss (fan use with skin wetting: FANwet), for mitigating physiological and perceptual heat strain during a simulated tennis match play in an environment representing the peak heat stress conditions experienced at the US Open (36°C, 50% RH). It was hypothesized that relative to a control condition without active cooling, the FANwet cooling strategy during mandated rest breaks throughout a simulated tennis match activity would best mitigate elevations in thermal and perceptual strain in a hot and humid environment representative of the peak conditions historically experienced at the US Open.

Back to Top | Article Outline



Nine healthy trained males, nonacclimatized to heat, were recruited to participate in the study. Their mean age, height, body mass, body fat, and maximum oxygen uptake (V˙O2max) were 25 ± 4 yr, 179.2 ± 5.7 cm, 77.3 ± 6.7 kg, 12.7% ± 6.1%, and 50.6 ± 6.3 mL·kg−1·min−1, respectively. Before any testing, participants gave written informed consent after they were informed about the risks of the study. All participants took part in one preliminary session and three experimental sessions on different days, separated by at least 48 h. The time of testing was kept constant within each subject. Participants refrained from strenuous physical activity and alcohol ingestion for 24 h before testing and from caffeine ingestion on testing days. All experimental procedures were first approved by the University of Sydney Human Research Ethics Committee.

Back to Top | Article Outline


Rectal temperature (T re) was measured in 8 of the 9 participants at 5-s intervals using a Mon-a-therm general-purpose thermistor probe 400TM (Covidien, Mansfield, MA) that was self-inserted 15 cm past the anal sphincter. Skin temperature (T sk) was measured every minute using wireless iButtons (DS1921H; Maxim Integrated Products Inc., San Jose, CA) taped to the chest, shoulder, thigh, and calf, and a four-point weighted mean value was determined according to the weighting of Ramanathan (24). Heart rate (HR) was monitored every 5 s by using a Polar RS 400 watch (Polar Electro Oy, Kempele, Finland).

For subjective measures, participants were asked to rate their thermal sensation (TS) using a bipolar 200-mm visual analog scale ranging from “very cold” (0 mm), to “neutral” (100 mm), to “very hot” (200 mm). Further, the effects on potential performance were assessed using the 6- to 20-point Borg scale RPE but with a modified question: “How easy it is to continue the exercise?”

Back to Top | Article Outline


Preliminary session

Upon arrival, the participant's physical characteristics were measured. Height and weight were measured using a wall-mounted stadiometer (HR-200; Tanita, Arlington Heights, IL) and digital scale (BWB-800, Tanita), respectively. Body composition was measured by dual-energy x-ray absorptiometry (GE-LUNAR Prodigy module; GE Medical Systems, Madison, WI). Then after a light warm-up, V˙O2max was assessed on a treadmill according to the procedures set forth by the Canadian Society for Exercise Physiology (9). All respiratory variables were measured throughout, using a stationary pulmonary gas analyzer system (Quark CPET; Cosmed, Rome, Italy). Participants also completed five games (~16 min) of simulated tennis match-play activity (cf. Experimental session section) to familiarize themselves with the experimental protocol and to ensure that the average V˙O2 during the exercise protocol was comparable with the V˙O2 values (and therefore metabolic heat production) previously reported (23–29 mL·kg−1·min−1) during live professional tennis match play (11,27). Accordingly, a mean V˙O2 of 26.3 ± 1.2 mL·kg−1·min−1 was measured.

Back to Top | Article Outline

Experimental session

Participants performed three different trials in a counterbalanced order. All trials were completed in a climate-controlled chamber in the Thermal Ergonomics Laboratory at the University of Sydney. Upon arrival before each session, urine specific gravity (USG) was determined with a refractometer UG-α (Atago Co., Ltd., Tokyo, Japan). A cutoff USG of >1.020 was used to ensure that the participants were adequately hydrated (1) before starting the experiment. None of our participants exceeded this USG cutoff value on any occasion. Participants then inserted their T re probe and were weighed. After the remaining instrumentation was completed, they entered the climate chamber regulated at an air temperature of 36.5°C ± 0.4°C and an RH of 51.0% ± 1.8%; these conditions were selected to simulate the most extreme weather reported at the US Open in recent years. The participant conducted a 5-min warm-up on the treadmill at a self-chosen intensity, which was the same between all trials. Then the exercise protocol—designed to simulate the metabolic load of a tennis match—began with the preprogramed exercise profile controlled via hp-paragraphics (H/P/Cosmos Sports & Medical GmbH, Nussdorf-Traunstein, Germany). To simulate one “point,” the participant ran at 16 km·h−1 for a variable period of 3–7 s (note: this time excludes 3 s of acceleration time), randomly selected for every “point” but with a mean “point” duration of 6 s within each “game,” to align with the reported average point length for male elite tennis players at the US Open (16,20). Although V˙O2 was not measured during the experimental trials because of concerns that the addition of a facemask would disproportionately affect perceptual, and possibly physiological, thermal strain, the maintenance of the same average point length within each “game” ensured a similar cumulative V˙O2 throughout. Running at top speed was followed by a 3-s deceleration phase, a 3-s walk at 2 km·h−1, and a standing phase of 5 s. A total of six “points” constituted one “game,” and eight “games” constituted one “set.” In total, four “sets” were completed in each trial. According to the current regulations of the ITF, after every odd-numbered game (i.e., game 1, 3, 5, etc.), a rest break of 90 s was enforced, and after every “set,” a rest break of 120 s was enforced, during both of which one of the three cooling strategies (see next section) were applied. TS was assessed three times per break: 1) directly before applying the cooling intervention, 2) immediately after removal of the cooling intervention, and 3) after they completed the next “game” of the protocol. In total, the simulated four-set match lasted a maximum of 108 min and 12 s for each trial. Trials were terminated early if T re exceeded 39.5°C or volitional exhaustion occurred.

Back to Top | Article Outline

Cooling Strategies

During each of the experimental trials, one of the following three cooling strategies was applied in each of the rest breaks.

Back to Top | Article Outline

Ice towel (ICE)

For one trial, two sets of three towels (70 × 140 cm, 100% cotton) were soaked with cold water and wrung. Per towel set, one damp towel, for application on the neck, was filled with 3 kg of ice and separated into three equal sections with a shoelace, and two damp chilled towels without ice were placed on the thighs and over the head. This strategy was performed according to current practice in elite tennis (E. Brady, personal communication, WTA, ITF Medical Commission, 2015). The towel set used in sets 1 and 2 was replaced at the beginning of set 3 and used for the remainder of the trial.

Back to Top | Article Outline

Fan with skin wetting (FANwet)

Participants moistened the surface of their uncovered neck, face, arms, and thighs (for ~5 s) with a damp sponge that was soaked in water (initially at 15°C) and wrung out with both hands for ~2 s over a bucket before application, so no excess water dripped. Simultaneously, a commercially available 45-cm diameter fan (MFE745S2DQ, Moretti) placed 1.5 m in front of the participant, set at its highest speed, generated an air velocity of 6.4 m·s−1. The water was not replaced throughout the trial and consequently warmed by ~8°C–10°C by the end of the trial.

Back to Top | Article Outline

Control (CON)

Participants sat quietly and applied no cooling intervention for the duration of each break, but they were allowed to ingest cold water.

Throughout each participant's first experimental trial (which was randomized between participants), they were allowed to drink cold water (~7°C) ad libitum during each mandated rest break. The amount of water ingested was measured and subsequently fixed for the remaining two trials to prevent differences in internal heat loss and any differences in sweat capacity because of the stimulation of abdominal thermoreceptors (18).

Back to Top | Article Outline

Statistical Analyses

Rectal temperature, HR, and RPE were averaged over the last 15 s of every odd game (i.e., games 1, 3, 5, and 7), and the last game, in each “set.” Values for T re, HR, and RPE were analyzed using a three-way repeated-measures ANOVA using the independent variables of “cooling intervention” (three levels: CON, ICE, and FANwet), “set” (four levels: sets 1, 2, 3, and 4), and “game” (five levels: games 1, 3, 5, 7, and the end of the set [game 8]). Note that the “cooling intervention” independent variable was reduced to two levels (ICE and FANwet) for analysis of data in sets 3 and 4 as a high dropout rate because of exhaustion, or a high core temperature was observed in the CON trial. TS and skin temperature from each “set” were separately assessed using a similar three-way repeated-measures ANOVA with the same independent variables of “cooling intervention” and “game,” but with the additional independent variable of “time” (three levels: start of each break [“precool”], end of each break [“postcool”], and end of the following game [“next game”]). If the Mauchly's test indicated a violation of sphericity, a Greenhouse–Geisser correction was used. The alpha error level was set at P < 0.05 to establish statistical significance and maintained at that level during pairwise comparisons using a Holm–Bonferroni correction. The data are presented as mean ± SD, unless otherwise indicated. All statistical analyzes were performed using SPSS (Version 22.0; IBM SPSS Inc., Chicago, IL).

Back to Top | Article Outline


Duration/dropout rate

Of the nine participants, seven completed the FANwet trial, five completed the ICE trial, and only one completed the CON trial. Reasons for withdrawal are illustrated in Figure 1. Compared with the CON trial (79.8 ± 22.4 min), mean exercise time was longer (P = 0.021) in the FANwet trial (106.7 ± 4.4 min) and tended to be longer (P = 0.057) in the ICE trial (96.2 ± 23.9 min).



Back to Top | Article Outline

Rectal temperature (T re)

Mean baseline absolute T re was similar (P = 0.586) between ICE (36.9°C ± 0.3°C), FANwet (36.8°C ± 0.4°C), and CON (36.8°C ± 0.2°C). In the initial phase of set 1, the rise in T re was similar between all trials. However, by game 7 of set 1, ΔT re in the FANwet trial was lower relative to the CON trial (Fig. 2). Moreover, by game 1 of set 2, ΔT re values in the FANwet and ICE trials were lower compared with CON (Fig. 2). These differences in ΔT re persisted throughout the remainder of the “match” (P < 0.001) with a very similar ΔT re evident between ICE and FANwet.



Back to Top | Article Outline

Skin temperature (T sk)

Mean baseline T sk was similar (P = 0.582) between the ICE (35.6°C ± 0.8°C), the FANwet (35.8°C ± 0.6°C), and the CON (35.8°C ± 0.5°C) trials. In set 1, T sk was lower in the ICE trial compared with CON before cooling (P = 0.002); however, immediately after cooling, T sk was lower in both ICE (P = 0.001) and tended to be lower in FANwet (P = 0.076) relative to CON. By the end of the following game after each break, T sk was lower in FANwet compared with CON (P = 0.009), but T sk was lower in ICE compared with FANwet (P = 0.013; Fig. 3). In set 2, T sk was lower in the ICE and FANwet trials compared with CON, both precooling and postcooling. These lower skin temperatures compared with CON remained until at least the end of the next game (Fig. 3). The same response as set 2 was evident in set 3, but T sk became progressively higher in CON, yet T sk was at similar levels to set 2 in ICE and FANwet and was not different between these two cooling trials (Fig. 3). Similarly, T sk was highest in CON in set 4, and T sk remained similarly low in the FANwet and ICE trials (Fig. 3).



Back to Top | Article Outline


In set 1, TS for FANwet and ICE was cooler immediately after cooling compared with precooling. Moreover, TS was cooler postcooling compared with CON (Fig. 3). However, TS had returned to similar levels to CON in both the FANwet and the ICE trials by the end of the next game. In sets 2 and 3, TS was already lower at the start of the break in ICE and FANwet compared with CON. This difference became greater postcooling, and although TS returned to precooling levels by the end of the next game, it remained lower compared with CON in ICE and FANwet (Fig. 3). In set 4, TS postcooling remained cooler relative to precooling in ICE and FANwet.

Back to Top | Article Outline

Exercise perception (with RPE scale)

Exercise perception throughout simulated match-play activity is illustrated in Figure 4A. RPE tended to be lower in the ICE trial (8.3 ± 2.2, P = 0.051) but not the FANwet trial (9.0 ± 2.5, P = 0.58) compared with CON (9.4 ± 2.3) after the first game of set 1. At the end of set 1, RPE was lower (11.4 ± 2.3, P = 0.023) in the ICE trial and tended to be lower in the FANwet trial (11.9 ± 2.3, P = 0.063) compared with CON (12.9 ± 2.0). At the end of set 2, RPE was lower in the ICE trial (13.6 ± 1.8, P = 0.004) and the FANwet trial (13.9 ± 2.2, P = 0.008) compared with CON (16.6 ± 1.8). RPE was similar between ICE and FANwet at the end of set 3 (15.4 ± 3.3 vs 15.3 ± 2.4, P = 0.23) and set 4 (15.7 ± 3.2 vs 15.8 ± 2.4, P = 0.19).



Back to Top | Article Outline


Mean HR throughout the simulated matches are given in Figure 4B. HR tended to be higher in CON compared with both FANwet (P = 0.09) and ICE (P = 0.10) by the end of set 1. HR in CON was higher compared with the ICE (P = 0.016) and FANwet (P = 0.020) trials by the end of set 2, but similar between ICE and FANwet (P = 0.81). HR was similar between ICE and FANwet at the end of set 3 (162 ± 19 vs 163 ± 20 bpm, P = 0.57) and set 4 (159 ± 15 vs 162 ± 17 bpm, P = 0.27).

Back to Top | Article Outline


The present study is the first to our knowledge to directly assess the efficacy of different in-play cooling interventions applied during mandated breaks in simulated tennis match-play activity for mitigating physiological and perceptual heat strain in hot/humid conditions. Compared with a control (CON) condition during which only cold water ingestion was permitted, the currently recommended (and used) “ice towel” cooling strategy that augments conductive heat loss by wrapping an ice-filled damp towel around the neck and placing two cold, damp towels on the head and thighs (ICE) led to a ~0.5°C smaller elevation in core temperature, an approximately two- to three-point lower RPE, a ~15-bpm lower HR, and a cooler whole-body TS. Promoting evaporative heat loss using the different cooling strategy of wetting the skin of the arms, neck, face, and lower legs with a damp sponge in front of an electric fan (FANwet) diminished physiological and perceptual heat strain to similar levels as the ICE trial compared with CON.

Under the very hot/humid ambient conditions assessed in the present study, four of the nine participants could not complete the four-set simulated tennis match because of exhaustion without an active cooling strategy (cold fluid ingestion only); and even with the ICE strategy, two participants stopped before the end of the trial because of exhaustion (Fig. 1). Given that the rate of heat-related incidents in tournaments such as the US Open is relatively low, this may indicate that the combination of climate and exercise in the present study represents the very worst-case scenario. The exercise intensities were selected to replicate rates of metabolic heat production and activity profiles previously measured in elite tennis players (11,27); however, the environmental conditions (36°C, 50% RH) have only been reached on one occasion (in 2014), with peak air temperatures for the US Open in most years typically closer to ~30°C–32°C. Moreover, the exercise protocol, while separating bouts of short duration fast running with intermediate periods of walking and standing (based on the activity profiles of elite tennis match play), required participants to run at the same speed (and generate the same V˙O2) in all sets, which was necessary to isolate the independent influence of cooling strategy on thermal strain. In a real-world scenario, however, an athlete would be free to behaviorally adapt as RPE and HR increases in later sets (Fig. 4A and B) before reaching exhaustion. Recent research has demonstrated that such behavioral thermoregulation occurs in tennis players competing in hot environments in the form of longer pauses in play between points (23) as opposed to a downregulation in exercise intensity as indicated by running speed (13) and point duration (23). Nevertheless, both the ICE and the FANwet strategies successfully dampened the progressive rise in RPE, which would have presumably reduced any heat-related decrements in performance in a competition setting (14). The ability to subjectively tolerate the heat during exercise is also directly influenced by aerobic fitness (7). The mean V˙O2max of our participants of ~51 mL·kg−1·min−1 is slightly lower than the V˙O2max values reported in elite male tennis players (~58 mL·kg−1·min−1 [5]), but no association was observed within our data between V˙O2max and exercise duration. Moreover, there seems to be no reason to believe that the efficacy of any one cooling intervention would be altered by differences in V˙O2max.

Although the present study was designed to assess the efficacy of different cooling strategies for elite tennis players in a worst-case US Open setting, the findings are also relevant to subelite athletes, as well as elite tennis tournaments scheduled annually in various geographical regions with hot/humid conditions; for example, Australia (Sydney and Brisbane: January), Brazil (Rio de Janeiro and Sao Paulo: February), UAE (Dubai: February), Mexico (Acapulco: February), USA (Miami: March; Washington DC, Atlanta, Cincinnati, and Winston-Salem: August), Morocco (Marrakech: April), and China (Shenzen: September). The findings associated with the FANwet trial in particular are potentially most applicable to lower resource settings (e.g., amateur, club, and junior tennis matches) that would most likely not have the infrastructure necessary to prepare and replenish “ice towels.” Indeed, the FANwet strategy presents a lower logistical burden requiring only a bucket of water, a sponge, and an electric fan, which could be battery powered. One potential practical disadvantage of the FANwet strategy could be increased risk of excess water soaking the athlete's clothing/shoes, which would probably be undesirable. During pilot testing, we found that the optimal method for avoiding this scenario was to wring out the sponge for 2 s over the bucket before applying to the skin. The water temperature in the FANwet strategy was ~15°C (typical cold tap water temperature), and the benefits were observed without the need to maintain a specific water temperature, as the water was not replaced throughout the trial. Given that this method primarily liberates heat via evaporation, one would not expect large differences in net heat dissipation with warmer or cooler water. However, the application of cooler water directly to skin may elicit a cooler TS because of a greater activation of cold-sensitive cutaneous thermoreceptors, which are densely populated on the neck and face (21,26).

Although TS was not different between the ICE/FANwet and the CON trials, both before each cooling period (precool) and by the end of the game after each cooling period (next game) throughout set 1 (Fig. 3), differences in T re progressively developed across the same timeframe (Fig. 2). This observation demonstrates the potential benefit of beginning to apply either ICE or FANwet cooling strategies from the beginning of a tennis match, even before an athlete starts to feel the effects of the hot environment. Moreover, there was a tendency for a lower RPE even after the application of the ICE intervention after the very first game of set 1. Weighted mean skin temperature (T sk) was lowest with the ICE strategy (Fig. 3). Thirty percent of T sk is determined by a sensor placed on the chest, which would have been covered and therefore directly cooled by the ice-filled towel wrapped around the neck. However, because 30% of the total body surface area was not cooled by the “ice towel,” the reductions in whole-body T sk in the ICE trial are probably slightly overestimated. Nevertheless, reanalysis of whole-body T sk estimated without the chest still yielded a lower T sk in the ICE trial relative to CON in all sets. Both the ICE and the FANwet cooling strategies had a more acute influence on TS compared with the changes observed in T sk (Fig. 3), which can be explained by the strong influence of local skin temperature of the face/neck and hands on whole-body TS despite high skin temperatures elsewhere (2,3).

It was initially hypothesized that the greater conductive cooling with the ICE strategy would elicit skin cooling and a cooler TS but not necessarily lead to sufficiently large differences in overall heat dissipation to lower T re because of the short application time (90-s breaks in play) and the greater barrier to evaporation from towels covering a large portion of skin surface. On the contrary, T re was lower by the beginning of set 2 compared with the CON trial, indicating that the volume of conductive heat loss possible during the short breaks exceeded any reductions in latent heat loss that may have been minimized by evaporation from the towels themselves as they were moistened before applying. Cooling of the neck and thus blood ascending along the carotid artery has been postulated to lower brain temperature (28), which may have also occurred in the ICE trial. However, this would mostly likely not have been reflected by our core temperature measurements as they were measured in the rectum. Even if any cooling of the brain did occur, it did not lead to a cooler TS in the ICE trial compared with the cooling trial that did not involve local cooling of the neck (FANwet trial), and heat tolerance, which is postulated to be higher with brain cooling (19), was actually slightly lower in the ICE trial compared with FANwet (Fig. 1). Nevertheless, considering that the primary purpose of these cooling strategies is to prevent heat-related illnesses, the temperature elevation of the deep viscera, which is well characterized by T re (12,25), was arguably the most important indicator of thermal status to assess because these are the temperature-sensitive tissues that are directly implicated in the development of exertional heat illness (17).

An experimental condition using fan use only without supplemental skin wetting was not assessed in the present study, so we therefore cannot dissociate the relative benefit of applying additional moisture to the skin surface from additional convection in the FANwet trial. Indeed, during maximal sweating, full skin area sweat coverage would be expected in a fully heat-acclimated individual (8), but such coverage would not be attained until steady-state sweating was achieved, which would be unlikely at the start of a match. The benefits of skin wetting were therefore probably most pronounced during the early stages of the simulated match. Augmented airflow could not have had any detrimental effects from a heat transfer perspective in the present environmental conditions because dry heat gain would not have occurred as air temperature was not greater than T sk, and greater evaporation because of a much higher evaporative heat transfer coefficient with additional airflow would have certainly occurred—as evidenced by the smaller elevations in T re. As the ICE and FANwet interventions were found to be equally effective at mitigating thermal strain, tennis players should be encouraged to adopt the strategy that suits their own personal preference and/or that which is available at specific events. Considerations include the time required to prepare “ice towels” and the resources needed (e.g., freezer for >3 kg of ice; power supply for fan [unless battery powered]). Players may also wish to avoid wet hands from the sponge in the FANwet trial; however, the nondominant hand can be used, the application time is relatively short (~5–10 s), water only needs to be applied once, and drying of the wet hand(s) will be accelerated by airflow from the fan for the remainder of the break (>60 s). The findings of the present study are only applicable to hot/humid environmental conditions and not necessarily hot/dry conditions characteristic of tournaments such as the Australian Open (e.g., ~44°C, <10% RH) because amplified airflow would (a) probably increase dry heat gain via convection because of a negative temperature gradient between the skin and the air and (b) potentially not increase evaporation (particularly without supplemental skin wetting) as sweat readily evaporates in dry climates. It also worth considering that the present environmental conditions do not account for any additional heat load that may arise from direct or reflected solar radiation. Finally, although almost identical, physiological, and perceptual responses were observed between ICE and FANwet trials, two participants withdrew because of exhaustion in the ICE trial, but none withdrew for this reason in the FANwet trial. The underlying reason for this difference is unclear, and it is possible that factors contributing to exhaustion other than those measured in the present study may have been altered by the cooling intervention used.

Back to Top | Article Outline


Relative to a control condition, during which only cold water ingestion was permitted during mandated breaks in play, the application of the currently recommended ICE cooling strategy successfully mitigated the development of physiological and perceptual heat strain during a simulated four-set tennis match in hot/humid environmental conditions representative of the peak heat stress historically recorded at the US Open. The FANwet cooling strategy during breaks in play was equally effective as the ICE strategy for reducing thermal strain and may present a more practical alternative in lower resource settings.

This research was supported by funding provided by the Royal Dutch Tennis Federation (Koninklijke Nederlandse Lawn Tennis Bond). Ms. Schranner and Ms. Scherer were supported by an exchange scholarship from the Bavarian State Ministry for Education, Science and the Arts.

The authors thank the participants who volunteered for this study as well as Mr. Nathan. Morris, Dr. Davide Filingeri, and Mr. Mu Huang for their assistance.

The authors declare no conflict of interest. The results of the present study do not constitute endorsement by the American College of Sports Medicine. The results of the study are presented clearly, honestly, and without fabrication, falsification, or inappropriate data manipulation.

Back to Top | Article Outline


1. American College of Sports Medicine, Sawka MN, Burke LM, et al. American College of Sports Medicine Position Stand: exercise and fluid replacement. Med Sci Sports Exerc. 2007;39(2):377–90.
2. Arens E, Zhang H, Huizenga C. Partial- and whole-body thermal sensation and comfort—part 1: uniform environmental conditions. J Therm Biol. 2006;31(1–2):53–9.
3. Arens E, Zhang H, Huizenga C. Partial- and whole-body thermal sensation and comfort—part II: non-uniform environmental conditions. J Therm Biol. 2006;31(1–2):60–6.
4. Bergeron MF. Hydration and thermal strain during tennis in the heat. Br J Sports Med. 2014;48(1 Suppl):i12–7.
5. Bergeron MF, Maresh CM, Kraemer WJ, et al. Tennis: a physiological profile during match play. Int J Sports Med. 1991;12(5):474–9.
6. Binkley HM, Beckett J, Casa DJ, et al. National Athletic Trainers' Association Position Statement: exertional heat illnesses. J Athl Train. 2002;37(3):329–43.
7. Cheung SS, McLellan TM. Heat acclimation, aerobic fitness, and hydration effects on tolerance during uncompensable heat stress. J Appl Physiol (1985). 1998;84(5):1731–9.
8. Cramer MN, Jay O. Biophysical aspects of human thermoregulation during heat stress. Auton Neurosci. 2016;196:3–13.
9. Canadian Society for Exercise Physiology. Certified Fitness Appraiser Resource Manual. Ottawa (ON): Canadian Society for Exercise Physiology; 1986. pp. 1–32.
10. Ellenbecker TS, Stroia KA. Heat research guides current practices in professional tennis. Br J Sports Med. 2014;48(1 Suppl):i5–6.
11. Fernandez J, Mendez-Villanueva A, Pluim BM. Intensity of tennis match play. Br J Sports Med. 2006;40(5):387–91; discussion 91.
12. Gagnon D, Lemire BB, Jay O, et al. Aural canal, esophageal, and rectal temperatures during exertional heat stress and the subsequent recovery period. J Athl Train. 2010;45(2):157–63.
13. Girard O, Christian RJ, Racinais S, et al. Heat stress does not exacerbate tennis-induced alterations in physical performance. Br J Sports Med. 2014;48(1 Suppl):i39–44.
14. Hargreaves M, Febbraio M. Limits to exercise performance in the heat. Int J Sports Med. 1998;19(2 Suppl):S115–6.
15. Jay O, Cramer MN, Ravanelli NM, et al. Should electric fans be used during a heat wave? Appl Ergon. 2015;46(Pt A):137–43.
16. Kovacs MS. A comparison of work/rest intervals in men's professional tennis. Med Sci Tennis. 2004;9(3):10–1.
17. Leon LR, Helwig BG. Heat stroke: role of the systemic inflammatory response. J Appl Physiol (1985). 2010;109(6):1980–8.
18. Morris NB, Bain AR, Cramer MN, et al. Evidence that transient changes in sudomotor output with cold and warm fluid ingestion are independently modulated by abdominal, but not oral thermoreceptors. J Appl Physiol (1985). 2014;116(8):1088–95.
19. Nybo L, Secher NH, Nielsen B. Inadequate heat release from the human brain during prolonged exercise with hyperthermia. J Physiol. 2002;545(Pt 2):697–704.
20. O'Donoghue P, Ingram B. A notational analysis of elite tennis strategy. J Sports Sci. 2001;19(2):107–15.
21. Ouzzahra Y, Havenith G, Redortier B. Regional distribution of thermal sensitivity to cold at rest and during mild exercise in males. J Therm Biol. 2012;37(7):517–23.
22. Parsons K. Human Thermal Environments: The Effects of Hot, Moderate, and Cold Environments on Human Health, Comfort, and Performance. 3rd ed. Boca Raton (FL): CRC Press/Taylor & Francis; 2014.
23. Périard JD, Racinais S, Knez WL, et al. Thermal, physiological and perceptual strain mediate alterations in match-play tennis under heat stress. Br J Sports Med. 2014;48(1 Suppl):i32–8.
24. Ramanathan NL. A new weighting system for mean surface temperature of the human body. J Appl Physiol. 1964;19(3):531–3.
25. Shiraki K, Konda N, Sagawa S. Esophageal and tympanic temperature responses to core blood temperature changes during hyperthermia. J Appl Physiol (1985). 1986;61(1):98–102.
26. Stevens JC. Variation of cold sensitivity over the body surface. Sens Processes. 1979;3(4):317–26.
27. Torres-Luque G, Sánchez-Pay A, Bazaco MJ, et al. Functional aspects of competitive tennis. J Hum Sport Exerc. 2011;6(3):528–39.
28. Tyler CJ, Sunderland C, Cheung SS. The effect of cooling prior to and during exercise on exercise performance and capacity in the heat: a meta-analysis. Br J Sports Med. 2015;49(1):7–13.
29. Tyler CJ, Wild P, Sunderland C. Practical neck cooling and time-trial running performance in a hot environment. Eur J Appl Physiol. 2010;110(5):1063–74.
30. Wenger CB. Heat of evaporation of sweat: thermodynamic considerations. J Appl Physiol. 1972;32(4):456–9.


© 2017 American College of Sports Medicine