Tennis performance is multifactorial, reflected by a sophisticated integration of physiological, biomechanical, psychological, and perceptual elements, all having considerable influence on the outcome of a match.¹ In competition, the integration of these performance facets is challenged by variable environmental conditions, playing surfaces, match durations, and playing strategies. The combination of these factors predisposes players to physiological strain that evolves either involuntarily (for example, hyperthermia, central fatigue)^2,3 or through player negligence (dehydration, hypoglycaemia).^4,5

The consequences of homeostatic disruptions are often overt in elite tennis. However, a review of published reports showed that integrated physiological and performance analyses of professional tennis had not been attempted. Rather, simulated match and tournament conditions were the prevalent research approach.^5,6 Those who analysed match play generally applied a notational approach, recording rally duration, shots per rally, inter-point time, inter-serve time, and comparisons across different court surfaces.^7–9

This investigation describes physiological responses to professional tournament tennis. The influence of physiology on match notation and performance variables (serve velocity, accuracy, serve kinematics, and error rates) was also examined. Both court surface and environmental conditions have a significant influence on physiological responses to match play and thus potentially on notational analyses and performance markers, and so must be considered when interpreting findings. It was hypothesised that during competition professional tennis players experience various types and degrees of physiological strain. A negative association between measures of physiological strain and performance was also proposed.

METHODS

Participants

Fourteen internationally ranked male tennis players participated in the investigation. Their average ranking was 512, range 125 to 882); mean (SD) age, 21.4 (2.6) years; height, 183.0 (6.9) cm; and body mass, 79.2 (6.4) kg. Participants trained on average for 32 hours a week during out-of-competition phases and had at least five years of national level tournament experience. Participants received explicit details of the experimental protocol before voluntarily consenting. Between tournaments a battery of physical tests was carried out and maximum heart rate was determined.

The study was reviewed and approved by the Australian Institute of Sport and the University of Ballarat ethics committees.

Study design

Tournament details

The investigation was carried out during two hard court tournaments and one clay court tournament on the 2003/2004 Australian summer and autumn tennis circuits (37 datasets were produced from 33 matches). Matches in all tournaments were best of three sets, and were contested between 0900 and 1700 hours. Matches were played outdoors and all facilities and equipment were defined according to the International Tennis Federation.¹⁰

Pre-match determinations

Participants carried out their routine game preparation with a few minor modifications. They arrived to the courts with their upon-waking urine sample, having previously ingested (at least three hours before start of play) a core temperature measurement capsule. Immediately before starting each match, participants emptied their bladders before being weighed nude. Subsequently a Polar Team System (Polar Electro Oy, Finland) heart rate chest strap was fitted, and baseline physiological measurements conducted. Nude body mass was measured on electronic scales (UC-300 Precision Health Scale, and Mercury Pty Ltd, South Australia). Core body temperature (T_C) was measured by short range telemetry from the single use capsule residing in the gastrointestinal tract to a data logger, BCTM3 (Fitsense Technology, USA). Urine specific gravity (U_sg) was measured from the upon-waking sample, using either a Bayer Multistix test strip (8SG, Bayer Diagnostics Manufacturing, South Wales) or a digital urine SG refractometer (UG-1, Atago Co, Tokyo, Japan). Capillary blood samples, including blood glucose (measured using a HemoCue (Angelholm, Sweden)), creatine kinase, creatinine, urea, and uric acid were analysed using a Reflotron (Roche Diagnostics, Indianapolis, Indiana, USA) or a Hitachi 911 automatic analyser (Hitachi Ltd, Tokyo, Japan). Reduced blood glucose and raised creatine kinase, creatinine, urea, and uric acid have been reported in athletes carrying out high volume training.¹¹ The total weights of the players’ food and fluids carried to the court were ascertained, using digital electronic kitchen scales (Portion Power, InterTAN Australia).

During match play

During game and set breaks, T_C and fluid consumption were monitored. Water and carbohydrate loaded (6%) sports drinks were available for participants. Serve velocity, serve kinematics, match notation, and environmental conditions (ambient temperature, court temperature, relative humidity (RH), wet bulb temperature (T_WB), and wind speed) were recorded. Matches were filmed from one end of the court, using a digital video camera (Sony DCR-VX2000E PAL, Japan). The footage was used to conduct the retrospective notational analysis using customised software (TenTimer V1.1, AIS, Canberra). A second digital video camera (Sony DCR-TRV950E, Japan) filmed the anterior plane of the participant executing the serve. The footage of individual serves of each game was qualitatively analysed by two elite coaches using a standardised checklist, producing a consistency percentage. The checklist assessed whole body technical components, including smoothness of action, hip and shoulder rotation (separation angle), sequence of body segments (kinetic chain), knee flexion and extension, height of the tossing arm at ball release, position of the ball toss, and the landing distance inside the court after racquet ball impact. First and second serve velocity was measured using a radar gun (Stalker Professional Sports Radar, Radar Sales, Plymouth, Minnesota, USA) placed 5 m behind the server, aligned with ball–racquet contact. Environmental conditions were monitored at five minute intervals using a digital weather tracker (Kestrel 4000, Nielsen-Kellerman, Australia) placed 1 m above the ground on the side court fencing. Court temperature was measured using a digital temperature analyser (Testo 610, Testo Sense, Victoria, Australia) placed on the court surface 50 cm inside the court fencing. Both apparatuses were positioned so that they were unimpeded by shade.

Post-match

Participants returned immediately to the testing area for assessment of nude body mass (towelled dry) and capillary blood samples (creatine kinase, creatinine, and blood glucose). Final weighing of food and fluid consumed was carried out. This ensured that all possible influences on body mass change were accounted for.

Statistical analysis

Data from the two hard court tournaments were pooled, whereas the single clay court tournament data are presented individually. Six participants competed in tournaments on both surfaces. Data from only those six participants were used for statistical analyses. Data were analysed using SPSS version 11.0 (SPSS Inc, USA). Standard descriptive statistics characterised the 14 participants. As a result of significant differences in environmental conditions, comparisons across the different court surfaces were not warranted. Instead, the purpose was to present physiological profiles of tournament tennis and in doing so, infer a role of physiological perturbation on match outcomes. However, in some cases paired t tests compared environmental conditions, notational analyses, and physiological responses between matches played on different court surfaces and pre- versus post-match responses. Where data expressed significant non-normality, non-parametric statistics were used to examine statistical significance (Wilcoxon signed rank test). A correlation matrix was used to identify relations between physiological responses, performance variables, and match notation over the duration of a match. Correlations were conducted using only measurements up to the duration of the shortest match from pooled data of hard and clay court matches. Statistical significance was identified at a probability (p) value <0.05.

RESULTS

Environmental conditions

Ambient temperature (t = −2.988, p = 0.031), T_WB (t = −3.59, p = 0.016), and court temperature (t = −2.731, p = 0.041) were significantly higher during the hard court tournaments than during the clay court tournament (table 1).

Physiological responses

Physiological responses to hard court and clay court tournaments are presented in table 2. The estimated playing intensity averaged 78% and 75% of maximum heart rate for hard and clay court matches, respectively. Post-match body mass deficits were significantly greater during hard court matches (hard courts v clay courts: 1.05 (0.49)% v 0.32 (0.56)%, z = −2.201, p<0.05). Only eccrine and urinary/faecal losses were considered in the calculation of changes in body mass. Pooling data from both hard and clay court tournaments, pre- versus post-match analysis of blood variables showed that blood glucose (t = −3.202, p = 0.008), creatine kinase (t = 4.835, p = 0.005) and creatinine (Wilcoxon z = −3.062, p = 0.002) were significantly higher after the match.

Performance characteristics

Quantitative

Match performance variables are presented in table 3. No significant differences were identified between court surfaces and no correlations were revealed with physiological manifestations. First serve velocity negatively correlated with court temperature (r = −0.772, p<0.05). Second serve velocity correlated with RH (r = 0.708, p<0.05) and negatively with ambient temperature (r = −0.739, p<0.05) and T_WB (r = −0.734, p<0.05).

Qualitative

Longitudinal assessment of serve kinematics revealed correlations (listed below under the headers “First serve” and “Second serve”) between technical inconsistencies and a several match variables, using pooled data from hard and clay court matches.

First serve

Consistency in the height of the tossing arm at ball release inversely correlated with both progressive match time (r = −0.741, p<0.05) and body mass deficit (r = 0.730, p<0.05).

Second serve

Consistency in position and height of the ball toss inversely correlated with both progressive match time (r = −0.729, p<0.05) and body mass deficit (r = 0.725, p<0.05).

Match notation

Match statistics for hard and clay tournaments are presented in table 4. Environmental conditions appeared to influence player recovery time between points, which was significantly longer during hard court matches (hard v clay: 25.1 (4.3) s v 17.2 (3.3) s, t = −6.886, p<0.01). No other match statistic differed significantly between court surfaces. Increasing T_C was the only physiological strain index that displayed a relation with match notation: time between games (r = 0.908, p<0.01), time between points (r = 0.815, p<0.05), shots per rally (r = 0.728, p<0.05), and rally duration (r = 0.794, p<0.05).

DISCUSSION

Physiological profile and potential performance implications

This investigation showed that professional tennis players competed while experiencing moderate thermoregulatory strain and hypohydration. Alongside these findings were mild increases in cardiovascular demand and indicators of muscle damage. Collectively, the effects of hot and humid ambient conditions, repeated ballistic actions, and intermittent exercise over a prolonged period are predisposing factors to homeostatic disruption and ultimately to performance impairment. Specifically, core temperatures approaching hyperthermia levels (exceeding 38.5°C) were common to both tournaments, irrespective of environmental conditions. As a result of the onset of exercise and increased metabolic heat production, this response generally occurred shortly after the start of the match, influencing several match notation variables and predisposing players to premature fatigue,¹² performance decrements,¹³ and heat illness.¹⁴ In this instance, performance implications were technical modifications of the service motion; however, the effects on outcome measures are not clear. It is shown that players experience various facets of physiological strain for substantial periods of play. This generates promise for future investigations of strategies to reduce physiological compromise.

Match preparation and hydration status

Analysis of pre-match hydration status reaffirmed the poor hydration practices of tennis players.¹⁵ Urine specific gravity below 1.020 is recommended for optimum pre-exercise hydration, particularly for prolonged activities in hot and humid conditions.¹⁶ Pre-exercise hyperhydration ascribes thermoregulatory benefits, such as earlier onset of sweating and mitigated increases in core temperature, relative to euhydration.¹⁷ In the current investigation some athletes returned upon-waking U_sg values as high as 1.030, showing the need to drink more in preparation for and recovery from matches. Whether players’ hydration status changed between waking and the start of play was not, however, assessed.

Hydration behaviours during match play under hot and humid conditions further emphasised the need for player education. During 13 of the 17 hard court matches studied participants incurred greater than 1% body mass deficit, consuming only 66 (18)% (2329 (1104) ml) of total fluid loss. On three occasions body mass deficits greater than 2.0% were reported. Comparatively, during only one of the 20 clay court matches studied (played in mild ambient conditions) did participants incur equivalent body mass deficits, consuming 97 (26)% (1908 (694) ml) of total fluid loss. Tennis matches are often played under thermally challenging conditions which would augment the above responses, placing players at risk of performance deterioration and compromised health. There is extensive research demonstrating compromised cognitive¹⁸ and motor/physical¹⁹ performance under states of hypohydration. Body mass deficits observed in this investigation are comparable to previous tennis research^20,21 and are easily offset through adherence to prescriptive hydration regimens.

Substrate availability and hypoglycaemia as a fatigue mechanism

In agreement with previous investigations of moderate²² and prolonged tennis play,²³ the current investigation showed that hypoglycaemia does not manifest during match play. Participants displayed relatively high post-match blood glucose concentrations which are comparable to those in studies of simulated match play.^3,24 It appears that ingestion of carbohydrate beverages is sufficient to maintain blood glucose concentration to match the increased uptake of glucose by active musculature. However, the moderate duration matches studied in the current investigation may not have been sufficient to challenge and lower blood glucose levels. Significant rises in core body temperature were recorded during both tournaments which would increase carbohydrate oxidation.²⁵ Further research of prolonged tennis match play, equivalent to five-set matches, is required before hypoglycaemia can be ruled out as a mechanism inducing performance deterioration.

Match intensity

Average heart rate values support the playing intensity estimates of previous research.^6,24 Maximum heart rate attained during play on hard and clay courts, respectively, equated to 93% and 94% of maximum heart rate previously obtained when participants undertook a 20 m shuttle run test. These observations, in addition to confirming the high intensity intermittent nature of tennis, confirm that cardiovascular strain also presents a physiological challenge to sustained performance during match play. Despite no measurement of recovery rate in this study, it appears that the time between points and games is sufficient for players to recover from the intermittent high intensity periods of play, nullifying the previously reported dramatic effects of cardiovascular duress on tennis skill performance.²⁶

Muscle damage associated with match play

Creatine kinase is an enzyme important for muscle energy production, but its increase in the circulation post-exercise is also indicative of muscle damage.²⁷ In the current study, post-match creatine kinase increased mildly but significantly relative to pre-match values. This response, which could be attributed to changes in plasma volume, occurred despite the players being conditioned to the ballistic movements (disruptive stimulus) of tennis, which is said to reduce muscle trauma.²⁸ It is hypothesised that greater tissue damage would accumulate during a prolonged match and over the course of a tournament. This response could increase a players’ susceptibility to performance impairment and acute or chronic injuries. These observations emphasise the need for players to employ effective recovery strategies between matches.

Match notation

The standard match statistics of the current investigation are comparable to those of previous researchers.^7–9 O’Donoghue and colleagues reported average rally lengths 6.3 (1.8) s and 7.7 (1.7) s, inter-point times of 18.7 (1.9) s and 19.5 (2.1) s, and inter-serve times of 9.3 (1.5) s and 9.2 (1.3) s for men’s Grand Slam tournament matches played at the Australian Open (hard courts) and the French Open (clay courts), respectively. One exception was the length of inter-point time on different court surfaces. O’Donoghue and Ingram⁸ reported longer recovery times during clay court matches and reasoned this was because of the longer rally lengths, whereas players in the current investigation required longer recovery times during the hard court matches. The different findings are likely to be related to the different environmental conditions. Additionally, this investigation revealed that increasing T_C also influenced match notation, highlighting the need to consider the many variables that determine match notation. Overall, the similarity in match statistical data between Grand Slam and entry level professional tennis tournaments adds merit to inferences applied from the overall findings of this investigation to elite tennis.

Correlations over time between physiological responses and performance

This is the first investigation identified in tennis that has attempted to correlate the increasing physiological demand of match play with skills that underscore performance. The qualitative analysis of first and second serves indicated that as matches progressed and adverse physiological conditions developed (for example, increased heart rate, body mass deficits, heat strain), elements of the service action were negatively affected. This finding is consistent with other investigations of elite sports²⁹ and is suggestive of a relation between physiological strain and diminished technical skill performance. Specifically, consistency in the height of the arm at ball release, and the position and height of the ball toss were the most overt technical inconsistencies. General fatigue, possibly modulated through hyperthermia or hypohydration, may also explain performance deterioration, but this possibility could not be confirmed nor discarded on the basis of the available data.

While it could be expected that modifications to coordination and timing of the movement pattern would compromise the outcome of the stroke,³⁰ this was not realised under match conditions as there were no prevailing reductions in serve velocity and accuracy. These findings are in agreement with those of Royal et al,²⁹ who observed altered technical proficiency but unchanged shooting velocity and accuracy with increasing levels of physiological strain in water polo players. This suggests either that the kinematic components that demonstrated inconsistency with physiological strain are not integral to sustained stroke velocity and accuracy, or that highly skilled athletes are able to maintain performance proficiency under duress.

CONCLUSIONS

This investigation provides a comprehensive profile of entry level professional tennis (three-set matches), where match notation reflected that of elite tournaments. Simultaneous measurement of performance and physiological variables infers implications of physiological status on match outcomes. While not directly measured in this investigation, it is proposed that standardised assessment of performance variables in a controlled setting would identify skill deterioration associated with physiological strain. Inadequate pre-match hydration, body fluid deficits incurred during play, heat strain, and early signs of acute muscle damage stand as the dominant adverse responses. It is envisaged that the magnitude of these responses would increase during Grand Slam tournaments (five-set matches). In some instances stroke kinematics were negatively impacted by physiological stress warranting a more detailed examination of this relation.