INTRODUCTION
The Davis Cup is an annual national team competition with more than 100 years of history organized by the International Tennis Federation (ITF), in which male players represent their respective national teams in a knock-out format.1 Until 2020, the format of the competition was structured in four groups based on the teams’ results in previous years. These national teams were promoted or relegated according to the results obtained in the different phases of the competition in a home and away schedule. In each tie, five matches were played to the best of five sets in three consecutive days (Friday to Sunday); with two singles matches played on the first day, the doubles match on the second day, and the last two singles matches on the third day. The playing surface for each tie was selected by the hosting team. This format was used over a long period of time, with players competing in different countries and playing surfaces at different periods of the season. A new format of the competition, which was introduced in 2019, consists of two World Groups (I and II) which include an additional play-off round, and provides opportunities for more nations to compete.2
Although the competitive load in men’s professional tennis is variable depending on the result or round reached in the tournament, due to the extensive competitive calendar of the Association of Tennis Professionals (ATP) and the ITF professional circuit, high-performance players are exposed to a high weekly volume of training and competition (17-19 hours per week).3 In addition to the professional tour tournaments, many of the players also play in the Davis Cup for their respective national teams. Unlike other sports, a tennis match does not have a predefined duration, it varies according to the number of games and sets to be played to determine the winner of the match, and in particular in Davis Cup and Grand Slam matches can be well over five hours of play.4 There is also a broad consensus that over the last two decades there has been a continuous trend in the increase in the speed of strokes and movements and consequently in the overall intensity of play in professional tennis.5–7 Players cover about 3 m per stroke, 8-12 m per point and a total of 600-800 m per set played, with about 70% of strokes being the sum of first serves, forehand and backhand rallies.4,8,9 This activity occurs at physiological intensities ranging from 12% to 80% of maximal oxygen consumption (VO2max) and 47% to 86% of maximum heart rate (HRmax), depending on the specific demands of each point or rally.10,11
Regarding health problems associated with the sport, it has been found that playing long matches on consecutive days generates neuromuscular and perceptual fatigue in the player, as well as an increase in muscle damage markers.12 There are also factors that increase the competitive stress of the player, as matches are usually played outdoors in high temperatures or humidity conditions, which have been associated with increased medical problems and retirements in Grand Slam players.13 A number of authors have explored the incidence of injuries in professional male tennis.14–18 Incidence rates of 8.25, 3.34 and 6.64 per 1000 match exposures (EP) have been observed in ATP tournaments, on the United States Tennis Association (USTA) professional circuit and in Davis Cup respectively.15,17,19 A retirement in tennis means that a player stops playing once the match has started, and the victory is automatically awarded to the opponent. Retirements in professional tennis and their causes have also been studied.15,17,19,20 However, these studies included data from 1968 to 2013. Given the specific competition features of the Davis Cup, only one of the studies analyzed the reasons for retirements in this competition.19 Consequently, it is relevant to determine whether the increase in the intensity of play, as described above, may have led to a change in the patterns of retirements in the Davis Cup. Therefore, the aims of this study were to describe the pattern of the epidemiology of retirements, and to determine the factors that influence match retirements over the last twenty years in Davis Cup.
METHODS
Study Design and Participants
The Strengthening the Reporting of Observational Studies in Epidemiology statement21 was utilized to control potential biases in the design phase (Table S1 in Supplemental material). The data set contains 6060 Davis Cup singles matches played between 2000 and 2019. Six thousand thirty-six matches were analyzed after exclusion of those matches that were walk-overs (W/O; n = 24). These matches were played by professional male tennis players from all qualifying groups and regions in which the ITF structures this competition. The matches were played throughout the year (i.e., at different times of the season), and on all four types of surfaces (i.e., clay, hard, carpet, and grass). The study focused on Davis Cup singles matches and relied solely on publicly available data. As there was no direct interaction with human subjects or animals, ethical approval was not applicable for this research.
Definitions
According to the current rules, a tennis match is officially recorded as a retirement when a player cannot continue due to illness or injury after the match has started.22 Only retirements due to medical conditions (injuries and illnesses) were included, not taking into account a considerable number of retirements that occur due to non-medical reasons, such as match suspensions due to weather conditions, unsportsmanlike behavior of a player, and other reasons.
Study procedure and Variables
For this study, data of 20 years of Davis Cup singles matches played between 2000 and 2019 were included from the datasets collected by GitHub, which are available at https://github.com/JeffSackmann. The four competition groups were unified into a world group (WG I and WG II) and a non-world group (G III and G IV) for the analysis. Specifically, the following variables were evaluated: anthropometric data (players’ age and height), match characteristics (surface, number of sets played: best of three or five), group (World Group or not), round (final rounds: finals, semi-finals, quarterfinals, or other), number of matches, total number of games played per match, average age of the players in the match, and retirements (including reasons, and anatomy and body region of the injuries).
Furthermore, new variables were added to the original data set: type of retirements (medical conditions, non-medical conditions, unknown reasons), sources of information on retirements (location of the player: home, away, neutral) which were obtained following an exhaustive search for each retirement using the Google search engine and specialized websites such as https://www.daviscup.com/; https://www.savannahsworld.blogspot.com; https://www.eurosport.com, and websites such as https://www.newspapers.com and https://www.theguardian.com. For the search strategy, keywords such as “Davis Cup”, “injury”, “retirement” (accompanied by the identifying data of the match for which information was sought) were used, and sometimes adding the full name of the retired player.
For the complementary analysis of medical conditions (i.e., injuries and diseases), the classification determined in the tennis-specific extension of the International Olympic Committee consensus statement, developed specifically for tennis was utilized.23 Medical conditions were classified in terms of location in four body regions: head/neck, upper extremities, trunk, and lower extremities. The type of injury was classified as: muscle/tendon, bone, joint/ligament, skin, central/peripheral nervous system or other. Illnesseswere classified according to the organ system affected as: cardiovascular, respiratory, gastrointestinal, renal, urogenital, metabolic/endocrinological, haematological, dermatological, neurological, psychiatric, ophthalmic/otorhinolaryngological, dental, rheumatological, allergic, infectious, environmental and others.24
Statistical analyses
A descriptive analysis was conducted using the data from 6036 Davis Cup matches played between 2000 and 2019. This involved the calculation of both absolute and relative frequencies for all qualitative variables. Measures of central tendency (mean and median) and dispersion statistics (standard deviation and range) were determined for the quantitative variables. Furthermore, various potential influencing factors associated with match outcomes were explored. These factors included the type of playing surface, the ‘best of sets’ format, group classifications, stages of the competition (round), and the sequential number of each match. These variables were examined at a bivariable level, with group comparisons examined using chi-squared tests or Fisher’s exact tests, depending on data distribution and sample size. The significance level was set at 0.05.
Epidemiological measures
Retirement incidence rate was calculated as the number of retirements per 1000 games. To calculate the unadjusted incidences, the epi.2by2 function of the R package epiR25 was mainly used, setting method as cohort.time. The number of retirements, exposure (no. games), incidence rate and the 95% confidence interval (95%CI) are provided. The raw incidence of retirements over time was studied representing their values during the range of years included in this study. In addition, adjusted relative and absolute measures of association are given as incidence rate ratio (IRR) and risk difference (with 95%CI’s). At the multivariable level, a generalized linear model was fitted, assuming the frequency of the retirements followed a Poisson’s distribution. The relationship of the incidence of retirements with age was studied, observing different behavior in two groups (< 27 years, ≥ 27 years; (Figure S1 in Supplementary Material). Consequently, two different slopes for both groups were considered in the model.
The model expression for i-th match is the following: log (λi) = log(gi)+ Xi β+ ui where λi is the expected number of retirements following a Poisson distribution, gi is the number of game exposures of matches, which in turn is the offset of this model, and Xi is the matrix containing all the independent variables of interest. Model selection was based on the Akaike Information Criterion (AIC). The fitted models were tested for overdispersion with the dispersion test function in the R package AER.26 The final model considered the interaction between the variables related to the number of sets (three or five) and the number of matches (1st; 2nd; 3rd; 4th). In addition, a logistic model was adjusted by variables prior to the match with the aim of providing the probabilities of retirement under different scenarios.
All analyses were carried according to the checklist for statistical assessment of medical papers (CHAMP) statement.27 All analyses were performed using version 4.1.3 of the R statistical software.28 The R code of the analysis carried out in this study are available at: https://github.com/marticasals/Davis_Cup_Retirements/tree/main.
Data availability statement
Data and the reproducible code used in this study are available on a publicly accessible GitHub repository (https://github.com/marticasals/Davis_Cup_Retirements/tree/main), allowing for the transparency and replicability of the statistical analysis. This ensures that other researchers can reproduce the findings of the study and build upon it for further research.
RESULTS
General characteristics of Davis Cup matches according to retirements
A total of 6036 Davis Cup matches were included during the period from 2000 to 2019. The players had a median age of 24 years, with ages ranging from 15 to 46 years, and a median height of 185 cm, varying from 160 to 208 cm. Table 1 presents a descriptive overview of Davis Cup matches during this 20-year period, with a categorization based on the presence or absence of retirements. Of all matches, 5845 (96.8%) concluded without retirements, while 191 (3.2%) had retirements.
In terms of playing surfaces, most matches were played on clay (37.6% of matches: 97.0% without retirements, 2.96% with retirements) and hard (49.5 % of matches: 96.4% without retirements, 3.62% with retirements), with grass (2.9% of matches) and hard courts experiencing a higher frequency of retirements. The most common competition format was to best of five sets (96.0% without retirements, 4.02% with retirements). Regarding the competition stage, most of the analysed matches were seen in the preliminary rounds, although there were similar percentage of retirements in the qualifying and the final rounds (3.18% and 3.10%, respectively).
The analysis of the distribution of the retirements according to the match order in each tie showed a higher percentage of retirements occurring in the last two matches played (3.72% and 3.74%) compared to the first two matched played (2.42% and 2.87%). On the other hand, a noticeable disparity in the number of games played per match in the matches with (20.6 ± 11.3 games played) and without (30.4 ± 10.7 games played) retirements was observed. In fact, there were 34% fewer games played in matches with retirements, highlighting the potential impact of retirements on match length and intensity. The average age of the players in the matches with retirements was 24.4 years (SD: 3.36) while players in the matches without retirement 24.7 (SD: 3.04) (Table 1).
Incidence of retirements in the Davis Cup
Through the period analyzed, there was an exposure time of 181414 games. The general retirement incidence was of 1.05 retirements per 1000 games (95% CI: 0.90, 1.21) during Davis Cup play. Three specific years (2012, 2018 and 2019) had a remarkably lower retirement incidence (Figure 1).
Both relative and absolute measures of retirement incidence in Davis Cup matches reveal a higher risk of retirements on matches played on hard courts, to best-of-five-set matches, and in the last two Davis Cup matches of the tie (Table 2). For instance, the fourth match showed the highest incidence rate in the Davis Cup, with 1.54 retirements per 1000 games. Conversely, the first match displayed the lowest retirement incidence, with 0.73 retirements per 1000 games. This indicates an IRR of retirements that is 2.10 (95% CI: 1.35, 3.28) times higher, which implies a greater risk of experiencing a retirement in the fourth match compared to the first. The risk difference between them was 0.80 (95% CI: 0.32, 1.29) retirements per 1000 games (Table 2).
Multivariable analysis of Davis Cup retirements
The IRR for the different factors of the final model chosen in matches of Davis Cup during 2000-2019 is shown in Figure 2.
Regarding play surfaces, the incidence of retirements on clay, hard, or grass court was more than double than on the carpet surface, which was statistically significantly different. The incidence of retirements is also different depending on the age of the players. If they are younger (mean age below 27), the incidence rate decreases by half over five-year period. Meanwhile, if the mean age of the players is at least 27 years, the incidence rate doubles over a five-year period, but without reaching a statistically significant difference.
Finally, the number of the singles matches also plays a role in the number of retirements, depending on the number of sets for the tie: If the matches are at best of three sets, retirements are less usual in the third and fourth matches. On the contrary, if the matches are at best of five sets, retirements are more likely to happen in the last matches of the tie.
A table showing the retirement probabilities per match based on the variables available prior to the start of the match was generated to assess the likelihood of retirements in tennis matches (Table 3). For instance, on hard courts, approximately one out of every 50 matches results in a retirement, whereas on carpet, retirements occur in approximately one out of every 125 matches. In terms of age, the highest probability of retirement is observed at the extremes of the age spectrum, such as 18 and 36 years old, with a probability of approximately one retirement in every 18 matches. Conversely, for ages between 26 and 28, the probability of retirement decreases to one in every 56 matches. Regarding the number of matches, the highest retirement probability is found in a five-game tie during the last match, where there is one retirement in every 15 matches.
Reasons for Retirements in Davis Cup Matches: Injuries, Illnesses, and Non-Medical Causes
Information on the reason for retirements was obtained for 74 (38.7%) of the 191 retirements in the period analyzed, while 61.3% of the retirements (n=117) were due to unknown causes. Of these 74 retirements, 73.0% (n=54/74) were due to injury, 14.9% (n=11/74) were for non-medical reasons, and 12.2% (n=9/74) were due to illness. Of the injuries, 53.7% (n=29/54) occurred when the injured player was playing away, and 46.3% (n=25/54) when playing at home. The most frequent injuries were to the lower extremities (n=34/54; 63%), were to the muscle/tendon (n=29/54; 54%) and were located in the thigh (n=15/54; 28%), as shown in Table 4. In terms of retirements due to illness (n=9), the most recurrent type of illness was metabolic causes with 55% (n=5/9) of the total number of illnesses, followed by infectious diseases (n=2/9; 22%), gastrointestinal diseases (n=1/9; 11%), and respiratory diseases (n=1/9; 11%).
DISCUSSION
This study aimed to describe the frequency of injuries and the main match factors that influence player retirements in Davis Cup matches. The analysis of 6036 Davis Cup matches spanning two decades (2000-2019) revealed a player retirement incidence of 3.2% or 1.05 retirements per 1000 games which was reasonably constant during the analyzed period. The hard-court surfaces, the matches played in best-of-five-sets, and the last two matches played in the tie were the variables that showed the highest incidence rate. Moreover, the youngest and the oldest players had higher probability of retirement. The most common reason for the retirement was injury, especially in the lower-extremities due to muscular and tendon injuries.
Previous authors have reported that the playing surface significantly affects the external and the internal match load,11,29 and the different court characteristics and game styles of tennis players might affect their retirement rate.30 The present study highlights hard court surfaces as a key factor contributing to the high incidence of retirements, as has been previously observed.31 Playing on hard courts leads to high-intensity changes of direction, resulting in high rates of acceleration and deceleration, which consequently increases the load on muscles and tendons.32 These surfaces, coupled with the endurance challenge of the best-of-five set matches, seem to increase the physical strain on players. These factors have a clear impact on retirements, as the hitting and movement workload in best-of-five matches is significantly greater than in best-of-three-set matches.9 Furthermore, the knockout format played througout the year seems to also play a critical role, underlining the multifaceted nature of challenges faced by athletes in this elite competition.
The injury pattern found, which predominantly affected the lower extremities (62.96%), and was often related to muscular or tendinous (53.7%) and joint or ligament (42.6%) issues, aligns with the existing literature on injury trends in professional tennis.19,33 However, it is important to note that while acute injuries are more prevalent in the lower body, chronic injuries tend to affect the upper extremity.34 Most tennis injuries occur in the lower extremity (31%–67%), followed by the upper extremity (20%–49%), and lastly, the trunk (3%–21%),14 highlighting the importance of understanding the specific regions of the body that are more vulnerable to injuries in tennis. While the current study focused on retirements in Davis Cup matches, the broader context of tennis injuries highlights the need for adopting comprehensive injury prevention measures that consider the unique demands tennis places on different body regions.
A novel finding is that the incidence of retirements occurred during the preliminary rounds (3.18%) and the final phase (3.10%) of the Davis Cup are similar. This finding might show that the pressures faced by players in different competitive stages (preliminary vs. final stages) in which retirements occur are similar. On one hand, preliminary rounds often serve as a gateway for players striving to secure a spot in the main draw of the Davis Cup, and the pressure to perform at their best can be immense. Furthermore, players in the preliminary rounds may have already navigated through multiple matches to reach this stage, potentially leading to physical and mental fatigue.35 On the other hand, retirements during the final phase of the Davis Cup may reflect the culmination of a long and gruelling journey over the course of a whole season. The fact that the final phase is played late in the season, when players may already be experiencing cumulative physical and mental fatigue, could be considered a contributing factor. Players who have advanced to this stage have already demonstrated their skill and endurance, but they now face the added pressure of representing their country on a prestigious international stage. The desire to perform exceptionally for their nation, coupled with the accumulated physical and psychological toll of earlier matches, could be contributing factors to retirements during this phase. Future research in this area should delve deeper into understanding the specific factors that influence retirements during these critical phases of the Davis Cup. This could include examining the psychological aspects of player performance, assessing the impact of match scheduling and recovery strategies, and exploring the role of team dynamics and national pride in influencing players’ decisions to retire.
Limitations
Limitations of this study include lack of comprehensive data regarding the reasons for retirement in 110 cases. One specific aspect that warrants attention is the documentation of injuries. The authors acknowledge that this study lacks detailed information on how injuries were documented during the matches. Understanding the precise nature and severity of injuries is essential for a comprehensive analysis of retirements in professional tennis. While efforts were made to gather data on retirements, the absence of detailed injury documentation is a limitation that future research should aim to address. This data gap suggests an opportunity for more granular data collection in future research to enable a more detailed exploration of retirement causes in professional tennis. Ensuring consistency in injury surveillance in tennis would greatly improve the understanding and management of injuries in the sport.
Additionally, it is important to note that while the current findings suggest an association between certain match conditions (such as hard courts and best-of-five-set formats) and a higher incidence of retirements, the study design does not allow the determination of specific factors underlying these associations. For example, hard courts might increase the risk due to their surface properties, but other factors such as match duration, player fatigue, or tournament schedule could also contribute to the higher retirement rates observed. Future research should aim to explore these specific factors to better understand the underlying causes of retirements.
Practical Applications
This study presents a comprehensive analysis of retirements in Davis Cup tennis matches over a 20-year period (2000-2019), filling a notable research gap. By examining this extensive timeframe, it unveils potential shifts in retirement patterns and sheds light on the evolving landscape of player injuries and match retirements, contributing significantly to the understanding of athlete health in professional tennis tournaments. The findings from this study can help coaches better understand the risk factors associated with high-level team competitions and lays the groundwork for more targeted and effective policies aimed at safeguarding athlete health in the competitive realm of Davis Cup tennis.
CONCLUSIONS
This results of this study indicate a retirement incidence of 3.2% or 1.05 retirements per 1000 games, with a consistent trend throughout the analyzed period. Higher retirement rates were observed on hard courts, during best-of-five-set matches, and in the last two matches of elimination rounds, highlighting specific conditions that increase the risk of retirements. Younger and older players were found to have a higher probability of retirement, indicating that the extremes of the age spectrum are more vulnerable. The most common cause of retirement was injury, particularly affecting the lower extremities, muscles, and tendons, underlining the need for targeted injury prevention.
Conflicts of interest
The authors report no conflicts of interest.
Funding Statement/Acknowledgments
This research was funded by the Ministerio de Ciencia e Innovación (Spain) (PID2019-104830RB-I00) and the Departament de Recerca i Universitats de la Generalitat de Catalunya (Spain) [2021 SGR 01421 (GRBIO)]. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.