Tennis is a global sport, with participation in more than 200 countries affiliated with the International Tennis Federation.¹ It is also a professional sport in which millions of dollars in prize money are at stake for both men and women players. In the Netherlands, it is the second most popular sport, with more than one million participants from a population of 16 million.² Among Dutch women it is actually the most popular sport.³ In other European countries tennis also ranks high on the list of popular sports.³

Like many other sports, playing tennis—at either a recreational, collegiate, or professional level—places participants at risk of injury. Though many injuries that occur in tennis are common to other sports, tennis does have a unique profile of injuries.⁴ Differences in equipment, biomechanics, and physical demands result in an injury profile that differs from other racquets and throwing sports.⁴ Sports injuries, including tennis injuries, are a common cause of disability and, in some cases, absence from work.^5–7 This can have substantial socioeconomic consequences, both on a personal and a societal level.⁸ For these reasons it is important to develop effective measures for the prevention of tennis injuries.

To develop prevention strategies, both the incidence and severity of tennis injuries must be determined. The severity of an injury can be described on the basis of the nature of the injury, the duration and nature of treatment, time lost from sports participation or work, permanent disability, and cost.⁹ Another important step is to determine risk factors and other mechanisms that are associated with these injuries.^9,10 This aetiological research entails understanding the causes of injury, with the goal that modification or removal of these causes can prevent the occurrence of the injuries.¹⁰ The next step consists of the formulation of preventive measures.¹⁰ These measures must be evaluated with regard to their effectiveness before implementation. Ideally, evaluation should include randomised controlled trials.^10–12

Our aim in this review of published reports was to provide an overview of the available scientific knowledge on the occurrence, aetiology, and possibilities for prevention of tennis injuries. We asked the following three questions. First, what are the most common tennis injuries, based upon the reported prevalence and incidence figures? Second, what associated risk factors and mechanisms are described with regard to the aetiology of tennis injuries? And third, what is known about the efficacy of prevention efforts designed to reduce the occurrence of tennis injuries?

Another purpose of this review was to identify gaps in knowledge with respect to the occurrence, aetiology, and prevention of tennis injuries and to encourage further methodologically sound epidemiological research in this field.

METHODS

We undertook a literature search to retrieve potentially relevant articles published since 1966. The following electronic databases were explored: Pubmed (from 1966 to October 2005), Embase (from 1989 to October 2005), and Cumulative Index to Nursing and Allied Health Literature (CINAHL) (from 1982 to October 2005). A priori defined search terms (Medical subject heading (Mesh) and text words) that were used in this search were: “injury”, “injuries”, “prevalence”, “incidence”, “incidence density”, “proportion”, “distribution”, “population”, “aetiology”, “etiology”, “mechanism”, “risk factor”, “risk factors”, “prevention” and “intervention”. These terms were combined with “tennis”. Reading titles and abstracts identified potentially relevant articles. Citation tracking of the articles retrieved was also performed to identify additional relevant articles.

To be included in this review studies had to meet the following inclusion criteria: they must contain data on tennis injuries; they must investigate the frequency of tennis injuries, the aetiology (for example, risk factors) of tennis injuries, the efficacy of prevention strategies, or a combination of these purposes; and they must have been published in English, German, or Dutch. Studies focusing on treatment for tennis injuries and literature reviews were excluded. For the purpose of this review we defined a tennis injury as a musculoskeletal problem requiring reduction or interruption of tennis activity for any length of time, with or without evaluation or treatment by a health care provider.¹³

We did not expect to find many cohort or randomised controlled studies in this field. Furthermore, with the expected heterogeneity in study designs and methods, we elected not to follow a formal meta-analytic approach. The studies retrieved were classified as case reports, laboratory studies, descriptive epidemiological studies, analytic epidemiological studies, or intervention/prevention trials. A similar approach was conducted earlier by Pollack et al^14,15 with regard to the available evidence for the prevention of softball injuries. For reasons of clarity we defined descriptive epidemiological studies a priori as cohort studies (either cross sectional or longitudinal), describing the frequency (that is, prevalence or incidence or both) of tennis injuries in a cohort or subcohort. Analytic epidemiological studies were defined as cohort studies (either cross sectional or longitudinal) which aimed to estimate a measure of association (that is, odds ratio, risk ratio, hazard ratio) between risk factors and the occurrence of tennis injuries. The results of the selected studies will be described and summarised to formulate answers to the research questions posed above. The emphasis lies on the results of descriptive and analytic epidemiological studies, and intervention/prevention studies rather than laboratory studies or case series and reports.

RESULTS

Our search in the Pubmed, Embase, and Cinahl databases resulted in, respectively, 1368, 1617, and 2460 potentially relevant hits. To identify appropriate papers for the present review, the titles and abstracts were read and, if considered relevant, selected by two persons (BMP and JBS). In cases of disagreement further discussion was undertaken to achieve consensus. We found 39 case reports, 49 laboratory studies, 28 descriptive epidemiological studies, three analytic epidemiological studies, and no intervention study which met the inclusion criteria of the present review. Table 1 provides an overview of the distribution of study type and body region within the relevant articles.

Case reports

Of the 39 case reports, 29 dealt with injuries of the upper extremity,^16–44 eight with injuries of the lower extremity,^45–52 and two with the trunk.^53,54 Of the case reports, the most common condition in the upper extremity injury section was stress fractures (14 case reports).^16–29 Stress fractures in the upper extremity included the metacarpals, hamate bone, radius, ulna, and humerus. The suggested causal mechanism involved repeated loading on the upper extremity during the tennis stroke,^27,28 and included the impact of the racket butt against the palm of the hand,^16–18 and high torsional stresses.^25,27,28

Vascular injury in the upper extremity was mentioned four times.^32–35 Though an uncommon injury, it has also been reported in overhead batting and racquet sports. Vascular injury may result from compression of the large vessels in the axilla during the service motion, resulting in aneurysm formation.³⁵ Distal embolisation may occur.³⁴ Endothelial injury caused by repeated microtrauma to the hand by the racket was also reported.^32,33

Regarding the lower extremity, injuries were more equally distributed and included case reports on tendon injuries,^46,48 plantar fascia tears,^49,52 muscles tears,⁴⁵ stress fractures,^47,50 and intra-articular knee injury.⁵¹

Laboratory studies

Forty nine laboratory studies were identified: 36 involved the upper extremity,^55–90 six the lower extremity,^91–96 three the trunk,^97–99 and four the whole body.^100–103 In the upper extremity articles, the following topics were discussed most often: range of motion (seven studies)^{55,58,61–65} and strength (nine studies)^{55–60,62,65,66} of the shoulder, and biomechanical analysis of the stroke (six studies).^{69–71,74,75,77}

In the studies examining range of motion of the shoulder, internal and external rotation was measured using a goniometer. In six^58,61–65 of seven studies^{55,58,61–65}, a significant decrease of internal rotation and total range of motion was demonstrated in the dominant arm. Kibler et al⁶⁴ showed that the loss of total range of motion was progressive with age and years of tournament play.

Muscular strength of the shoulder was determined by isokinetic testing. Five^{58–60,62,65} of seven studies^{55,56,58–60,62,65} showed an imbalance of muscle strength, with significantly greater isokinetic strength in the dominant arm than in the non-dominant arm for internal rotation, leading to a reduced external/internal rotation ratio. Both the loss of internal rotation motion and the muscle strength imbalance were hypothesised to increase the risk of shoulder injuries.

In three^91,95,96 of the six articles^91–96 focusing on the lower extremity, the interaction between shoe and court surface was examined. The main conclusion of these studies was that lateral stability of the shoe is important in the prevention of injuries.

Descriptive epidemiological studies

Twenty eight descriptive epidemiological studies were identified, including 19 on tennis injuries in general,^104–124 seven on injuries of the upper extremity,^125–132 one on the lower extremity,¹³³ and one on the trunk.¹³⁴

Injury incidence

Injury incidence varied from 0.05^122–124 to 2.9¹¹⁹ injuries per player per year (table 2). Per hour of play, the reported incidence varied from 0.04 injuries/1000 hours¹⁰⁸ to 3.0 injuries/1000 hours.¹⁰⁵ Incidence and prevalence rates for tennis elbow were quite high, with reported incidence varying from 9%¹²⁸ to 35%¹³⁰ and prevalence varying from 14%¹²⁸ to 41%.^131,132

Injury localisation

Ten^{104–106,108,110–113,117,119} of 13 studies^{104–106,108,110–113,116,117,119,121–124} showed a preponderance of injuries of the lower extremity compared with the upper extremity.

Injury type

Four^{108,112,113,118} of six studies^{105,108,112,113,116,118} reported more acute than chronic injuries. Most acute injuries occurred in the lower extremities, whereas most chronic injuries were located in the upper extremities. Injuries to the trunk comprised 5% to 25% of all injuries.^{104–106,108,110,111,113,116–119,121–124}

Injury severity

Injury severity was expressed in various ways in the different studies, including number of injuries requiring hospital admission^108,113 or operative treatment,¹⁰⁶ average medical costs per injury,¹⁰⁸ time loss,^114,115 or the percentage requiring medical treatment.^104,109,112 Injuries sustained while playing indoors tended to be more severe than outdoor injuries, with a higher percentage requiring medical treatment.^104,109 In the study by Kuhne et al, 3.3% of acute and 2.2% of chronic injuries required surgery.¹⁰⁶ Five per cent of the injuries in the Letsel Informatie Systeem (LIS) study required an average of five days of hospital admission.¹⁰⁸ In the studies on juniors, injury severity was significantly less, with only one player of 1440 being taken to hospital and two injuries of 176 requiring surgery.¹¹³

Sex

Injury rates between men and women were compared in a prospective cohort study of intercollegiate tennis.¹¹⁵ In this study, 1.6 injuries per 1000 hours were recorded for male tennis players versus 1.0 injury per 1000 hours in female players. This was not a statistically significant difference (p = 0.37).

Sallis et al¹¹⁰ studied injury patterns in 18–22 year old tennis players. This was a retrospective cohort study of injury reports compiled by certified athletic trainers. The incidence was 0.46 injuries per male player per year and 0.42 injuries per female player per year. The differences was not statistically significant.

Hutchinson et al¹¹³ compared injury patterns in elite junior players (male and female) during a three year period (1996–1998) at the United States Tennis Association (USTA) tennis championships. There was no significant difference in the overall rate of injury (new and recurrent) between male and female players.

Winge et al¹¹⁶ found a higher injury rate in men (2.7 injuries per 1000 hours) than in women (1.1 injuries per 1000 hours). This was a statistically significant difference (p<0.05).

Age

The Letsel Informatie Systeem¹⁰⁸ is a continuous registration of injuries treated in the emergency departments of a selection of 15 hospitals and medical centres in the Netherlands. These injuries are generally acute and more serious. In this study, injury risk in tennis has been shown to gradually increase with age, from 0.01 injuries per player per year in the 6–12 year age group to 0.5 injuries per player per year in those over 75 years of age. An increased incidence with age was consistently shown for tennis elbow.^127–132

Level of play

In these general descriptive epidemiological studies, the study populations can be characterised as recreational/general, elite, or junior competitive. The studies include a wide distribution of retrospective, cross sectional, prospective cohort, and prospective longitudinal study designs. There were nine studies^{104–106,108,109,111,112,118,121} involving recreational players or the general population, seven studies^{110,115–117,119,120,122–124} involving elite players, and three studies^107,113,114 relating to junior tournament players. Study designs of junior and elite players often involved recording of injuries as medical consultations at tournaments or training centres. This method of injury reporting may inflate injury rates, therefore making it difficult to make direct comparisons with studies involving recreational players that often involve self reporting of their injuries.

We were able to identify only two studies that compared injury rates between players of different ability. Baxter-Jones et al¹¹⁴ studied elite young athletes. They found that performance success was significantly related to injury rate. Jayanthi et al¹⁰⁵ described the incidence and prevalence of injuries in recreational players of different skill levels, ranging from International Tennis Number 3 to 8. Despite trends, there were no statistical differences in overall injury incidence and prevalence rates across all skill levels.

Volume of play

Studies describing the risks associated with volume of play are scarce. Increased playing time was associated with increased incidence of new cases of tennis elbow in recreational players playing more than two hours a day versus those playing less than two hours a day.¹²⁸ However, total incidence and prevalence of all tennis related injuries was not different among recreational players who played less than four hours a week, four to six hours a week, or more than six hours a week.¹⁰⁵

Analytic epidemiological, and intervention studies

Three studies were found that investigated risk factors for tennis injuries (table 3). Two of these^135,136 had a longitudinal study design and one¹³⁷ was cross sectional. Adjustments for confounding variables was made in one longitudinal cohort study.¹³⁵ The type of sports injury described in the investigated studies was variable and consisted of discomfort or pain from wearing tennis shoes, sport related injuries in general, low back pain, and osteoarthritis.

Llana et al¹³⁷ described in a cross sectional study a significant correlation (p = 0.02) between perceived incorrect arch support and plantar discomfort. Spector et al¹³⁵ found that long term weight bearing sports activity was associated with the development of osteoarthritis.

No intervention study was retrieved investigating the effects of prevention measures on tennis injuries.

DISCUSSION

The principal findings of our study are first, that there is a great variation in the reported incidence rate of tennis injuries; second, that most injuries occur in the lower extremities, followed by the upper extremities and then the trunk; third, that there are very few cohort studies available that estimate a measure of association between risk factors and occurrence of tennis injuries; and fourth, that there are no randomised controlled trials on preventative measures in tennis.

The variation in the reported incidence rates of tennis injuries most probably reflects variation in injury definition, study design, populations under study, methods of data collection, and the duration of follow up or recall period. The lowest incidence rate (0.04 injuries per 1000 players per year) was reported in the LIS study.¹⁰⁸ Injuries in this study included only those for which the player was treated at a hospital casualty department. This implies that predominantly more acute and serious injuries will be reported, as players with less serious and chronic injuries are more likely to visit their general practitioner, physiotherapist, or sports physician. The other study with a relatively low injury rate (0.11/1000 hours of play) was by Weijermans et al.¹¹² In that study, injuries sustained by tennis players at a club had to be reported to a contact person in order to be recorded. This may have resulted in underreporting of injuries. Biener et al^122–124 also reported a very low injury rate, which can be explained by their long recall period of 17.5 years.

The highest injury rates were found by Hutchinson et al¹¹³ and Silva et al.¹⁰⁷ This is undoubtedly related to their rather inclusive injury definitions: “any medical problem that required physical or medical assistance”¹¹³ and “any consultation and/or treatment given to a player during a tournament on site”,¹⁰⁷ respectively. Using these definitions, injuries which may not have had any effect on tennis play, time loss, or work were also included. Kuhne et al addressed this problem by making a separate category for “Bagatellverletzungen” (minor injuries), which included sunburns, abrasions, and blisters. We were not able to find any study that identified the relation of match volume within a tournament or through a season and the risk of injury. Prospective studies of independent risks associated with increased playing time in junior tournament players are lacking and necessary in order to counsel parents, coaches, and tournament directors with appropriate evidence based recommendations.

Despite the wide variation of reported injury rates and study designs, comparisons of injury rates in tennis can be made with the rates in other sports. In order to make optimal comparisons, similar study designs and injury definitions should be used. There were 377 injuries in 456 matches involving all team sports studied during the 2004 summer Olympic Games.¹³⁸ There was a total injury incidence of 0.8 injuries per match and 54 injuries per 1000 player matches, where injury was defined as any physical complaint incurred during the match that received medical attention regardless of consequence. Handball players (114/1000 player matches) and soccer players (108/1000 player matches) had the highest injury rates, while volleyball players (7.7/1000 player matches) had the lowest. With a similar definition of injury and comparable study design, Hutchinson et al¹¹³ reported 21.5 injuries/1000 athletic exposures and Silva et al¹⁰⁷ reported 6.9 medical appointments/1000 games in prospective studies of junior national tournament tennis players.

It may be more appropriate to compare tennis with other individual non-contact sports rather than contact team sports. A retrospective cohort survey study in golfers reported 3.06 injuries/player injured in professional players of average age 36.5 years, and 2.07 injuries/player injured in amateur players of average age 47.2 years.¹³⁹ This study did not report the total prevalence of injuries/100 players but had a total of 637 injuries in 703 golfers surveyed. In a cross sectional survey of recreational tennis players with an average age of 46.9 years, there were 299 injuries in 528 players giving a prevalence of 52.9 injuries/100 players.¹⁰⁵ In a cross sectional survey study of recreational runners, 45.8% of 4358 male joggers sustained jogging injuries in the previous one year period.¹⁴⁰ A prospective study of recreational runners training for a 10 km race reported that 29.5% of runners experienced an injury that caused at least some pain after exercise.¹⁴¹ Comparable prospective studies of recreational tennis players over a six month period reported injury rates of 0.11/1000 hours played¹¹² and 1.5 injuries/player/year.¹⁰⁶

Despite some variation in study design and definition of injury, tennis appears to have lower injury rates than contact team sports and also in some comparisons with non-contact individual sports such as golf and running. However, no known study has made direct comparisons between risks of injury or lifetime prevalence of injury between tennis and other sports.

Unfortunately, we were not able to identify any intervention studies on tennis injuries. An intervention study by Kibler et al,¹⁴² in which 51 tennis players undertook a specific programme of stretching exercises showed that the exercises improved the range of motion. Although they did not record the rate of injury, the authors hypothesised that this stretching programme would reduce injury risk. There is currently no evidence that limited flexibility is associated with an increased risk for tennis injuries. In a systematic review of intervention studies on the effect of stretching, Herbert and Gabriel¹⁴³ showed that stretching before exercise did not result in a reduction of injury risk. However, they noted that generalisation of this conclusion required further testing. It therefore may be worthwhile to investigate the effects of the type of programme designed by Kibler et al¹⁴² on the occurrence of tennis injuries.

The aim of the present literature review was to provide an overview of available knowledge on the occurrence, aetiology, and prevention of tennis injuries. For practical reasons we refrained from doing a formal methodological quality assessment of individual studies or a quantitative data synthesis. However, by presenting studies with different study designs, a picture emerges that represents the current base of knowledge in this field. It is clear from the results that further studies on injury rates, risk factors, and prevention of tennis injuries are needed. Researchers should, if possible, choose a prospective study design in order to decrease the risk of recall bias.

A comparison of injury rates across studies will be facilitated when similar definitions of injuries are used and are clearly stated in the studies. The injury definitions in the studies in this review can be categorised as “time loss”, “medical assistance”, and “tissue injury” definitions.^144,145 Each definition has advantages and disadvantages¹⁴⁵ and delivers its own scope at the problem of tennis injuries. A clear benefit of using a “time loss” definition is that it will generally result in the recording of injuries which substantially affect the player’s health or performance, or both.¹⁴⁵

Few studies have been carried out on the reliability of injury recording systems, and this should therefore also be priority for future research.^145–147 To improve the reliability of data collection it would be wise to use instruction manuals for observers. Another important epidemiological variable is exposure time. This is a measure of participation time in training and matches. It represents the amount of time the player is at risk of injury. The exposure time should, if possible, be recorded individually for each player. Individually recorded exposure times enable researchers to study risk factors in an advanced way by using a Cox proportional hazards regression model.¹⁴⁸ However, because of practical limitations in many cases—especially in large cohort studies—estimates of exposure time must be used.

Currently, based on this literature review, we were unable to identify measures proven to prevent tennis injuries. There are no randomised controlled trials available, and the limited results of the studies on risk factors for tennis injuries fail to provide a clear perspective. Clinical experience and also the results of prevention studies in other sports^114,115 suggest that physical training specifically targeting injury-prone movement patterns may lead to beneficial results. It would be very interesting to test the hypothesis put forward by Kibler associating loss of glenohumeral internal rotation (GIRD) and shoulder strength imbalance with an increased risk of shoulder injuries.¹⁴²

Other possibilities for prevention include: education of players, parents, and coaches about tennis injuries, interval musculoskeletal screening of players to identify problem areas before injuries occur, and adjustment of equipment including shoes, racquets, strings, and balls as well as court surfaces. However, further research is needed to move from a stage of clinical expertise and speculation to real evidence based prevention of tennis injuries.