Background Shoulder injuries are common among handball players and predominantly characterised by overuse characteristics. Reduced total glenohumeral rotation, external rotation weakness and scapular dyskinesis have been identified as risk factors among elite male handball players.
Aim To assess whether previously identified risk factors are associated with overuse shoulder injuries in a large cohort of elite male and female handball players.
Methods 329 players (168 male, 161 female) from the two upper divisions in Norway were included and tested prior to the 2014–2015 season. Measures included glenohumeral internal and external rotation range of motion, isometric internal and external rotation strength, and assessment of scapular dyskinesis. Players were followed prospectively for one competitive season, with prevalence and severity of shoulder problems registered monthly using the Oslo Sports Trauma Research Center Overuse Injury Questionnaire. A severity score based on players’ questionnaire responses was used as the outcome measure in multivariable logistic regression to investigate associations between candidate risk factors and overuse shoulder injury.
Results No significant associations were found between total rotation (OR 1.05 per 5° change, 95% CI 0.98 to 1.13), external rotation strength (OR 1.05 per 10 N change, 95% CI 0.92 to 1.20) or obvious scapular dyskinesis (OR 1.23, 95% CI 0.25 to 5.99) and overuse shoulder injury. A significant positive association was found between greater internal rotation (OR 1.16 per 5° change, 95% CI 1.00 to 1.34) and overuse shoulder injury.
Conclusion None of the previously identified risk factors were associated with overuse shoulder injuries in a mixed-sex cohort of elite handball players.
- throwing athletes
- overuse shoulder injuries
- shoulder pain/epidemiology
- risk factors
Statistics from Altmetric.com
Shoulder injuries are common among handball players,1 2 particularly at the elite level. In the Norwegian elite division, 52% of male players experienced shoulder problems at some point during the season,3 and 58% of female players reported a history of shoulder injury.4 Among elite players, a majority of shoulder problems are thought to be related to overuse.3–5 Recent attention has therefore been directed towards the prevention of overuse shoulder injuries in elite handball,6 and identification of risk factors is a key step to develop successful prevention programmes.7
Several studies have investigated internal modifiable risk factors for shoulder injuries among overhead athletes, with particular focus on glenohumeral range of motion (ROM),3 8–10 shoulder strength3 11 12 and scapular control.3 13–15 In handball, cross-sectional studies have suggested several potential risk factors for shoulder injuries, including reduced glenohumeral internal rotation (IR), excessive glenohumeral external rotation (ER), low ratios of concentric ER to concentric IR strength and high ratios of eccentric IR to concentric ER strength.9 12 16 In a prospective risk factor study of 206 male Norwegian elite handball players, Clarsen et al 3 observed significant associations between obvious scapular dyskinesis, total rotation (TROM), ER strength and the risk of shoulder injury. In a recent critical review, it was argued that identified risk factors should be confirmed in relevant populations.17
Therefore, the aim of the current study was to investigate if the risk factors reported by Clarsen et al 3 could be confirmed in a large, mixed-sex cohort of Norwegian elite handball players using the same methods. Our hypothesis was that the risk of overuse shoulder injury would be associated with scapular dyskinesis, reduced total rotation and low ER strength.
Study design and participants
This was a prospective cohort study involving the control group arm of a recently published randomised controlled trial.6 The cohort consisted of 23 handball teams (12 male; 11 female) from the two upper divisions in Norway. We visited each team during a preseason training session and invited every player present to participate in the study. All players with a team contract were eligible for participation, irrespective of their baseline injury status or history (n=333). Players who consented to participate (n=329) completed questionnaires and shoulder testing at baseline (figure 1), and were followed for the duration of the regular season (September 2014–March 2015). Six times during the season, players reported any shoulder problems using the Oslo Sports Trauma Research Center (OSTRC) Overuse Injury Questionnaire,18 as described in previous studies of shoulder problems among elite handball players.3 6
Six different test teams of two sports physiotherapists conducted baseline testing. Each test team visited between three and four teams in the period from August to mid-September 2014. Within each test team, one tester was responsible for administering questionnaires and evaluating scapular control, and the other measured players’ strength and glenohumeral ROM. Prior to baseline testing, 3 days of training was carried out to ensure that examiners were familiar with the measurement techniques.
Players reported baseline injury status and history using a modified version of the Fahlström questionnaire, as used in previous studies of elite handball players.3 4 6 In addition, players reported any shoulder problems during the week prior to baseline testing using the OSTRC Overuse Injury Questionnaire.6
Range of motion
Glenohumeral joint IR and ER ROM was measured bilaterally using a digital inclinometer attached to a 30 cm Perspex ruler (Acumar Digital Inclinometer, Lafayette Instrument, Lafayette, Indiana, USA), with the player in supine and with their shoulder abducted to 90° with 0° rotation and elbow flexed to 90°.3 19–21 If necessary, a folded towel was used to align the upper arm in the frontal plane. The examiners palpated the scapula with their thumb on the coracoid process and four fingers on the spine of the scapula to control scapular compensation. The end of IR and ER ROM was defined as the point at which the scapula was felt to move, as previously described by Wilk et al. 19 The examiners performed two repeated measurements and the average was recorded as the players’ IR and ER values. TROM was calculated by summing these values.
Isometric IR and ER shoulder strength was measured on the dominant side using a handheld dynamometer (microFET, Hoggan Health Industries, Salt Lake City, Utah, USA), with the player in supine and with their shoulder abducted to 90° with 0° rotation and elbow flexed to 90°.20–22 The opposite arm was placed resting on the hip. The handheld dynamometer was externally fixed to limit measurement error related to manual fixation from multiple examiners.21 Players were verbally and manually assisted to stabilise their scapula prior to testing. We used no external fixation of the scapula during the actual testing. Players performed the strength measures three times and the best attempt was recorded.
The examiners observed players performing five repetitions of flexion and abduction in the glenohumeral joint while holding an external weight: 5 kg for male players and 3 kg for female players.3 21 The examiners were situated 3 m behind the players and rated shoulders bilaterally as having normal scapular control, slight scapular dyskinesis or obvious dyskinesis for each of the two motions individually, according to the method proposed by McClure et al 23 and previously used in studies on handball players.3 21
The methods used to measure ROM, isometric strength and scapular control has been described in detail in the appendices of previous publications.3 21
Monitoring of shoulder problems
We emailed the OSTRC Overuse Injury Questionnaire to all players in the study on the last Sunday of each month from October 2014 to March 2015, six times in total, using online survey software (Questback V.9692, Questback AS, Oslo, Norway). Automatic reminders were sent to non-responders after 3 and 7 days per email and short message service (Pling, Front Information DA, Oslo, Norway). In addition, we visited teams throughout the season and asked non-responders to complete a paper version of the questionnaire. The questionnaire, used in a previous study on shoulder problems in elite handball,3 6 addresses the extent to which overuse shoulder injuries, expressed as shoulder problems, affect participation, training volume and performance, as well as the level of shoulder pain experienced during the past week.18 Players reported shoulder problems only in their dominant shoulder, with shoulder problems defined as any pain, ache, stiffness, instability, looseness or other symptoms related to their shoulder.3 6 Acute shoulder injuries were recorded as previously described and excluded from the analyses.6
For each player response to the OSTRC Overuse Injury Questionnaire, the response enabled the calculation of a severity score ranging from 0 to 100.18 At the end of the study, we calculated the individual average severity score by summing each player’s scores and dividing by their number of questionnaire responses. The average severity scores were dichotomised using a cut-off value of 40 to distinguish players with an overuse shoulder injury from uninjured players. This was used as the outcome measure in the risk factor analyses, as previously described.3 In addition, we calculated the prevalence of shoulder problems for the dominant shoulder for both sexes each time the questionnaire was administered by dividing the number of players who reported any problem (ie, anything but the minimum value in any of the four questions) by the number of questionnaire respondents.18 To filter out problems with fewer functional consequences, we calculated the prevalence of substantial shoulder problems in the same way, including only shoulder problems leading to moderate or severe reductions in training volume or performance, or a total inability to participate.18 At the end of the study, the average prevalence of shoulder problems and substantial shoulder problems was calculated for both sexes.
By the end of each month, players reported their exposure to handball training, match play and additional strength training during the past week. We calculated the mean weekly exposure in each measure for both sexes by summing up the number of minutes reported and dividing by the number of respondents. At the end of the study, we calculated the individual average weekly exposure by summing each player’s exposure data and dividing by their number of responses.
Players with no injury data and players with missing test results or pain during baseline testing were excluded from the risk factor analyses. We used multivariable logistic regression models to investigate associations between candidate risk factors and overuse shoulder injury (average severity score ≥40).3
The following were analysed as potential risk factors for injury to the dominant shoulders: IR strength, ER strength, ratio of ER to IR strength (ER:IR ratio), ER:IR ratio of <75%, <80% and <85%, IR ROM, ER ROM, TROM, >5° TROM difference between shoulders, <5° ER gain and glenohumeral IR deficits of ≥5°, ≥10°, ≥15° and ≥20°, obvious scapular dyskinesis during flexion and/or abduction, slight or obvious scapular dyskinesis during flexion and/or abduction, average weekly exposure to handball training, match play, and additional strength training. A range of cut-off values were used for ER:IR ratio and IR deficits, as previous studies in throwing sports have reported associations with shoulder injury with different cut-offs for these variables.3 8 10–12 24
We adjusted strength measures for body mass, and demographic variables possibly associated to shoulder injury (p<0.2) were added to each model using a forward selection procedure. We compared dominant and non-dominant shoulder ROM using paired-samples t-tests for both sexes. Isometric shoulder strength and shoulder ROM were compared between sexes using independent-samples t-tests. Scapular control was compared between sexes using Χ2 test.
To assess the reliability of the baseline tests, we performed a pilot prior to the study including a convenience sample of 19 asymptomatic adults (10 male and 9 female). The reliability of strength and ROM measures was assessed by calculating the intraclass correlation coefficient (ICC), using a two-way mixed single measure model (absolute agreement) for inter-rater reliability and two-way random single measure model (absolute agreement) for intrarater reliability.3 20 Spearman’s r (Rs) was used to assess the inter-rater and intrarater reliability of subjective rating of scapular control.
Players had played handball for an average of 14 years (SD 5, range 4–37) and 78% were right-handed. There were no sex differences in playing position distribution, with 41% backs, 25% wings, 15% line players, 13% goalkeepers and 6% reporting multiple positions. Dropout during the study and the number that was tested and included in each analysis are presented in figure 1.
Shoulder injury status and history at baseline
At the time of testing, 87 male players (52%) and 68 female players (43%) reported a history of shoulder pain during the previous handball season. Current shoulder pain was reported by 47 male players (28%) and by 49 female players (31%). Based on the OSTRC Overuse Injury Questionnaire, 82 male players (49%) and 74 female players (46%) reported a shoulder problem during the previous 7 days. Of these, 21 male players (13%) and 25 female players (16%) reported substantial shoulder problems. There were no sex differences in the prevalence of shoulder pain or problems reported at baseline.
Range of motion
Women and men both had less IR in their dominant shoulders than their non-dominant shoulders, male players with a mean difference of 4° (95% CI 3° to 5°, p<0.01; figure 2) and female players with a mean difference of 6° (95% CI 5° to 8°, p<0.01; figure 2). Female players had significantly more IR in their non-dominant shoulders compared with male players (mean difference: 5°, 95% CI 1° to 8 °, p<0.01). A total of 16 players (5%), 8 male and 8 female, had greater than 20° glenohumeral IR deficit. The ER ROM was greater in the dominant shoulders for both sexes, male players with a mean difference of 2° (95% CI 0.2° to 3°, p<0.03; figure 2) and female players with a mean difference of 3° (95% CI 2° to 5°, p<0.01; figure 2). One hundred and eighty-seven players (59%), 103 male and 84 female, had <5° ER gain in their dominant shoulder. Significantly less TROM in the dominant shoulders was observed in both sexes, male players with a 2° mean difference (95% CI 1° to 4°, p<0.01; figure 2) and female players with 3° mean difference (95% CI 1° to 4°, p<0.01; figure 2). A total of 135 players (42%), 71 male and 64 female, had >5° TROM loss on their dominant side. There were no significant sex differences in ER ROM and TROM measures.
Isometric strength dominant shoulder
Compared with female players, male players were significantly stronger in both ER (mean difference: 0.23 N/kg, 95% CI 0.33 to 0.12, p<0.01; figure 3) and IR (mean difference: 0.13 N/kg, 95% CI 0.25 to 0.02, p<0.01; figure 3) in their dominant shoulders. The average ER:IR ratio in the dominant shoulders was 91% (SD 18%) among female players and 96% (SD 17%) among male players (mean difference: 5%, 95% CI 1% to 9%, p=0.017; figure 3). A total of 71 players (25%), 43 male and 28 female players, had an ER:IR ratio of less than 80%.
A total of 161 players (56%), 81 male and 80 female players, were rated as having slight scapular dyskinesis in their dominant shoulders during flexion and 100 players (35%) during abduction (43 male, 57 female). Thirty-two players (11%), 13 male and 19 female, were rated as having obvious scapular dyskinesis in their dominant shoulders during flexion and 22 (8%) during abduction (11 male; 11 female). There were no significant sex differences in scapular control.
Reliability of shoulder tests
The inter-rater and intrarater reliability (ICC) of strength and ROM measures is presented in table 1. The inter-rater reliability (Rs) of subjective rating of scapular control into three groups (normal, slight and obvious dyskinesis) varied from 0.57 to 0.82 for flexion and from 0.32 to 0.55 for abduction. The intrarater reliability (Rs) was 0.68 for flexion and 0.85 for abduction. As shown in table 1, the inter-rater reliability of ROM measures was fair.
The average response rate for the OSTRC Overuse Injury Questionnaire during the season was 85% (range 82%–87%). Complete injury data were available from 65% (215 players) of the cohort, while 16% (53 players) had no injury data during the season. Female players had a higher average response rate (90%, range 88%–93%) compared with male players (79%, range 78%–80%). The average response rate for the exposure data was 49% (range 30%–67%).
There were no sex differences in the average weekly exposure to match play or strength training (table 2). However, male players reported higher average weekly exposure to handball training (mean difference: 47 min, 95% CI 22 to 72, p<0.01; table 3).
Shoulder problems during the season
The average prevalence of shoulder problems during the season was 23% (95% CI 21% to 26%). The average prevalence of substantial shoulder problems was 8% (95% CI 7% to 9%). Female players reported a higher prevalence of both shoulder problems (mean difference: 6%) and substantial shoulder problems (mean difference: 2%) compared with male players (table 3).
Risk factor analyses
No associations were identified between overuse shoulder injury and sex, age, height, body mass, dominant arm, player position, team affiliation, competition level, years of handball participation, shoulder pain at baseline or history of shoulder pain last season.
No associations were detected between overuse shoulder injury and obvious scapular dyskinesis, total rotation or external rotation strength (figure 4). As shown in the figure, increased IR ROM was significantly associated with overuse shoulder injury.
No associations were observed between overuse shoulder injury and average weekly exposure to handball training, match play or additional strength training.
Based on the findings of our previous study exploring potential risk factors for shoulder problems among elite male handball players,3 we hypothesised that the risk of overuse shoulder injury would be associated with obvious scapular dyskinesis, reduced total rotation and low external rotation strength. However, in this study, none of these three factors were associated with injury. In fact, players with greater IR range of motion had a higher probability of experiencing overuse shoulder injuries throughout the season.
Our prospective injury data extend previous epidemiological studies reporting that shoulder injuries, predominantly from overuse, are common among handball players.1–4 25 These studies, conducted in a range of player populations, have used a variety of designs, measurement methods and injury definitions. The injury registration method used in the current study was designed specifically to capture overuse problems,18 and has previously been used to study shoulder problems among elite male handball players.3 Similar to Clarsen et al, 3 we found that the prevalence of shoulder problems was among the highest reported with this method, regardless of anatomical region or sport.25 26
Our results revealed a higher average prevalence of shoulder problems (26% vs 20%) and substantial shoulder problems (9% vs 7%) among female elite handball players than male. This indicates that female sex may represent a risk factor for overuse shoulder injury, although this was not confirmed in the univariate statistical analysis. For the whole cohort, our results showed lower average prevalence of shoulder problems (23% vs 28%) and substantial shoulder problems (8% vs 12%) compared with Clarsen et al. 3 This may be due to an increased awareness of shoulder problems, since Clarsen et al 3 monitored male handball players in the top division, who also formed part of our cohort. In addition, our prevalence results may be affected by a crossover effect, since our study population represented the control group in a randomised intervention study aiming to reduce the prevalence of shoulder problems.6 Nevertheless, our results reiterate the importance of targeting the throwing shoulder with preventative efforts and may provide guidance when determining which risk factors to target in prevention programmes.
Glenohumeral ROM and overuse shoulder injury
In throwing sports, reduced IR and increased ER have been reported in the dominant arm of asymptomatic athletes.3 4 8–10 24 27 This is considered as a normal soft tissue and/or bony adaptation to repeated throwing,28 and has even been suggested to prevent shoulder injuries.29 However, several studies have reported an association between reduced IR and total ROM in the dominant shoulder and throwing-related shoulder injuries.3 8 9 We did not identify any associations between IR deficits or TROM differences and overuse shoulder injury, despite using a range of cut-off values to define these terms.
Reduced ER ROM has also been proposed as a potential risk factor,27 but we did not identify any association with shoulder injury. However, we did find that increased IR was a significant risk factor. This contrasts with the results of Clarsen et al,3 who suggested that stretching should be considered in the development of injury prevention programmes. As the magnitude of the association appears to be limited (16% increased risk per 5° increase in IR) and the reliability of the IR measurements may be questioned, this result must be interpreted with caution; IR stretching should therefore not be abandoned as a prevention strategy based on our data.
Glenohumeral rotation strength and overuse shoulder injury
Weakness in ER is a risk factor for shoulder injury in elite male handball.3 In the current study, we found no association between ER strength and overuse shoulder injury, despite excellent reliability of the strength measures performed. However, non-significant trends in our data suggest that lower ER:IR ratios may also be worth considering as a risk factor. Similar findings have been reported among elite youth handball players and baseball pitchers, where lower ER:IR ratios have been associated with shoulder injury.11 12 24 In addition, Møller et al 21 recently reported that reduced ER strength exacerbated the association between handball load and shoulder injury among elite youth handball players increasing their load by 20% or more per week. Based on this overall body of evidence, it appears reasonable to suggest that exercises to strengthen ER should be included in injury prevention programmes.
Scapular dyskinesis and overuse shoulder injury
Scapular dyskinesis is common among overhead athletes with shoulder pain, across a variety of shoulder pathologies.13 15 30–32 However, it has also been demonstrated to be common among asymptomatic overhead athletes,14 15 33 34 and there is conflicting evidence from prospective cohort studies on the association between scapular dyskinesis and shoulder pain among overhead athletes.3 14 15 In contrast to Clarsen et al, 3 we did not find any association between scapular dyskinesis and shoulder injury. However, Møller et al 21 recently reported that scapular dyskinesis exacerbated the association between handball load and shoulder injury among elite youth handball players increasing their load between 20% and 60% per week. Due to methodological limitations of this study (discussed below) and variable reliability, the relationship between scapular dyskinesis and shoulder injury remains unclear in elite handball players.
Handball load and overuse shoulder injury
There is growing evidence supporting a rapid increase in training load as a risk factor for overall injury.35 36 Recently, Møller et al 21 reported that a large weekly increase in handball load represents the primary risk factor for shoulder injuries among elite youth players. Our results show that both female and male elite handball players have high exposure to handball training, match play and strength training. However, we did not find any association between any of these exposure measures and overuse shoulder injury. These results must however be interpreted with caution, as the exposure measures are only average weekly approximates based on self-reporting, do not include any measure of intensity and clearly are vulnerable to recall bias.
Future risk factor studies should strive to investigate the association between load and shoulder injuries prospectively, and examine whether the association is influenced by internal modifiable risk factors. Ideally, the exact individual throwing workload should be monitored prospectively. However, accurate and feasible methods to complete this are not yet available.37
A major strength of this study is the use of a prospective cohort design with a large representative sample of elite male and female handball players. We employed an injury surveillance method developed, validated and recommended to study overuse injuries.18 38 39 The method was previously used in the study of shoulder injuries among elite handball players and allows comparability.3 6 Players and team medical staff reported acute injuries alongside their reports of shoulder problems to avoid misreporting acute injuries as overuse injuries; this allowed us to assess the association between risk factors and overuse injuries alone.
A key consideration is the choice of tests and measurement techniques to assess risk factors. We used the same tests as Clarsen et al,3 with minor modifications, to ensure comparability and maximise the clinical relevance. As previously discussed, the reliability of the strength measurements was observed to be excellent. However, the reliability of the ROM measurements and subjective rating of scapular control varied from fair to good and from fair to excellent, respectively.
The validity of the strength measurements can be questioned. They are isometric and performed with the player in a supine position for IR and ER with players’ shoulder abducted to 90° and elbows flexed to 90°. This position is reliable20 22 and was selected for its resemblance to the throwing position in handball compared with the neutral shoulder position used by Clarsen et al.3 This difference should be borne in mind when comparing the results. To limit measurement error related to manual fixation, we externally fixated the handheld dynamometer. However, to which degree isometric testing in this position relates to shoulder strength in a throwing motion is unknown.
Similar to Clarsen et al,3 we used single testers with a digital inclinometer rather than two testers with a bubble goniometer for ROM measurements. Both methods are reliable.20 However, there may be systematic differences in the results.40 41 Therefore, our ROM values in this study may be compared directly with previous results on elite male handball players,3 but not to previous research in general. The use of multiple testers to perform ROM measurements and the fair inter-rater reliability observed represent limitations, indicate that our results must be interpreted with caution and may explain the difference in results between this study and Clarsen et al, 3 where only two testers performed all measurements.
When evaluating the presence of scapular dyskinesis, we used subjective assessment based on criteria recommended in a consensus statement and previously used on elite male handball players.3 42 The method used consists of three rating options, and proved valid and reliable for assessing three-dimensional scapular motion in overhead athletes.23 34 It has been suggested that a two-option rating (normal or abnormal) is more reliable.43 Clarsen et al 3 did, however, not find this, and we therefore used the three-option rating to ensure comparability. Due to the use of multiple testers and the inter-tester reliability, ranging from fair to excellent when assessing scapular control, our results must be interpreted with caution and may explain the difference in results between the current study and Clarsen et al, 3 where one experienced physiotherapist performed all evaluations.
Inclusion of players irrespective of injury status or history
Traditionally, risk factor studies exclude players injured at baseline and only record new cases throughout the study, allowing for an assumption of cause and effect. However, applying such an approach in the current study would have resulted in a biased cohort, not representative of elite handball players, where overuse shoulder injuries with periods of remission and exacerbation are common. Therefore, we included all players present at training sessions, irrespective of their injury status or history, and only excluded players experiencing pain during actual testing from analyses. Consequently, we are limited to assess associations between risk factors and overuse shoulder injury and causation cannot be assumed.
In contrast to Clarsen et al 3, we excluded acute injuries from the analyses. However, we were not able to differentiate between specific shoulder injury diagnoses, since we did not have diagnostic information on each case.3 6 Our definition of overuse shoulder injury encompassed all physical symptoms and the condition may have had multiple causes, such as subacromial and internal impingement, tendon pathology, glenoid labrum injuries, glenohumeral joint instability and acromioclavicular joint dysfunction, all commonly observed in throwing athletes.28 29 44 The risk factors may differ among these conditions, but our study design meant we were unable to link risk factors with specific conditions.
The response rate and the number of players with complete injury data are high compared with previous studies using the same surveillance method.25 Nevertheless, we excluded 53 players from the risk factor analyses, as they had not reported their injury data. In addition, players with missing test results and pain during testing were excluded (figure 1). Consequently, our statistical power decreased and this may have affected the accuracy of our coefficient estimates. Another limitation was the low response rate for the exposure data, which may have limited our ability to detect any association between exposure and shoulder injury.
Our prospective cohort study of over 300 elite handball players did not confirm previously identified, so-called ‘established’ risk factors for overuse shoulder injuries, including total ROM, ER strength and scapular dyskinesis.
What are the findings?
Greater glenohumeral internal rotation range of motion was associated with increased probability of experiencing overuse shoulder injuries.
Reduced glenohumeral rotation, external rotation weakness and scapular dyskinesis were not associated with overuse shoulder injuries in a mixed-sex cohort of elite handball players.
There was a trend to higher probability of experiencing overuse shoulder injuries among players with a ratio of external to internal rotation strength below 80%.
The prevalence of overuse shoulder injuries was greater among elite female handball players than male players.
How might it impact on clinical practice in the future?
The role of glenohumeral internal rotation stretching, external rotation strengthening and scapular stability training in preventing overuse shoulder injuries in elite handball remains unclear.
The authors thank A Aune, A Skjølberg, C Skovly, D Major, E E Eriksen, H K Vileid, K Røed, K Søderstrøm, M Fagerheim, M J Olsen, O T Østvold, R M Kristensen and R Nesje for their assistance in data collection. We also thank all players and coaches who participated in the study, as well as the team medical staff. The Oslo Sports Trauma Research Center has been established at the Norwegian School of Sport Sciences through generous grants from the Royal Norwegian Ministry of Culture, the South-Eastern Norway Regional Health Authority, the IOC, the Norwegian Olympic and Paralympic Committee and Confederation of Sport, and Norsk Tipping AS.
Contributors SHA drafted the manuscript and performed the data analysis. SHA and the main supervisor, GM, are responsible for the overall content as guarantors.
Competing interests None declared.
Ethics approval The study was reviewed by the Regional Committee for Medical and Health Research Ethics (REK 2014/653 A), which concluded that, according to the Act on Medical and Health Research (the Health Research Act 2008), the study did not require full review by REK. The study was approved by the Norwegian Social Science Data Service (NSD 2014/38187).
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement All data are available upon request to SHA (email@example.com).
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.