Background

Football is the most popular sport worldwide. Unfortunately, football-related knee injuries are common and constitute a serious problem regardless of the playing level. The injury that draws most attention is the anterior cruciate ligament (ACL) injury. This injury usually causes long absence from football and may even force some players to give up their career [1]. ACL injury is also associated with an increased risk of new knee injury [2, 3], as well as long-term medical disability related to osteoarthritis [4, 5].

There are, however, only few investigated risk factors for ACL injury in football. First, many studies have shown a higher incidence among female than male players [6]. Second, low age is a risk factor for female players and the risk seems to be highest during the late pubertal or first postpubertal years [7]. Third, match play is associated with a considerably higher risk of ACL injury compared to training [8–11]. Several other potential risk factors, for example anatomical and biomechanical factors, have repeatedly been discussed in the literature, but lack scientific evidence [12]. Recently, it has been speculated that a high match frequency as well as match play at senior level could be a risk factor among female adolescent football players [1], but evidence for such a relationship is currently lacking.

Several studies have been conducted with the aim of preventing serious knee injuries and/or ACL injuries in female football [13–20]. However, only two of these studies have shown that the risk of ACL injury can be reduced using a comprehensive warm-up program [13, 16]. In one study, the injury risk may actually have been increased in the group allocated to balance board training [20]. Unfortunately, many of the studies have serious methodological limitations, such as having few included participants or high drop-out rates resulting in low statistical power [14, 15, 17, 20]. Furthermore, the studies commonly do not account for true exposure (playing time) resulting in a less accurate estimation of the actual injury risk [13–17]. Finally, another well-known problem is a poor compliance with the intervention [19], probably due to very ambitious programs taking too much time from the ordinary football training.

A structured warm-up program specifically developed for youth and adolescent team sports has been recently introduced in Sweden (Knäkontroll, SISU Idrottsböcker©, Sweden, 2005). This commercially available program contains six exercises focusing on neuromuscular knee control and core stability and is intended to be administrated by coaches. The effectiveness of the program in reducing the incidence of knee injury has, however, not been evaluated. The purpose of the present study is to conduct a cluster randomized controlled trial (RCT) evaluating the effectiveness of the program in adolescent female football players. Our hypothesis is that the structured warm-up program reduces the incidence of ACL injury as well as the incidence of any acute knee injury (except contusion) and severe knee injury (defined as injury resulting in absence of more than 4 weeks).

Methods

Study design and definitions

The study is a two-armed RCT designed in accordance with the CONSORT statement guidelines [21]. The definitions follow the international guidelines for football injury research published recently (Table 1) [22, 23]. The study was registered in ClinicalTrials.gov at the start of the study http://clinicaltrials.gov.

Table 1 Operational definitions used in the study.
Full size table

Club recruitment

Eight regional districts of the Swedish Football Association (FA) located in the south and middle of Sweden were invited to participate in the study being carried out during the 2009 campaign. Leading officials from the selected districts were contacted by telephone in August 2008 and all agreed to support club recruitment to the study. Only female players aged 13–17 years (born 1992 to 1996) were eligible for inclusion in the study. Clubs were informed about the trial by advertisement on the websites of each regional district and through invitational e-mails and letters in December 2008. Club enrolment registries for the 2009 season were obtained for the U-14 to U-18 series from all districts in February 2009 and the coaches of all eligible teams were contacted by the authors (MH and MW). Some coaches of teams belonging to the senior league systems responded to the invitation and these teams were included if the squads consisted mainly of adolescent girls in the age groups eligible for study. Teams that did not respond to the invitation or declined participation, those that had less than two scheduled training sessions per week, and those already using the structured warm-up program or a similar preventive program were not included in the study.

Randomization

All clubs (516 teams) that agreed to participate were randomized into an intervention or control group (Figure 1). All teams from the same club were assigned to the same group. The computer-generated cluster randomization was performed by one of the authors (IA) who is not involved in the intervention.

Figure 1
figure 1

Flow of teams and clusters (all teams from the same club are assigned to the same cluster) through the trial.

Full size image

Study therapist education

Physical therapists were recruited to the study to educate the coaches and assist the clubs with data collection. The study therapists were members of the Swedish Society of Football Physical Therapists or the Swedish Society of Sports Medicine. A 3-hour educational meeting for the study therapists was held by the authors (MH and MW) in November 2008. The meeting consisted of a presentation of the theoretical background to knee injury epidemiology in female adolescents and description of the study design followed by a practical education in the program given by an experienced supervisor. The participating therapists also received a specific manual describing the study methodology in detail as well as the tasks associated with the study. A few therapists, who were not yet recruited at the time of the educational meeting, received the same information on a later occasion before the start of the study. A total of 68 therapists accepted to participate in the trial.

Coach education

All randomized teams were invited to regional instructional meetings held by the authors (MH and MW) before the start of the trial. In the intervention group, one coach or team leader from each club together with one player ("team captain") took part in a 3-hour meeting. Knee injury epidemiology in female adolescents, the study design and the data collection procedures were presented. This theoretical session was followed by a practical education in the structured warm-up program given by the study therapists. Each coach/team leader received a CD and a leaflet describing the warm-up exercises. In addition, they received a study manual describing the methodology, operational definitions and data collection procedures. The manual contains several examples of how to fill in the study forms. Coaches or team leaders that could not attend the educational meetings were sent the same theoretical session as a slide show presentation and received the practical education by their study therapists before the start of the trial.

Coaches of clubs assigned to the control group received a one-hour lecture about study background and methodology as well as the study manual but were informed to train and play as usual throughout the season. They were also informed that if the study shows that the warm-up program is successful in preventing knee injuries they will be offered the same practical education of the exercises before the start of the 2010 season.

Intervention

The structured warm-up program (Knäkontroll, SISU Idrottsböcker©, Sweden, 2005) was developed for youth and adolescent team sports and is intended to be administrated by coaches. It was designed by experienced physical therapists involved in the medical organisations of the Swedish Football Association (FA) in collaboration with the Swedish Handball Federation, the Swedish Basketball Federation and the Swedish Floorball Federation. It contains six exercises focusing on neuromuscular knee control and core stability: (1) one-legged knee squat, (2) pelvic lift, (3) two-legged knee squat, (4) the bench, (5) the lunge, and (6) jump/landing (Table 2). Each exercise is subdivided into four steps of progressing difficulty (A to D) and each association/federation adjusted the progression of the exercises to be sport-specific. In addition, a pair-exercise was developed for intermittent use with the aim of making the training more varying and fun (Table 2). The structured warm-up program is commercially available on a CD with a corresponding leaflet describing the execution and purpose of each exercise.

Table 2 The structured warm-up program used in the intervention group
Full size table

The clubs assigned to the intervention group were instructed to perform the exercises during the warm-up at two training sessions per week throughout the entire 2009 competitive season. The exercises are preceded by 5 minutes of low-intensity running and take about 10–15 minutes to complete after familiarization. Coaches and players are encouraged to watch each other closely and give feedback about the execution of exercises. All players start on the first level of difficulty (A) and the coach clears each player's progress to the next level when she performs the exercise with good control, mainly focusing on core stability and proper knee alignment.

Data collection

Data is collected during the competitive season from April until the end of the season in October (approximately 7 months). At the start of the study, a baseline form containing information about name, social security number, stature, body mass, menarche, family history of ACL injury, and the player's previous knee injuries and any current complaints is filled in by the player and parents/guardians. During the season, individual player exposure for each training session and match is reported by a coach or team leader on a computer-based registration form that is e-mailed to the research centre monthly. If a player has additional exposures outside the team, such as training or match play with a national team or with a senior team of the club, these are also registered on the exposure form. If a player sustains an acute knee injury during football training or match, the coach contacts the study therapist assigned to the club for evaluation of the injury. All acute knee injuries, except contusions, resulting in absence from play (time loss) are documented soon after the event by the study therapist on a standardized injury report form. If a severe knee injury is suspected, the player can be either advised to visit the nearest emergency department or referred to a study physician involved in the study. One physician experienced in diagnosing knee injuries is assigned to each regional district to assist the study therapists in the evaluation of severe knee injury. All players sustaining possible first-time ACL injuries (based on history and/or clinical examination) will be routinely examined with magnetic resonance imaging (MRI) by the study physicians to verify the diagnosis and to evaluate associated injuries. Suspected re-ruptures or secondary intraarticular injuries in knees with a previously reconstructed ACL will be assessed individually according to best practice. Injuries occurring at school, in other sports or during leisure time are not registered. Injuries are followed until full return to play or until 1 year post-injury.

Compliance

At the start of the study, coaches in the intervention group deliver a preliminary training schedule specifying the two weekly sessions where the structured warm-up program is planned to be carried out. Any subsequent changes in the schedule need to be reported to the study therapist. The coach documents on the exposure form if the program has been carried out as scheduled. The study therapists will make unannounced visits to their teams in the intervention group twice during the study period (once during the spring season and once during the autumn season) to monitor compliance to the program. Study therapists are encouraged to contact teams with low compliance (fewer than 6 prevention sessions over a 4-week period) to motivate them and increase their compliance.

Ethics

The study was approved by the Regional Ethical Review Board in Linköping (# M197-08). All players are verbally informed about the study by their coaches before the start of the study and also receive information to read together with their parents/guardians. Participation in the study is voluntary and all players sign written informed consent together with their parents/guardians according to the Declaration of Helsinki. None of the authors are involved in diagnosis or treatment of player injuries.

Sample size

The sample size was based on previous data showing an average annual/seasonal incidence of ACL injury among female adolescent football players of 1.15% as reported in one prospective study [24] or in the control groups of previous knee injury prevention studies (Table 3) [14–17, 19]. With a power of 80%, significance level of 5%, and an estimated risk reduction of 50% in the intervention group, randomization of approximately 4000 players to each group would be required [25].

Table 3 Annual/seasonal incidence of ACL injury in prospective cohort or controlled intervention studies in adolescent female football used for the sample size calculation.
Full size table

Statistical analysis

The incidence will be calculated as number of injuries per 1000 playing hours. The primary outcome of the study is the incidence of ACL injury. The secondary outcomes are the incidence of any acute knee injury (except contusion) and the incidence of severe knee injury (defined as injury resulting in absence of more than 4 weeks). All outcomes will be evaluated according to the mechanism of injury (contact or non-contact). The exposure and injuries for a team in the intervention group is calculated from the first interventional session performed during the season (matched with the same date to a team in the control group). All analyses will be performed both according to the intention-to-treat principle and according to per protocol. The analyses will be performed with group identity concealed. Clustering will be accounted for in the analyses and the intracluster correlation coefficient will be computed. A Cox regression model with cluster as a random effect will be performed as well as a multilevel mixed model analysis based on Poisson regression. The relative risk reduction (RRR) is calculated using the rate in the control group divided by the rate in the intervention group and will be reported as a ratio with 95% confidence interval. Covariates in the analyses will include age, menarche, previous knee injury, match frequency and match play with other teams than the player's own. The absolute risk reduction (ARR) is calculated using the rate in the control group minus the rate in the intervention group. The number needed to treat (NNT) will be calculated based on the ARR. All statistical tests will be 2-sided and a significance level of < 0.05 will be used.

Discussion

Knee injuries in football are common regardless of age, gender or playing level, but adolescent females seem to have the highest risk. Prevention of knee injury would be beneficial for players, clubs, insurance companies, and society. This cluster randomized controlled trial is designed to evaluate the effect of a commercially available structured warm-up program in adolescent female football. The strengths of our study include the large study sample compared to most of previous research in the field [13–15, 17–20] and the use of a recently validated methodology [22, 23]. In addition, the documentation of background information and individual player exposure will allow us to evaluate the risk of knee injury and control for possible interaction or confounding factors.

One limitation of the present trial is that it was not possible to perform a power calculation for a cluster RCT from exposure-based incidence rates of ACL injury, since valid data is lacking in the literature. Consequently, we cannot know for sure that the study sample is large enough. Another limitation is that the sports physical therapists supporting the teams and evaluating knee injuries are not blinded to group allocation. We decided to favour a professional evaluation of players' acute knee injuries by the physical therapists and it was not possible to recruit additional therapists specifically to assist teams with data collection and to instruct and monitor compliance to the preventive program. Due to the nature of the intervention, blinding of players or coaches to group allocation is not possible. Study physicians evaluating severe knee injuries are, however, blinded to group allocation.