The purpose of “Evidence in Practice” is to illustrate how evidence is gathered and used to guide clinical decision making. This article is not a case report. The examination, evaluation, and intervention sections are purposely abbreviated.
A collegiate-level soccer player was instructed by her coach to incorporate a proprioceptive component into her training program. He suggested that she purchase a balance board and immediately begin a program that he designed. She approached her physical therapist (SJF) for more information. I immediately recognized that, because of her sex and sport of choice, she would be at high risk for an anterior cruciate ligament (ACL) injury. Hewett et al1 estimated that as many as 2,200 ACL ruptures per year occur in female collegiate athletes in both the recreational and competitive ranks. Treatment and rehabilitation costs are estimated at $17,000 per ACL injury, which do not take into account the potential loss of long-term participation, loss of scholarship funding, and future disability from arthritic changes in a reconstructed knee.1 For these reasons, a shift toward injury prevention is warranted.
Injury prevention for the ACL can take many forms, including a variety of training protocols, athlete education, and bracing. Current studies focus on neuromuscular training as a preventive measure, with programs that include strength, flexibility, plyometrics, sport-specific agility drills, speed enhancement, balance, and athlete education.1–7 A clinician who understands the individual components of these programs could optimize injury prevention and aid athletes in appropriate program design and equipment purchases. In the case of this athlete, my colleagues and I focused on the use of proprioception or balance training and its effect on incidence of ACL injury.
We searched the literature to answer our clinical question “Is there evidence that proprioception or balance training can prevent ACL injuries in athletes without previous ACL insufficiencies?” and our secondary question “What balance training protocols are used in ACL injury prevention programs?”
Database used for search: PubMed
We began our search using PubMed (http://www.pubmed.gov). The largest component of PubMed is MEDLINE, which provides access to citations from over 4,800 biomedical and health journals from 1966 onward and is updated almost daily. The search was performed on July 28, 2006.
Initial keywords: anterior cruciate ligament injury, ACL injury, knee injury, balance, proprioception, prevention
We selected keywords that were specific to the question raised by our client (“anterior cruciate ligament injury”) and the intervention we were being asked to assess (“balance” or “proprioception”). In order to ensure that studies abbreviating anterior cruciate ligament were not missed, the keyword “ACL” also was used. We also used the term “prevention” because our intervention would focus on this outcome. To ensure a thorough search, we also used the term “knee injury” because studies may have been conducted with analysis of ACL injuries as only a portion of the study but still contain relevant information. Throughout the literature, the terms “proprioceptive” and “balance training” are used synonymously. Paterno et al6 described balance training as including exercises focusing on postural awareness and the maintenance of equilibrium without changing the base of support. Proprioception has been defined as the awareness of a body segment's orientation and positioning.8 Both terms were included in this literature search.
Because our question dealt with a clinical scenario, we chose to use the Clinical Queries link accessed in the left-hand column of the PubMed home page. This option allows the retrieval of clinical trials with the use of filters designed for either a broad, sensitive search or a narrow, specific search. By selecting the Search by Clinical Study Category, checking the boxes entitled therapy and broad, sensitive search, the search string (anterior cruciate ligament injur* OR ACL injur* OR knee injur*) AND (proprioception OR balance OR prevention) retrieved 162 results. In these searches, we used an asterisk to truncate the word “injury” to search for all words beginning with “injur-” including “injury” and “injuries.” We then clicked on the Limits tab and applied the limits: English language; human; and published in the last 10 years. This reduced our results to 99 citations. Because we also were interested in systematic reviews published on our topic, we input the same search string into the Find Systematic Reviews section of Clinical Queries page. When the same limits were applied as above, 10 citations were retrieved for a total of 109 citations.
Selection of articles for review:
We were primarily interested in articles reporting on prevention of ACL injuries in athletes without previous ACL pathology. Our inclusion criteria were: (1) the training program could not be a combination program but proprioception or balance training only, and (2) the article must present original research or be a systematic review. The exclusion criterion was subjects with previous ACL injury. Of the 109 citations retrieved, only 4 met our inclusion and exclusion criteria as determined by 2 reviewers who independently scanned the titles, and, if necessary, the abstracts (Figure).
The fourth article, by Hewett et al,9 was a systematic review that focused on 6 published ACL injury prevention studies. Of those 6 studies, only the article by Soderman et al10 met our inclusion and exclusion criteria. This article by Soderman et al was 1 of the 3 relevant citations that will be discussed in detail. Because Hewett et al did not offer any new evidence to answer our clinical question, we decided not to include it in our summation of relevant articles.
The first article we selected was a systematic review on knee injury prevention by Thacker et al.11
Thacker SB, Stroup DF, Branche CM, et al. Prevention of knee injuries in sports: a systematic review of the literature. J Sports Med Phys Fitness. 2003;43:165–179.
The authors of this systematic review looked at evidence for measures clinicians may recommend to coaches and athletes to prevent knee injuries. They compared alternative methods to prevent knee injury: equipment (bracing), footwear, and conditioning (training). The studies retrieved under the conditioning category were of interest to us. Although it did not exclusively discuss the use of proprioception or balance training to prevent ACL injuries, the review provided a background on knee injury prevention research and gave context to our topic. Because systematic reviews are peer reviewed and use scientific strategies to search the literature in an exhaustive, unbiased manner and to critically appraise all relevant studies, we felt this article would be useful.
This systematic review searched the literature concerning knee injury prevention. The search included the electronic databases MEDLINE (1996–2001), Current Contents (1996–2001), Biomedical Collection (1993–2001), the Cochrane Collaboration database, and Dissertation Abstracts as well as citations from the reference lists of relevant textbooks. References also were hand searched and first authors of randomized controlled trials and cohort studies on knee injury prevention were contacted. Articles were excluded if they did not provide primary research data, addressed treatment or rehabilitation rather than prevention, or provided previously published data. All articles were screened by the same reviewer. Thirteen articles met the criteria. Using a modified standardized scoring instrument, 3 reviewers evaluated each study independently.
The article, however, did not describe the credentials of the reviewers. The median quality scores ranged from 11 to 56 out of 100. Thacker et al laid out the methods and results of each study clearly in a table, including the study population (eg, 20 adult soccer teams over 3 seasons), intervention, study groups (including number of athletes), and outcomes (eg, ACL injury incidence, knee injuries, medial collateral ligament [MCL] injuries).
Of the 6 studies included under the conditioning part of the review, only 2 used balance training as an intervention. One of these studies, by Wedderkopp et al,12 found the use of ankle disk training effective in reducing injuries in both upper and lower extremities in female team handball players. Thacker et al ranked this study as 56/100, the highest quality of the studies included in their systematic review. Eleven of 22 teams with 111 players were randomized to the intervention group, and age, practice time, playing level, floor composition, and injury incidence during the previous season were controlled. The authors, however, did not specify the methods they used to control for these variables. Subjects ranged in age from 16 to 18 years of age and ranged in skill from recreational to elite. The coaches were instructed in the intervention program, including the use of an ankle disk for 10 to 15 minutes at all practice sessions, and each player was given their own ankle disk.
There was no further detail about the ankle disk training program other than that the design was based on principles from elite athlete training programs and those designed for rehabilitation of injured athletes with ankle or knee instability. Only 2 knee injuries were noted in the intervention group versus 8 in the control group; however, the difference was not statistically significant.11
This study's strengths related to our clinical question were that it directly addressed balance board or proprioceptive training and the intervention was with young female athletes. We believed that the running, planting, and cutting maneuvers as well as jumping in soccer and team handball were comparable, which made the study relevant to our client. Conversely, the sport of team handball varies from soccer in the playing surface, court and field dimensions, and footwear. Anterior cruciate ligament injuries were not distinguished from other knee injuries (eg, MCL tears). In addition, upon further investigation, Wedderkopp et al included all athletes with and without previous knee injury in their study. Although injury incidence was one of the factors controlled for in the randomization of 11 teams to the intervention group, it was not stated whether the randomization was specific for each type of injury (eg, ankle sprain versus ACL sprain). With more practice time, the risk of injury increased in the control group and decreased in the intervention group12; this points to a positive effect of the training program.
Thacker et al concluded that neuromuscular and proprioception training have the most encouraging evidence for knee injury prevention. Furthermore, they stated that (1) the quality of the research to date was not high enough because of methodological flaws, and (2) further research in this area was needed. We believed that the systematic review by Thacker et al was a comprehensive search of the literature in the area of knee injury prevention and made conclusions in accordance with the assessment of the quality of the selected articles. We therefore believed that the systematic review did provide limited evidence of the positive effects of proprioception or balance training on knee injury prevention. Specifically, the Wedderkopp et al study, with the relatively high quality score of 56/100, pointed to decreased knee injury incidence with ankle disk training. The training protocol was not described in detail, which also prevented us from making a recommendation to our client. In the context of our clinical question, we needed more evidence before recommending the incorporation of a proprioception or balance training program into our client's training regimen.
The second study identified in the systematic review as relevant to our clinical question also was retrieved by our own search.
Caraffa A, Cerulli G, Projetti M, et al. Prevention of anterior cruciate ligament injuries in soccer: a prospective controlled study of proprioceptive training. Knee Surg Sports Traumatol Arthrosc. 1996;4:19–21.
This study by Caraffa et al13 interested us because it examined the effect of proprioception or balance training in prevention of ACL injuries over a 3-year period. Six hundred players in 40 semi-professional and amateur soccer teams in Italy volunteered to participate and were divided into 2 groups: group A received the proprioception training program, and group B followed their normal training. According to the authors, both groups had equal numbers of semi-professional and amateur teams, the same number of players according to field position, and similar age ranges, training hours, and playing hours. The authors, however, did not provide data on the age of the participants or on the number of training or playing hours. We assumed that all participants were male because it was not otherwise stated by the authors. It also was unclear how the participants were allocated to the control and intervention groups. Players with previous knee injuries were excluded from the study.
The proprioception training group (group A) trained daily for a minimum of 20 minutes per day throughout the preseason and at least 3 times per week during the regular season. Training occurred in 5 phases, progressing from balance training without a balance board to balance training on a multiplanar balance board. Any players who were injured in either group underwent physical examination, and suspected ACL injuries were confirmed through arthroscopy. The incidence of ACL injury was 1.15 injuries per team per season in group B and 0.15 in group A. The authors concluded that proprioceptive training reduces the number of ACL injuries in soccer.
Although the authors mentioned that the 2 groups were comparable, no data were provided to support this contention. They also did not provide details on the training protocol. Teams may have incorporated neuromuscular programs (plyo-metrics, weights) that may have helped lower the incidence of ACL injuries.3 Team roster changes, dropouts from the program, and level of adherence to the training protocol also were not mentioned. There was no indication of monitoring throughout the season, such as average training frequencies or instruction in the proprioception training techniques.
Because no training protocol details were provided in the article, we remembered a commentary on this article by the same author group (Cerulli et al14) that was retrieved in our initial search results. We initially had excluded the commentary from our review because it did not present original research. This commentary outlines the training protocol used by Caraffa et al in detail. The 5-phase proprioceptive training program involved anterior and posterior step-down exercises while standing on balance boards of varying difficulty. To progress to the next phase, each participant was instructed to assess his own ability to perform the exercise without the need for additional support in a consistent manner and without the fear of falling. All subjects in the experimental group (group A) participated in proprioceptive neuromuscular facilitation exercises for the lower limbs, assisted by a trained technician. These exercises were meant to increase subject awareness of limb position; however, this effect was not measured, which and unfortunately confounds our search for the effectiveness of proprioceptive balance board training.
Initially, the evidence in favor of adding proprioception or balance training to traditional training looked positive. For our client, this study of elite male soccer players may have some degree of carryover; however, more information on the study method was needed before we could endorse a proprioception training program. Furthermore, the results of the study were confounded by the simultaneous use of proprioceptive neuro-muscular facilitation.
Soderman K, Werner S, Pietila T, et al. Balance board training: prevention of traumatic injuries of the lower extremities in female soccer players? A prospective randomized intervention study. Knee Surg Sports Traumatol Arthrosc. 2000;8:356–363.
This study was appealing because it focused specifically on balance board training to prevent ACL injuries in women. A total of 221 female soccer players from 13 different teams in the second and third divisions in Sweden volunteered to participate in the study. Seven teams were randomized to an intervention group and 6 teams to a control group. The authors do not explain the details of how the randomization was done other than stating that approximately the same number of players from the second and third Swedish soccer divisions were represented in each group. Both groups had a mean age of 20, and averaged 12 years of soccer training.
The balance board training protocol was not specified other than a statement that the program was devised and instructed by one of the authors. Each player was provided with her own balance board and a printed handout of instructions, and the players were instructed to perform the exercises for 10 to 15 minutes per day at home. The program consisted of 5 exercises with a progressively increasing degree of difficulty. One exercise had the players stand on 1 leg at a time in a slightly flexed position for 3 sets of 15 seconds per leg. There was no further information provided regarding the exercises and how progression was determined except a statement that the balance board height could be elevated to increase the degree of difficulty.
Soderman et al10 included all members of the soccer team, regardless of pre-existing lower-extremity pathology. Although our search criteria excluded studies that had participants with prior ACL injury, we decided to include this study because the authors distinguished between the incidence of first-time ACL injury and repeat ACL injury. According to the authors, there were no significant differences between the control and intervention groups regarding the number of first-time injuries or reinjuries. Specifically, 4 out of 5 ACL injuries occurred in the intervention group. Upon closer examination, among the players who had been injured prior to the study, more players from the control group than from the intervention group sustained new injuries during the study period.10 Soderman et al provided a clear flow chart of dropouts in the intervention and control groups and made efforts to monitor adherence of the participants (players recorded training time in a special protocol). The dropout rate was 37%.
The incidence of ACL injuries, in fact, was greater in the intervention group compared with the control group. These results contradict the findings of Caraffa et al,13 who reported a considerable decrease in ACL injuries in male soccer players following proprioception or balance training. Possible causes for this disagreement could include differences in intervention protocol (Soderman et al had the players complete 3 × 15 seconds of each exercise per leg for a total of 15 minutes, whereas Caraffa et al had the players perform 4 × 2.5 minutes per leg for a total of 20 minutes), inclusion criteria (Caraffa et al excluded players with an earlier knee injury), sex, duration of the study (Caraffa et al studied the soccer players over 3 seasons versus 1 season by Soderman et al), and training environment and motivation (Soderman et al had the athletes perform exercises independently at home, whereas Caraffa et al did not explain where the teams performed the protocol).
According to the PEDro scale for assessing study quality, 85% of subjects initially allocated to groups must be measured for at least 1 key outcome at the conclusion of the study.16 Because dropouts can seriously distort the study's findings, the treatment effect found by Soderman et al cannot be considered strong evidence. In addition, the article's description of the training protocol was vague, which limited our ability to make recommendations for a training protocol. Therefore, we concluded that this study did not provide evidence that would help us make a clinical decision about whether to include proprioception or balance board training in our client's training regimen.
In our examination of the 3 articles retrieved by our search, we found a lack of strong evidence demonstrating the effectiveness of isolated proprioception or balance training in preventing ACL injury. Two of 3 studies, however, showed a decrease in ACL injury incidence in the treatment group, and no studies demonstrated detrimental effects from proprioceptive or balance training that were statistically significant.10,11,13 Therefore, despite the methodological flaws we identified in one of these studies,13 we believed that our client might benefit from proprioception or balance training.
Thacker et al11 supported our reasoning by stating that studies of neuromuscular and proprioception training show the most encouraging evidence of effective ACL injury prevention when compared with bracing, footwear, and ski boot binding systems.
Caraffa et al13 found a significant decrease in ACL injuries in elite male soccer players as a result of proprioception or balance training with a balance board. However, methodological weaknesses in the study do not allow us to base our recommendation on the work of this study alone.
Soderman et al10 showed no significant change in incidence of ACL injury following balance board training in female soccer players; it was the only study to find no significant effect of proprioception or balance training on number of ACL injuries. Although there were a greater number of incidents of injury in the intervention group, they were not statistically significant. Therefore, this study does not suggest a significant benefit or detrimental effect of balance board training on ACL injury prevention.
We informed our client that there was limited evidence to support the use of balance board training alone. Furthermore, none of the articles from our search provided detailed, replicable training protocols that would allow us to produce a training program for her. We could not recommend balance board training alone in order to reduce the risk of ACL injury. Although no significant detrimental effects have been found with balance board training, we advised our client that her time may be better spent on other efficacious training components that are backed by evidence. As described in the meta-analysis published by Hewett et al16 in 2006, a program combining multiple neuromuscular training components appears to provide a reduced risk of ACL injury. For example, plyometric training combined with technique training and biomechanical analysis has shown consistent effectiveness in reducing the incidence of ACL injuries.16 A training protocol encompassing these principles may be the most effective for our client's goal; however, without review of the literature on this topic, we could not make a specific recommendation for our client at that time.
We recommend that future research on ACL injury prevention should investigate training protocols for proprioception or balance in isolation and as a component of a neuromuscular training program. In addition, detailed outlines of these programs need to be provided so that clinicians may follow them with their clients.
All authors provided writing. The authors would like to gratefully acknowledge the advice and support of Donna L MacIntyre, PT, PhD, Susan R Harris, PT, PhD, and Tyler L Dumont, PT, MSc, of the School of Rehabilitation Sciences, University of British Columbia, Canada.
- Received October 14, 2005.
- Accepted June 23, 2006.
- Physical Therapy