Patrick Sadoghi, MD, Arvind von Keudell, MD, and Patrick Vavken, MD, MSc

Transcription

1 1 COPYRIGHT Ó 2012 BY THE JOURNAL OF BONE AND JOINT SURGERY, INCORPORATED Effectiveness of Anterior Cruciate Ligament Injury Prevention Training Programs Patrick Saoghi, MD, Arvin von Keuell, MD, an Patrick Vavken, MD, MSc Investigation performe at the Department of Orthopaeic Surgery, Meical University of Graz, Graz, Austria; the Cartilage Repair Center, Brigham an Women s Hospital, Harvar Meical School; an the Sports Meicine Research Laboratory, Chilren s Hospital Boston, Harvar Meical School, Boston, Massachusetts Backgroun: The objective of this stuy was to systematically review the literature on anterior cruciate ligament (ACL) injury prevention programs an to perform a meta-analysis to aress three questions: First, what is the effectiveness of ACL injury prevention programs? Secon, is there evience for a best program? Thir, what is the quality of the current literature on ACL injury prevention? Methos: We conucte a systematic review with use of the online PubMe, MEDLINE, EMBASE, CINAHL (Cumulative Inex to Nursing an Allie Health), an Cochrane Central Register of Controlle Trials atabases. Search terms were anterior cruciate ligament, knee, injury, prevention, an control. Data on stuy esign an clinical outcomes were extracte inepenently in triplicate. After assessment of between-stuy heterogeneity, DerSimonian-Lair ranom-effect moels were use to calculate poole risk ratios an risk ifferences. The risk ifference was use to estimate the number neee to treat (the number of iniviuals who woul nee to be treate to avoi one ACL tear). Results: The poole risk ratio was 0.38 (95% confience interval [CI], 0.20 to 0.72), reflecting a significant reuction in the risk of ACL rupture in the prevention group (p = 0.003). The number neee to treat range from five to 187 in the iniviual stuies. Stratifie by sex, the poole risk ratio was 0.48 (95% CI, 0.26 to 0.89) for female athletes an 0.15 (95% CI, 0.08 to 0.28) for male athletes. Conclusions: Our stuy inicate strong evience in support of a significant effect of ACL injury prevention programs. Our poole estimates suggest a substantial beneficial effect of ACL injury prevention programs, with a risk reuction of 52% in the female athletes an 85% in the male athletes. Level of Evience: Therapeutic Level II. See Instructions for Authors for a complete escription of levels of evience. Approximately 100,000 injuries of the anterior cruciate ligament (ACL) occur per year in the Unite States, an costs associate with the evaluation, therapy, rehabilitation, an possible loss of funing an scholarships of affecte athletes total $625 million to $1 billion annually 1-4. ACL tears are serious injuries an preispose patients to subsequent osteoarthritis, which in turn causes further pain, immobility, an reuction in quality of life an ability to work. Fifty to ninety percent of untreate patients evelop osteoarthritis within ten to fifteen years after the ACL injury 1-4. Disclosure: None of the authors receive payments or services, either irectly or inirectly (i.e., via his or her institution), from a thir party in support of any aspect of this work. One or more of the authors, or his or her institution, has ha a financial relationship, in the thirty-six months prior to submission of this work, with an entity in the biomeical arena that coul be perceive to influence or have the potential to influence what is written in this work. No author has ha any other relationships, or has engage in any other activities, that coul be perceive to influence or have the potential to influence what is written in this work. The complete Disclosures of Potential Conflicts of Interest submitte by authors are always provie with the online version of the article. This article was chosen to appear electronically on March 28, 2012, in avance of publication in a regularly scheule issue. A commentary by Rick W. Wright, MD, is linke to the online version of this article at jbjs.org. J Bone Joint Surg Am. 2012;94:1-8

2 2 The current treatment of ACL tears involves reconstruction with use of autogenous or allogeneic grafts 5-9. However, espite promising short an intermeiate-term outcomes, questions have arisen regaring the long-term efficacy of current treatments of ACL tears with regar to preventing osteoarthritis 2, Long-term stuies have inicate that 41% to 75% of patients unergoing ACL reconstruction still evelop osteoarthritis by fourteen years postoperatively; in aition, the rate of graft failure is high, peaking at 20% to 25% in aolescent patients, resulting in knee instability as well as osteoarthritic progression 1,25,26. New treatment options, such as regenerative treatments base on tissue engineering methos, are being evelope but are not available for clinical use A ifferent approach to the problem of ACL tears involves prevention. Approximately 80% of all ACL tears are noncontact injuries, suggesting that a substantial percentage of tears coul be avoie 28,29. Inee, a number of ACL injury prevention programs have been evelope an have been shown to reuce ACL injury rates significantly These programs usually target high-risk groups, such as young female athletes, an aim to improve angerous motion patterns. For example, a program may aim to improve laning technique from flat foot laning with an extene lower extremity to laning with eep hip an knee flexion. However, the effectiveness of these programs has not been comprehensively analyze an escribe. The objective of the present stuy was to systematically review the literature on programs for prevention of ACL injury in female an male athletes an to perform a meta-analysis to aress three questions: First, o ACL injury prevention programs ecrease the risk of ligament injuries? Secon, is there an inication for a best program? Thir, what is the valiity of these finings? Materials an Methos This stuy was conucte in accorance with the PRISMA (Preferre Reporting Items for Systematic reviews an Meta-Analyses) statement 33,34. Systematic Search an Strategy We conucte a systematic review of the literature with use of the PubMe, MEDLINE, EMBASE, CINAHL (Cumulative Inex to Nursing an Allie Health Literature), an Cochrane Central Register of Controlle Trials atabases. We searche these electronic atabases online for (anterior cruciate ligament OR knee) AND (control OR prevention) AND injury, using these terms as keywors an exploe MeSH (Meical Subject Heaings) terms. The search was not restricte by language or year of publication. We then performe a meta-analysis of the controlle trials that were ientifie, pooling ata on the effectiveness of ACL injury prevention algorithms to answer our three stuy questions. We inclue only prospective, controlle stuies that irectly compare ACL injury prevention programs with no treatment in human subjects. Stuies with partially overlapping ata were merge to the extent possible; uplicate stuies with completely overlapping ata were exclue. Stuies that i not focus on clinical treatment or outcome, animal stuies, stuies without any intervention, an stuies with a reporte attrition of >20% were also exclue. Eligible interventions involve proprioceptive neuromuscular training techniques (e.g., Prevent injury an Enhance Performance [PEP] programs) with or without a balance boar, roun boar, or wobble boar. Extraction of Relevant Data Eligibility of stuies was assesse inepenently in triplicate an crosschecke to avoi errors. Disagreement was resolve by iscussion or, if necessary, by the ecision of the senior author (P.V.). The bibliographies of all inclue stuies were reviewe for aitional relevant stuies. Data concerning stuy esign, participant characteristics, an the number of ACL ruptures at the time of final follow-up were abstracte from the inclue stuies in triplicate an crosschecke. All searches were conclue by December The level of evience (I through V), use of ranomization (yes/no), use of blining (yes/no), an reporting of attrition (yes/no) were extracte from the inclue stuies to escribe the stuy quality. Other extracte ata involve the type of prevention program, the total number of participants enrolle in the prevention program an in the control group, the number of athletes with suitable follow-up in each group, the number of uninjure athletes an the number of athletes with an ACL tear at the time of follow-up in each group, the uration of follow-up, the type of sport performe, an the sex of the participants. Assessment of Valiity We etermine the level of evience for each inclue stuy. The internal valiity of each stuy was further assesse with use of a moifie Jaa scale, which assigns one point each for use of ranomization, use of blining, an reporting of attrition (with zero representing the poorest possible result an three points representing the best result) 35. Publication Bias Another important problem that jeoparizes the valiity of a meta-analysis is publication bias, or so-calle file-rawer bias, leaing to the omission of unpublishe stuies 36. These stuies often remain in a file rawer because their results are not statistically significant, resulting in an erroneously high proportion of stuies with significant results among publishe stuies 37,38. Publication bias among the inclue stuies was assesse graphically with use of a funnel plot an mathematically with use of the Egger weighte regression technique 39. Stuy Heterogeneity The presence of between-stuy heterogeneity was assesse qualitatively with use of the Cochrane Q test (with a p value of 0.10 being consiere significant because of the low power of this test in small samples) an quantitatively with use of the I 2 inex 40. Meta-regressions were performe to assess the potential sources of such heterogeneity 41,42. Quantitative Data Synthesis The risk ifference was use to calculate the number neee to treat (equal to 1 ivie by the risk ifference) in this case, the number of participants who have to be enrolle in the prevention program to result in one less ACL tear. (The number neee to harm, rather than the number nee to treat, was reporte for stuies that showe a better outcome in the group that i not participate in the prevention program.) All analyses were performe on an intention-to-treat basis i.e., participants were inclue in the analysis accoring to their initial allocation an their injury status at the time of final follow-up 43. Such an analysis prouces a more conservative result, with a larger p value compare with an astreate analysis, but it is also more realistic. Data were poole to calculate the poole risk ratio an the poole risk ifference by constructing ranom-effects moels with use of the DerSimonian- Lair metho 44. Such moels postulate that the observe heterogeneity among the stuies in a meta-analysis is attributable to normally istribute iniviual effects aroun a common effect. This assumption was assesse graphically with use of a forest plot. All calculations were performe with use of Intercoole Stata (version 10; StataCorp, College Station, Texas). A p value of 0.05 was consiere significant for the poole estimates. Source of Funing No funing was obtaine for this stuy.

3 3 Fig. 1 Ientification of stuies eligible for inclusion in the meta-analysis. Results Stuy Characteristics Our search strategy ientifie 909 stuies. After exclusion of uplicate stuies, stuies that i not focus on clinical treatment or outcome, animal stuies, an stuies without any intervention, nine stuies remaine eligible for analysis (Fig. 1). The nine inclue stuies were publishe between 1996 an 2008 in English an German (see Appenix). Five stuies focuse exclusively on soccer players, two focuse exclusively on hanball players, an two involve a combination of soccer, basketball, an volleyball players (see Appenix). The stuy by Söermanetal. 53 ha 22% attrition in the control group an 49% in the prevention group. That stuy was therefore use only for a sensitivity analysis i.e., ata were poole with an without that stuy, an the two poole estimates were compare to assess the magnitue of the change that resulte from aition of the potentially biase stuy. The magnitue of this change shows how robust our results were (how resistant they were to flaws in the esigns of the iniviual stuies). Outcomes of the Inclue Prevention Programs Caraffa et al. 45 compare 300 Italian soccer players who use a special training program for at least twenty minutes per session for at least thirty ays with 300 players who were aske to train as usual. Training was ivie into five phases an inclue the use of a balance boar with a focus on proprioceptive training. The authors reporte significantly fewer ACL injuries in the intervention group (ten of 300) than in the control group (seventy of 300), an they conclue that proprioceptive training shoul become stanar uring preseason training as well as uring the actual playing season. Gilchrist et al. 46 ranomly assigne National Collegiate Athletic Association (NCAA) Division I women s soccer teams to a group of 583 athletes who performe special training or to a control group of 852 athletes who performe normal training. The program was use for twelve weeks, with 11.6% attrition overall. The training focusing on stretching, strengthening, plyometrics, agility, an avoiance of high-risk positions epicte on a vieo, an replacement exercises were also provie to minimize boreom. The overall ACL injury rate was 1.7 times lower in the intervention group than in the control group (p < 0.198). The authors conclue that athletes with a history of ACL injury obtaine an especially great benefit from the prevention program. Heit et al. 47 ranomly selecte forty-two of 300 female high school soccer players to participate in the Frappier Acceleration

4 4 Training Program, which combines cariovascular conitioning, plyometric work, sport cor rills, strength training, an flexibility exercises over a seven-week perio. The authors observe significantly fewer injuries in the intervention group than in the control group, an they conclue that this type of conitioning can lower the rate of ACL injury in female aolescent soccer players. Hewett et al. 48 prospectively observe three groups of aolescents: 366 girls on fifteen high school soccer, volleyball, an basketball teams (97, 185, an 84 girls, respectively) participate in a six-week preseason neuromuscular training program; 463 girls on the same fifteen high school soccer, volleyball, an basketball teams (193, 81, an 189 girls, respectively) i not receive the training; an 434 boys (soccer an basketball players) on thirteen teams who i not receive the training serve as an aitional control population. A total of fourteen serious knee injuries were observe in the 1263 athletes. The girls who unerwent the training sustaine significantly fewer injuries compare with the girls who i not train, but not compare with the boys who i not train. The authors conclue that before young female athletes participate in sports that entail jumping, pivoting, an cutting, they shoul unergo a jump training program of proven effectiveness that inclues progressive resistance weight training for the lower extremity. Manelbaum et al. 49 performe a prospective, nonranomize cohort stuy of 1041 female soccer players from fiftytwo teams who receive a sports-specific training intervention. An age an skill-matche group of 1905 players from the same league serve as the control group. The authors conucte a secon stuy with the same esign, incluing 844 players in the intervention group an 1913 in the control group, one year later. A total of 189 players were lost to follow-up. The intervention consiste of eucation, stretching, strengthening, plyometrics, an sports-specific agility rills. The authors observe an 88% ecrease in ACL injuries uring the first season an a 74% ecrease uring the secon season compare with the control groups, an they conclue that neuromuscular training programs may have a irect benefit by ecreasing the number of ACL tears. Petersen et al. 50 conucte a prospective case control stuy of female hanball players to investigate the effect of injury prevention training consisting of balance boar exercises an jump exercises. The stuy inclue 134 players in the intervention group an 142 players in the control group. The authors foun knee injuries to be the secon most frequent type of injury after ankle sprains, an they observe five ACL tears in the intervention group compare with nine in the control group. The authors conclue that proprioceptive an neuromuscular training is appropriate for the prevention of knee an ankle injuries among female hanball players. In a secon investigation, Petersen et al. 51 compare a hanball team that performe proprioceptive an neuromuscular training uring the preseason with another team that was traine as usual. The intervention consiste of proprioceptive training, jump training, an information on injury mechanisms. The authors foun that the training significantly reuce the risk of ACL tears, an they conclue that prevention strategies shoul be inclue in routine hanball training programs. Pfeiffer et al. 52 compare 577 female athletes who unerwent a knee ligament injury prevention (KLIP) program with 862 female athletes in a control group. The rills in the KLIP program consiste of running, jumping, an laning in forwar an backwar irections; the athlete initially lane on both feet an later progresse to laning on a single foot. This program was esigne to require less time than the prevention program propose by Hewett et al. 48. The authors observe no significant ifference in the rate of knee injuries between the groups, an they conclue that the rate of noncontact ACL tears will not be lowere by use of the KLIP program. Fig. 2 Funnel plot showing the eight poole ACL injury prevention stuies. The funnel plot is a tool to graphically assess publication bias by plotting the stanar error (se) of the logarithm of the relative risk (RR) against the RR. The vertical ashe line shows the poole RR, an the iagonal ashe lines show the corresponing 95% confience interval. Although the stuies are somewhat asymmetrically arrange aroun the poole effect, all stuies lie within the funnel, suggesting a low risk of publication bias. Furthermore, mathematical assessment showe no evience of publication bias (p = 0.131).

5 5 Söermanetal. 53 conucte a prospective stuy in which 121 female soccer players were ranomize to the intervention group an 100 to the control group. The intervention consiste primarily of exercises involving use of a balance boar for ten to fifteen minutes three times per week at home for at least thirty ays. The injury rate i not iffer significantly between the groups, an the authors conclue that balance boar training i not prevent ACL injuries in female soccer players. As mentione above, this stuy suffere from massive, ifferential attrition of participants prior to the follow-up evaluation. Publication Bias The funnel plot (Fig. 2) was slightly skewe, suggesting that a few negative stuies (i.e., stuies that showe no significant ifference between the groups) are missing from the literature. However, quantitative assessment of this possibility with use of Egger regression showe no significant evience of publication bias in the inclue stuies (p = 0.131). Heterogeneity Heterogeneity among the results of the stuies was significant (p = 0.011), with an I 2 inex of 64.0%. However, as shown in the forest plot (Fig. 3), the stuies were uniformly istribute about a common effect, suggesting that ranom-effect moeling coul be use to pool the iniviual results. Poole Effect The risk ifferences varie consierably among the iniviual stuies, with the corresponing estimates of the number neee to treat ranging from five to 187 except in the stuy by Pfeiffer et al. 52, which showe a slightly lower risk in the control group (leaing to an inversion of the number neee to treat to a number neee to harm [582]) (see Appenix). However, the poole DerSimonian-Lair ranom-effect risk ratio was 0.38 (95% confience interval [CI], 0.20 to 0.72), an the reuction in the risk of ACL rupture in the prevention group was significant (p = 0.003) (Fig. 3). Fig. 3 Forest plot showing the finings of the meta-analysis. The stuy by Caraffa et al. analyze all ACL tears, an the other seven poole primary stuies analyze only noncontact ACL tears. The risk ratio (RR) is plotte on a logarithmic scale on the x axis, an the ashe vertical line an the iamon represent the poole RR. The sizes of the squares for the inclue stuies represent the sizes of those stuies, an the horizontal lines an the with of the iamon represent the associate 95% confience intervals (CIs). A significant treatment effect is inicate if the CI oes not inclue the soli vertical line enoting the absence of an effect.

6 6 Stratifie by sex, the poole risk ratio was 0.48 (95% CI, 0.26 to 0.89) for female athletes an 0.15 (95% CI, 0.08 to 0.28) for male athletes. The reuction in the number of ACL tears was significant in both female athletes (p = 0.021) an male athletes (p < 0.001). Inclusion of the probably biase stuy by Söerman et al. 53 change the poole DerSimonian-Lair ranom-effect risk ratio from 0.38 to 0.45 (95% CI, 0.23 to 0.89), which was still consistent with a significant reuction in the risk of ACL rupture in the prevention group (p = 0.021). Meta-Regression Meta-regression of the effect of iniviual variables on the poole risk ratio (i.e., on the effect of preventive treatment) showe no effect of a focus on balance boar use (p = 0.712), use of vieo assistance (p = 0.914), uration of follow-up (p = 0.437), or the year of publication (p = 0.358). The regression showe that conucting prevention programs uring the preseason compare with the playing season reuce the risk of ACL injury by 19.1%, but this ifference was not significant (p = 0.691). Stuy Quality The mean Jaa score for the inclue stuies was one point (95% CI, 0.43 to 1.57 points). Only three of the nine stuies use a ranomize esign, an only two use bline outcome assessment. One of the stuies with the highest Jaa score (two points), which was also the only stuy with an a priori sample size calculation, was the stuy by Söerman et al. 53 that ha to be exclue because of selection bias. Discussion Summary of Evience The authors conclusions regaring the effectiveness of ACL injury prevention programs iffere among the inclue manuscripts. Caraffa et al. 45,Heitetal. 47,Petersen et al. 50,51, an Hewett et al. 48 reporte positive effects with injury prevention programs involving various training methos, resulting in significantly lower rates of ACL injury. Gilchrist et al. 46 an Manelbaum et al. 49 conclue only that there might be an avantage to training, without making efinite recommenations. In contrast, Pfeiffer et al. 52 an Söerman et al. 53 conclue that a twenty-minute plyometric-base exercise program twice a week 52 or training involving the use of balance boars for ten to fifteen minutes three times a week for at least thirty ays 53 woul not significantly reuce the risk of ACL injury in female high school athletes in noncontact sports, but these authors i not clearly pinpoint the reason for the ifference between these finings an those of prior stuies. Pooling the results from the iniviual stuies showe strong evience for a significant, positive effect of prevention programs. Furthermore, the poole risk ratio of 0.38 inicate that athletes in the prevention programs ha a 62% reuction in the risk of ACL rupture compare with athletes in the control groups. Stratification by sex showe that female athletes actually benefite less than male athletes, with a risk reuction of 52% compare with 85% in male athletes. The substantial magnitue of this protective effect was supporte by the estimates of the number neee to treat; the number of participants who woul nee to be enrolle in a prevention program to result in one less ACL tear range from five to 187, which is fairly small consiering the size of an average high school or varsity team. Of note, two of the stuies, by Caraffa et al. 45 an by Söerman et al. 53, inclue all ACL injuries rather than focusing on noncontact ACL tears, which coul bias our fining. However, the outcomes reporte in those two stuies i not eviate from those of the other inclue stuies. Although we attempte to ientify a best training program to avoi ACL injury, we were unable to fin conclusive evience supporting any one specific type of intervention because of the consierable heterogeneity of the inclue stuies. Despite this heterogeneity, there is agreement among the results of the successful trials that an ACL injury prevention program shoul inclue at least ten minutes of exercises three times per week, with a focus on neuromuscular training, as a bare minimum. However, we foun no evience supporting the superiority of balance boar exercises, vieo assistance, or new protocols over oler ones. In general, the scientific quality of the inclue stuies was low, with only two stuy esigns having appropriate blining (Petersen et al. 50 an Pfeiffer et al. 52 ) an only three having aequate ranomization (Gilchrist et al. 46, Heit et al. 47,an Söerman et al. 53 ). Matche-pair analysis was performe in only three stuies: groups were matche with regar to age an number of training units by Caraffa et al. 45 ; with regar to age, height, weight, muscle flexibility, balance an postural sway of the lower extremities, an the number of years in soccer training by Söerman et al. 53 ; an with regar to age an athletic skill by Manelbaum et al. 49. Finally, only one stuy inclue an a priori power analysis 53, an this stuy was biase by very high an ifferential rop-out rates of 49% in the intervention group an 22% in the control group. Limitations Our stuy has potential shortcomings. As with all systematic reviews an meta-analyses, the valiity of our finings epens on the valiity of the primary stuies. Although the overall quality of the inclue stuies was low, such low stuy quality graes are not unusual for surgical an musculoskeletal trials. Furthermore, the fact that the poole estimates change only moestly (by 18%) after inclusion of the biase stuy by Söerman et al. 53 showe that our finings were quite robust. Another shortcoming is the heterogeneity of the results of the inclue primary stuies. Although this heterogeneity was normally istribute about a common effect, an thus pooling with use of a ranom-effect moel is allowe, the heterogeneity prevente us from ientifying eterminants of the success of preventive training. Once more stuies involving the same type of prevention are available, subgroup analyses can be conucte to assess eterminants of the outcome. It is

7 7 worth noting that most stuies (five of nine) focuse exclusively on soccer players, an this might therefore be a source of bias in the present stuy. However, such bias woul result in clustering in the forest plot (Fig. 3), with soccer stuies on one sie an hanball stuies on the other. Instea, the forest plot showe a uniform istribution of all stuies about the poole relative risk. Conclusion In conclusion, our stuy prouce strong evience in support of a significant effect of ACL injury prevention programs. Our poole estimates suggeste a substantial beneficial effect of ACL prevention programs, with a 52% reuction in the risk of an ACL tear in female athletes but an 85% reuction in male athletes. Although the current literature was foun to be of average quality, sensitivity analyses suggeste that our finings are robust. However, we are not able to recommen a specific type of prevention program on the basis of the current publishe evience. Appenix Tables showing the characteristics of the inclue stuies an the estimates of the number neee to treat are available with the online version of this article as a ata supplement at jbjs.org. n Patrick Saoghi, MD Department of Orthopaeic Surgery, Meical University of Graz, Auenbruggerplatz 5, 8036 Graz, Austria. aress: Patrick.saoghi@klinikum-graz.at Arvin von Keuell, MD Cartilage Repair Center, Brigham an Women s Hospital, Harvar Meical School, 850 Boylston Street, Suite 112, Chestnut Hill, MA aress: avonkeuell@rics.bwh.harvar.eu Patrick Vavken, MD, MSc Department of Orthopeic Surgery, Chilren s Hospital Boston, 300 Longwoo Avenue, Eners 270, Boston, MA aress: patrick.vavken@chilrens.harvar.eu References 1. von Porat A, Roos EM, Roos H. High prevalence of osteoarthritis 14 years after an anterior cruciate ligament tear in male soccer players: a stuy of raiographic an patient relevant outcomes. Ann Rheum Dis. 2004;63: Lohmaner LS, Englun PM, Dahl LL, Roos EM. The long-term consequence of anterior cruciate ligament an meniscus injuries: osteoarthritis. Am J Sports Me. 2007;35: Daniel DM, Stone ML, Dobson BE, Fithian DC, Rossman DJ, Kaufman KR. Fate of the ACL-injure patient. A prospective outcome stuy. Am J Sports Me. 1994;22: Lohmaner LS, Roos H. Knee ligament injury, surgery an osteoarthrosis. Truth or consequences? Acta Orthop Scan. 1994;65: Biau DJ, Tournoux C, Katsahian S, Schranz PJ, Nizar RS. Bone-patellar tenonbone autografts versus hamstring autografts for reconstruction of anterior cruciate ligament: meta-analysis. BMJ. 2006;332: Poolman RW, Abouali JA, Conter HJ, Bhanari M. Overlapping systematic reviews of anterior cruciate ligament reconstruction comparing hamstring autograft with bone-patellar tenon-bone autograft: why are they ifferent? J Bone Joint Surg Am. 2007;89: Poolman RW, Farrokhyar F, Bhanari M. Hamstring tenon autograft better than bone patellar-tenon bone autograft in ACL reconstruction: a cumulative metaanalysis an clinically relevant sensitivity analysis applie to a previously publishe analysis. Acta Orthop. 2007;78: Saoghi P, Müller PE, Jansson V, van Griensven M, Kröpfl A, Fischmeister MF. Reconstruction of the anterior cruciate ligament: a clinical comparison of bonepatellar tenon-bone single bunle versus semiteninosus an gracilis ouble bunle technique. Int Orthop. 2011;35: Saoghi P, Kröpfl A, Jansson V, Müller PE, Pietschmann MF, Fischmeister MF. Impact of tibial an femoral tunnel position on clinical results after anterior cruciate ligament reconstruction. Arthroscopy. 2011;27: Casteleyn PP. Management of anterior cruciate ligament lesions: surgical fashion, personal whim or scientific evience? Stuy of meium- an long-term results. Acta Orthop Belg. 1999;65: Linko E, Harilainen A, Malmivaara A, Seitsalo S. Surgical versus conservative interventions for anterior cruciate ligament ruptures in aults. Cochrane Database Syst Rev. 2005;2:CD Neuman P, Kostogiannis I, Frién T, Roos H, Dahlberg LE, Englun M. Patellofemoralosteoarthritis 15 years after anterior cruciate ligament injury a prospective cohort stuy. Osteoarthritis Cartilage. 2009;17: Lin M, Menhert F, Peersen AB. The first results from the Danish ACL reconstruction registry: epiemiologic an 2 year follow-up results from 5,818 knee ligament reconstructions. Knee Surg Sports Traumatol Arthrosc. 2009;17: Lién M, Sernert N, Rostgår-Christensen L, Kartus C, Ejerhe L. Osteoarthritic changes after anterior cruciate ligament reconstruction using bone-patellar tenonbone or hamstring tenon autografts: a retrospective, 7-year raiographic an clinical follow-up stuy. Arthroscopy. 2008;24: van er Hart CP, van en Bekerom MP, Patt TW. The occurrence of osteoarthritis at a minimum of ten years after reconstruction of the anterior cruciate ligament. J Orthop Surg Res. 2008;3: Neuman P, Englun M, Kostogiannis I, Frién T, Roos H, Dahlberg LE. Prevalence of tibiofemoral osteoarthritis 15 years after nonoperative treatment of anterior cruciate ligament injury: a prospective cohort stuy. Am J Sports Me. 2008;36: Kessler MA, Behren H, Henz S, Stutz G, Rukavina A, Kuster MS. Function, osteoarthritis an activity after ACL-rupture: 11 years follow-up results of conservative versus reconstructive treatment. Knee Surg Sports Traumatol Arthrosc. 2008;16: Cohen M, Amaro JT, Ejnisman B, Carvalho RT, Nakano KK, Peccin MS, Teixeira R, Laurino CF, Aballa RJ. Anterior cruciate ligament reconstruction after 10 to 15 years: association between meniscectomy an osteoarthrosis. Arthroscopy. 2007;23: Meunier A, Oensten M, Goo L. Long-term results after primary repair or non-surgical treatment of anterior cruciate ligament rupture: a ranomize stuy with a 15-year follow-up. Scan J Me Sci Sports. 2007;17: Aït Si Selmi T, Fithian D, Neyret P. The evolution of osteoarthritis in 103 patients with ACL reconstruction at 17 years follow-up. Knee. 2006;13: Seon JK, Song EK, Park SJ. Osteoarthritis after anterior cruciate ligament reconstruction using a patellar tenon autograft. Int Orthop. 2006;30: Lohmaner LS, Ostenberg A, Englun M, Roos H. High prevalence of knee osteoarthritis, pain, an functional limitations in female soccer players twelve years after anterior cruciate ligament injury. Arthritis Rheum. 2004;50: Jäger A, Welsch F, Braune C, Eberhart C, Kappler C. [Ten year follow-up after single incision anterior cruciate ligament reconstruction using patellar tenon autograft]. Z Orthop Ihre Grenzgeb. 2003;141:42-7. German. 24. Gillquist J, Messner K. Anterior cruciate ligament reconstruction an the long-term incience of gonarthrosis. Sports Me. 1999;27: Vavken P, Murray MM. Translational stuies in anterior cruciate ligament repair. Tissue Eng Part B Rev. 2010;16: Vavken P, Murray MM. The potential for primary repair of the ACL. Sports Me Arthrosc. 2011;19:44-9.

8 8 27. Vavken P, Murray MM. Treating anterior cruciate ligament tears in skeletally immature patients. Arthroscopy. 2011;27: Myklebust G, Maehlum S, Engebretsen L, Stran T, Solheim E. Registration of cruciate ligament injuries in Norwegian top level team hanball. A prospective stuy covering two seasons. Scan J Me Sci Sports. 1997;7: Myklebust G, Maehlum S, Holm I, Bahr R. A prospective cohort stuy of anterior cruciate ligament injuries in elite Norwegian team hanball. Scan J Me Sci Sports. 1998;8: Alentorn-Geli E, Myer GD, Silvers HJ, Samitier G, Romero D, Lázaro-Haro C, Cugat R. Prevention of non-contact anterior cruciate ligament injuries in soccer players. Part 2: a review of prevention programs aime to moify risk factors an to reuce injury rates. Knee Surg Sports Traumatol Arthrosc. 2009;17: Griffin LY, Agel J, Albohm MJ, Arent EA, Dick RW, Garrett WE, Garrick JG, Hewett TE, Huston L, Irelan ML, Johnson RJ, Kibler WB, Lephart S, Lewis JL, Linenfel TN, Manelbaum BR, Marchak P, Teitz CC, Wojtys EM. Noncontact anterior cruciate ligament injuries: risk factors an prevention strategies. J Am Aca Orthop Surg. 2000;8: Hewett TE, Myer GD, For KR. Reucing knee an anterior cruciate ligament injuries among female athletes: a systematic review of neuromuscular training interventions. J Knee Surg. 2005;18: Liberati A, Altman DG, Tetzlaff J, Mulrow C, Gøtzsche PC, Ioanniis JP, Clarke M, Devereaux PJ, Kleijnen J, Moher D. The PRISMA statement for reporting systematic reviews an meta-analyses of stuies that evaluate health care interventions: explanation an elaboration. PLoS Me. 2009;6:e Moher D, Cook DJ, Eastwoo S, Olkin I, Rennie D, Stroup DF. Improving the quality of reports of meta-analyses of ranomise controlle trials: the QUOROM statement. Quality of Reporting of Meta-analyses. Lancet. 1999;354: Jaa AR, Moore RA, Carroll D, Jenkinson C, Reynols DJ, Gavaghan DJ, McQuay HJ. Assessing the quality of reports of ranomize clinical trials: is blining necessary? Control Clin Trial. 1996;17: Vavken P, Dorotka R. The prevalence an effect of publication bias in orthopaeic meta-analyses. J Orthop Sci Mar 2 [Epub ahea of print]. 37. Okike K, Kocher MS, Mehlman CT, Heckman JD, Bhanari M. Publication bias in orthopaeic research: an analysis of scientific factors associate with publication in the Journal of Bone an Joint Surgery (American Volume). J Bone Joint Surg Am. 2008;90: Johnson RT, Dickersin K. Publication bias against negative results from clinical trials: three of the seven ealy sins. Nat Clin Pract Neurol. 2007;3: Egger M, Davey Smith G, Schneier M, Miner C. Bias in meta-analysis etecte by a simple, graphical test. BMJ. 1997;315: Higgins JP, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. BMJ. 2003;327: Vavken P, Culen G, Dorotka R. Management of confouning in controlle orthopaeic trials: a cross-sectional stuy. Clin Orthop Relat Res. 2008;466: Thompson SG. Why an how sources of heterogeneity shoul be investigate. In: Egger M, Smith GD, Altman DG, eitors. Systematic reviews in health care: meta-analysis in context. 2n e. Lonon: BMJ Publishing Group; p Bubbar VK, Kreer HJ. The intention-to-treat principle: a primer for the orthopaeic surgeon. J Bone Joint Surg Am. 2006;88: DerSimonian R, Lair N. Meta-analysis in clinical trials. Control Clin Trials. 1986;7: Caraffa A, Cerulli G, Projetti M, Aisa G, Rizzo A. Prevention of anterior cruciate ligament injuries in soccer. A prospective controlle stuy of proprioceptive training. Knee Surg Sports Traumatol Arthrosc. 1996;4: Gilchrist J, Manelbaum BR, Melancon H, Ryan GW, Silvers HJ, Griffin LY, Watanabe DS, Dick RW, Dvorak J. A ranomize controlle trial to prevent noncontact anterior cruciate ligament injury in female collegiate soccer players. Am J Sports Me. 2008;36: Heit RS Jr, Sweeterman LM, Carlonas RL, Traub JA, Tekulve FX. Avoiance of soccer injuries with preseason conitioning. Am J Sports Me. 2000;28: Hewett TE, Linenfel TN, Riccobene JV, Noyes FR. The effect of neuromuscular training on the incience of knee injury in female athletes. A prospective stuy. Am J Sports Me. 1999;27: Manelbaum BR, Silvers HJ, Watanabe DS, Knarr JF, Thomas SD, Griffin LY, Kirkenall DT, Garrett W Jr. Effectiveness of a neuromuscular an proprioceptive training program in preventing anterior cruciate ligament injuries in female athletes: 2-year follow-up. Am J Sports Me. 2005;33: Petersen W, Braun C, Bock W, Schmit K, Weimann A, Drescher W, Eiling E, Stange R, Fuchs T, Heerich J, Zantop T. A controlle prospective case control stuy of a prevention training program in female team hanball players: the German experience. Arch Orthop Trauma Surg. 2005;125: Petersen W, Zantop T, Steensen M, Hypa A, Wessolowski T, Hassenpflug J. [Prevention of lower extremity injuries in hanball: initial results of the hanball injuries prevention programme]. Sportverletz Sportschaen. 2002;16: German. 52. Pfeiffer RP, Shea KG, Roberts D, Granstran S, Bon L. Lack of effect of a knee ligament injury prevention program on the incience of noncontact anterior cruciate ligament injury. J Bone Joint Surg Am. 2006;88: Söerman K, Werner S, Pietilä T, Engström B, Alfreson H. Balance boar training: prevention of traumatic injuries of the lower extremities in female soccer players? A prospective ranomize intervention stuy. Knee Surg Sports Traumatol Arthrosc. 2000;8: