SURGICAL VERSUS CONSERVATIVE INTERVENTIONS FOR ANTERIOR CRUCIATE LIGAMENT RUPTURES IN ADULTS

Linko E, Harilainen A, Malmivaara A, Seitsalo S

This review should be cited as: Linko E, Harilainen A, Malmivaara A, Seitsalo S. Surgical versus conservative interventions for anterior cruciate ligament ruptures in adults (Cochrane Review). In:

The Cochrane Library, Issue 4, 2007. Oxford: Update Software.A substantive amendment to this systematic review was last made on  07 February 2005. Cochrane reviews are regularly checked and updated if necessary.

A B S T R A C T

Background Anterior cruciate ligament rupture is a common knee injury. Surgical treatment, usually involving reconstruction of the ligament, is widely used especially in active individuals.Objective Evaluation of the effect of surgical treatment compared with conservative treatment of anterior cruciate ligament (ACL) rupture.Search strategy We searched the Cochrane Musculoskeletal Injuries Group Specialised Register (January 2005), the Cochrane Central Register of Controlled Trials (The Cochrane Library Issue 1, 2005), MEDLINE (1966 to January Week 3 2005), EMBASE (1988 to 2005 Week 05), MEDIC (1978 to January 1999), Current Contents (9.2.1998 to 1.2.1999), BIOSIS (1970 to December 1998), reference lists of articles and consulted trialists and experts.Selection criteria All randomised and quasi-randomised trials that compared surgical with conservative treatment of ACL rupture in adults.Data collection and analysis Two authors independently performed study selection, data extraction and quality assessment.Main results Two poor quality randomised trials conducted in the early 1980s were included in the review. The two trials differed considerably and no data pooling was done for the few shared outcome measures.Reviewers' conclusions There is insufficient evidence from randomised trials to determine whether surgery or conservative management was best for ACL injury in the 1980s, and no evidence to inform current practice. Good quality randomised trials are required to remedy this situation.

S Y N O P S I S

Surgery versus conservative treatment to repair knee anterior cruciate ligament tears (ruptures) in adultsAnterior cruciate ligament (ACL) rupture is a common knee injury that often results in an unstable knee. Surgical treatment, usually involving reconstruction of the ligament, is widely used especially in active individuals. This review identified two randomised trials (324 people) testing outdated treatment methods. Surgical repair or augmented repair of the ligament gave better knee stability during daily activities and strenuous activities than non-surgical treatment. Functional recovery was similar with both treatments. Fewer people, however, went on to have a knee reconstruction or other knee additional surgery within the 13 to 55 months after surgical repair than following conservative treatment. No long-term information on development of arthritis was available. Overall, these studies presented some evidence that conservative treatment (immobilization, knee braces, physiotherapy) often gave acceptable recovery results. We found no randomised trials comparing modern methods of surgery with current methods of non-surgical treatment for these injuries.

B A C K G R O U N D

Each year around one in a hundred people suffer a serious knee injury (Kannus 1989). Anterior cruciate ligament (ACL) rupture is the most common cause of acute, traumatic hemarthrosis (bleeding) of the

knee, an isolated rupture being found in about 38% of patients with acute knee hemarthrosis (Harilainen 1990).The ACL is the primary passive restraint to forward movement of the tibia relative to the femur (Seitz 1996). A rupture of the ACL compromises the stability of the knee in active individuals resulting in chronic instability, recurrent injury, and associated intra-articular pathology (O'Neill 1996). ACL rupture may also lead to changes in lifestyle and disability. Both conservative and operative interventions are used in the treatment of ACL injuries. While operative treatment of ACL lesions in athletes is widely performed, conservative treatment has been considered to have a good outcome in the general population (Casteleyn 1996).Hey Groves 1917 presented the first report of a procedure to reconstruct torn ACL by using a tethered fascia lata graft through anatomically placed drill holes in the femur and tibia. Surgical techniques include simple repair by suturing or suturing with augmentation, intra- or extra-articular reconstruction, open or arthroscopic operations, use of different kinds of grafts as well as various graft fixation methods. The materials used for ACL replacement can be broadly classified as follows:1) autogenous (from the patient) or allogenous (for example cadaveric) grafts (for example fascia lata, hamstring tendons and patellar ligament);2) heterogenous (from different species) grafts (for example lyophilised fascia lata and dura mater);3) artificial ligaments (for example carbon fibres and Dacron) (Franke 1985; Shino 1993).Despite the growing and widespread use of ACL reconstruction and the many studies of different operative treatment options for the reconstruction of ruptured ACL, the benefit (both efficacy and cost-effectiveness) of ACL reconstruction surgery still needs to be established, as do the relative benefits of the different methods used. This review looks at the evidence from randomised trials of the surgical versus conservative treatment of ACL rupture.

O B J E C T I V E S

We aimed to examine the evidence from randomised and quasi-randomised controlled trials of relative effectiveness of surgical treatment versus conservative treatment for ruptured ACL in adults.

C R I T E R I A F O R C O N S I D E R I N G S T U D I E S F O R T H I S R E V I E W

Types of studies

Any randomised or quasi-randomised (for example alternation and dates of birth) clinical trials (RCTs) which compared surgical and conservative treatment of ACL lesions were considered for inclusion in the review.

Types of participants

Adults who have suffered an acute injury of the ACL or the long term effects of this. Trials that focused specifically on adolescents were omitted. We planned, where possible, distinctions between age (under 30, over 30 years), gender, sports participation (athletes, not athletes), activity levels, the duration of the disorder (acute: under one week, one to six weeks; chronic: six weeks to six months, over six months), previous knee injury (absence, presence), other knee injuries (absence, presence), and the type of rupture (partial, total).

Types of intervention

Any operative (e.g. repair, reconstruction) or conservative (e.g. immobilisation, knee braces, physiotherapy) intervention used for the management of ACL injuries.

Types of outcome measures

We anticipated that it would be difficult to compare the outcome of different treatment methods because there are so many different scoring systems available (Hefti 1993; Tegner 1985).The following outcomes were sought:return to ordinary daily activities or sports activity;pain intensity (visual analogue scale, ordinal scale);ability to work (e.g. sickness absences/ return to work/ number off days of work and subjective working ability);subjective instability (giving way);

objective measurement of knee stability (for example KT 1000);knee scores (e.g. IKDC, Tegner, Lysholm);objective measurement of muscle strength (isokinetic muscle torque);global status (e.g. overall improvement, objective and subjective evaluation);health care consumption and costs;satisfaction with treatment;complications (including mortality, post-traumatic arthritis, infection, thrombosis).

S E A R C H S T R A T E G Y F O R I D E N T I F I C A T I O N O F S T U D I E S

See: Cochrane Bone, Joint and Muscle Trauma Group search strategySee: methods used in reviews.We searched the Cochrane Musculoskeletal Injuries Group Specialised Register (January 2005), the Cochrane Central Register of Controlled Trials (The Cochrane Library Issue 1, 2005), MEDLINE (1966 to January Week 3 2005), EMBASE (1988 to 2005 Week 05), MEDIC (1978 to January 1999), Current Contents (9.2.1998 to 1.2.1999), BIOSIS PREVIEWS (1970 to December 1998), reference lists of articles and contacted experts in the field. We would have considered studies written in English, Finnish, Swedish, Danish, Norwegian or German for inclusion in the review.In MEDLINE (OVID WEB) a subject-specific search strategy was combined with the three stages of the optimal trial search strategy (Clarke 2001)(see ). In 2005 the search was re-run across MEDLINE and EMBASE in OVID WEB using a revised strategy combining subject-specific terms with the revised trial search strategy (Alderson 2004a) (see ).

M E T H O D S O F T H E R E V I E W

The methods for the systematic review were based on recommendations contemporary to the development of the protocol for the review (Mulrow 1997; NHS CRD 1996; Van Tulder 1997). The following methods were pertinent to the original scope of the review, which included comparisons of operative treatments (Linko 1999).Study inclusionTwo authors independently selected references from citations in MEDLINE, other databases and reference lists for retrieval of full articles. Where there was disagreement or doubt, the full article was retrieved. Two authors independently assessed the full study report to see if it met the review inclusion criteria. A third review author was consulted in cases of unsolvable disagreement. If needed, the trial authors were contacted for more information.Assessment of methodological qualityThe same two authors who extracted the data, independently assessed the methodological quality of the trials. A consensus method was used to solve disagreements and a third author consulted where disagreement persisted.We used a modified version of the criteria list used by Koes 1991. The criteria list and notes on scoring individual items are presented in . In our scoring scheme, each criterion was scored as positive (+), negative (-) or unclear (?). We rated a trial to be of high methodological quality if at least five out of 10 items were scored positive. If less than five items were scored positive, the trial was rated low quality.Clinical relevance of the included studies In conjunction with data extraction and methodological quality assessment, we assessed the clinical quality of included trials. We based this on five issues, modified from Shekelle 1994, from the point of view of the patient and primary care provider shown in .Analysis We planned, where clinically appropriate and where data were available, to perform meta-analysis presenting a treatment effect estimate with 95% confidence intervals. We intended to use a random-effects model where there was statistically significant heterogeneity as determined by the chi squared statistic (P < 0.10) in conjunction with the I squared statistic (Alderson 2004b). In such cases, we intended to examine the potential sources of variation between trials such as the study population (e.g. age, gender, type of disorder, duration of disorder) and content of intervention. However, regardless of any evidence of statistical heterogeneity, we planned to explore the influence of specific differences between pooled trials. We stipulated in the protocol the various sub-group analyses and sensitivity analyses we would have performed if the data had allowed. These are listed in .We stated in the protocol (Linko 1999) that if the included studies were clinically or statistically heterogeneous, which disallowed pooling, we would only conduct a qualitative review using a best evidence synthesis that assessed the strength of the evidence using a modified classification of the US Agency for Health Care Policy and Research guidelines on acute low back problems in adults (Bigos

1994). We have, however, presented some of the results of individual trials graphically, primarily to check the claims within the trial reports. In these case, where data from individual trials are presented to graphically, relative risks and 95% confidence intervals are presented for dichotomous outcomes, and mean differences and 95% confidence intervals for continuous outcomes.Rating of scientific evidenceWe planned to use a rating system with four levels of scientific evidence to draw an overall conclusion regarding the efficacy of interventions. We used an adaptation of the system used in the US Clinical Practice Guideline for Acute Low Back Problems in Adults (Van Tulder 1997). This is shown below.Level A: Strong research-based evidence: provided by generally consistent findings in multiple (more than one) high quality RCTs.Level B: Moderate research-based evidence: provided by generally consistent findings in one high quality RCT and one or more low quality RCTs, or generally consistent findings in multiple low quality RCTs.Level C: Limited research based evidence: provided by one RCT (either high or low quality) or inconsistent or contradictory evidence: findings in multiple RCTs.Level D: No research-based evidence: no RCTs.

D E S C R I P T I O N O F S T U D I E S

Of nine trials selected for inclusion in this review, we included two and excluded seven for reasons stated in the 'Characteristics of excluded studies' table.Both included trials (Andersson 1991; Sandberg 1987) were conducted in Sweden in the early 1980s and compared surgical versus conservative treatment of acute ACL rupture. There were five reports, comprising three full papers, a thesis and a conference abstract, for Andersson 1991 and three reports, comprising one full paper and two abstracts, for Sandberg 1987: all reports were in English. The numbers of trial participants, all with complete ACL rupture, and length of follow up varied over time in the reports of Andersson 1991 cumulating in a total of 167 people, aged between 13 and 59 years, 156 of whom were re-examined at 41 to 80 months follow up. Sandberg 1987 recruited 200 people with acute ACL and/or MCL (medial collateral ligament) injuries, aged between 15 and 61 years. Follow up at 11 to 18 months was available for 199 participants, and the outcome of the first 100 participants in the trial was re-examined at an average of 33 months (range 24 to 40 months). The one-year follow up for the first 100 patients was reported in a conference abstract.In Andersson 1991, all participants had arthroscopy within one week after their injury. The 75 operatively treated trial participants were divided into two groups. In 25 participants, the ACL repair was done with six to eight sutures placed in the tibial portion of the ACL remnant, these were passed over the top and through a drilled hole through the lateral condyle of the femur at the site of the attachment of the ACL and then tied over the lateral condyle. In the other 50 participants, ACL repair was combined with an augmentation with a distally based strip of the iliotibial band. In the 92 conservatively treated participants, 30 people had associated lesions of other knee structures that were treated operatively (other ligament than ACL or meniscal repair). All people treated by ligament repair wore a long leg cast for approximately five weeks without weight bearing. In the conservative group, people without any associated lesions that needed operative treatment, were allowed normal range of motion immediately after the arthroscopy. People who had repair of the ACL or who suffered from a symptomatic instability were provided with a derotation brace.Of the 200 participants with ACL and/or MCL injuries in Sandberg 1987, 102 had only an ACL injury (55 versus 47), 55 had both ACL and MCL injuries (26 versus 29), and 44 had only an MCL injury (19 versus 24). The maximum time from injury to the participant's entry into the study was two weeks. Fifty of the 100 participants who were treated conservatively underwent arthroscopy. Surgical treatment consisted of primary ligament suture without augmentation. Both conservative treatment and postoperative treatment consisted of plaster cast with full weight bearing for six weeks.Further details of these two studies are given in the 'Characteristics of included studies' table.

M E T H O D O L O G I C A L Q U A L I T Y

The results of the methodological quality assessment of the two trials are shown in . Out of a maximum total score of 10, Andersson 1991 scored 1, and Sandberg 1987 scored 3. Thus, both trials were rated as poor quality trials (see 'Methods').Andersson 1991 was a quasi-randomised trial where patients were allocated treatment according to their year of birth. Sandberg 1987 randomised treatment allocation by casting lots.As described in the above section, the number of trial participants reported in various publications of Andersson 1991 differed, and increased over time. There were also other important differences. For example, in Odensten 1985 all participants had midsubstance rupture of the ACL and the operation group was divided at random to simple repair and augmented repair. But in Andersson 1989, simple repair was performed in participants who had a proximal rupture of the ACL and augmented repair in the remaining participants. Andersson provided some clarification of the conduct of the study in his thesis,

and described some of the potential sources of bias. These include: the higher proportion of active people in the augmented repair group; the smaller proportion of women in the surgical treatment group (20% versus 36%); the incorrect assignment of six people to augmented repair; the assignment of 16 people to non-surgical treatment (9 because of injury older than 2 weeks). In addition, the indications for non-augmented repair became more restricted over the recruitment period of the trial: this also had an effect on the randomisation process.The clinical relevance, judged using the criteria in , was better than the methodological quality of the two trials (see). The treatment protocols were evaluated with a view to the options that were available during the study periods in the 1980s.

R E S U L T S

Many of the outcome measures listed in 'Types of outcome measures' were not reported in the two included trials (Andersson 1991; Sandberg 1987). Generally, Andersson 1991 presented the results split by three groups: augmented repair, repair only and non-surgically treated participants. Where appropriate and possible, we have combined the data from the two surgical groups. Sandberg 1987 included people with ACL injuries alone, with both ACL and MCL injuries, and with MCL injuries alone. Where possible, we present the results of people with ACL injuries for this trial. We did not pool the results from the two trials because of their differences in study protocols, interventions and often the lack of comparable data.Return to ordinary daily activities or sports activity (Graphs 01.01 and 01.02)In Sandberg 1987, the time required for return to sports activities was significantly longer in surgically treated participants with only ACL injury (mean difference (MD) 4.00 weeks, 95% confidence interval (CI) 0.44 to 7.56 weeks). Neither trial found a significant difference in the numbers of people failing to return to their previous level of sports activity or at all: Andersson 1991 (42/68 versus 42/71; relative risk (RR) 1.04, 95% CI 0.80 to 1.37); Sandberg 1987 (all ACL injuries: 32/76 versus 26/81; RR 1.31, 95% CI 0.87 to 1.98).Pain intensity (visual analogue scale, ordinal scale)There were no data available for this outcome from the included trials.Ability to work (e.g. sickness absences/ return to work/ number off days of work and subjective working ability) (Graph 01.03)In Sandberg 1987, the duration of sick leave ("disability") was significantly longer in those treated surgically for their ACL injury (MD 6.00 weeks, 95% CI 3.35 to 8.65 weeks). One person with an ACL rupture in each treatment groups had changed his or her job due to their injury.Subjective instability (giving way) (Graph 01.04)Both trials reported significantly fewer surgically treated participants with positive pivot shift at follow up (Andersson 1991: RR 0.44, 95% CI 0.29 to 0.67; Sandberg 1987: RR 0.46, 95% CI 0.31 to 0.68).Of those with positive pivot shifts in Andersson 1991, seven participants in the surgical groups had instability problems compared with 35 participants in the conservative group (RR 0.21, 95% CI 0.10 to 0.44); one versus seven of these had instability during daily activities (RR 0.15, 95% CI 0.02 to 1.18).Of those with ACL injury in Sandberg 1987, three in the surgical group versus 13 in the conservative group had episodes of giving way during strenuous activity (RR 0.22, 95% CI 0.06 to 0.73).In Andersson 1991 instability symptoms were reported by 14 participants in the surgical groups and by 44 participants in the conservative group (RR 0.33, 95% CI 0.20 to 0.55). Two participants in the surgical groups and ten in the conservative group had instability symptoms during daily activities such as walking (RR 0.21, 95% CI 0.05 to 0.92).Objective measurement of knee stability (for example, the KT 1000) (Graph 01.05)In Andersson 1991, sagittal laxity was measured with a laxity tester. With a 180 N load, the total laxity was 10 ± 3 mm in the augmented-repair group, 12 ± 4 mm in the repair-only group and 14 ± 4 mm in the non-surgically treated group. The combined results of the two surgical groups showed significantly lower laxity than that of the non-surgical group (MD -3.35 mm, 95% CI -4.57 to -2.13 mm). All groups had an abnormal laxity of the involved legs compared with the uninvolved legs.In Sandberg 1987, a caliper was used to record the anteroposterior motion of the tibial tuberosity. Data presented in a conference abstract (Bauer 1988) showed less laxity in those people with surgically treated ACL injury at one-year follow up (see Graph 01.05).Knee scores (e.g. IKDC, Tegner, Lysholm) (Graph 01.06)Both trials reported Lysholm scores, where 100 points is the maximum score and 85 or more is considered to be a good result. In Andersson 1991, the Lysholm score for the people treated surgically was statistically significantly higher compared with those treated non-surgically (MD 5.43, 95% CI 2.14 to 8.54). There was no difference in the mean Lysholm scores between the two treatment groups for people with ACL injuries in Sandberg 1987 (MD 1.00, 95% CI -2.92 to 4.92).Andersson 1991 recorded but did not report on the results of the Tegner activity score, used to determine the level of activity.

Objective measurement of muscle strength (isokinetic muscle torque)The Cybex-II dynamometer was used to measure muscle force and endurance in both studies. Andersson 1991 reported that quadriceps muscle strength was similar among the three groups (two surgical and one non-surgical) and observed that hamstring muscle strength was higher for men treated with repair and augmentation than for those in the other groups (reported P < 0.05). Sandberg 1987 reported that the isometric force of the hamstrings did not differ between the two treatment groups at any time. However, for people with ACL injuries, the isokinetic strength of the quadriceps was significantly greater (reported P < 0.05) after one year in those treated conservatively compared with those treated surgically.Global status (for example overall improvement, objective and subjective evaluation)There were no data available for this outcome from the included trials.Health care consumption and costsThere were no data available for this outcome from the included trials.Satisfaction with treatmentSandberg 1987 reported that 57 out of 84 trial participants with treated ACL injuries followed up for two years or more regarded their results as excellent at an average of 13 months follow up; however, after 33 months follow up the number decreased to 35. Sandberg 1987 reported there was no difference between the two treatment groups.Complications (including mortality, post-traumatic arthritis, infection, thrombosis)In Andersson 1991, 19 people (three in the repair-only group versus 16 in the non-surgical group) with severe symptomatic instabilities had an ACL reconstruction during the follow-up period. Three postoperative complications were reported in an earlier trial report (Andersson 1989). One person in the repair-only group had septic arthritis which responded to antibiotic treatment, and another in the same group with limited motion postoperatively had manipulation under anaesthesia 15 and 28 months after operation but at the latest follow up the function of the knee was still poor. One person in the conservatively treated group had venous thrombosis postoperatively.There were no complications in participants of the non-operative group of Sandberg 1987. In the operative group, 75 patients had an uncomfortable, localised loss of sensation and nine required postoperative mobilisation of the joint under general anaesthesia. One person had septic arthritis, which resolved after aspiration and antibiotic therapy. There was one hematoma that had to be drained. Two people had phlebography-verified deep vein thrombosis. In those people with treated ACL injuries who were followed for two years or more, nine (six of them in the non-operative group) had a reconstruction of the ACL. In addition, four meniscectomies had been performed in the operative group and 12 in the non-operative group.

D I S C U S S I O N

ACL rupture is a common knee injury. In recent years ACL surgery has become very popular especially among surgeons treating sports injuries and it is becoming common practice for young and active people with ACL rupture to be treated operatively.We found thousands of articles about ACL injuries but, among these, there were only two randomised or quasi-randomised studies that addressed the basic question as to whether ACL ruptures should be operated on or not. The methodological quality of both these trials was poor. In particular, there was clear indication of bias at randomisation in Andersson 1991. The overall numbers of trial participants were acceptable in both trials but the inclusion and treatment of a variety of other knee injuries in Andersson 1991 and of MCL ruptures in Sandberg 1987 both reduce the effective size of these trials and their ability to answer the basic question regarding ACL treatment. Follow up averaged 55 months in Andersson 1991 but only 13 months for the whole study population in Sandberg 1987, although the first 100 people recruited in the trial were followed up for longer (33 months). Even for shared outcome measures, we did not pool data for these two very different trials with dissimilar trial populations, interventions and study methods.The two methods of surgical treatment used in Andersson 1991 adds some complexity to the interpretation of the results of this trial. As our focus here is on the comparison of operative versus non-operative treatment, we have usually combined the results of both surgical groups. As reported in 'Methodological quality of included studies', the allocation of those assigned to surgical treatment of their ACL rupture by augmented repair or repair only was reported as random in one trial report but according to the location of the ACL injury in another. Overall, the evidence from Andersson 1991 showed no difference in the return to sports activities between people treated surgically and those treated conservatively. Measures of knee stability and the Lysholm knee scores were higher in surgically-treated participants. By the end of follow up, three people treated with repair only and 16 treated conservatively had had ACL reconstruction.In Sandberg 1987, conservatively treated participants recovered from their injury more rapidly but at the last follow up the functional outcome was similar in both treatment groups. A large proportion of participants experienced some temporary discomfort after surgery and there were some more serious postoperative complications. There was less knee instability in surgically-treated participants and a tendency to fewer subsequent operations in the longer term. In their conclusion, Sandberg 1987 suggested that "a possible candidate for ACL surgery could be a patient with high physical demands and

with a positive pivot shift. However, it has not been possible in the present study to prove that such a patient will do better with primary suture of the torn anterior cruciate ligament."Neither trial reported radiographical assessment of the knee joint nor reported on post-traumatic arthritis. Thus the question on whether the risk of post-traumatic arthritis could be reduced by the operation of an ACL injury remains unanswered here.Both studies included in this review were conducted in the early 1980s and the treatment of ACL injuries is usually quite different nowadays. Plaster casts are no longer being used in conservative treatment and operative treatment usually entails reconstruction of the ACL with a bone-patellar tendon-bone or hamstring tendon graft with fast postoperative mobilisation. Unfortunately, we found no randomised studies that compared conservative treatment with operative treatment used nowadays. However, it is important to note that in both studies conservative treatment often had a good functional result and so, even today, it can be a good alternative in some patients with acute ACL injury.

R E V I E W E R S ' C O N C L U S I O N S

Implications for practice

There is insufficient evidence from randomised trials comparing operative versus conservative treatment of ACL ruptures to inform current practice. While there is some limited evidence from two outdated, heterogeneous and poor quality trials that surgery improves knee stability, there was insufficient evidence to determine the relative effects of the surgical versus non-surgical treatment interventions performed in the early 1980s. However, there was some evidence that conservative treatment of acute ACL injuries can result in a satisfactory outcome.

Implications for research

ACL rupture is a common knee injury and operative treatment is widely used but we have not found any randomised trials comparing operative versus conservative treatment using currently used methods of treatment. Given this, there is a need for good quality, and well reported, randomised trials evaluating the effectiveness and cost-effectiveness of current methods of surgical treatment versus non-surgical treatment. The follow up of such trials should be at least 10 years so that the long term effects including degenerative changes can be established.

A C K N O W L E D G E M E N T S

We thank Lesley Gillespie, Leeann Morton, Merja Jauhiainen and Kaija Karjalainen for their help. We would also like to acknowledge the substantial contribution made by Helen Handoll who commented on, and edited, earlier drafts of this review and provided help with methodology. We are also grateful to the following for their helpful comments: A/Prof P Herbison (editor), Dr J Wale (editor), Prof WJ Gillespie (editor) and A/Prof R Pitto (consultant orthopaedic surgeon, Auckland, New Zealand).

P O T E N T I A L C O N F L I C T O F I N T E R E S T

None known.

N O T E S

The title of the protocol for this review was: "Operative treatment for anterior cruciate ligament ruptures in adults." This was amended to reflect the reduced scope of the review to a comparison between operative and conservative treatment.

T A B L E S

Characteristics of included studies

Study  Andersson 1991 

Methods  Allocation to treatment groups: by year of birth.Overall methodological assessment score = 1 

Participants  167 with an acute and complete ACL rupture (75 in surgical groups, 92 in conservative group).Recruitment period: 1980-1983Location: Linkoping, SwedenAge: mean 26 years (range 13-59 years).Sex: 48 female, 119 male.All patients had arthroscopy within one week of injury.Follow up: 156 people 

Interventions 

(1) Surgical treatment of ACL injury: simple repair of ACL or augmented repair of ACL with a strip of the iliotibial band.(2) Conservative treatment of ACL injury.Surgical repair of all major injuries to knee structures other than ACL for all 3 groups.(1) In surgical group: 25 participants with ACL repair only: repair with sutured passed over the top and through a drilled hole in the lateral femoral condyle. In the other 50, ACL repair with augmentation of a distally based strip of the iliotibial band. Long leg cast for approximately 5 weeks.(2) In conservative group: when associated lesions treated operatively, 5 weeks long leg cast, otherwise patients were instructed to regain normal range of motion. 

Outcomes  Follow up: mean 55 months (range 41-80 months)Return to sports activityKnee instabilityLaxity testing (Stryker)LysholmTegnerMuscle strength (Cybex II)Complications including subsequent operationsAdditional outcomes: one-leg jump, figure-of-eight running test. 

Notes 

Highly active patients were more frequent in augmented group. Six people were incorrectly assigned to have an augmented repair; and 16 to non-surgical treatment (9 because of injury older than 2 weeks; others, possibly because of age and minor instability). There were fewer women in the surgical group (20% versus 36%). Indications of non-augmented repair became more restricted with time.In a previous report of the same study (Odensten 1985) all patients were reported to have midsubstance rupture of the ACL and the surgical group was randomly divided into two different surgical groups. 

Allocation concealment  C - Inadequate 

Study  Sandberg 1987 

Methods  Allocation to treatment groups: by casting lots.Overall methodological assessment score = 3 

Participants  200 with acute ACL and/or MCL injuries (100 in each group). (ACL only: 102; ACL + MCL: 55; MCL only: 44)Recruitment period: 1982-1984Location: Malmo, SwedenAge: mean 29 years (range 15-61 years).Sex: 87 female, 113 male.Injury to entry into study was less than 2 weeks.Follow up: 199 people 

Interventions  (1) Suture of the ligament or ligaments and plaster cast with full weight bearing for six weeks.(2) Immobilisation in a plaster cast with full weight bearing for six weeks. 

Outcomes 

Follow up: mean 13 months (range 11-18 months); also 33 months (for first 100 participants)Return to sportsDuration of sick leaveKnee instabilityLysholmMuscle strength (Cybex-II)SatisfactionComplications including subsequent operationsAdditional outcomes: range of motion, Lachman, anterior drawer, pivot shift, valgus stability, level of activity, swelling, circumference of the thigh. 

Notes  There were no differences between the groups with regard to stability tested under general anaesthesia.In the second half of the study some participants were also examined with arthroscopy. 

Allocation concealment  B - Unclear 

Characteristics of excluded studies

Study Reason for exclusion

Andersson 1992

A sub-group analysis of a previous study of ACL injuries focusing on combined ACL and MCL injuries.

Fink 1993 A non-randomised study comparing sports activity after ACL rupture treated operatively or non operatively.

Fink 1994a A non-randomised study comparing operative and conservative treatment of ACL rupture.

Fink 1994bA non-randomised study comparing neuromuscular changes in the knee stabilizing muscles after operative or conservative treatment of rupture of the anterior cruciate ligament.

Fink 1996 A non-randomised study about long-term outcome of conservative or surgical therapy of anterior cruciate ligament rupture.

Halinen 1998 A randomised study of grade 3 MCL and ACL injuries of the knee. Excluded because it doesn't compare different treatment methods for ACL injuries.

Kruger-Franke 1997

A study comparing acute versus delayed surgery in ACL rupture. Excluded because it is not a randomised study.

A D D I T I O N A L T A B L E S

Original MEDLINE search strategy MEDLINE

1 randomized controlled trial.pt.2 controlled clinical trial.pt.3 randomized controlled trials.sh.4 random allocation.sh.5 double blind method.sh.6 single blind method.sh.7 or/1-68 animal.sh. not human.sh9 7 not 810 clinical trial.pt.

11 exp clinical trials12 (clin$ adj25 trial$).ti,ab.13 ((singl$ or doubl$ or trebl$ or tribl$) adj25 (blind$ or mask$)).ti,ab.14 placebos.sh.15 placebo$.ti,ab16 random$.ti,ab.17 research design.sh.18 or/10-1719 18 not 820 19 not 9

21 comparative study.sh.22 exp evaluation studies23 follow up studies.sh.24 prospective studies.sh.25 (control$ or prospectiv$ or volunteer$).ti,ab.26 or/21-2527 26 not 828 26 not (9 or 20)29 9 or 20 or 2830 anterior cruciate ligament.sh.

31 knee injuries/32 exp joint instability/33 exp joint instability/rh,th,su34 knee instabil$.tw.35 knee joint.sh.36 knee.tw.37 exp ligaments articular/38 exp ligaments/39 knee reconstruc$.tw.40 31 or 34 or 35 or 36 or 39

41 33 or 37 or 3842 30 or 40 or 4143 (acl and knee).tw.44 anterior cruciate ligament$.tw.

45 or/42-4446 29 and 4547 surgery.tw.48 rehabilitation.tw.49 treatment$.tw.50 therap$3.tw.

51 46 and 4752 46 and 4853 46 and 4954 46 and 5055 or/51-5456 acl.tw.57 cruciate.tw.58 30 or 43 or 44 or 56 or 5759 55 and 5860 59

Revised MEDLINE search strategy MEDLINE (OVID WEB)

1. Anterior Cruciate Ligament/2. Joint Instability/3. Ligaments, Articular/4. Knee Injuries/5. Knee Joint/6. or/2-57. (anterior cruciate$1 or acl).tw.8. and/6-79. or/1,810. su.fs.11. (surg$ or operat$ or reconstruct$ or repair$ or graft$ or screw$).tw.12. or/10-1113. (non-surg$ or nonsurg$ or non-operat$ or nonoperat$ or conserv$).tw.14. and/12-1315. and/9,14

16. randomized controlled trial.pt.17. controlled clinical trial.pt.18. Randomized Controlled Trials/19. Random Allocation/20. Double Blind Method/21. Single Blind Method/22. or/16-2123. Animals/ not Human/24. 22 not 2325. clinical trial.pt.26. exp Clinical Trials/27. (clinic$ adj25 trial$).tw.28. ((singl$ or doubl$ or trebl$ or tripl$) adj25 (blind$ or mask$)).tw.29. Placebos/30. placebo$.tw.31. random$.tw.32. Research Design/33. or/25-3234. 33 not 2335. 34 not 24

36. Comparative Study/37. exp Evaluation Studies/38. Follow Up Studies/39. Prospective Studies/40. (control$ or prospectiv$ or volunteer$).tw.41. or/36-4042. 41 not 2343. 42 not (24 or 35)44. and/15,2445. and/15,3546. and/15,4347. or/44-46

Criteria list for the methodological assessment of the included studies

Item Notes

(1)(a) Method of randomisation(b) Concealment of treatment allocation

Item (a) is scored positive if a random (unpredictable) assignment sequence is used. Examples of adequate concealment procedures are some form of centralised randomisation scheme, numbered or coded containers, an on-site computer system providing allocations in a locked, unreadable file that can be assessed only after inputting the characteristics of an enrolled participant, and sequentially numbered, sealed, opaque envelopes. Clearly inadequate are procedures such as alternation, or reference to case record numbers, dates of birth, day of the week or any such other approach. If the concealment of treatment allocation is described only as random or randomised it must be considered as unclear.

Item (b) is scored positive if the assignment is generated by an independent person not responsible for determining the eligibility of the patients.

(2) Withdrawals < 20%

Item is scored positive if the withdrawal rate (that is the number of randomised patients minus the number of patients at the main moment of effect measurement divided by all randomised patients and multiplied by 100) is less than 20%.

(3) Co-interventions avoided or equal

Item is scored positive if co-interventions are avoided in the design of the study or are equally divided among the intervention groups.

(4) Blinding of patients

Item is scored positive if patients are blinded regarding treatment allocation and the method of blinding is appropriate. If blinding of patients is not feasible or if the trial is pragmatic, the credibility of the treatment modalities (interventions) should be evaluated and the treatment modalities should be equally credible and acceptable to patients.

(5) Blinding of observers

Item is scored positive if the observers are blinded regarding treatment allocation and the blinding is evaluated and adequate. It is scored negative if only self-reported (questionnaire) outcomes are used and no observer outcomes.

(6) Intention-to-treat analysis

Item is scored positive if all patients are included in the analysis as part of the intervention group allocated by randomisation, irrespective of non-compliance and co-interventions. If loss to follow up is substantial (20% or more), an intention-to-treat analysis as well as an alternative analysis, which accounts for missing values (for example, a worst-case analysis) should be performed.

(7) Compliance Item is scored positive if, according to the reviewer, compliance is measured and satisfactory in all study groups.

(8) Similarity of baseline characteristics

Item is scored positive if, according to the reviewer, the study groups are comparable at baseline. (The following characteristics were considered: age, duration of complaints, recurrence status and value of outcome measure.)

(9) Blinding of therapists

Item is scored positive if therapists are blinded regarding treatment allocation and the method of blinding is appropriate. If blinding of therapists is not feasible, the credibility of the treatment modalities (interventions) should be evaluated and the treatment modalities should be equally credible to the therapists.

Aspects of clinical relevance Aspect Notes

(1) The comparability of the study patients to those seen in the usual practice

Restrictions imposed by the setting and inclusion/ exclusion criteria

(2) The intervention's suitability when applied to the usual practice The cost, time and special skills required

(3) The appropriateness of the comparison group

Comparability of the subjects with the intervention group and suitability of the intervention

(4) The clinical relevancy of measured  

outcomes for the patients and society

(5) Reporting of the side effects of the programme  

Planned subgroup and sensitivity analyses Description

Subgroup analyses:Based on the following:(1a) age, gender(1b) athlete versus non-athlete(1c) activity level(2) Duration of disorder (<1 week, 1 to 6 weeks, 6 weeks to 6 months, >6 months)(3) Participants with or without previous knee surgery(4) Participants with multiple knee injuries(5) Type of intervention (conservative, the type of operative reconstruction, open versus arthroscopic reconstruction)(6) Co-interventions - postoperative treatment. Degree of mobilisation (immobilisation, orthoses, weight bearing)(7) Follow-up time (<1 year, 1 to 2 years, 2 to 5 years, >5 years after randomisation)

Sensitivity analyses:(1) Comparison of the results of high versus low methodological quality trials.(2) Comparison of the treatment effect depending on the year of surgery.

Methodological quality assessment results Criteria Andersson 1991 Sandberg 1987

(1) (a) Randomisation method 0 1

(1) (b) Allocation concealment? 0 ?

(2) Withdrawals < 20% 1 1

(3) Co-interventions equal? 0 ?

(4) Patient blinding? 0 0

(5) Observer blinding? ? ?

(6) Intention-to-treat analysis? ? 1

(7) Compliance ? ?

(8) Baseline similarity? 0 0

(9) Therapists blinding? 0 0

Total score 1 3

Clinical relevance of the included studies Study Type of

patientsInterventio

nComparison

groupOutcome

sSide

effectsTotal score

Andersson 1991 1 1 ? 1 0 3

Sandberg 1987 1 1 ? 1 1 4

R E F E R E N C E S

References to studies included in this review Andersson 1991 {published data only} Andersson C. In: Acute anterior cruciate ligament injuries: Evaluation of surgical and non-surgical treatment [Linköping University Medical Dissertations No. 333] Linköping: Linköping University, 1991:-. Andersson C, Odensten M, Gillquist J. Knee function after surgical or nonsurgical treatment of acute rupture of the anterior cruciate ligament: a randomized study with a long-term follow-up period. Clinical Orthopaedics and Related Research 1991;(264):255-63. Andersson C, Odensten M, Good L, Gillquist J. Surgical or non-surgical treatment of acute rupture of the anterior cruciate ligament. A randomized study with long-term follow-up. Journal of Bone and Joint Surgery - American Volume 1989;71:965-74. Odensten M, Hamberg P, Nordin M, Lysholm J, Gillquist J. Surgical or conservative treatment of the acutely torn anterior cruciate ligament. A randomized study with short-term follow-up observations. Clinical Orthopaedics and Related Research 1985;(198):87-93. Odensten M, Hamberg P, Nordin M, Lysholm J, Gillquist J. Treatment of the acute torn anterior cruciate ligament: A randomized study with a short-term follow-up [abstract]. Acta Orthopaedica Scandinavica 1984;55:474-. Sandberg 1987 {published data only} Bauer M, Sandberg R. Surgical versus non-surgical treatment of knee ligament injuries [abstract]. Surgery and arthroscopy of the knee. Second European Congress of Knee Surgery and Arthroscopy; 1986 Sept 29 Oct 4; Basle (Switzerland). Berlin: Springer-Verlag, 1988:121-2.Sandberg R, Balkfors B, Edwards P, Nilsson B, Westlin N. Surgical or non-surgical treatment of ligamentous knee injuries. A randomized controlled study - early results [abstract]. Acta Orthopaedica Scandinavica 1986;57:254-. Sandberg R, Balkfors B, Nilsson B, Westlin N. Operative versus non-operative treatment of recent injuries to the ligaments of the knee. A prospective randomized study. Journal of Bone and Joint Surgery - American Volume 1987;69:1120-6. * indicates the major publication for the study References to studies excluded from this review Andersson 1992 Andersson C, Gillquist J. Treatment of acute isolated and combined ruptures of the anterior cruciate ligament. A long-term follow-up study. American Journal of Sports Medicine 1992;20:7-12. Fink 1993 Fink C, Hoser C, Benedetto KP. Sports activity after ACL rupture - Operative versus non operative treatment. Aktuelle Traumatologie 1993;23:371-5. Fink 1994a Fink C, Hoser C, Benedetto KP. Development of arthrosis after rupture of the anterior cruciate ligament. A comparison of surgical and conservative therapy. Unfallchirurg 1994;97:357-61. Fink 1994b Fink C, Hoser C, Benedetto KP, Judmaier W. (Neuro)muscular changes in the knee stabilizing muscles after rupture of the anterior cruciate ligament. Sportverletzung Sportschaden 1994;8:25-30. Fink 1996 Fink C, Hoser C, Benedetto KP, Hackl W, Gabl M. Long-term outcome of conservative or surgical therapy of anterior cruciate ligament rupture. Unfallchirurg 1996;99:964-9. Halinen 1998 Halinen J. A randomised prospective study about grade 3 MCL and ACL knee injuries. Finnish Journal of Orthopaedics and Traumatology 1998;21:332-4. Kruger-Franke 1997 Kruger-Franke M, Martin J, Trouillier HH. Acute VS. delayed surgery in ACL rupture [abstract]. Journal of Bone and Joint Surgery - British Volume 1997;79 Suppl 2:188-. Additional references Alderson 2004a Alderson P, Green S, Higgins JPT. MEDLINE highly sensitive search strategies for idenitifying reports of randomized controlled trials in MEDLINE. Cochrane Reviewers' Handbook 4.2.3 [updated November 2004]; Appendix 5b. :-. Alderson 2004b

Alderson P, Green S, Higgins JPT. Identifying and measuring heterogeneity. Cochrane Reviewers' Handbook 4.2.3 [updated November 2004]; Section 8.7.2. :-. Andersson 1989 Andersson C, Odensten M, Good L, Gillquist J. Surgical or non-surgical treatment of acute rupture of the anterior cruciate ligament. A randomized study with long-term follow-up. Journal of Bone and Joint Surgery - American Volume 1989;71:965-74. Bauer 1988 Bauer M, Sandberg R. Surgical versus non-surgical treatment of knee ligament injuries. Surgery and arthroscopy of the knee. Second European Congress of Knee Surgery and Arthroscopy; 1986 Sept 29 Oct 4; Basle (Switzerland). Berlin: Springer-Verlag, 1988:121-2.Bigos 1994 Bigos S, Bowyer O, Braen G, Brown K, Deyo R, Haldeman S. In: Acute low back problems in adults. Clinical Practice Guideline No. 14. AHCPR Publication No. 95-0642 Rockville: Agency for Health Care Policy and Research, Public Health Service, U.S. Department of Health and Human Services:-. Casteleyn 1996 Casteleyn P-P, Handelberg F. Non-operative management of anterior cruciate ligament injuries in the general population. Journal of Bone and Joint Surgery - British Volume 1996;78:446-51. Clarke 2001 Clarke M, Oxman AD. Optimal search strategy for RCTs. Cochrane Reviewers Handbook 4.1.2 [updated March 2001]; Appendix 5c. :-. Franke 1985 Franke K. Secondary reconstruction of the anterior cruciate ligament (ACL) in competitive athletes. Clinical Orthopaedics and Related Research 1985;(198):81-6. Harilainen 1990 Harilainen A. In: The diagnosis and treatment of acute traumatic hemarthrosis of the knee joint. A prospective study of 350 patients based on clinical, radiological, arthroscopic and operative findings [thesis] Helsinki: University of Helsinki, 1990:-. Hefti 1993 Hefti F, Müller W, Jakob RP, Stäubli H-U. Evaluation of knee ligament injuries with the IKDC form. Knee Surgery, Sports Traumatology, Arthroscopy 1993;1:226-34. Hey Groves 1917 Hey Groves EW. Operation for the repair of crucial ligaments. Lancet 1917;2:674-5. Kannus 1989 Kannus P, Järvinen M. Incidence of knee injuries and the need for further care: a one year prospective follow-up study. Journal of Sports Medicine 1989;29:321-5. Koes 1991 Koes BW, Bouter LM, Beckerman H, Heiden GJMG, Knipschild PG. Physiotherapy exercises and back pain: a blinded review. BMJ 1991;302:1572-6. Linko 1999 Linko E, Harilainen A, Malmivaara A, Seitsalo S. Operative versus conservative treatment for anterior cruciate ligament ruptures in adults (Protocol for a Cochrane Review). The Cochrane Library 1999:-. Mulrow 1997 Mulrow CD, Oxman AD. Cochrane Collaboration Handbook [updated September 1997]. :-. NHS CRD 1996 In: Undertaking systematic reviews of research on effectiveness. CRD's guidelines for those carrying out commissioned reviews. CRD Report No. 4 York (UK): NHS Centre for Reviews and Dissemination, University of York:-. O'Neill 1996 O'Neill DB. Arthroscopically assisted reconstruction of the anterior cruciate ligament. A prospective randomized analysis of three techniques. Journal of Bone and Joint Surgery. American Volume 1996;78:803-13. Odensten 1985

Odensten M, Hamberg P, Nordin M, Lysholm J, Gillquist J. Surgical or conservative treatment of the acutely torn anterior cruciate ligament. A randomized study with short-term follow-up observations. Clinical Orthopaedics and Related Research 1985;(198):87-93. Seitz 1996 Seitz H, Schlenz I, Müller E, Vécsei V. Anterior instability of the knee despite an intensive rehabilitation program. Clinical Orthopaedics and Related Research 1996;(328):159-64. Shekelle 1994 Shekelle PG, Andersson G, Bombardier C, Cherkin D, Deyo R, Keller R. A brief introduction to the critical reading of the clinical literature. Spine 1994;19:2028S-31S. Shino 1993 Shino K, Nakata K, Horibe S, Inoue M, Nakagawa S. Quantitative evaluation after arthroscopic anterior cruciate ligament reconstruction. Allograft versus autograft. American Journal of Sports Medicine 1993;21:609-16. Tegner 1985 Tegner Y, Lysholm J. Rating systems in the evaluation of knee ligament injuries. Clinical Orthopaedics and Related Research 1985;(198):43-9. Van Tulder 1997 Tulder MW, Assendelft WJ, Koes BW, Bouter LM. Method guidelines for systematic reviews in the Cochrane Collaboration Back Review Group for Spinal Disorders. Spine 1997;22:2323-30.

G R A P H S

Graphs and Tables

To view a graph or table, click on the outcome title of the summary table below. Surgical versus conservative treatment for ACL rupture

Outcome title No. of studies

No. of participants Statistical method Effect size

01 Time to return to sports activities (weeks)    

Weighted Mean Difference (Fixed) 95% CI 

Totals not selected 

02 Non-return or reduction in sports activity     Relative Risk (Fixed)

95% CI  Totals not selected 

03 Time to return to work (duration of disability) (weeks)    

Weighted Mean Difference (Fixed) 95% CI 

Totals not selected 

04 Subjective instability     Relative Risk (Fixed) 95% CI 

Totals not selected 

05 Sagittal laxity (stability) (mm)    

Weighted Mean Difference (Fixed) 95% CI 

Totals not selected 

06 Lysholm scores (maximum = 100)    

Weighted Mean Difference (Fixed) 95% CI 

Totals not selected 

Cover sheet

Surgical versus conservative interventions for anterior cruciate ligament ruptures in adults

Reviewer(s) Linko E, Harilainen A, Malmivaara A, Seitsalo S

Contribution of Reviewer(s)

Seppo Seitsalo and Antti Malmivaara initiated the review. The review was designed by all the authors. Eric Linko and Arsi Harilainen screened the retrieved papers against inclusion criteria and Seppo Seitsalo was consulted when a third opinion was needed. Eric Linko abstracted data from papers and wrote to some authors of papers for additional information. Data

management for the review and entering data into RevMan was done by Eric Linko. Methodological and clinical assessment were done by Eric Linko and Arsi Harilainen. Eric Linko wrote the review.

Issue protocol first published

1999 issue 1

Issue review first published

2005 issue 2

Date of last minor amendment

10 January 2005

Date of last substantive amendment

07 February 2005

Most recent changes Date new studies sought but none found

29 January 2005

Date new studies found but not yet included/excluded

Information not supplied by reviewer

Date new studies found and included/excluded

Information not supplied by reviewer

Date reviewers' conclusions section amended

Information not supplied by reviewer

Contact address Mr LinkoAlppikatu 2Helsinki

FINLANDFIN-00530Telephone: +358 9 7750800Facsimile: +358 9 77507480E-mail: eric.linko@welho.com

Cochrane Library number CD001356Editorial group Cochrane Bone, Joint and Muscle Trauma Group Editorial group code HM-MUSKINJ

S O U R C E S O F S U P P O R T

External sources of support

No sources of support supplied

Internal sources of support

Jorvi Hospital Research Foundation FINLAND Finnish Institute of Occupational Health FINLAND Orton Orthopaedic Hospital FINLAND

K E Y W O R D S

Adult; Humans; Anterior Cruciate Ligament[*injuries][surgery]; Joint Instability[etiology][surgery]; Randomized Controlled Trials; Recovery of Function; Rupture[surgery][therapy]Imprimir |  Fechar