Treatment for Acute Anterior Cruciate Ligament Tear in Young Active Adults

Lohmander LS, Roemer FW, Frobell R, Roos EM. Treatment for acute anterior cruciate ligament tear in young active adults. NEJM Evidence. 2023;2(8). doi:10.1056/evidoa2200287

Link to Original Article: https://evidence.nejm.org/doi/full/10.1056/EVIDoa2200287

Key Points

- The study followed 121 young active adults with an acute sports-related ACL tear for 11 years to compare patient-reported and radiographic outcomes between those randomized to receive early ACL reconstruction (ACLR) followed by exercise therapy and those treated with early exercise therapy plus optional delayed ACLR.

- At 11-year follow-up, there were no differences in patient-reported outcomes between the two treatment groups. About two thirds of the cohort met the patient-acceptable symptom state, while 44% had developed radiographic osteoarthritis of their injured knee.

- There was no clinically relevant change in mean Knee Injury and Osteoarthritis Outcome Score (KOOS4) score in either treatment group between 5 and 11 years. The primary outcome KOOS4 ranged between -4.5 and 0.7.

- The mean improvement in KOOS4 from baseline to 11 years was 46 points for patients assigned to exercise therapy plus early ACLR and 45 points for those assigned to exercise therapy plus optional delayed ACLR.

- There was no difference in the pivot shift test between groups, and mean summed incident radiographic osteoarthritis feature scores were higher in the group assigned to early ACLR compared to the group assigned to exercise therapy plus optional delayed ACLR.

- There was no between-group difference in the frequency of radiographic osteoarthritis between the two groups for either the index or contralateral knee after 11 years.

Introduction

The paper discusses the 11-year follow-up of the Knee Anterior Cruciate Ligament, Nonsurgical versus Surgical Treatment (KANON) trial and the Early Surgical Reconstruction versus Rehabilitation with Elective Delayed Reconstruction for Patients with Anterior Cruciate Ligament Rupture (COM-PARE) trial, both of which compared nonsurgical treatment with early surgical reconstruction for acute ACL tears. The 2-year results of the trials showed that early reconstruction was not superior to an initial nonsurgical strategy. The 5-year follow-up of the KANON trial reaffirmed this conclusion. Despite this evidence, early surgical reconstruction remains the dominant treatment, and concerns have been raised about the long-term rate of optional delayed ACL reconstruction for those initially treated nonsurgically. Observational studies have suggested that patients without or with delayed surgical reconstruction may exhibit higher rates of structural osteoarthritis and worse patient-reported outcomes over time. The research paper aims to present the 11-year outcomes of the KANON trial to further investigate the long-term effects of different ACL tear treatment strategies, particularly comparing initial nonsurgical treatment with early surgery.

Methods

Inclusion and Exclusion Criteria

This study focused on active adults with an average age of 26 who had recently sustained an ACL injury to an uninjured knee, with the injury being less than 4 weeks old. The majority of the patients (98%) were injured while participating in sports, with soccer being the most common sport involved (73%). Exclusion criteria for the study included professional athletes, individuals with a low level of activity based on the Tegner Activity Scale (TAS score <5), total collateral ligament rupture, full-thickness cartilage lesions on MRI, and extensive meniscus tears requiring repair that could interfere with the trial exercise therapy protocol. The Tegner Activity Scale ranges from 1 to 10, with a score of 5 indicating participation in recreational sports and a score of 9 indicating participation in competitive sports at a nonprofessional level.

Trial Procedures and Treatments

The study implemented a physiotherapist-supervised, goal-based exercise therapy protocol for all patients, which commenced before or at the time of randomization. Patients in the early-reconstruction group had an average of 63 visits, while those in the optional delayed-reconstruction group had an average of 53 visits. Patients who received early reconstruction underwent anterior cruciate ligament reconstruction (ACLR) within 10 weeks of injury by experienced knee surgeons, whereas those in the exercise therapy group with the option of delayed ACLR had the procedure when presenting with symptomatic knee instability as per the study protocol. All ACLRs were single bundle and utilized either patella tendon or hamstring tendon procedures. Meniscal tears were addressed with partial resection or fixation based on MRI findings and clinical signs, and additional concomitant meniscal surgery was performed during ACLR if unstable meniscal tears were identified. The treatment approach was consistent across both groups, with the main difference being the timing of ACLR.

Outcomes

• Patient-Reported Outcomes

The study's outcomes were evaluated through patient-reported measures including the Knee Injury and Osteoarthritis Outcome Score (KOOS), the Medical Outcomes Study 36-Item Short-Form Health Survey (SF-36), and the TAS score. The primary endpoint focused on the change in mean value of four KOOS subscales over an 11-year period: pain, symptoms, function in sports and recreational activities, and knee-related quality of life, with scores ranging from 0 (worst) to 100 (best). Secondary outcomes included all five separate KOOS subscales, SF-36 physical and mental component scores, and the TAS score. The fifth KOOS subscale pertained to function in activities of daily living, also scored on a range from 0 (worst) to 100 (best). The SF-36 component scores were assessed on a scale from 0 (worst) to 100 (best), and the TAS score was evaluated according to a previously defined measure.

• Body-Mass Index

The height and weight of the patients were self-reported and calculated as body weight in kilograms divided by the square of height in meters.

• Responder Criteria

The section discusses the challenges of interpreting mean score changes and mean absolute follow-up scores in the context of patient response to treatment. The concept of responder criteria is introduced, where each patient is categorized as either a responder or nonresponder to treatment. To aid in clinical interpretation, the study applied established thresholds for ACLR patients, specifically the KOOS 4 thresholds, to the 11-year KOOS intention-to-treat data. The three responder criteria utilized were: minimal important change (MIC) with a KOOS 4 MIC threshold value of 9 for baseline to follow-up score changes, patient-acceptable symptom state (PASS) with a KOOS 4 PASS threshold value of 79, and treatment failure with a KOOS 4 treatment failure threshold value of 42 at follow-up. By applying these responder criteria, the study aimed to provide a clearer understanding of how patients respond to treatment and to facilitate the clinical interpretation of the data.

• Mechanical Stability

The manual Lachman and pivot shift tests were used to test for knee laxity to identify number of knees without the signs of mechanical instability.

• Knee Radiography and Grading of Images

The research study involved obtaining knee radiographs at baseline, 2, 5, and 11 years using standardized methodology. A musculoskeletal radiologist, blinded to grouping and clinical data, graded the radiographs for osteophytes and joint space narrowing according to the Osteoarthritis Research Society International Atlas. The intrareader reliability for atlas-based scoring was reported as 0.67. The study reported the presence of radiographic osteoarthritis (OA) for the total knee joint and separately for the tibiofemoral and patellofemoral compartments. The criteria for defining radiographic OA included joint space narrowing grade 2 or more, sum of the two marginal osteophyte grades 2 or higher, and/or grade 1 joint space narrowing with grade 1 osteophytes. For the tibiofemoral compartment, the cutoff approximated grade 2 radiographic OA on the Kellgren and Lawrence scale. Additionally, the study reported the mean incident and prevalent individual OA radiographic features and graded joint space narrowing and osteophytes for the index and contralateral knees, with higher scores indicating more severe radiographic involvement.

• Adverse Events

All adverse events were self-reported. Injuries to the contralateral knee that occurred before study inclusion were self-reported at the baseline visit, and injuries to this knee that occurred between baseline and the 11-year visits were self-reported at the 11-year visit.

Statistical Analysis

The statistical analysis in the research paper utilized various methods to assess primary and secondary outcomes. The primary outcome, KOOS 4, was presented as means and 95% confidence intervals, while secondary outcomes were reported as differences of means or medians with 95% confidence intervals. Additionally, post hoc outcomes including body-mass index, treatment responders, and radiographic OA features were added and presented with proportions and 95% confidence intervals. Group comparisons were conducted using analysis of covariance, Mann-Whitney U test, and chi-squared test. A significance level of P < 0.05 was used for two-tailed tests. The statistical analyses were performed using Stata SE14. Furthermore, a post hoc analysis for dropouts was conducted, including an analysis of reasons for missingness and a sensitivity analysis for the primary outcome. It is noted that the 5-year report did not include a provision for correcting for multiplicity when conducting tests for secondary or other outcomes, thus caution should be taken when inferring definitive treatment effects for secondary outcomes from the reported confidence intervals.

Results

Characteristics of Patients and Treatments

The research paper reports on the follow-up of 120 out of 121 initially allocated patients at 5 years, and 107 (88%) at an average of 10.9 years after randomization. Out of 107 patients, 53 patients were assigned to receive early ACLR, and 54 to receive optional delayed ACLR. Over the 11-year period, 28 patients from the delayed ACLR group underwent the surgery, with one patient having the procedure since the 5-year follow-up, while 26 received exercise therapy alone. The demographic composition of the patients was predominantly of European descent, with less than 5% non-European, and the age and sex distribution was consistent with national ACL registers of Sweden, Norway, and the United Kingdom.

Full Set Analysis

Primary End Point

The study analyzed the mean improvement in the Knee injury and Osteoarthritis Outcome Score (KOOS) from baseline to 11 years for patients assigned to exercise. The mean improvement of KOOS from baseline to 11 years for patients in exercise therapy plus ACLR was 46 and for patients in exercise therapy plus optional delayed ACLR was 45. In addition, the study found there was no clinically relevant changes in KOOS at 5 years and 11 years. When accounting for the missing data, the between-group difference was 1.4, indicating similar results between the two randomized groups.

Secondary and Other End Points

• Self-Reported Outcomes

The mean differences in Knee injury and Osteoarthritis Outcome Score (KOOS) subscales between the two groups ranged from -4.5 to 0.7. Both groups showed similar improvement in Short Form 36 (SF-36) components and Tegner activity scale score over the 11 years. Around 90% of patients in both groups achieved the minimal clinically important change (MIC) threshold, and approximately two-thirds in each group achieved a KOOS 4 score of 79, indicating Patient Acceptable Symptom State (PASS). The proportion of patients reporting a KOOS 4 score of 42 or below, indicative of treatment failure, was less than 10% in both groups. The mean body-mass index increased by 1.2 units over the 11 years with no difference between the study groups.

• Mechanical Stability

The study found 88% of the patients in the early ACLR group and 73% of the patients in the optional delayed ACLR group had tested negative in the pivot shift test. 67% of the patients in the early ACLR tested negative with the Lachman Test, while 33% of the patients in the optional delayed ACLR group tested negative for the same test.

• Radiographic Osteoarthritis

The research paper examined the progression of radiographic osteoarthritis (OA) in patients who underwent early anterior cruciate ligament reconstruction (ACLR) compared to those who underwent exercise with optional delayed ACLR. The study found that mean summed incident radiographic feature scores increased over time for both index and contralateral knees, with the most significant increase observed in the index knee of patients with early ACLR within the first 2 years after injury. At 11 years, the difference in mean summed index knee scores between the early ACLR group and the exercise with optional delayed ACLR group was 1.0. Additionally, in the full analysis set, 44% had developed radiographic OA in their index knee, while 22% had developed it in their contralateral knee. The frequency of radiographic OA did not differ significantly between the two groups for either the index or contralateral knee after 11 years.

• Self-Reported Adverse Events

The self-reported adverse events section of the research paper discusses knee injuries reported by participants between 5 and 11 years following the initial assessment. It notes that knee injuries during this period were only self-reported at the 11-year visit, limiting the details available due to recall. The study found five definite index knee injuries reported during this period, two of which were severe enough to require evaluation by an orthopedic surgeon. One of these severe injuries was managed through ACL reconstruction (ACLR) and partial meniscus resection. The remaining three injuries were seen in primary care with unclear diagnosis, severity, or management. Additionally, a significant proportion of the patients self-reported injuries to the opposite knee.

As-Treated Analysis Set

This section presents key findings related to the outcomes of different treatments for anterior cruciate ligament reconstruction (ACLR) and exercise therapy. The study found that at 11 years, patients who initially underwent exercise therapy and then had delayed ACLR had the highest mean Knee injury and Osteoarthritis Outcome Score (KOOS) 4 scores, followed by those who had early ACLR, while those treated with exercise therapy alone had the lowest mean scores. This pattern was consistent across the KOOS 4 responder criteria, SF-36 components, and total aggregate score (TAS).

Additionally, the radiographic osteoarthritis (OA) feature scores showed that the early ACLR group had the highest mean summed value, followed by the delayed ACLR group, and then the exercise therapy-only group. The frequency of compartmental radiographic OA was also lowest in the exercise therapy-only group compared to the early ACLR and delayed ACLR groups.

These findings suggest that patients who underwent exercise therapy followed by delayed ACLR demonstrated better outcomes in terms of KOOS scores, SF-36 components, TAS score, and radiographic OA features compared to those with early ACLR or exercise therapy alone. The results highlight the potential benefits of a combined approach of exercise therapy and delayed ACLR in managing anterior cruciate ligament injuries.

Discussion

The 11-year follow-up results of a randomized clinical trial comparing early ACL reconstruction (ACLR) and exercise with optional delayed ACLR showed no significant differences in knee outcomes, physical or mental health status, or activity level between the two groups. There were no notable changes in patient-reported outcomes or incident radiographic osteoarthritis (OA) features between 5 and 11 years. The study also found no differences in knee osteoarthritis and patient-reported outcomes at 11 years, consistent with previous trial results at 2 and 5 years. Notably, 52% of patients initially allocated to receive exercise therapy and optional ACLR eventually underwent delayed ACLR. The summed incident individual tibiofemoral and patellofemoral radiographic OA features did not markedly change between 5 and 11 years. However, the widths of confidence intervals for these secondary outcomes were not adjusted for multiplicity and therefore cannot be used to infer treatment effects. Overall, the study indicates no substantial differences in knee outcomes and radiographic OA features between the early ACLR and exercise with optional delayed ACLR groups at the 11-year follow-up.

As-Treated Patient-Reported and Radiographic Outcomes

The section discusses the as-treated patient-reported and radiographic outcomes of individuals initially treated with exercise therapy and then with a delayed ACL reconstruction (ACLR), compared to those treated with only exercise therapy. Patients who initially received exercise therapy and later underwent delayed ACLR showed improvement over 5 to 11 years, while those treated with exercise therapy only deteriorated. Potential reasons for these differences include selection bias from crossover, reporting bias due to having multiple treatment options, and decreased physical activity in the exercise-only group at 11 years. Costly treatments such as surgery may lead to greater satisfaction, while lower physical activity is linked to more bodily symptoms. Radiographic feature scores for the index knees at 11 years revealed higher scores (indicating more severe involvement) in the early ACLR group, followed by the delayed ACLR group, and the exercise-only group with the lowest scores. Additionally, approximately half of the early and delayed ACLR groups showed compartment radiographic osteoarthritis (OA), compared to less than one-third of the exercise therapy-only group. This difference was primarily driven by a higher frequency of patellofemoral OA in those treated with ACLR, particularly with a patellar tendon graft.

Comparison with Patient-Reported Outcomes of Other Studies

The study's comparison of patient-reported outcomes with other trials and studies revealed consistent findings at 11 years with the KANON trial's 2- and 5-year reports and the recent COMPARE trial's 2-year outcomes. These results were also broadly consistent with long-term follow-up reports from register-based and case-control studies, aligning with the majority in finding no relevant differences in patient-reported outcomes between patient groups managed with or without tibiofemoral osteoarthritis (TFOA) and patellofemoral osteoarthritis (PFOA). The study also assessed radiographic osteoarthritis (OA) using specific criteria and noted that the widths of the confidence intervals for secondary outcomes were not adjusted for multiplicity. Additionally, the study reported individual-level responder outcomes at 11 years, revealing no differences in the proportions of patients between the two groups in terms of improvement, acceptable outcomes, or treatment failure. These individual-level responder outcomes are considered valuable for assigning clinical value and providing insight into the interpretation of group-based patient-reported outcome measure data. Finally, the study mentioned the loss of knee radiographs for one patient and emphasized the importance of early surgical reconstruction of the anterior cruciate ligament (ACLR) in their observations.

Comparison with Imaging and Biomarker Outcomes of Other Studies

The section compares the imaging and biomarker outcomes of the study with those of other research. Studies have suggested varying effects of ACL reconstruction (ACLR) on the risk of developing structural osteoarthritis (OA), with some indicating a decrease in risk, an increase, or no difference. The study found no difference in incident radiographic knee OA between individuals randomly assigned to surgical ACLR or not, consistent with observational studies. However, it observed greater mean incident-summed radiographic scores for the early ACLR group at 2, 5, and 11 years, which may be due to the categorizing inherent to the radiographic OA criteria. Within-patient comparisons were confounded by individuals reporting contralateral knee injury before or during the study. The study's results align with previous findings of early bone changes and rapid alterations in bone shape after ACL injury, particularly in those who underwent surgical reconstruction. Moreover, cartilage and biomarker changes were most pronounced early after injury and in individuals with surgical ACLR, suggesting that surgery may further impact the traumatized joint.

Limitations

The paper discusses the limitations of an extended follow-up trial analysis, which focused on the effects of treatment actually received. The as-treated analysis aimed to estimate these effects but was found to be subject to considerable bias. The decision to crossover to a different treatment was not straightforward and introduced selection bias that affected the randomization balance. Additionally, loss to follow-up over the 11 years of observation resulted in small groups, limiting the ability to draw conclusions for the as-treated groups. Analyses of secondary outcomes were not adjusted for multiplicity, further limiting their interpretation. Self-reported adverse knee events from 5 to 11 years were subject to recall bias, and the blinding of the expert reader of x-ray images to treatment allocation was compromised by apparent radiographic features of the treatment. While MRI was used for the 2-and 5-year follow-up visits, it was not available at 11 years. Post hoc analysis for dropouts did not appreciably change the primary trial outcome, and incomplete data on reasons for missingness affected sensitivity analysis assumptions.

Conclusions

After 11 years of conducting a trial, the researchers did not find any significant differences in the primary trial outcome, KOOS 4, or in any of the five KOOS subscales, KOOS responder criteria, general physical or mental health status, current activity level, or radiographic osteoarthritis (OA) between patients who underwent early anterior cruciate ligament reconstruction (ACLR) plus exercise and those who underwent initial exercise therapy with the option of later ACLR if needed. It was observed that the mean incident-summed radiographic OA scores for the early ACLR group were higher than for the exercise plus optional ACLR group; however, this difference did not lead to a disparity in patient-reported outcomes. This suggests that there were no substantial benefits in terms of patient-reported outcomes for patients undergoing early ACLR compared to those who underwent initial exercise therapy with the possibility of later ACLR. Therefore, the findings indicate that the decision between early ACLR and initial exercise therapy with the option of later ACLR should take into consideration factors beyond radiographic OA scores, as these scores did not correlate with differences in patient-reported outcomes.