The authors have identified two small errors in the published article.
The first error is in the Abstract, where the decimal point was shifted in four point estimates (for example, 0.01 where it should read 0.1) and one confidence interval limit lost its minus sign. All other numbers as published are correct. The correct abstract is as follows:
Objective: To determine bone shape changes over two years after acute anterior cruciate ligament (ACL) injury and to compare knees treated with ACL reconstruction (ACLR) vs no ACLR.
Design: We used prospective data involving 129 young adults (46 % female, mean age 25 years) with recent ACL injury from a subcohort within the prospective multicenter NACOX study. Patients were treated according to Swedish guidelines with supervised rehabilitation before considering ACLR. 3D MR images obtained at baseline, 3-, 6-, 12-, and 24-months post-injury were assessed for changes in the femur and tibia bone surface area using active appearance models. We used a linear mixed-effects model adjusted for age, sex, body mass index, and pre-injury Tegner physical activity level.
Results: We observed growth in bone areas over the first two years, particularly in the medial femur and medial tibia, with an average relative increase of 1.28 % (95 % CI 0.67 %, 1.88 %) and 1.04 % (0.44 %, 1.65 %), respectively. The lateral femur also showed growth, with an increase of 1.15 % (0.55 %, 1.75 %). We found little differences comparing ACLR patients to non-ACLR patients, adjusted relative percentage differences: medial femur, 0.1 % (95 % CI: −0.9 %, 1.1 %); medial tibia, −0.3 % (−1.2 %, 0.7 %); lateral femur, 0.4 % (−0.6 %, 1.4 %); and lateral tibia, −1.1 % (−2.1 %, −0.1 %).
Conclusion: After acute ACL injury, there is a growth primarily in the medial femoral and tibial condyles. The bone shape changes could hypothetically be a response to altered biomechanical conditions and may precede the development of knee OA.
The second error is a wording error in the Methods. The text states that participants were classified as ACLR when having surgery prior to the two-year follow-up. The analysis in fact used reconstruction within twelve months, as correctly reported in the Results.
The statistical models, the confidence intervals, Figure 2 and the conclusions of the article are unaffected.
The authors would like to apologise for any inconvenience caused.