Aside from potential overestimation of the knee OA incidence rate, we are concerned about the reliability of some other conclusions of the study. Culvernor et al state that their data extend those of recent MRI studies suggesting that ACLR does not restore a knee to normal. This conclusion does not seem to follow from their study, as they reported the incidence rate of early knee OA after ACLR, but not specific features such as stability, function, etc. In addition, they seem to be concluding that reconstruction may not be appropriate in a large proportion of patients with ACL injury. However, this conclusion cannot be supported from a study in which the control group consisted of uninjured asymptomatic individuals. Patients with ACL injury treated with conservative interventions rather than with reconstruction may have been a more appropriate control group.Besides the above, some other issues need to be mentioned. First, the authors did not report whether the ACL-injured patients also had collateral ligament injuries, which may have affected the results (3). Second, the authors conclude that the presence of OA 1 year following ACLR was more common than previously recognized (31% of individuals had either patellofemoral or tibiofemoral OA). However, it was already reported, in a recent meta-analysis, that of the 16 included studies, 28% of patients showed radiologic signs of OA (4), and that associated meniscal resection dramatically increased the risk of developing OA. Third, there may be other confounders, such as intraarticular injection of hyaluronic acid.We respect the contributions of the authors. We would look forward to seeing the results of an extension of this study over a longer followup period and would also be very interested in their responses regarding the issues raised herein.
ReplyTo the Editor: We appreciate Dr. Zeng and colleagues' interest in our article and would like to respond to their comments and expand on some key aspects of our study.The central point raised by Zeng et al appears to concern a possible overestimation of the rate of early knee OA 1 year following ACLR. In particular, they suggest that the absence of preoperative images in our ACLR group and certain demographic differences in our uninjured control group led to the higher-than-expected OA rate.We acknowledged that we did not have access to preoperative images and hence, some OA features may have preceded the knee trauma. However, several findings counter the likelihood that preexisting OA could have accounted for most of the OA observed at 1 year following ACLR, i.e., 1) our uninjured control group exhibited few features of OA, 2) no overt degeneration (i.e., bony changes such as osteophytes) was observed arthroscopically at the time of ACLR, 3) no association between older age and MRI-diagnosed OA or osteophytes was found, 4) more than one-third of the participants in the ACLR group who were age ,25 years had an osteophyte evidenced on MRI, which would be unexpected in the absence of trauma, and 5) radiographic osteophyt...