Abstract
Objectives:
Osteochondral injury is associated with pain and/or dysfunction and has been implicated as a factor leading to early degenerative joint disease. Fresh osteochondral allograft (OCA) transplantation is a well-described surgical replacement therapy for patients with large osteochondral defects. Optimal selection of patients and factors most associated with a successful outcome is of clinical importance. Previous studies have found age and prior surgery to be predictive of graft failure; however, the potential influence of BMI, coronal plane alignment, and degree of osteoarthritis as well as donor factors such as donor age, graft age and donor match for gender and specific graft locations on patient reported outcomes remains less well characterized.
Methods:
We utilized a prospective database of 97 consecutive patients receiving osteochondral allograft transplantation to the distal femur from a single surgeon practice over 5 years (2006-2011). Patient and donor characteristics were routinely collected, as were patient reported outcome measures (PRO) including KOOS and IKDC. Pre-operative full length standing lower extremity radiographs were measured. Values were considered as both continuous (degrees of varus/valgus) and dichotomous variables (WB axis through the recipient condyle). We used AP X-rays to grade osteoarthritis using the Kellgren-Lawrence scale. Responder analysis was applied to individual patient scores using solitary and dual threshold criteria. Logistic regression was then applied to determine relationships between the candidate variables and response according to the responder analysis threshold. We considered 19 candidate variables in the analysis (Figure 1) and used a forward process for variable selection and multivariate model building based on Akaike Information Criterion (AIC) as the measure of relative goodness-of-fit for the models.
Results:
53 patients (36 male, 17 female, mean age 38.6 years) with a minimum of 24-month data (mean 35.7 months) and without concurrent osteotomy are reported. A statistically significant improvement from baseline to final follow-up (p<.006) was seen for KOOS (Pain: +14.6, ADL: +15.6, Sports: +33.8, QOL: +27.8) and IKDC (+18.7). Patient factors independently associated with greater improvement included a BMI of less than 30 (p<.047), prior cartilage surgery (p=.019), and Kellgren-Lawrence score of 2-4 (p=.024). Valgus alignment trended to greater improvement (p=.08), but medial vs lateral condyle did not. We built predictive multivariate models for our six dependent variables (KOOS, IKDC) and identified 4 significant contributors within patient and donor characteristics. Amongst the factors, the most prominent predictors were valgus alignment (p=.03), smaller defect size (p=.03), younger age (p=.01), and no additional simultaneous procedure (p=.01).
Conclusion:
We confirm that at an average of 3 years after surgery, patients undergoing OCA to the distal femur can expect significantly improved PRO scores. Independent factors resulting in greater patient improvement in this cohort included the absence of obesity, a history of prior cartilage surgery, and, to our surprise, advanced osteoarthritis (Kellgren-Lawrence 2-4). In the multivariate model, younger age, valgus alignment, smaller defect size, and no simultaneous procedure were predictive of better PRO. Surgeons may use this data to select and advise potential patients under consideration for OCA transplant.