Noninferiority meta-analysis of autologous vs allogeneic osteochondral transplantation for the treatment of osteochondral defects of the knee confounded by defect size differences

Background: Autologous and allogeneic osteochondral transplantations are widely employed for knee osteochondral defects (OCDs). However, there has been a lack of consensus on the efficacy of autografts vs allografts. Objectives: To evaluate the complication profile following autologous vs allogeneic osteochondral transplantation for knee OCDs. Data sources: PubMed and Embase. Study eligibility criteria, participants, and interventions: Clinical studies that reported complication rates following autologous or allogeneic osteochondral transplantation for knee OCDs. Study appraisal and synthesis methods: Meta-analysis of proportions using the restricted maximum-likelihood method after the Freeman-Tukey double-arcsine transformation. Results: There were 35 studies with 2647 and 2569 OCDs included before and after lost to follow-up, respectively. Graft failure rates were not significantly lower in autografts vs allografts in all analyses: as reported (6.23% vs 10.81%, respectively; P = .63), best-case (6.00% vs 10.24%, respectively; P = .66) and worst-case analyses (7.46% vs 14.43%, respectively; P = .42). In the subgroup of mean follow-up time <36 months, graft failure rates were significantly lower in autografts vs allografts in all analyses. Limitations: The meta-analysis was confounded by the inherent differences in mean defect sizes of the autograft and allograft arms. Conclusions and implications of key findings: Overall graft failure and revision rates were lower in autograft vs allograft for the treatment of knee OCDs but this may have been confounded by the inherent differences in mean defect sizes between the groups. Further studies on autografts for larger-sized knee OCDs and allografts for smaller-sized knee OCDs are warranted to better complete the noninferiority analysis.