Conditional Survivorship of Osteochondral Allograft Transplantation in the Knee
Nathanial W. Linstrom, William D. Bugbee, Tim WangBackground:
Osteochondral allograft (OCA) transplantation has demonstrated reliable long-term outcomes in treating chondral and osteochondral lesions of the knee. Although several pre- and intraoperative factors associated with graft failure have been identified, the temporal distribution of graft failures after OCA transplantation has not been described.
Purpose:
To evaluate the timing and rate of graft failure after knee OCA transplantation and to report conditional survivorship (CS) to inform postoperative patient counseling.
Study Design:
Case series; Level of evidence, 4.
Methods:
We identified 288 knees (267 patients) that underwent OCA transplantation by a single surgeon between 1997 and 2015, with a minimum follow-up of 10 years. Graft failure was defined as any reoperation requiring removal of the allograft. Nonfailure reoperation was defined as any surgical procedure on the affected knee that did not involve graft removal. Rates of graft failure and nonfailure reoperation were estimated using piecewise linear regression applied to inverse Kaplan-Meier curves. Kaplan-Meier analysis estimated overall and 10-year CS.
Results:
The median patient age was 33.1 years (range, 11-68 years), and 58% were male. Overall graft survivorship was 78% at 10 years. Graft failure occurred in 64 knees (22%) at a median of 2.7 years (range, 0.3-23.6) postoperatively. Annual failure rates were 4.9% in the first 3 years postoperatively and decreased to 1.5% from 3 to 10 years. Nonfailure reoperations occurred in 48 knees, with a median of 1.5 years postoperatively, and followed a similar temporal pattern, with most occurring in the early postoperative period. CS analysis demonstrated that grafts remaining in situ at 4 years postoperatively had a 90% probability of survival to 10 years.
Conclusion:
OCA transplantation of the knee demonstrated an overall 10-year graft survivorship of 78%. Graft failures were not uniformly distributed over time, with the highest failure rates occurring during the early postoperative period and substantially decreasing after 3 years postoperatively to a low steady-state of 1.5% annually. This temporal pattern was also noted with nonfailure reoperations, which occurred predominantly during the early postoperative period. CS analysis offers a patient-centered approach for interpreting these findings and may inform postoperative patient counseling.