Analyzing Learning Curve Effects: Total Ankle Replacement Design Switch and Long-Term Survival.

ABSTRACT

Different aspects of the learning curve in total ankle replacement (TAR) have been studied in the short to mid-term, with 30 cases often considered critical. However, its impact on long-term (10- and 15-year) survival remains unclear. Therefore, we retrospectively analyzed 77 consecutive TARs performed by one orthopedic surgeon. The main outcome was long-term survival between cases 1-30 and 31-77 using the Kaplan-Meier with Competing Risk Analyses. Secondarily, we used Moving Average Method with LOESS regression to confirm the learning curve based on the perioperative complications. Thirdly, associations between perioperative complications and operation time on long-term survival were assessed using Cox proportional hazard models. The 10-year survival of cases 1-30 was 89.9% (95% CI 70.4-96.5), and of 31-77, 92.4% (95% CI 7745- 97.5) (p = .58). The 15-year survival was 81.8% (95% CI 59.5-91.8) and 74.8% (95% CI 52.4-86.6), respectively (p = .97). The long-term survival rate for the TAR that endured perioperative complication was 96.70% (95% CI 90.28-103.12), and for the uncomplicated TAR 87.50% (95% CI 77.12-97.88%) (p = .24). Operating time nor occurrence of perioperative fractures were significantly associated with long-term survival (p = .11 and 0.26, respectively). However, moving average method revealed a significant decreasing trend with a cut-off value of 33 procedures regarding the marginal probability of perioperative osseous complications (p < .01). In conclusion, surgeons should note a learning curve when adapting arthroplasty procedures. After the prosthesis design switch, the learning curve regarding perioperative osseous complications was confirmed at 33 TAR. The switch did not affect long-term survival.