Abstract
AIMS: To assess the survival function of cementless total trapezium metacarpal prostheses (TTMPs) at 20 years, to compare survival functions by trapezium size, and to evaluate the association between the instantaneous risk of TTMP failure and small trapezium size using a multivariate Cox regression model. METHODS: This observational cohort study included 221 consecutive patients with a mean follow-up after TTMP of 137.3 months (maximum of 246 months). Kaplan-Meier and actuarial life-table methods were used to evaluate the survival function of thecohort. Kaplan-Meier survival curves were compared by trapezium size. Multivariate Cox regression analysis was used to determine the effect of potential confounders on the association between small trapezium and the instantaneous risk of TTMP failure. RESULTS: At the end of follow-up, there was a 89.01% chance of the TTMP surviving for 246 months or more. There was an association between TTMP survival time and trapezium size showing a significant trend such that the survival curves weresignificantly higher with larger trapezium size (Mantel-Cox test, p = 0.0001; WilcoxonBreslow test, p = 0.0002; Tarone-Ware test, p = 0.0001).The unadjusted Cox regression model showed a significant association between small trapezium size (smaller than 9 mm) and the instantaneous risk of TTPM failure (HR: 7.37, 95% CI: 2.46-22.07). In the multivariate Cox analysis, "age", "trapezium morphology", and "complications" were confounders in the association between small trapezium size and the hazard ratio of prosthetic failure (HR = 3.76; 95% CI 0.96 to 13.82). CONCLUSION: These results confirm the long-term functional survival of TTMP prostheses and reveal a significant increase in trend of the survival curve with larger trapezium size. Patient age, trapezium morphology, and the presence of post-surgical complications are confounders in the association between small trapezium size and the hazard ratio of TTMP failure.