Predictive Factors for Revision and Survivorship Analysis of a Prevalent 36-mm Metal-on-Metal Total Hip Replacement System: A Large Single-Center Retrospective Cohort Study

Abstract

BackgroundTo our knowledge, this is the largest single-center cohort of the 36-mm Corail-Pinnacle metal-on-metal total hip replacements system, aiming to determine 10-year survivorship and identify predictors of revision. We further assessed year of implantation given reports of manufacturing variations affecting shells made after 2006 predisposing these components to increasing wear. MethodsAll Corail-Pinnacle 36-mm metal-on-metal hips implanted in a single center (2005-2012). The effect of patient and implant-related variables, and year of implantation on revision risk was assessed using Kaplan-Meier, Cox regression, and interrupted time series analysis. ResultsIn total, 1212 metal-on-metal total hip replacements were implanted with a 10-year survival rate of 83.4% (95% confidence interval [CI] = 81.3-85.5). Mean follow-up duration was 7.3 years with 61% of patients reaching a minimum of 7 years of follow-up. One hundred nineteen patients required revision surgery (9.8%). Univariate analysis identified female gender (hazard ratio [HR] = 1.608, CI = 1.093-2.364, P = .016), age at implantation (HR = 0.982, CI = 0.968-0.997, P = .019), smaller 50-mm to 54-mm cup diameter (HR = 1.527, CI = 1.026-2.274, P = .037), and high-offset stems (HR = 2.573, CI = 1.619-4.089, P < .001) as predictors of revision. Multivariate modeling confirmed female gender and high-offset stems as significant predictors of revision. For components implanted after 2007, the number of revisions showed no statistically significant step increase compared to pre-2007 implantation. ConclusionWe observed a high 10-year failure rate (16.6%) with this implant, mostly due to adverse reaction to metal debris. Female gender and high femoral offset stems were significant predictors for all-cause revision. Year of implantation was not significantly associated with an increasing number of revisions from 2007 onwards, although further studies to validate the impact of manufacturing discrepancies are recommended.