Abstract
PURPOSE: Decisions on the treatment of periprosthetic joint infection (PJI) are typically guided by established algorithms. However, as these algorithms often lack substantial supporting evidence, this study aimed to evaluate 1-year survival rates and compare different surgical approaches. METHODS: In this single-centre retrospective cohort study, all revisions of the knee due to PJI with at least 1 year of follow-up performed between January 2008 and September 2021 were identified. In total, 141 debridement, antibiotics, and implant retentions (DAIRs), 98 one-stage, and 120 two-stage revisions were performed. Infections were classified as early, acute hematogenous, or chronic infections. Survival was calculated using the Kaplan-Meier method and the cumulative incidence function. Predictors of outcomes were examined with Fine-Gray regression and Cox proportional hazards regression, and subdistribution hazard ratios (sdHR) and adjusted hazard ratios (aHR) with 95% confidence intervals (CIs) were calculated. RESULTS: At 1-year follow-up, 23% (CI 19%-27%) of patients had undergone a reoperation, and 4% (CI 2%-6%) had died. The risk of reoperation was largest after two-stage revision (28%, CI 20%-36%) and smallest after one-stage revision (15%, CI 9%-23%). For every infection type, the failure rates at one-year follow-up favoured one-stage revision over two-stage revision. Higher ASA-scores increased the risk of death (aHR 1.7, CI 1.1-2.5 per one-unit increase). CONCLUSION: The risk of failure after one-year follow-up is high after revision for periprosthetic joint infection. The lowest risk was observed after one-stage revision; however, this may partly reflect patient selection, as one-stage revision may not be suitable for all patients. LEVEL OF EVIDENCE: Level III, retrospective comparative study.