Abstract
PURPOSE: This study investigates the outcomes of two-stage exchange arthroplasty (EA) for periprosthetic joint infection (PJI) following initial or unplanned repeat debridement antibiotics, and implant retention (DAIR). METHODS: We retrospectively reviewed cases of knee arthroplasty infection treated with two-stage EA after DAIR, spanning from January 1994 to December 2010. A total of 138 patients were included, comprising 112 with initial DAIR and 26 with an unplanned second DAIR. Data on demographics, comorbidities, infection characteristics and causative organisms were analyzed. The primary outcome was implant failure or reinfection, observed over a minimum follow-up of 10 years. RESULTS: The overall success rate for two-stage EA was 87% (119/138 patients). Factors identified for treatment failure included reinfection with the same pathogen for unplanned second DAIR (hazard ratio [HR] = 3.41; 95% confidence interval [CI] = 1.35-4.38; p = 0.004), higher reinfection rates in patients undergoing EA after an unplanned second DAIR, especially with a prior history of PJI within 2 years (HR = 4.23; 95% CI = 2.39-5.31; p = 0.002), pre-first DAIR C-reactive protein (CRP) levels over 100 mg/dL (HR = 2.52; 95% CI = 1.98-3.42; p = 0.003) and recurrence with the same pathogen (HR = 2.35; 95% CI = 1.32-4.24; p = 0.007). Additional factors such as male gender (HR = 3.92; 95% CI = 1.21-5.25; p = 0.007) and osteoporosis (T score < -2.5; HR = 3.27; 95% CI = 1.23-5.28; p = 0.005) were identified as risk factors for implant failure in all EA cases. CONCLUSIONS: This study identifies key risk factors for worse knee EA outcomes following DAIR, including a pre-first DAIR CRP level over 100 mg/L, same pathogen recurrence, and PJI history within 2 years. It shows implant failure rates remain constant across EA cases, regardless of DAIR sequence, particularly with risk factors like male gender and severe osteoporosis (T score < -2.5). These results underscore the need for careful evaluation before an unplanned second DAIR, given its significant impact on EA success. LEVEL OF EVIDENCE: Level III.