Treatment Failure Rates Decrease Significantly at Approximately One Year Following Two-Stage Exchange Arthroplasty for Periprosthetic Joint Infection

BACKGROUND: The two-stage exchange arthroplasty is a preferred treatment for chronic periprosthetic joint infection (PJI). A gap in knowledge remains surrounding the optimal follow-up period for staged revision. This study aimed to determine the length of time PJIs should be monitored after two-stage exchange arthroplasty and the timepoints by which most failures occur. METHODS: This retrospective study evaluated 148 patients who underwent two-stage exchange arthroplasty for chronic total knee arthroplasty PJI from 2015 to 2022. Dates of failure were recorded using the 2019 Musculoskeletal Infection Society Outcome Reporting Tool criteria. Bayesian regression with Monte Carlo Markov Chain simulations was used to generate simulation models with one or two statistically significant breakpoints in cumulative failure curves. "Tier 1 failure" included patients on suppressive antibiotics; "tier 1 plus 2 failure" excluded them, with tier 3 and 4 cases classified as failures in both models. RESULTS: For tier 1 failure, we identified a significant breakpoint at 1.20 years (95% confidence interval: 1.03 to 1.47), with a 441% greater failure rate before this point (P < 0.001). A two-breakpoint model identified breakpoints at 1.11 and 2.42 years; failure was 3.43 times higher before 1.11 years than between 1.11 and 2.42 years and 7.40 times higher than after 2.42 years (P < 0.001). For tier 1 plus 2 failure, a single breakpoint at 0.73 years (95% confidence interval: 0.62 to 1.06) was associated with 997% higher failure than before this time (P < 0.001). Single breakpoint models demonstrated superior overall fit. CONCLUSIONS: Failures clustered early, with steep first-year accrual for the tier 1 plus 2 endpoint, while Tier 1 reoperation failures continued to accumulate into the second year. Breakpoints mark the onset of declining failure rates, rather than complete elimination of risk. Thus, 1-year follow-up may be adequate for studies using broad failure definitions, while extended follow-up into the second postoperative year remains essential to fully characterize late failures.