Predicting Outcomes of Early and Late Acute Periprosthetic Joint Infections Treated with Debridement, Irrigation, and Implant Retention: An External Validation of the CRIME80 (COPD, C-reactive protein, Rheumatoid arthritis, Index surgery, Male, Exchange of mobile components, age greater than 80 years) Score.

Abstract

Debridement, antibiotics, and implant retention (DAIR) is a common treatment for periprosthetic joint infection (PJI), though it requires careful patient selection. The CRIME80 (COPD, C-reactive protein, Rheumatoid arthritis, Index surgery, Male, Exchange of mobile components, age greater than 80 years) score is a tool initially designed to predict treatment failure in patients who have late-acute PJI treated with DAIR. We aimed to assess the predictive performance of the CRIME80 score in patients who have early and late acute PJI. We conducted an international multicenter validation study using retrospectively collected data from patients who had acute infections of the hip or knee treated with DAIR. Failure was defined as any of the following events: (i) removal of the prosthesis; (ii) the need for suppressive antibiotics one or more years after surgery for PJI; (iii) relapse of infection as determined by the treating clinician; (iv) repeated DAIR more than seven days after the initial procedure; or (v) death. Non-parametric estimation of the area under the receiver operating curve (AUC) was used to establish the discriminative performance of the CRIME80 score. The cohort included 281 knee and 99 hip infections treated with DAIR. Although the CRIME80 score showed a linear trend with failure, the score only had modest discriminative performance (AUC in hip PJI: 0.63, 95% confidence interval (CI) 0.51 to 0.74; AUC in knee PJI: 0.57, 95% CI 0.51 to 0.64). There was no significant difference in performance in early acute (AUC: 0.62, 95% CI 0.55 to 0.69) and late acute (AUC: 0.55, 95% CI 0.45 to 0.65) infections. This finding was consistent across several sensitivity analyses. Our study found that the CRIME80 tool had limited predictive performance. This finding was consistent across patients who had PJI of the hip or knee and early or late acute infection. Given these findings, the benefit of implementing the CRIME80 score in clinical practice remains uncertain.