Background: Periprosthetic joint infection (PJI) are treated with Debridement, Antibiotics and Implant Retention (DAIR) whenever possible. KLIC and CRIME80 scores have been created to estimate the risk of DAIR failure in acute PJI based on patient-dependent factors. The objective of this study was to analyze the performance of both the KLIC and CRIME80 scores. Patients and method: This study includes patients diagnosed with early, delayed or late acute PJI of the hip and knee that were treated by DAIR procedure at Lausanne University Hospital from January 2009 until December 2018. Patients with missing data to calculate the KLIC and CRIME 80 scores, with unknown 2-year outcome due to loss of follow-up and those who refuse to give their consent to participate were excluded. Both scores were calculated retrospectively in each patient, regardless of the type of acute PJI. Scores were analyzed using different cut- offs: < 4 vs. ≥ 4 for KLIC score and < 3 vs. ≥ 3 for CRIME80 score.
Results: This study included 101 patients, with 51% of males and a median age of 71 years. The rate of early failure was 10% and the 2-year failure was 23%. KLIC score had a sensitivity of 60%, a specificity of 49.5% and an area under ROC curve of 0.56. CRIME80 score had a sensitivity of 34.8%, a specificity of 75.6% and an area under ROC curve of 0.61. Conclusion: In this study, KLIC and CRIME80 scores did not predict accurately the risk of early failure and 2-year failure after DAIR procedure for acute PJI, respectively. Due to the retrospective nature of these data and the small sample size, it might be interesting to do a prospective study with a larger sample size to obtain better statistical power.