Predictive factors for 30-day mortality after polymyxin B treatment of carbapenem-resistant Gram-negative bacilli infections

Objectives: This study aimed to investigate the associated factors predicting 30-day mortality of carbapenem-resistant Gram-negative bacilli (CR-GNB) infections, with focus on polymyxin B AUC(ss,24)(h)/MIC. Methods: This prospective cohort study included patients with CR-GNB infections from January 2022 to January 2024. The primary outcome was 30-day mortality. Classification and regression tree (CART) analysis was used to calculate the AUC(ss,24)(h)/MIC threshold for 30-day mortality. Multivariate analysis was conducted to assess risk factors affecting this outcome. Results: We enrolled 107 patients, among which 30-day mortality occurred in 31 (29.0%) cases. The CART-derived AUC(ss,24)(h)/MIC breakpoint was 51.3, and patients in the below-breakpoint group had 3.4-fold higher 30-day mortality than those in the above-breakpoint group (58.1% vs 17.1%, P < 0.001). On multivariate analysis, polymyxin B AUC(ss,24)(h)/MIC of >= 51.3 (aOR 0.08, P = 0.001) predicted a lower risk for 30-day mortality. In subgroup analysis, the survival benefit of AUC(ss,24)(h)/MIC target attainment remained in patients with high risk of mortality or carbapenem-resistant Acinetobacter baumannii infections. Conclusions: Polymyxin B AUC(ss,24)(h)/MIC of >= 51.3 independently predicted lower 30-day mortality in treating CR-GNB infections. Further studies should verify the AUC(ss,24)(h)/MIC target associated with survival outcomes in larger randomized controlled trials.