A comparison of strategies for identifying patients at risk for carbapenem-resistant or extended β-lactam-resistant Pseudomonas aeruginosa

Abstract Objectives To assess risk factors for carbapenem-resistant Pseudomonas aeruginosa (CR) and extended-β-lactam-resistant P. aeruginosa (EBR) infection/colonization, and to develop and compare tools for predicting isolation of CR and EBR from clinical cultures. Methods This retrospective study analysed hospitalized patients with positive P. aeruginosa cultures between 2015 and 2021. Two case–control analyses were performed to identify risk factors and develop scoring tools for distinguishing patients with CR versus carbapenem-susceptible (CS) P. aeruginosa and EBR versus CS P. aeruginosa. The performance of institutionally derived scores, externally derived scores and the presence/absence of key risk factors to predict CR and EBR were then compared. Results A total of 2379 patients were included. Of these, 8.3% had a positive culture for CR, 5.0% for EBR and 86.7% for CS P. aeruginosa. There was substantial overlap in risk factors for CR and EBR. Institutional risk scores demonstrated modestly higher area under the ROC curve values than external scores for predicting CR (0.67 versus 0.58) and EBR (0.76 versus 0.70). Assessing the presence/absence of ≥1 of the two strongest predictors (prior carbapenem use or CR isolation within 90 days) was slightly inferior to scoring tools for predicting CR, and comparable for predicting EBR. Conclusions Clinicians concerned about CR in P. aeruginosa should consider the likelihood of EBR when making treatment decisions. A simple approach of assessing recent history of CR isolation or carbapenem usage performed similarly to more complex scoring tools and offers a more pragmatic way of identifying patients who require coverage for resistant P. aeruginosa.