Predictive model for carbapenem-resistant Klebsiella pneumoniae bloodstream infection based on a nomogram: a retrospective study.

objectiveCarbapenem-resistant Klebsiella pneumoniae bloodstream infections (BSIs-CRKP) are associated with high mortality rates, necessitating early risk stratification tools. This study aimed to identify risk factors and develop a nomogram model to predict BSIs-CRKP in hospitalized patients. A single-center retrospective case-control study was conducted at Shaanxi Provincial People's Hospital (2017-2024). Patients with Klebsiella pneumoniae bacteremia were stratified into CRKP (n = 154) and carbapenem-susceptible (CSKP, n = 233) groups. Clinical data, including demographics, comorbidities, treatments, and antimicrobial exposure, were analyzed. Risk factors were identified via multivariate logistic regression, and a nomogram model was developed using R software. Model performance was evaluated using ROC curves, calibration plots, and the Hosmer-Lemeshow test.

resultsIndependent risk factors for BSIs-CRKP included indwelling urinary catheterization (OR = 2.531, P = 0.038), central venous catheterization (OR = 2.673, P = 0.015), immunosuppressant use (OR = 3.782, P = 0.006), carbapenem exposure (OR = 4.470, P < 0.001), and age ≥ 65 years (OR = 3.740, P = 0.002). A nomogram model was constructed based on these risk factors, which demonstrated good predictive performance with an area under the curve (AUC) of 0.878 (95% CI: 0.845-0.910) and exhibited excellent calibration (Hosmer-Lemeshow, P = 0.234). This nomogram model effectively stratifies BSIs-CRKP risk using five clinically accessible variables. External validation and integration of genomic data are warranted to enhance generalizability and precision.