OBJECTIVE  To explore the risk factors for mortality in patients with Klebsiella pneumoniae bloodstream infection （KP-BSI） in hepatopancreatobiliary surgery, and to construct a prediction model.

METHODS  A retrospective analysis was conducted on the clinical data of 234 patients with KP-BSI admitted to a hospital from 2013 to 2022. The patients were randomly divided into a training set of 164 cases and a validation set of 70 cases in a ratio of 7∶3. The training set was further divided into a death group of 38 cases and a survival group of 126 cases based on prognosis. Lasso regression and multivariate logistic regression analysis were used in the training set to screen risk factors construct a prediction model for mortality in patients with KP-BSI, and visualize the results with a nomogram. The predictive ability of the model was evaluated through the receiver operating characteristic （ROC） curves, calibration curves and decision curves in both the training and validation sets.

RESULTS  Among 234 patients with blood cultures for KP, 121 （51.71%） were identified as having carbapenem-resistant K. pneumoniae （CRKP）. The use of ventilators （OR = 5.816, 95%CI: 1.520–22.259, P = 0.032）, interventional embolization for hemostasis （OR = 2.472, 95%CI: 1.811–7.534, P = 0.021） and hypotension （OR = 4.823, 95%CI: 1.608–14.470, P＜0.001） were identified as risk factors for mortality in patients with KP-BSI. General anesthesia during surgery （OR = 0.203, 95%CI: 0.054–0.763, P = 0.003） and the use of antibacterial agents during blood culture were identified as protective factors. The constructed nomogram demonstrated good predictive performance in both the derivation cohort （area under the receiver operating characteristic ROC curve AUC = 0.874） and the validation set （AUC = 0.878）.

CONCLUSIONS  This study identifies the risk factors for mortality in patients with KP-BSI in hepatopancreatobiliary surgery and constructs an effective prediction model. This model can assist clinicians in early identification of high-risk patients and optimize treatment strategies.