Abstract
BACKGROUND The combination of vancomycin (VAN) and piperacillin-tazobactam (TZP) is commonly used to treat sepsis, but it is associated with a high risk of acute kidney injury (AKI). This article describes our development of a nomogram to predict the probability of AKI caused by the combination of the VAN and TZP in the treatment of sepsis. MATERIAL AND METHODS Patients with sepsis treated with VAN and TZP from the MIMIC-IV database were included. The patients were randomly divided into a training set and a validation set at a 7: 3 ratio. Key variables were identified through the integration of least absolute shrinkage and selection operator (LASSO) and multivariate logistic regression analysis. The performance of the nomogram was evaluated using area under the receiver operating characteristic curves (AUC), calibration curves, and decision curve analysis in the training set, and was further assessed in the validation set. RESULTS We included 618 patients, with 469 developing AKI. Six risk factors - body mass index, SOFA score, mechanical ventilation, antihypertensive drugs, serum potassium, and total vancomycin dosage - were identified as predictors of AKI occurrence. The AUC was 0.75 in the training set and 0.74 in the validation set. Calibration curves showed good consistency. Decision curve analysis indicated the nomogram worked well for AKI risk prediction if the threshold in the training set was 3-80% and that in the validation set was 10-95%. CONCLUSIONS This study was the first attempt to develop and validate a model that could predict the risk of AKI caused by the combination of VAN and TZP in the treatment of sepsis, providing a reference for clinical decision-making.