Development and Validation of a Novel Scoring Model Integrating Clinical Risk Factors and Pharmacokinetic Parameters to Predict Vancomycin‐Induced Nephrotoxicity

BACKGROUND: Vancomycin (VCM), a first-line treatment option for infections caused by methicillin-resistant Staphylococcus aureus, has been reported to cause nephrotoxicity even within therapeutic concentration ranges. Traditional therapeutic drug monitoring strategies rely primarily on the area under the concentration-time curve (AUC), without adequately accounting for multiple clinical risk factors associated with nephrotoxicity. OBJECTIVE: The present study aimed to develop a novel scoring model that integrates clinical risk factors and pharmacokinetic parameters to predict VCM-induced nephrotoxicity and validate its predictive performance. METHODS: We conducted a single-center retrospective cohort study on patients who received VCM therapy between April 2021 and March 2023. A multivariable logistic regression analysis was performed to identify independent risk factors for VCM-induced nephrotoxicity, and regression coefficients were used to construct the scoring model. The predictive performance of the proposed model was compared with a conventional AUC-based model using the area under the receiver operating characteristic curve (ROC AUC), net reclassification improvement (NRI), and integrated discrimination improvement (IDI). RESULTS: The scoring model consisted of the following components: the steady-state area under the concentration-time curve (0-8 points), the concomitant use of tazobactam/piperacillin (2 points), the use of loop diuretics (1 point), and the presence of chronic liver disease (2 points). The proposed model demonstrated high predictive performance, with ROC AUC values of 0.79 (95% confidence interval [CI]: 0.71-0.87) in the derivation cohort and 0.84 (95% CI: 0.72-0.96) in the validation cohort. Furthermore, the proposed model showed significantly better performance than the conventional model in terms of NRI (derivation cohort: 0.78, 95% CI: 0.47-1.08; validation cohort: 1.23, 95% CI: 0.79-1.67) and IDI (derivation cohort: 0.07, 95% CI: 0.04-0.11; validation cohort: 0.27, 95% CI: 0.15-0.39) (p < 0.001). CONCLUSION: The scoring model developed in the present study may enhance risk stratification for VCM-induced nephrotoxicity and contribute to advances in individualized dosing strategies in clinical practice.