External validation of a flowchart related to achieving the target area under the concentration-time curve for vancomycin: A retrospective multicenter study

Abstract

During therapeutic drug monitoring (TDM) for vancomycin (VCM), the area under the concentration-time curve (AUC) is important for balancing efficacy versus toxicity. In a previous study, we developed a decision tree (DT) model to achieve the target AUC during TDM over the follow-up period (AUC_follow-up). This study aimed to validate the DT model for achieving the target AUC_follow-up. Patients who received VCM for at least 72 h and had an initial TDM within four doses between January 2023 and December 2023 were analyzed. The AUC of the initial TDM was calculated over 2-point (peak/trough) concentrations via Bayesian estimation. The target AUC_follow-up was defined as 400–600 μg h/mL.

Among 188 patients (median age [interquartile range], 66 [56, 79] years; 50 % female), the target AUC_follow-up was achieved in 70 % (132/188). When the predicted AUC was 400–600 μg h/mL, 84 % (102/121) achieved the target AUC_follow-up. In a 12 h dosing interval subgroup, 86 % (88/102) achieved the target AUC_follow-up. Conversely, when the predicted AUC was <400 or >600 μg h/mL, the proportion who achieved the target AUC_follow-up dropped to 44 % (30/67). Only 30 % (3/10) of those with creatinine clearance rates of >130 mL/min/1.73 m² achieved the target. The area under the receiver operating characteristics curve was 0.74, and the R² value was 0.15.

Our findings confirmed the external validity of the DT model and supported its use for optimizing VCM dosing. Overall, the DT model offers a reliable framework for achieving target AUC values during follow-up for TDM, aiding safe and effective treatment.