[Predictive performance of population pharmacokinetic software on vancomycin steady-state trough concentration]

Area of Science:

• Pharmacokinetics and Therapeutic Drug Monitoring
• Clinical Pharmacy
• Computational Drug Dosing

Background:

• Accurate vancomycin dosing is crucial for therapeutic efficacy and minimizing toxicity.
• Population pharmacokinetic (PopPK) models aid in optimizing drug regimens.
• JPKD-vancomycin is a software tool for predicting vancomycin concentrations.

Purpose of the Study:

• To evaluate the predictive performance of JPKD-vancomycin software for vancomycin steady-state trough concentrations.
• To identify factors influencing the accuracy of JPKD-vancomycin predictions.

Main Methods:

• Retrospective analysis of 310 patients receiving vancomycin.
• Comparison of JPKD-vancomycin predictions against measured concentrations for initial and adjusted regimens.
• Weight residual (WRES) calculation to assess prediction error.
• Logistic regression and ROC curve analysis to identify influencing factors.

Main Results:

• JPKD-vancomycin showed higher accuracy for adjusted regimens (82.80% WRES < 30%) compared to initial regimens (63.87% WRES < 30%).
• Factors associated with inaccurate predictions included female gender, low body weight, short height, low vancomycin clearance (CLVan), and postoperative status.
• Multivariate analysis identified female gender, low CLVan, and postoperative status as independent risk factors for inaccurate predictions.

Conclusions:

• JPKD-vancomycin demonstrates improved predictive accuracy for adjusted vancomycin regimens.
• Predictive performance is compromised in female patients, those with low CLVan, and postoperative patients.
• Accuracy significantly decreases when CLVan is below 0.065 L×h⁻¹×kg⁻¹.