Predictability of Vancomycin Trough Concentrations Using Five Predictive Methods

Abstract

OBJECTIVES: Five methods for estimating vancomycin pharmacokinetic parameters were studied to determine the accuracy of each method in predicting vancomycin concentration. METHODS Patient vancomycin pharmacokinetic data from a prior study were retrospectively reviewed and used in five methods to calculate vancomycin clearance and volume of distribution in order to determine the accuracy in prediction of the measured value. RESULTS: The coefficients of determination ranged from 0.167 to 0.224, bias ranged from -5.18 to 2.13, and precision ranged from 5.98 to 7.64. The Buelga method had the highest percentage of predictions within 2.5 and 5 mg/L of the measured trough at 42% and 65% respectively, while the Thomson method had the highest percentage of predictions within 50% of the measured trough at 52%. The highest percentage of predictions within 25% of the measured trough were found with both the Thomson and Yasuhara method at 29%. CONCLUSION: There was wide variation in the prediction of vancomycin trough concentrations from the five methods for estimating vancomycin pharmacokinetic parameters. None of the methods provided adequate reliability to recommend discontinuation of therapeutic drug monitoring for vancomycin.