Individualized adjustment of vancomycin dosage: comparison with two dosage nomograms

An individualized method of vancomycin dosage adjustment using steady-state serum concentrations was assessed in 50 patients (86 sets of vancomycin serum concentrations). The predictive accuracy of this method was compared with that of two published nomograms (Moellering, Matzke). Peak and trough serum concentrations predicted from previously drawn vancomycin serum concentrations (individualized method) using a one-compartment pharmacokinetic model were compared with measured steady-state peak and trough serum concentrations. Predicted peak and trough serum concentrations were also generated for both the Moellering and Matzke nomograms by using the elimination rate constant derived from each of the respective nomograms and the fixed volume of distribution (0.9 L/kg) assumed by the nomograms, and the actual administered vancomycin dose and dosage interval. These predicted concentrations were also compared with the measured peak and trough concentrations. Statistical measures of bias and precision indicated that the individualized method of dosage adjustment more closely predicted vancomycin serum concentrations following a dosage change than did either of the nomograms. Overall, the Moellering nomogram was the least accurate of the three methods in predicting vancomycin serum concentrations, and this nomogram should not be used to titrate vancomycin dosages in a clinical setting. Adjustment of vancomycin dosages should be individualized based on pharmacokinetic data derived from measured serum concentrations. In situations where quantitative analysis of vancomycin concentrations is not available, the Matzke nomogram appears to be a reasonable method of adjusting vancomycin dosages.