The importance of knowing how vancomycin is measured when interpreting its pharmacokinetic results.

摘要

To the Editors: In our recent study, 6 published neonatal population pharmacokinetic models of vancomycin were evaluated with an independent external dataset, which consisted of 112 vancomycin trough concentrations (Co) obtained from 78 neonates, to identify the potential study-related factors that might limit transferability of published models in different clinical settings.7 The results demonstrated notable differences in predictive performances of models, which could be partly explained by taking into account the method used to measure serum creat-inine concentrations. When a conversion factor of 1.3 (enzymatic to Jaffe) was evaluated, an important improvement of predictive performance as illustrated by the simulation-based diagnostic normalized prediction distribution errors (NPDE) was observed (Table 1). However, the model developed by Allegaert et al still underpredicted the vancomycin concentrations in the external evaluation dataset (mean NPDE of 0.63, Table 1).