Teicoplanin population pharmacokinetic analysis in hospitalized patients.

The goal of this study was to build a population pharmacokinetic (PK) model to characterize the population PK parameters in our hospitalized patients. Teicoplanin serum concentrations from clinical routine were used. Antibiotic dose history and blood collection times were recorded and analyzed with NONMEM-V. Demographic and biologic data creatinine clearance (CLcr), weight (WT), and albumin (Alb) were tested for inclusion as covariates in the basic model. Intraindividual and residual variability were modeled. One hundred seven sparse samples (mainly trough levels), from 79 patients, were included. A 2-compartment PK model characterized by clearance (CL), central compartment volume of distribution (Vc), intercompartment clearance, and steady-state volume of distribution (VSS) with first-order elimination adequately described the data. CLcr and WT significantly influenced teicoplanin CL (CL = 0.57[0.15]*(1+0.0048[0.39]*(CLcr - averageCLcr)*WT) L/h). VSS was not affected by any covariate (VSS = 50.2[0.13]L). A negative trend between Alb and individual VSS estimates was observed without statistical significance. In a new data set, bias and precision resulted in mean values of -3.24% and 9.42%, respectively. In conclusion, CLcr and WT are significant covariates on teicoplanin CL. Results from predictive accuracy and precision show the usefulness of this model for implementation in a therapeutic drug monitoring program in the near future.