Population PK/PD modelling of meropenem in preterm newborns based on therapeutic drug monitoring data

Background: Preterm neonates rarely participate in clinical trials, this leads to lack of adequate information on pharmacokinetics for most drugs in this population. Meropenem is used in neonates to treat severe infections, and absence of evidence-based rationale for optimal dosing could result in mismanagement. Aim: The objective of the study was to determine the population pharmacokinetic (PK) parameters of meropenem in preterm infants from therapeutic drug monitoring (TDM) data in real clinical settings and to evaluate pharmacodynamics (PD) indices as well as covariates affecting pharmacokinetics. Materials and methods: Demographic, clinical and TDM data of 66 preterm newborns were included in PK/PD analysis. The NPAG program from the Pmetrics was used for modelling based on peak-trough TDM strategy and one-compartment PK model. Totally, 132 samples were assayed by high-performance liquid chromatography. Meropenem empirical dosage regimens (40–120 mg/kg/day) were administered by 1–3-h IV infusion 2–3 times a day. Regression analysis was used to evaluate covariates (gestation age (GA), postnatal age (PNA), postconceptual age (PCA), body weight (BW), creatinine clearance, etc.) influenced on PK parameters. Results: The mean ± SD (median) values for constant rate of elimination (Kel) and volume of distribution (V) of meropenem were estimated as 0.31 ± 0.13 (0.3) 1/h and 1.2 ± 0.4 (1.2) L with interindividual variability (CV) of 42 and 33%, respectively. The median values for total clearance (CL) and elimination half-life (T1/2) were calculated as 0.22 L/h/kg and 2.33 h with CV = 38.0 and 30.9%. Results of the predictive performance demonstrated that the population model by itself gives poor prediction, while the individualized Bayesian posterior models give much improved quality of prediction. The univariate regression analysis revealed that creatinine clearance, BW and PCA influenced significantly T1/2, meropenem V was mostly correlated with BW and PCA. But not all observed PK variability can be explained by ...