Pharmacokinetic prediction for intravenous beta-lactam antibiotics in pediatric patients

A method for predicting pharmacokinetics in pediatric patients for intravenous beta-lactam antibiotics is proposed. We focused on the allometric relationships of pharmacokinetic parameters with individual body weights (BW) in human including healthy adults and pediatric patients. Drug concentration data for 15 intravenous beta-lactam antibiotics were collected retrospectively from the published articles and the individual pharmacokinetic parameters were re-calculated. A mixed effect modeling (MEM) was applied for the allometric relationship for those beta-lactam antibiotics, and mean and variances of inter-drug variability for the allometric parameters and also variance for intra-drug (residual) variability were estimated. Then drug-specific allometric parameters were estimated by an empirical Bayesian method using the pharmacokinetic parameters for a drug only in healthy adults as observations, and finally the individual pharmacokinetic parameters in pediatric patients were predicted. The predictability of the method was evaluated by the leave-one-out method. We also demonstrated a method for simulating plasma concentration-time profiles in pediatric patients, and the predicted time-course curves generally coincided well with the actual plasma concentration data for the tested drugs.