Potential of population pharmacokinetics to reduce the frequency of blood sampling required for estimating kinetic parameters in neonates.

Data obtained from neonates receiving zidovudine as part of a phase I study were used to estimate the population pharmacokinetic parameters of this drug and to determine the minimum number of data points necessary to provide accurate estimates of the kinetic parameters and their variability. Analysis was performed with 541 concentrations of zidovudine, obtained from 32 infants and with a variety of reduced data sets using NONMEM (nonlinear mixed effect model). The reduced data sets were derived by randomly reducing the number of sampling time points per dosing interval and/or by randomly reducing the number of available subjects. We determined that accurate estimates of pharmacokinetic parameters and their variability were obtained with the inclusion of all 32 patients using only two concentration-time points per dose interval, provided that one of the points was obtained during the first 2 h after administration of the drug. The parameters themselves were adequately estimated with only 24 subjects and two concentration-time points per dose interval. We suggest that NONMEM should be used in addition to the traditional pharmacokinetic analysis to obtain more precise information directly in the population of interest with a minimum of blood sampling from each patient. This is especially critical in infants whose blood volumes are limited.