Abstract
Sampling constraints and ethical concerns are two of the issues that prompt the conduct of pediatric/animal cross-sectional pharmacokinetic studies. A simulation study was carried out to investigate the effect of optimization of design features (arrangement of concentrations (n) in time and sample size (N)) on the accuracy and precision of parameter (especially intersubject/interanimal variability) estimation in cross-sectional population pharmacokinetic studies involving drugs administered by single or multiple intravenous bolus input(s). Drugs exhibiting one and two compartment (i.e., 1 CMT and 2 CMT) pharmacokinetics were investigated using N of 20 to 100 and 30 to 150, respectively, n was optimized, using the profile (block) randomized sampling design, for accurate and precise estimation of population pharmacokinetic parameters. N of 50 and 80 was found to be adequate for the estimation of clearance (CL) and its variability ((ωCL) needed for dosage optimization) for the 1 CMT and 2 CMT drugs, respectively. Estimates of ωV (1 CMT), and most of the parameters of the drug exhibiting 2 CMT pharmacokinetics were biased and imprecise. the bootstrap resampling of weighted residuals was found to provide only a limited (= 15%) bias correction in the estimation of ωV and this may not be an adequate correction for the study setting. Cross-sectional pharmacokinetic study was not useful in the estimation of population pharmacokinetic parameters of highly variable drugs. However, it was found to be useful in the estimation of the relevant population pharmacokinetic parameters required for dosage optimization of non-highly variable drugs in the pediatric/animal population.