抄録
We performed a simulation for the clinical pharmacokinetic study, in which blood was sampled at two time points corresponding to the peak concentration (Cpeak) and trough concentration (Ctrough) following repetitive oral drug administration to subjects. We estimated the approximate oral clearance (CL/Fapprox) as 2·D/(Cpeak·τ+Ctrough·τ), where D is the dose, and τ is the dosing interval. The CL/Fapprox value was accurate for drugs with a long-elimination half-life, and the estimation error of the CL/F value was slightly increased for drugs with a shorter elimination half-life. The accuracy of CL/Fapprox in each subject was not affected by the magnitude of the interindividual pharmacokinetic variability, but was significantly decreased by the larger measurement error of drug concentrations (or intraindividual pharmacokinetic variability). We further performed several computer simulations to mimic statistical hypothesis testing following the clinical repeated-dose pharmacokinetic trials. The statistical power to detect the difference of oral clearance between two groups was marginally dependent on the measurement error of drug concentration, but was highly dependent on the interindividual pharmacokinetic variability. These findings suggested that the peak-and-trough sampling design to estimate the CL/Fapprox value is useful for clinical repeated-dose pharmacokinetic trials, and that the study design and protocol should be evaluated carefully by computer simulation prior to a real clinical trial.
