An evaluation of population pharmacokinetics in therapeutic trials. Part III. Prospective data collection versus retrospective data assembly.

The ability to estimate population pharmacokinetic parameters of alprazolam from fragmentary clinical data collected during clinical efficacy trials was evaluated in this series of studies. In the first study modifications to the protocol and case report forms were employed to allow for the prospective collection of accurate dosing and blood sampling time data. In the second study only total daily dose and measured drug concentration data were recorded. Dosage and blood sampling time data were assembled retrospectively by assuming strict adherence to protocol guidelines. Comparisons of the results of these analyses demonstrate the marked impact of prospective data collection efforts on data quality and the ability to determine subsequent pharmacokinetic parameters. Analysis of the prospectively collected data yielded pharmacokinetic estimates for alprazolam that were nearly identical to previously reported values obtained according to traditional methods, whereas the analysis of retrospectively collected data yielded biased estimates of the mean pharmacokinetic parameters and markedly upward biased estimates of both interindividual and residual variability. These analyses demonstrate the ability to estimate population pharmacokinetic parameters from data collected during clinical efficacy trials, provided that critical issues pertaining to the design and use of data collection forms to enhance data quality and the education of patients and staff are satisfactorily addressed.