Case Studies for Statistical Analysis of Toxicokinetic Data

The present paper discusses aspects of the statistical analysis of toxicokinetic data by reference to 102 case studies. In toxicokinetic studies, dose dependence of exposure is the most important subject. For this purpose, and because variability in concentration is closely related to the mean value, before beginning the standard procedures of analysis of variance and regression analysis concentration data should be transformed into logarithms. This data transformation also assists assessment of nonlinear kinetics. The paper discusses a special issue in the analysis of experiments for which measurement on each subject has been repeated. A simple procedure is proposed for estimating the daily exposure level in rodent studies when test chemicals have been mixed in diet and exposure level is measured on different animals at each time point. Experimental designs are classified into three types, and these are illustrated by data from representative case studies. Conclusions from 102 case studies are summarized, with attention to how often sex differences, change in exposure level during repeated dosing, and how often evidence is strong for nonlinear kinetics. Various key points of study design are discussed.