Karlsson M.(1) , Milligan P.(2) , Nichols D.(2)
Division of Biopharmaceutics and Pharmacokinetics, University
of Uppsala, S-75123 Uppsala, Sweden(1)
Early Clinical Research Group, Pfizer Central Research,
Sandwich, Ken CT13 9NJ, United Kingdom(2)
One of the tasks in drug development is to assess what molecular species are contributing to the effect and toxicity of a drug and the relative importance of such entities. In vitro assays, in situ models and in vivo animal experiments may provide qualitative and semi-quantitative information, but only experiments in man can provide the quantitative information necessary to assess the contribution of any metabolites to the therapeutic outcome. Theoretically appealing as it may be to administer the metabolite alone, this is often not practically possible. At least one criterion has to be fulfilled if one is to determine the drug/metabolite potency ratio in vivo after the administration of the drug alone: drug/metabolite ratios have to vary between observations of the effect in an individual. For a continuous, frequently measured effect variable, this will generally not constitute a problem. Using real data and a population model, we have estimated the in vivo potency ratio of drug and metabolite when only oral doses were administered. This was done from multiple oral doses in the presence of intraindividual variability in the pharmacodynamic parameters and a delay in the observed response compared to the concentration-time profile. The estimated relative potency and its confidence interval was consistent when different models for the delay (no delay, the link model and an indirect effect model) of the parameter variability were used. It could also be shown that the estimated potency ratio resulted in better predictions of new effect data than the assumption of equipotency that was supported by preclinical data.