Estimation of absorption ratios and their variation for orally administered drugs.

Differences in the plasma concentrations of drugs after oral administration are caused by two main factors: variation in absorption ratios and in the distribution processes in the body. A new method for the dissection of both types of factors is discussed. The method uses a reference regression of the AUC-values to the corresponding values after intravenous infusion of graded doses. The reference regression is estimated from an appropriate trial. Deviation of the determined AUC-values from the regression curve afford an estimate of the residual variance due to varying distribution volumes or similar random biological effects. For the estimation of absorption ratios after oral administration the drug is given orally to another sample of subjects and their AUC-values are calculated. The deviation of these AUC values due to the above mentioned random effects are simulated using the residual variance of the reference regression, and are subtracted from the observed AUC-values. Then, the differences in the corresponding absorbed doses are transformed by inverting the reference regression. From these doses the empirical distribution function and statistical parameters (e.g. quantiles) are determined. The method has the advantage that no restrictive assumptions are required, such as first order processes, dose linearity, homogeneity of variance or normal distribution of absorption ratios. Its applicability to substances with qualitative differences in their pharmacokinetics is demonstrated by appropriate examples.