Overview of oral absorption models and modelling issues
Karlsson, Mats O. and Radojka Savić
Dept of Pharmaceutical Biosciences, Uppsala University, Sweden
Clinical trials with description of pharmacokinetics as an objective often have designs that do not allow a detailed description of the oral absorption process. First, studies that lack an intravenous reference dose will not allow estimation of the extent of absorption. Neither will such studies, without supplemental information and assumptions, unambiguously allow an estimation of rate of absorption. Second, most studies allocate only few samples in the time period of ongoing absorption. In this sense, most studies can be considered as “sparse” with respect to drug absorption.
Some of the processes that are involved in the absorption of drugs (e.g. dissolution, diffusion across the gut wall and hepatic first-pass) can often be characterized as first-order processes. First-order models are also by far the most common in the description of drug absorption. This, however, is probably a consequence of lack of data, as other processes contributing to the absorption may show other properties. Saturable processes that may show non-linearity include drug dissolution, drug transport (both uptake from and transport back into the intestine) and metabolism (both hepatic and in the gut wall). Such non-linearities are often identified from administration of different dose sizes and modeled as “dose-dependent”, although dependence on concentration can be a mechanistically more appropriate manner to handle such phenomena. Another type of complexity is that the rate of absorption may change in a discontinuous manner as a consequence of for example formulation disintegration, gastric emptying, food intake, bile release, pH changes and transitions from one intestinal region to another. Such discontinuities may cause rapid changes, multiple peaks and other, seemingly erratic, patterns in the plasma concentration-time profile. Thus, for the absorption, simple models are generally used to characterize a complex system. In comparison, the disposition model is often more complex for a better behaved system. The result is usually high variability in absorption parameters (accommodating model misspecification) and a high residual error magnitude in the absorption phase, although the latter is seldom identified.
Models that attempt to accommodate complex absorption have spanned from highly empirical models, involving series of variable rate zero-order inputs and spline functions, to mechanism-based models. This presentation will briefly review these models and modelling issues.