A multi-mechanistic drug release approach in a bead dosage form and in vivo predictions.

Our previous study has successfully prepared a combination of immediate release, enteric coated, and controlled release (CR) beads and mathematically modeled in vitro drug release characteristics of the combination based on the release profiles of individual beads. The objectives of the present study are to evaluate the combination and individual beads in vivo and to mathematically model in vivo drug input characteristics of the combination based on the in vivo input of individual beads. Beagle dogs were used as an animal model, and theophylline as a model drug. In vivo percent drug absorbed at different times (input function) after administration of a capsule bead dosage form was calculated using the Wagner-Nelson deconvolution method using intravenous injection of theophylline in each dog as a reference. The in vivo input functions of individual beads were each fitted to appropriate mathematical equations. The in vivo input function of the bead combination dosage form was calculated based on the individual mathematical equations (expected), and verified experimentally in vivo (experimental). The results showed that all bead dosage forms behave in vivo as defined in vitro. Enteric coated beads significantly delay the time to reach the maximum concentration of drug (tmax=4.9 h) compared to uncoated immediate release beads (2 h). The lag time of enteric coated beads is 1.1 h. CR beads showed both longer tmax (6 h) and mean residence time (MRT=9.7 h) compared to the uncoated immediate release beads (tmax=2 h and MRT=7.1 h) as designed in vitro. The in vivo input functions for the three individual beads can be fitted to equations as a function of square root of time. The combined bead dosage form showed tmax of 2.4 h and MRT of 7.9 h. The experimental and expected in vivo input profiles agreed to within +/- 12% (residues at individual data points). Our results suggest that the drug input function of a combined multi-mechanism oral dosage form can be predicted from the in vivo performance of individual formulations using the dog as an in vivo model.