On the Analysis of Dissolution Data

Abstract

Dissolution profiles for solid dose forms represent several observations over time on an experimental unit such as a tablet or capsule. The pharmaceutical scientist is interested in a (statistical) comparison of these profiles under a variety of conditions relating to formulation characteristics, and lot-to-lot and brand-to-brand variation.

This paper discusses the analysis of dissolution profiles using an analysis of variance approach. la particular, the profiles are tested for differences in level and shape. The latter characteristic is potentially important with respect to learning about differences in the dissolution mechanism.

An approximate F test is discussed with the possibility of an arbitrary covariance matrix in mind, and alternative conservative analysis methods are presented.

Results from the analysis of variance of typical dissolution data show several important features. First, the analysis does not rely on curve fitting procedures and the data are used in their native form or as a simple transform. Second, the data can be analyzed using the fraction dissolved and/or the time for a particular fraction to dissolve, e.g., t_50%, as the variable of interest. Finally, the analysis is capable of showing differences between profiles where a realistic variation from profile-to-profile exists.