Assessment of similarity factor using different weighting approaches

Abstract

The objective of the present work was to determine the value of similarity factor (f2) using different values of the optional weight (w) to consider the variability in dissolution data. Three approaches are proposed for establishing similarity of dissolution profiles using the Moore and Flanner equation. In the first approach,the optional weight (w) was calculated by taking the ratio of 50 to f2Th,where 50 was selected as it is the borderline value for similarity or dissimilarity of the batches and f2Th is the conversion factor which takes into account variability between samples at each timepoint. In the second approach,the optional weight (w) was calculated by taking the ratio of the absolute difference of the percentage of drug dissolved between reference (R) and test (T) formulations to 10% of percentage of drug dissolved from the reference formulation at each timepoint to consider the variability between samples with more specificity. In the third approach,the weight was calculated from the equation (1+SD/ maximum allowed SD) where standard deviation (SD) was calculated from 144 data points of absolute difference between R and T of 12 reference and 12 test formulations. The maximum allowed SD was arbitrarily chosen as 10 to consider within-sample variability as well as variability between samples. The calculation steps for estimating f2 are explained using the literature reported values of dissolution study. The results of the proposed approaches are compared with the classical approach of calculating f2 (w=1). In some cases,the status of similarity changed to dissimilarity when the proposed approaches were adopted. The use of all three approaches is recommended in borderline cases of similarity. On the basis of consideration of variability,the three approaches are given preference in the order of Approach 3> Approach 2> Approach 1. Approach 3 may arouse new interest in f2 among industrial pharmacists and regulatory personnel as the within-samples variability and between-samples variability in dissolution data are considered in calculation of f2.