Estimators and Confidence Intervals of f_2 Using Bootstrap Methodology for the Comparison of Dissolution Profiles

Resumen

Dissolution tests are essential in the development of medicinal product, but there are many methods to compare dissolution profiles. The most widely used one is the similarity factor f_2. Nevertheless, the f_2 method has several drawbacks, which lead to certain restrictions described in regulatory guidelines, e.g. when variability of dissolution data is more than 20 % and 10 % for early and later time points, respectively, the f_2 method cannot be used. In such circumstance, alternative methods are recommended in several regulatory guidelines, for instance, the model-independent multivariate statistical distance methods (MSD). However, members from US FDA indicated in 2013 that the model-independent MSD method is less discriminative and sensitive than the f_2 method. Therefore, they recommended the confidence interval of f_2 approach using bootstrap. Recent studies comparing the MSD method with the confidence interval of f_2 approach with bootstrap method confirmed these findings. However, the guidelines neither specify the estimator nor the type of confidence interval to be used and literature with this regard is scarce. Therefore, we investigated the accuracy and precision of several estimators and types of confidence intervals by simulation.