A Novel Simulation-Based Approach for Comparing the Population Against Average Bioequivalence Statistical Test for...

The aim of this work is to evaluate average bioequivalence (ABE) and population bioequivalence (PBE) statistical approaches so as to identify which approach is most suitable for in vitro bioequivalence (IVBE) testing of nasal spray products. For droplet size distribution (DSD) and spray pattern (SP), in vitro data were collected using a well-established nasal spray on the market (Nasonex®, manufactured by Merck Sharp & Dohme Limited). Simulations were performed using in vitro data to comparatively investigate ABE and PBE tests. For highly variable parameters such as SP area, this study clearly demonstrates that the level of agreement between ABE and PBE test conclusions is much smaller as compared with that of DSD Dv(50), which was found to have moderate variability. PBE approach dictates equivalence for both means and variances, and was found to handle both SP and DSD parameters with similar passing rates compared to the passing rates from the ABE approach. However, pronounced asymmetric behavior of PBE empirical power curves for highly variable SP area was observed. A modified PBE statistical approach is proposed for DSD span and Dv(50) in vitro parameters, where acceptance criteria would be based on comparison of reference/branded product to itself as part of "pre-IVBE study" via innovative statistical bootstrap simulations. Due to inherent high variability of the SP area parameter driving pronounced asymmetric behavior of PBE power curves, and due to unclear in vivo relevance for SP area and ovality, authors propose that SP parameters be used as development and quality control tools rather than for demonstration of IVBE.