Comparison of the reference scaled bioequivalence semi-replicate method with other approaches: focus on human...

To compare the performance of the reference scaled average bioequivalence (scABE(R)) method proposed by FDA scientists [Haidar et al., 2008. Pharm. Res. 25, 237-241] with other approaches focusing on the human exposure expressed as the product sample size x periods of drug administration. Simulated bioequivalence studies were generated assuming the partial replicate 3-way crossover design and the classic (2 x 2) crossover design. Intrasubject variability (CV(W)) values ranged from 15% to 60% and sample sizes from 16 to 54. The procedures examined include: the scABE(R) method, the classic 0.80-1.25 approach, a levelling-off scaled BE limit (BELscW), and some other scaled bioequivalence limits. To assess the performance of the aforementioned approaches, the typical as well as novel three-dimensional modified power curves were constructed. A new index, termed %Mean Relative Difference (MRD%), was introduced in order to quantitatively compare the performance of the bioequivalence limits. The recently proposed scABE(R) approach showed the lowest producer risk in particular for highly variable drugs. When exposure was taken into account scABE(R) resulted in a desired behaviour when CV(W) was low. For high CV(W) values the overall performance diminished when geometric mean ratio (GMR) substantially deviated from unity. Application of the MRD% index clearly revealed that the effect of lowering the producer risk at GMR=1 was totally counterbalanced by the rise of consumer risk at high GMR values. The classic 0.80-1.25 limits were favoured at low intrasubject variability and high exposure, whereas the levelling-off limits demonstrated a preferred overall performance when variability was high and exposure was limited.