Statistical Methodology for Highly Variable Compounds: A Novel Design Approach for the Ofatumumab Phase 2 Bioequivalence Study

This article describes a novel mixed-scaling testing strategy, in combination with an adaptive parallel-groups bioequivalence trial, to test pharmacokinetic equivalence of two formulations of a drug with highly variable pharmacokinetics. The methodology was applied to the Phase 2 APLIOS study in relapsing multiple sclerosis patients, where the bioequivalence of subcutaneous ofatumumab 20 mg administered via an autoinjector pen (test formulation) versus prefilled syringe (reference formulation) in the abdomen has been studied. Due to a high coefficient of variation (CV) of the relevant pharmacokinetic metrics (AUC(tau) and C-max) for the reference formulation (>30%), a reference-Scaled bioequivalence (RSABE) approach was applied but modified for a parallel-groups design. In the absence of regulatory guidance for applying RSABE in parallel-group designs, and in contrast to the available regulatory guidance for RSABE in cross-over trials, the suggested testing strategy uses the between-subject variability of the reference drug instead of the corresponding within-subject variability that would be available if a standard cross-over bioequivalence trial had been possible. Moreover, due to high uncertainty in the initial CV estimate for the sample size determination, a two-stage adaptive design was used, allowing for a sample size adjustment based on the pharmacokinetic variability observed at an interim analysis. The interim analysis timing was pre-specified based on simulations which included re-estimation of the final sample size at the end of the first stage to ensure sufficient power of the RSABE test at the end of the second stage. Using this approach, bioequivalence was shown between the test and reference formulations. The APLIOS trial is registered at : NCT03560739.