Comparison of Innovative and Conventional Methods in Biosimilar Bridging Studies with Multiple References

Annpey Pong,1 Susan S Chow,2 Shein-Chung Chow3 1Biostatistics, Merck & Co Inc., Rahway, NJ, USA; 2Internal Medicine, The Wright Center for Community Health, Scranton, PA, USA; 3Department of Biostatistics and Bioinformatics, Duke University School of Medicine, Durham, NC, USACorrespondence: Annpey Pong, Biostatistics, Merck & Co Inc., Rahway, NJ, USA, Email [email protected]: For assessment of biosimilar drug products, if there are multiple-reference products (eg, a US-licensed product and an EU-approved product), a biosimilar bridging study with a 3-way pairwise comparison is often conducted. In our paper, two innovative methods in biosimilar bridging study are compared with the conventional method of pairwise comparisons. For parallel study design, the simultaneous confidence interval (CI) method is compared to the convention method. For crossover study design, the multiplicity-adjusted Schuirmann’s two one-sided tests (MATOST) is considered. This paper conclude that the simultaneous CI method achieves the similar statistical power to the conventional approach in biosimilarity assessment. However, the MATOST method using the conservative Holm and Bonferroni approaches is not favorable since it leads to a large sample size although it controls the type I error rate.Keywords: bioequivalence and biosimilarity, multiple references, simultaneous confidence interval approach, multiplicity-adjusted two one-sided tests, MATOST