Objective For assessment of biosimilarity of biosimilar products, the U. S. Food and Drug Administration proposed three evaluation approaches based on the levels of stringency in analytical studies. The evaluation approaches are applied to the three tiers of critical quality attributes(CQAs) based on the criticality ranking and other factors. In this article, we focus on discussing the statistical methods applicable and sample size nT, nR in tier 1. Methods We introduce an adapted F-test for homogeneity of variances and the expression of the hypothesis and test statistics to assess biosimilarity in variability. We propose a power equation based on the two one-sided tests(TOST) procedure and bioequivalence(BE) for highly variable drugs. Exact calculation of sample size can be obtained through numerical iteration of the equation. Results We proposed the idea for the selection of nR:nR ∈ [nT, 1.5nT] in analytical similarity assessment. Based on the results of simulation for various parameter combinations, we also propose the method for the selection of nT, nR and the process of analytical similarity assessment. Conclusion The trategy for the selection of nT, nR and the process in this article are practically applied in analytical similarity assessment. We can get the other sample size easily for various parameter combinations through the exact formulas for the power and sample size.%目的 在生物类似药的质量比对研究中,为评价候选药与参照药的分析相似性,FDA建议对3个不同层级的关键质量属性采用不同的评价方法.本文建立第1层级的关键质量属性分析相似性评价的步骤及方法,重点讨论如何确定候选药与参照药的样本量nT,nR.方法 利用判断方差齐性的F检验介绍方差相似性检验假设及统计量的建立;利用高个体差异药物的生物等效性评价方法及双单侧检验的性质等建立分析相似性检验假设及统计量,并由推导得到的检验效能的精确公式确定样本量.结果 在分析相似性评价过程中将nR控制在[nT,1.5nT]内,根据数值模拟得到不同参数组合的样本量,提出了确定nT,nR的策略,在此基础上给出了评价生物类似药分析相似性的具体步骤和方法.结论 本文提出的样本量确定策略及评价分析相似性的步骤与方法可应用到实际的分析相似性研究中;同时由推导得到的检验效能精确公式还可获得其他参数组合的样本量.
展开▼
