摘要
由于生物制品具有复杂的特征和性质,为评估申请者的生物相似性证据,美国FDA建议使用"基于风险"和"证据链完备性"(totality-of-the-evidence)的方式来评估已递交的全部特征域间的所有可用数据和信息。因此,需要建立一个生物类似性指数,使该指数能将各个具体特征域中的生物相似性整合成为一个跨指标领域的总体指数(global index)。Shein-Chung Chow等人已经在2010年根据受试品-参考品(T-R)与参考品-参考品(R-R)比较研究中重现性概率(reproducibility probability)的概念,提出了一个评估相似性通用的方法,该方法可用于"证据链完备性"的评估。本文提出另外一个生物相似性评价指数,该指数基于配对随机设计下随机系数的线性回归模型,并提供了相应的统计检测步骤。文中用一个数值型实例对所提出的方法进行了证明。
As the complexity of characteristics and nature of biologic products, for evaluating a sponsor＇s demonstra- tion of biosimilarity, the United States Food and Drug Administration（FDA） suggests using a risk-based, totality- of-the-evidence approach to evaluate all available data and information submitted in support of the regulatory sub- mission across different domains. Thus, the development of a biosimilarity index which can integrate local biosimilar- ity in each specific domain into the global index across domains is required. Shein-Chung Chow et al. proposed a general approach for assessing biosimilarity based on the concept of the reproducibility probability in a T-R study as compared to that of an R-R study, which can be used for assessing totality-of-the-evidence. In this article, an alter- native biosimilar index based on linear regression model with random coefficients under a paired-randomized design and corresponding statistical testing procedures are proposed. A numeric example is provided to illustrate the pro- posed methods.
出处
《药物分析杂志》
CAS
CSCD
北大核心
2015年第5期837-843,共7页
Chinese Journal of Pharmaceutical Analysis
关键词
证据链完备性
通用相似性
生物相似性指数
线性回归
随机系数
totality- of- the- evidence
global biosimilarity
biosimilarity index
linear regression
random coeffi-cients
