多水平模型在生物等效性评价中的应用(Ⅰ)

摘要

探讨多水平模型在生物等效性评价中的应用价值.以2×4试验设计的抗高血压药生物等效性评价为研究实例,研究多水平模型对效应指标值的变异即方差的分解方式,并与FDA推荐的矩法所获得的方差分量进行比较.对比传统FDA推荐的生物等效性评价标准,研究利用多水平模型直接进行平均等效性、群体等效性和个体等效性评价的可行性.对于2×4试验设计的单变量两水平模型获得ln(AUC)指标的方差分量如试验药T的总方差σ_(TT)~2、个体间方差σ_(BT)~2和个体内方差σ_(WT)~2以及参比药R的总方差σ_(TR)~2、个体间方差σ_(BR)~2和个体内方差σ_(WR)~2,与FDA推荐的矩法所获得的结果非常接近.实际应用中,根据FDA提出的生物等效性评价的标准和程序进行评价,直接用多水平模型的估计值进行平均、群体和个体等效性评价,两者结果一致.多水平模型适合于交叉设计的生物等效性评价,相对于FDA推荐的方法,多水平模型对于复杂的有影响因素的交叉试验设计更容易估计方差分量,进而可以评价平均、群体和个体等效性,实际应用上更具有灵活性,为生物等效性评价提供了新的思路和方法.%This study aims to explore the application value of multilevel models for bioequivalence evaluation.Using a real example of 2×4 cross-over experimental design in evaluating bioequivalence of antihypertensive drug,this paper explores complex variance components corresponding to criteria statistics in existing methods recommended by FDA but obtaines in multilevel models analysis.Results are compared with those from FDA standard Method of Moments,specifically on the feasibility and applicability of multilevel models in directly assessing the bioequivalence(ABE),the population bioequivalence(PBE)and the individual bioequivalence (IBE).When measuring ln(AUC),results from all variance components of the test and reference groups such as total variance(σ_(TT)~2 and σ_(TR)~2),between-subject variance(σ_(BT)~2 and σ_(BR)~2)and within-subject variance(σ_(WT)~2 and σ_(WR)~2)estimated by simple 2-level models are very close to those that using the FDA Method of Moments.In practice,bioequivalence evaluation can be carried out directly by multilevel models,or by FDA criteria,based on variance components estimated from multilevel models.Both approaches produce consistent results.Multilevel models can be used to evaluate bioequivalence in cross-over test design.Compared to FDA methods,this one is more flexible in decomposing total variance into sub components in order to evaluate the ABE,PBE and IBE.Multilevel model provides a new way into the practice of bioequivalence evaluation.