針對非對數常態資料的生體相等性檢定

Bioequivalence test for non-lognormally distributed data

詹國松(Kuo-sung Chan)

指導教授 ： 陳玉英

國立中央大學/理學院/統計研究所/碩士(2012年)

摘要

生體相等性試驗在學名藥的審核中扮演重要的角色，一般是針對自願
參與的健康的成人進行兩期雙序列交叉設計(2×2 crossover design)之下進行實驗。每一位參與者分別在不同時期施予兩種不同的藥物，一為原廠藥，另一為待評估的學名藥。然後參與者在服藥後一段時間內重複抽血，測量血液中含有該藥物的濃度，稱為藥物濃度時間側寫，一般都是根據藥物濃度時間側寫下的面積（記作AUC），檢定這兩種藥物的生體相等性。針對這些參與者個別的AUC資料，傳統上都假設服從對數常態分布。鑒於實務上的需要本文考慮為逆伽瑪分布。因此，本文將在兩期雙序列交叉設計下，配適對數AUC的隨機效應模式，除了考慮實驗對象之間的變異外，也假設誤差項為逆伽瑪分布。本文使用SAEM演算法(stochastic approximation expectation maximization algorithm )估計模式中的參數，然後根據估計的AUC均值進行這兩種藥物是否具有生體相等性。本文除以模擬方式研究所提檢定的型I誤差率及檢定力表現，最後藉由分析一筆實例說明所提模式及檢定的應用。

關鍵字

SAEM演算法 ； 逆伽瑪分布 ； 兩期雙序列交叉設 ； 生體相等性 ； 生物可用率

並列摘要

In a pharmacokinetic (PK) study, to claim a test drug under study as a generic drug, proof of the bioequivalence between the test drug and a comparative reference drug is needed. To do so, some healthy volunteers are recruited and administered with the two drugs in a 2×2 crossover design with a reasonable wash-out time period, where the volunteers in one sequence receive the reference drug and then the test drug in two different periods, while the volunteers in the other sequence take the drugs in reverse order in the two periods. After the drug is administered to each volunteer, the drug concentrations in blood or plasma at different time points are then measured, which is referred to as the drug concentration–time curve or profile. The average bioavailability parameters such as the area under the drug concentration–time curve (AUC) is conventionally of interest for assessing the bioequivalence of the test drug to the reference drug. Conventionally, however, the distribution of the logarithm of individual AUC (denoted by logAUC) is followed the lognormal distribution. In practice, this assumption is violated and hence, in this thesis, we propose an alternative distribution, inverse gamma distribution, to satisfy the assumption of the distribution of individual logAUC. In this thesis, we consider to construct the model of individual logAUC which has subject variation and the error term are distributed by normal distribution and inverse gamma distribution, respectively, under the 2x2 crossover design. We consider using the stochastic approximation expectation- maximization algorithm to find the maximum likelihood estimates of the parameters. Then, the bioequivalence test of two drugs is inducted by estimated mean AUC. We further present some results of a simulation study investigation of the level and power performances of the purposed method and the application of the proposed test is finally illustrated by using a real data.

並列關鍵字

SAEM algorithm ； inverse gamma ； Bioequivalence ； bioavailability ； 2×2 crossover design

參考文獻

2.Cox,C. , Chu,H. , Schneider,F. and Munoz,A. (2007). “Parametric survival analysis and taxonomy of hazard functions for the generalized gamma distribution.” Statist. in Medicine, 26, 4352-4374.

3.Cristian Meza • Felipe Osorio • Rolando De la Cruz(2010). Estimation in nonlinear mixed-effects models using heavy-taileddistributions. Springer Science+Business Media, LLC 2010

7.Louis, T.A. (1982): Finding the observed information matrix when using theEM algorithm. J. R. Stat. Soc. Ser. B 44, 226–233

8.Meza, C., Jaffrézic, F., Foulley, J.-L. (2009):Estimation in the probit normalmodel for binary outcomes using the SAEM algorithm. Comput.Stat. Data Anal. 53, 1350–1360

9.Wei, G., Tanner, M. (1990): A Monte Carlo implementation of the EM algorithm and the poor man’s data augmentation algorithms. J. Am.Stat. Assoc. 85, 699–704