药物长期毒性试验血液和生化指标的统计处理  

Statistics for hematological and serum biochemical parameters in chronic toxicity studies of drug safety evaluation

作　　者：张晓冬[1] 宗英[1] 陆国才[1] 黄矛[1] 袁伯俊[1] 

机构地区：[1]第二军医大学基础部新药评价中心,上海200433

出　　处：《中国新药杂志》2010年第4期270-272,289,共4页Chinese Journal of New Drugs

摘　　要：目的:探讨药物长期毒性试验血液和生化指标的统计处理。方法:分析长期毒性试验血液和生化指标错误的统计处理,为药物的临床前研究提供合适的统计方法。结果:应用t检验进行啮齿类长期毒性试验多个样本间的两两比较,会加大犯I类错误的概率。非啮齿类长期毒性试验中血液和生化指标是重复测量的资料,应先用球对称检验判断资料之间的相关性,然后用重复测量的方差分析进行组间和不同时间点的比较。结论:啮齿类长期毒性试验多个样本间的两两比较应采用方差分析。非啮齿类长期毒性试验中血液和生化指标应用重复测量的方差分析较为合适。Objective： To study the statistical method in chronic toxicity studies of drug safety evaluation.Methods： We analyzed the mistakes in statistics for hematological and serum biochemical parameters in chronic toxicity studies,and offered a proper statistical method for preclinical medicine researchers.Results： Type I errors in comparing differences among groups in chronic toxicity studies of rodents were increased when t test was used.The hematological and serum biochemical parameters in chronic toxicity studies of nonrodents are repeatedly measured data.For these data,Mauchly′s test of sphericity was used to judge whether there were relations.Then,repeated measures analysis of variance was used to compare differences among groups and in different measure time.Conclusion： Analysis of variance should be taken for multiple comparisons in chronic toxicity studies of rodents.Repeated measures analysis of variance is suitable for variance analysis of hematological and serum biochemical parameters in chronic toxicity studies of nonrodents.

关 键 词：统计 长期毒性 方差分析 

分 类 号：R965.1[医药卫生—药理学]