线性回归模型在盐酸二甲双胍缓释片降糖效能非劣效性检验中的应用  

APPLICATION OF LINEAR REGRESSION MODEL IN NON-INFERIORITY TEST TO EVALUEATE THE EFFICACY OF METFORMIN HYDROCHLORIDE SUSTAINED-RELEASE TABLET IN TYPE 2 DIABETIC PATIENTS

作　　者：陈卫[1] 徐利娜[1] 迭敏[1] 刘丽珍[1] 陈卫中[2] 

机构地区：[1]成都医学院医学检验系医学检验专业2005级,成都610083 [2]成都医学院公共卫生学教研室

出　　处：《现代预防医学》2010年第3期407-409,共3页Modern Preventive Medicine

基　　金：2008年成都医学院大学生创新性实验项目(SYSCX2008-46)

摘　　要：[目的]探讨盐酸二甲双胍缓释片对2型糖尿病病人的临床效果和安全性。[方法]以二甲双胍普通片为对照,对76例患者在2家医院进行了6周的随机对照试验研究。分别在基线和疗后6周,随访观测了血糖、血脂等糖尿病疗效相关指标。[结果]在控制了中心、基线水平的混杂后,利用线性回归模型进行非劣效性检验表明,各指标组间均有统计学意义(P≤0.05)。两组不良反应发生率差异无统计学意义(P﹥0.05)。[结论]盐酸二甲双胍缓释片在疗效可以认为非劣于普通片,安全性上和普通片相当,但由于剂型优势,其依从性要高于普通片,因此具有较强的临床推广价值。同时探索了一条在临床研究中利用线性回归模型,通过SAS软件包进行统计学检验的途径,并为类似问题的解决提供了理论参考。[Objective] To evaluate the efficacy and safety of metformin hydrochloride sustained-release tablet on blood glucose control in type 2 diabetic patients.[Methods] A randomized controlled multicenter clinical trial was conducted.76 Chinese type 2 diabetic patients were enrolled in 2 clinical centers.The patients were divided randomly into experimental group（metformin hydrochloride sustained-release tablet 1 500 mg,q.d,n=38） or control group（metformin hydrochloride tablets 500mg,t.i.d,n=38）.At baseline and 6 weeks after treatment,related laboratory and physical examinations were performed.[Results] There were significant differences between the two groups in efficacy indexes and no significant difference in the incidence rate of adverse reactions.[Conclusion] Results from the trial have shown that the efficacy of metformin hydrochloride sustained-release tablet in lowering hyperglycemia is non-inferior to metformin hydrochloride tablets.The safety of metformin hydrochloride sustained-release tablet is comparable to metformin hydrochloride tablet.At the same time,it has explore a right way to put forward equivalence,non-inferiority and superiority test using SAS software package,and provided a theoretical reference for similar problems.

关 键 词：2型耱尿病 盐酸二甲双胍缓释片 随机对照试验 非劣效性检验 线性回归模型 

分 类 号：R195.1[医药卫生—卫生统计学]