以都可喜为对照药的血管性认知功能障碍随机对照试验的效应值评价

Effect evaluation in randomized controlled trials for vascular cognitive impairment taking Duxil as known effective medicine

关键词:  都可喜  血管性认知功能障碍  血管性痴呆效应值

Key words:  Duxil  vascular cognitive impairment  effect value of vascular dementia

摘要:

[目的] 新的研究证据表明都可喜对血管性认知功能障碍无效，故重新评价以都可喜为阳性对照的临床试验中试验组治疗方法治疗血管性认知功能障碍随机对照试验的疗效。[方法] 通过计算机检索国内外医学数据库（截止到2012年8月），收集所有以都可喜作为阳性对照药的随机对照试验，根据Cochrane Handbook 5.0中的偏倚风险评估工具进行质量评价，由2名研究者独立提取资料与评价质量，并应用EpiCalc2000、SPSS 18.0进行各研究中试验组与都可喜对照组比较的效应值Cohen’d系数计算，以0.2为等效与优效临床试验的界值，0.2~0.3为临界区间。[结果] 共纳入19个血管性痴呆（VD）的临床研究（包括两个三臂试验共21个试验措施），其中6个试验措施与都可喜比较的效应值Cohen’d≥0.3，7个试验效应值位于0.2~0.3临界区间，5个小于0.2，3个对照组疗效优于治疗组。[结论] 19个研究21个试验措施中，6个干预措施优效于都可喜，8个试验措施的临床疗效值得商榷，其对血管性认知障碍的疗效有待高质量的安慰剂对照临床试验予以证实；另有7个组间比较的效应值位于0.2～0.3临界区间，其治疗措施的临床有效性亦值得进一步探讨。

Abstract:

[Objective] New evidence has showed that Duxil is not effective for treating vascular cognitive impairment (VCI), so it is necessary to re-evaluate the clinical effectiveness of clinical trials using Duxil as the positive control. This study aims to re-evaluate the efficacy of therapies for VCI in randomized Duxil controlled trials. [Methods] Chinese and English electronic databases were searched for randomized Duxil controlled trials for VCI till August, 2012. The evaluating tool of risk in bias assessment of Cochrane Handbook 5.0 was used for quality assessment. Two investigators extracted the data and assessed its quality independently. EpiCalc 2000 and SPSS18.0 were used for calculating the effect size (Cohen's d). The cutpoint between superiority and equivalence trial was set as less than 0.2, and 0.2～0.3 was taken as the margin interval. [Results] In total, 19 trials for vascular dementia were enrolled (two 3-armed trials, including 21 interventions). Among them, six interventions had a effect size (Cohen'D) not less than 0.3 or equal to 0.3. The 7 effect values lied in the margin interval between 0.2 to 0.3, 5 effect values less than 0.2, and the efficacy of the other 3 control groups were superior than treatment group. [Conclusion] In 21 experiments performance comparisons of 19 trials, 6 interventions are superior than Duxil control group, while 8 experiments performance are in need for further discussed. High quality placebo-controlled clinical studies on treatments of VCI are needed. For the 7 effect values lied in the margin interval between 0.2～0.3, their clinical effectiveness also needs further discussing.