国内脑卒中临床试验疗效判断方法中改良Rankin评分的应用现状

The Application of m RS in the Methods of Outcome Assessment in Chinese Stroke Trials

目的分析国内脑卒中临床试验疗效判断方法中改良Rankin评分应用的现状和存在的问题。方法人工检索中国知网、万方、维普三大中文数据库中2013年9月之前发表的文献。纳入所有研究脑卒中治疗措施、自述为随机或半随机临床对照试验、疗效评价指标中使用改良Rankin评分(modified Rankin scale,m RS)的人体研究报告。排除关于蛛网膜下腔出血和TIA的试验。对脑卒中类型、m RS分值的统计分析方法、随访时间、是否采用盲法判断疗效、干预类型及阳性结果率等方面进行分析。结果共有298个试验使用m RS,其中缺血性卒中212个(71.14%),出血性卒中64个(21.48%),包含缺血性卒中和出血性卒中的22个(7.38%)。急性脑卒中(发病时间<14d)272个(91.28%)。在对m RS分值的统计分析中按连续性资料进行分析并采用t检验或方差分析的有149个(50.00%),按等级资料采用秩和检验或多分类卡方检验进行分析的有66个(22.15%),将资料转换为二分类资料进行统计分析的有76个(25.50%)。将m RS与其他评分量表一起作为卒中疗效判定指标的有37个(12.42%)。随访时间10d至2年,中位时间90(30,90)d。仅15个试验(5.03%)描述使用盲法对疗效进行评价。药物治疗179个(60.07%),康复治疗23个(7.72%),外科治疗31个(10.40%%),综合治疗43个(14.43%),管理方式7个(2.35%),护理2个(0.67%),其他治疗方式13个(4.36%)。259个(86.91%)试验显示出有统计学意义的阳性结果。结论国内脑卒中临床试验疗效判断方法中m RS的应用与国外相比,在随访时间、盲法判断疗效、mRS的统计分析方法上存在较大差异。

Objective To evaluate the current situation and problems of the application of modified Rankin scale （mRS） in the outcome assessment in Chinese stroke trials. Methods Randomised and quasi-randomised controlled trials on stroke therapy published before September 2013 in 3 Chinese databases were included. All clinical trials applied mRS as the method of outcome assessment. Subarachnoid hemorrhage and transient cerebral ischemia were excluded. Types of stroke, statistical methods used for data analysis, duration of follow up, blinding of outcome assessment, types of intervention and the significance of the results were evaluated. Results Two hundred and ninety-eight trials were included in this analysis. 71.14% was for ischemic stroke, 21.48% for hemorrhagic stroke, 7.38% for both ischemic and hemor- rhagic stroke and 91.28% was for acute stroke（onset time 〈 14d）. Regarding to statistical methods used for data analysis, 50.00% of the trials used t-test or variance analysis which treated the mRS score as continuous data, while 22.15% used rank sum test or Chi-square test which regarded the mRS score as ranked data or multiply variable data. Dichotomous data was applied in statistical analysis accounts for 25.50% of trials. 12.42% trials applied mRS with other scales as the methods of outcome assessment. Duration of follow up ranged from 10d to 2 years （median 90 d, interquartile range 50-90 d）. Only 5.03% assessed outcome blindly. 60.07% of the trials were drug therapy, 7.72% was rehabilitation therapy, 10.40% were surgical treatment, 14.43% were combined therapy, 2.35% were management mode, 0.67% were nursing, and 4.36% other therapy. Results in 86.91% of the trials were favorable to the tested interventions. Conclusions In aspects of, there is large difference between domestic and foreign clinical stroke trials in methodology of mRS including duration of follow up, blinding of outcome assessment and statistical methods used for data analysis.