Meta-Analysis在随机临床试验安全性数据中的应用

发布时间：2019-06-25 10:06

【摘要】：本论文通过运用Meta-Analysis方法研究与安慰剂相比西替利嗪10mg/天在6岁及以上患有季节性过敏性鼻炎(SAR)、常年性变应性鼻炎(PAR)和慢性荨麻疹的病人嗜睡方面的安全性,说明了Meta-Analysis在临床试验安全性方面的应用。论文首先以Cetirizine和Somnolence为关键词系统检索了PubMed以及PMC、ScienceDirect、Wiley以及Google,并对所有检索到的文献仔细阅读并进行筛查、评价其质量及按照预定的入选标准进行资料提取,然后运用统计软件SAS9.3对所有数据进行分析：用DerSimonian and Laird(DL)方法分别计算基于固定效应和随机效应模型的合并效应量,用Q检验、H检验以及I^2进行异质性检验并进行亚组分析、亚组异质性检验、Galbraith图形分析、敏感性分析以及Meta-Regression分析来探索异质性的根源。共有12个符合入选标准的随机临床试验进入本研究。根据本研究的结果,基于12个临床试验的异质性检验表明试验间存在异质性,相应的亚组分析、亚组异质性检验、Galbraith图形分析、敏感性分析以及Meta-Regression分析的结果一致表明安慰剂导入期试验是异质性的根源,这对于今后类似药物的临床试验的试验设计具有重要的指导意义。根据这一重要发现,我们认为在研究西替利嗪10mg/天与安慰剂在导致嗜睡方面的差别的时候,应该对有安慰剂导入期和无安慰剂导入期的试验分别计算合并效应量,分别下结论。根据Meta-Analysis亚组分析的结果,没有安慰剂导入期的临床试验基于固定效应的西替利嗪10mg/天相对于安慰剂在嗜睡方面的合并危险度差值为7.03,95%CI为(5.01,9.05),95%CI不包含0,说明没有安慰剂导入期的试验中西替利嗪10mg/天与安慰剂在导致嗜睡方面的差别有统计学意义；有安慰剂导入期的临床试验的合并危险度差值为1.09,95%CI为(-0.16,2.33),95%CI包含0,说明有安慰剂导入期的试验中西替利嗪10mg/天与安慰剂在导致嗜睡方面的差别没有统计学意义,这一点将极大的颠覆行业内人士既往对西替利嗪的观点,也很有可能改变医生对处方的选择。由于克服了单个研究样本量的局限,Meta-Analysis应用于安全性数据可以更好地提供关于西替利嗪与安慰剂相比在导致嗜睡方面差别的真实情况。
[Abstract]:In this study, Meta-Analysis was used to study the safety of cetirizine 10mg/ in patients with seasonal allergic rhinitis (SAR), perennial allergic rhinitis (PAR) and chronic urticaria compared with placebo. The application of Meta-Analysis in clinical trials was demonstrated. In this paper, PubMed, PMC,ScienceDirect,Wiley and Google, are searched with Cetirizine and Somnolence as keywords, and all the searched literature is carefully read and screened, the quality of which is evaluated and the data are extracted according to the predetermined selection criteria. Then all the data are analyzed by statistical software SAS9.3: the combined effects based on fixed effect and random effect model are calculated by DerSimonian and Laird (DL) method, and Q test is used. H test and I ^ 2 were used for heterogeneity test and subgroup analysis, subgroup heterogeneity test, Galbraith graphic analysis, sensitivity analysis and Meta-Regression analysis to explore the root causes of heterogeneity. A total of 12 randomized clinical trials that met the selection criteria were enrolled in this study. According to the results of this study, the heterogeneity test based on 12 clinical trials showed that there was heterogeneity between the trials. The results of corresponding subgroup analysis, subgroup heterogeneity test, Galbraith graphic analysis, sensitivity analysis and Meta-Regression analysis showed that placebo introduction trial was the root of heterogeneity, which had important guiding significance for the design of clinical trials similar to drugs in the future. On the basis of this important finding, we suggest that when studying the difference between cetirizine 10mg/ days and placebo in causing lethargy, we should calculate the combined effects of trials with placebo introduction and no placebo introduction, and draw conclusions respectively. According to the results of Meta-Analysis subgroup analysis, the combined risk difference of cetirizine 10mg/ days based on fixed effect in clinical trials without placebo was 7.03, 95% CI (5.01, 9.05), and 95%CI did not include 0, indicating that there was significant difference between cetirizine 10mg/ days and placebo in patients without placebo introduction. The combined risk difference for clinical trials with placebo introduction was 1.09, 95% CI was (- 0.16, 2.33), and 95%CI included 0, indicating that there was no significant difference between cetirizine 10mg/ days and placebo causing lethargy in placebo-induced trials, which would greatly subvert the previous view of cetirizine in the industry and was likely to change doctors' choice of prescription. By overcoming the limitations of a single study sample size, the application of Meta-Analysis to safety data can better provide a true picture of the difference between cetirizine and placebo in causing lethargy.
【学位授予单位】：华东师范大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：O211.6;R969.4

【共引文献】