上海交通大学学报（医学版）

上海交通大学学报（医学版）

• 论著（临床研究） • 上一篇    下一篇

5种药物在中国精神分裂症人群中的有效性与网络meta分析

顾文洁1，张五芳2，王超1，张晓艳1   

  1. 1.同济大学生命科学与技术学院， 上海 200082； 2.北京大学精神卫生研究所， 北京 100091
  • 出版日期:2015-11-28 发布日期:2016-01-13
  • 通讯作者: 张晓艳, 电子信箱: xyzhang@tongji.edu.cn。
  • 作者简介:顾文洁（1990—）， 女， 硕士生; 电子信箱: drguwenj@163.com。

Therapeutic effect of 5 drugs on Chinese patients with schizophrenia by using network meta-analysis

GU Wen-jie1, ZHANG Wu-fang2, WANG Chao1, ZHANG Xiao-yan1   

  1. 1.School of Life Science and Technology, Tongji University, Shanghai 200082, China; 2.Mental Health Institute, Peking University, Beijing 100091, China
  • Online:2015-11-28 Published:2016-01-13

PDF (PC)

1371

摘要：

目的  以网络meta分析方法综合评价阿立哌唑（Ari）、奥氮平（Ola）、帕利哌酮（Pal）、喹硫平（Que）和利培酮（Ris）5 种药物对中国精神分裂症患者的疗效。方法  计算机检索PubMed、Cochorane Library、中国知网（CNKI）、中文科技期刊数据库、中国学术期刊数据库和中国科学引文数据库，按照入选标准纳入文献，并进行质量评价。在WinBUGS 1.4.3软件中构建贝叶斯网络化模型，用药物的OR值来估计推荐的优先级。结果  纳入的60项临床试验共5 280例患者。网络meta分析显示：相比Pal、Ari、Ola和Ris，Que能显著提高临床有效率，OR值(Pal、Que)=0.4154（95%CI：0.2529~0.6367），OR值(Pal、Ris)=0.6744（95%CI：0.5217~0.8555），OR值(Ari、Que)=0.5491（95%CI：0.2988~0.9096），OR值(Ola、Que)=0.4746（95%CI：0.2805~0.7558），OR值(Que、Ris)=1.694（95%CI：1.115~2.495），差异具有统计学意义。结论  由分析结果可知，尽管Que有较高的推荐等级，但由于纳入的研究较少，其药物有效性还需进一步的证据支持。

关键词: 精神分裂症, 贝叶斯模型, 网络meta 分析

Abstract:

Objective  To comprehensively assess the therapeutic effect of 5 antipsychotic drugs, i.e. aripiprazole (Ari), olanzapine (Ole), paliperidone (Pla), quetiapine (Que), and risperidone (Ris), on patients with schizophrenia by using the network meta-analysis. Methods  Papers were retrieved from PubMed, Cochorane Library, CNKI, China Science and Technology Journal Database, China Science Periodical Database, and Chinese Science Citation Database according to the predefined inclusion and exclusion criteria and quality assessment was conducted. The Bayesian network model was established by WinBUGS1.4.3 and the priority of recommendation levels was estimated by the OR value of drugs. Results  A total of 5280 patients were enrolled in 60 clinical tests. The results of network meta-analysis showed that compared with Pal, Ari, Ola and Ris, Que significantly improved the clinical efficiency. OR (Pal, Que)=0.415 4 (95%CI: 0.252 9-0.636 7), OR (Pal, Ris)=0.674 4 (95%CI: 0.521 7-0.855 5), OR (Ari, Que)=0.5491 (95%CI: 0.298 8-0.909 6), OR (Ola, Que)=0.474 6 (95%CI: 0.280 5-0.755 8), and OR (Que, Ris)=1.694 (95%CI: 1.115-2.495). The differences were statistically significant. Conclusion  Based on the results of network meta-analysis, the recommendation level of Que is higher than that of other drugs. But further evidences are needed to verify the therapeutic effect due to lack of studies.

Key words: schizophrenia, bayesian model, network meta-analysis