摘要
目的 通过对利妥昔单抗相关不良反应进行信号挖掘研究,为临床合理安全用药提供参考。方法 采用报告比值比法（reporting odds ratio,ROR）和贝叶斯置信度递进神经网络法（Bayesian confidence propagation neural network,BCPNN）,对美国不良事件报告系统（FDA Adverse Event Reporting System,FAERS）2014年第一季度~2015年第四季度收集的报告进行分析,挖掘利妥昔单抗产生的相关不良反应信号。结果 ROR法和BCPNN法分别挖掘利妥昔单抗可疑ADR不良反应相关信号657个和43个,其中包括药品说明书中未出现的新的可疑ADR信号分别为68个（ROR法,筛选条件：95%CI-排名前300位且ROR值大于2.5）和10个（BCPNN法）。结论 利妥昔单抗可疑ADR信号的挖掘,可以为国内临床合理安全用药提供参考依据。
OBJECTIVE To provide the reference for clinical rational safe drug use by the signal mining of adverse drug reactions caused by rituximab. METHODS Reporting odds ratio and Bayesian confidence propagation neural network methods were used to make a signal mining of suspected ADRs caused by rituximab from FDA Adverse Event Reporting System, where the reports collected from the first quarter in 2014 to the fourth quarter in 2015 can be used in this study. RESULTS 657 and 43 warning signals were ob- tained by ROR method and BCPNN method,including 68 new suspected ADR signals by ROR（ limited with 95% CI value ranked top 300 ADR signals and ROR value greater than 2. 5 ） and 10 new ADR signals by BCPNN, which are not mentioned in instructions. CONCLUSION Mining the signals of the adverse drug reactions of rituximab can provide reference for domestic clinical rational safe drug use.
出处
《中国药学杂志》
CAS
CSCD
北大核心
2016年第22期1976-1981,共6页
Chinese Pharmaceutical Journal
