基于医院不良反应报告数据的妊娠期心脏和神经系统药物不良反应的信号挖掘

Signal mining of adverse drug reactions in the heart and nervous system during pregnancy based on adverse reaction reports in the hospital

目的对医院药物不良反应(ADR)自发报告数据进行分析,挖掘与妊娠期妇女心脏和神经系统ADR可能相关的药物风险信号。方法从中国医院药物警戒系统数据库中下载福建省妇幼保健院2000年1月1日至2022年12月31日的ADR报告,选取关于妊娠期妇女的报告,对其中系统器官分类为“心脏器官疾病”和“各类神经系统疾病”ADR的占比和临床表现进行描述性统计分析。采用报告比值比(ROR)法和贝叶斯置信传播神经网络(BCPNN)法对妊娠期妇女ADR报告中与心脏器官疾病和各类神经系统疾病ADR可能有关的药物进行数据挖掘。ROR法中目标药物的目标不良反应/事件报告数(a)≥3且ROR的95%置信区间(CI)下限>1,BCPNN法中信息成分-2倍标准差(IC-2SD)>0定义为一个风险信号。当2种方法检测结果均符合上述条件时,则判断该信号为可疑药物信号。结果共有783例妊娠期ADR报告纳入分析,其中涉及心脏器官疾病的报告94例(12.0%),主要表现为心悸;涉及各类神经系统疾病的报告121例(15.5%),主要表现为颤抖和头晕。心脏器官疾病和各类神经系统疾病的ADR多发生在孕晚期。ROR法及BCPNN法的结果均显示,利托君注射液(a=61,ROR=14.64,95%CI下限=9.08;IC-2SD=1.36)和利托君片(a=24,ROR=8.34,95%CI下限=4.64;IC-2SD=1.10)为导致妊娠期心脏器官疾病的风险药物,硫酸镁注射液(a=37,ROR=6.66,95%CI下限=4.10;IC-2SD=0.98)和利托君注射液(a=45,ROR=3.72,95%CI下限=2.44;IC-2SD=0.56)是导致妊娠期各类神经系统疾病的风险药物。结论利托君和硫酸镁与妊娠期的心脏和神经系统ADR可能存在一定的关联性,应当引起临床警惕。

Objective To analyze the spontaneous reports on adverse drug reaction(ADR)in a hospital and to mine the risk signals of drug that might be associated with cardiac and neurological ADRs in pregnant women.Methods The ADR reports in the database of China Hospital Pharmacovigilance System reported by Fujian Maternal and Child Health Hospital from January 1st,2000 to December 31st,2022 was downloaded.ADR reports about pregnancy women were collected,and the proportions and clinical manifestations of ADRs classified as"cardiac disorders"and"nervous system disorders"according to systems and organs were analyzed by descriptive statistic method.Data mining was conducted on drugs that might be associated with cardiac disorders and nervous system disorders in ADR reports using the reporting odds ratio(ROR)method and Bayesian confidence propagation neural network(BCPNN)method.The definition of risk signals in ROR method was the number of adverse reaction/event reports on the target drug(a)≥3,and the lower limit of the 95%confidence interval(CI)of ROR>1.The definition of risk signals in BCPNN method was the information component minus twice the standard deviation(IC-2SD)>0.When the calculation results of an adverse reaction/event was in accordance with above-mentioned conditions in both methods,a suspected drug risk signal was determined.Results A total of 783 ADR reports about pregnancy women were included in the analysis.Ninety‑four reports(12.0%)were about cardiac disorders,mainly manifested as palpitations;121 reports(15.5%)were about nervous system disorders,mainly characterized by trembling and dizziness.These ADRs of cardiac disorders and nervous system disorders often occurred in the third trimester of pregnancy.Through the ROR and BCPNN methods,results showed that ritodrine hydrochloride injection(a=61,ROR=14.64,the lower limit of 95%CI=9.08;IC-2SD=1.36)and ritodrine hydrochloride tablets(a=24,ROR=8.34,the lower limit of 95%CI=4.64;IC-2SD=1.10)were risk signals of drug leading to cardiac disorders during pregnancy,while magnesium sulfate injection(a=37,ROR=6.66,the lower limit of 95%CI=4.10;IC-2SD=0.98)and ritodrine hydrochloride injection(a=45,ROR=3.72,the lower limit of 95%CI=2.44;IC-2SD=0.56)were risk signals of drug leading to nervous system disorders during pregnancy.Conclusion Ritodrine and magnesium sulfate may be associated with cardiac and nervous system ADR in women during pregnancy,which should arouse clinical vigilance.