Mining and analysis of adverse event signals of non-steroidal anti-inflammatory drugs based on the FAERS database

Yang Qingqing; Gong Yan; Wang Xiaolong

doi:10.3969/j.issn.1672-8157.2025.04.037

Abstract：Objective To mine the signals of adverse drug events(ADE)associated with non-steroidal anti-inflammatory drugs(NSAIDs)in the U.S.Food and Drug Administration's Adverse Event Reporting System(FAERS)and provide evidence-based basis for clinical medication safety.Methods ADE reports from the FAERS database of 12 commonly used NSAIDs(e.g.,ibuprofen,celecoxib,diclofenac)as primary suspect drugs from January 2016 to December 2024 were extracted.After standardizing drug names and deduplication,reports were categorized into severe ADE(n=9,664)group and non-severe ADE(n=15,804)group based on the U.S.FDA's definition of"death,hospitalization,disability,etc."for severe ADEs.The signals were detected respectively by using the reporting odds ratio(ROR)method,Bayesian confidence propagation neural network(BCPNN)method,and comprehensive standard method,and system organ class(SOC)and standardized terminology mapping were performed with the aid of MedDRA v27.0.The observed indicators included the number of valid signals detected by each method,the classification of the involved SOC,the asso-ciation strength between different NSAIDs and specific ADEs,the annual distribution of ADE report,and the time distribution of ADE occurrence.Results A total of 25 468 ADE reports were included,with females accounting for 59.80% (15,230/25,468)and patients aged 45-64 years representing the largest proportion(41.46%,10,560/25,468).Oral administration predominated(69.30%,17,650/25,468),with ibuprofen(24.49%,6,238/25,468)and celecoxib(23.13%,5,890/25,468)being the most reported.The severe ADE group comprised 9,664 cases(hospitalization:79.30%,7,664/9,664;death:20.70%,2,000/9,664),and celecoxib-associated deaths accounting for 30.00% (600/2,000)of the total deaths,which was significantly lower than that of other NSAIDs(χ2=320.00,P<0.001).A total of 1,580 signals were detected,with gastrointestinal disorders(30.13%,476/1 580),various examinations abnor-malities(19.24%,304/1,580),and cardiovascular system diseases(13.16%,208/1,580)ranking the top three.High-strength signals included ibuprofen-upper gastrointestinal hemorrhage(ROR=8.12,IC025=3.45)and celecoxib-myocardial infarction(ROR=5.68,IC025=3.02),etc.The annual distribution results showed that the number of reports peaked in 2020-2021(an increase of 37.2% com-pared with 2016-2019),with notable rises for ibuprofen and celecoxib;reports declined slightly in 2022-2024 but remained above the baseline.ADEs most occurred within 1-14 days post-administration(47.99%,12,221/25,468),with a median time of 10 days(IQR:5-30).The results of Weibull distribution test indicated early risk accumulation(β=0.92,P<0.05).Conclusion NSAID-associated ADE risks exhibit significant heterogeneity.In clinical practice,individualized risk monitoring should be implemented based on drug properties,with attention to the risk changes in the early treatment phase and specific years.