Evaluation Of Cardiovascular Risk With Nsaids: Limitations Arising From Inconsistent Reporting In Pharmacoepidemiological Studie

Concerns about cardiovascular risks with NSAIDs arose from randomized studies of gastro-intestinal safety, Alzheimer’s disease and tumor protection. Subsequent pharmacoepidemiological studies explored risks associated with NSAID use in general populations. Factors important in judging risk include: dose, underlying cardiovascular risk, timeframe for risk onset. We investigated how these factors were reported in the pharmacoepidemiological studies. Methods: Review of studies included in a systematic review and metaanalysis of cardiovascular risk. Results: Of 51 eligible studies, 28 case–control studies included 184,946 cardiovascular events, 23 cohort studies described outcomes in >2.3 million exposed individuals. Dose: Of 38 studies reporting on rofecoxib and/or celecoxib, 14 and 8 respectively reported risk with different doses; all used the same low/ high dose stratifications. Of 42 studies reporting on one or more of three commonly-used NSAIDs (ibuprofen, diclofenac, naproxen), only nine reported low/high dose risk estimates and three different dose stratifications were used. Background: cardiovascular risk: Fourteen studies provided information on ‘high’ and ‘low’ risk populations. ‘High’ and ‘low’ risk definitions varied between studies. Timeframe for risk onset: Thirteen studies, reported events among newusers of NSAIDs; nine found risk elevated within 30 days; three within 14 days. Conclusion: In assessing NSAID-associated cardiovascular risk, a minority of studies reported three factors that are important potential effect modifiers. Among studies providing this information, betweenstudy variability in defining factor-related parameters impeded assessment of variation in cardiovascular risk. Agreement on reporting criteria would greatly improve the value of data from pharmacoepidemiological studies of NSAID-related risk and permit greater precision in risk estimates.