Population Factors Not Addressed in Cohort Study

Garcia Rodriguez et al have conducted a cohort study to investigate differential effects of aspirin and non-aspirin NSAIDs in the primary prevention of myocardial infarction in post-menopausal women. 1 They reported that aspirin, but not non-aspirin NSAIDs, offers protection against development of first non-fatal myocardial infarction in women. We have some reservations regarding the validity of their data. Information about drug use was obtained from the UK General Practice Research Database. Approximately 3 million residents in England and Wales are registered with general practitioners who participate in this scheme, which is a computerised system of recording clinical diagnoses and prescribed drugs. The database does not contain information regarding over-the-counter medication or compliance with prescribed drugs. We question their assumption that United Kingdom patients would not take prophylactic aspirin without a prescription from their general practitioner. Additionally, the authors did not consider the possible effects of socio-economic status differences in the study population. The prevalence of Coronary Heart Disease (CHD) is much higher in lower socio-economic classes. 2 Higher socio-economic classes are generally more aware of risk factors for CHD and perhaps more likely to request prophylactic medication. Furthermore, there are differences in price between aspirin and non-aspirin NSAIDs, which may lead to variations in self-medication practice. Aspirin is significantly cheaper over the counter than on prescription, but the difference is not so marked for non-aspirin NSAIDs. There is also likely to be a difference in lifestyle factors. The only indication that the authors may have addressed this issue is their finding that there were 30% more smokers among women taking aspirin. It is generally accepted that smoking is more prevalent among those of lower socio-economic class. 3 Although the GP Research Database is a valuable and easily accessible source of information, we suggest that further studies should incorporate data on over-the-counter sales.