Responses to USPSTF Guideline on Aspirin for Prevention of Cardiovascular Disease

IN RESPONSE:

We appreciate the thoughtful letters from Dr. Budhraja and Dr. Mohan and colleagues regarding the USPSTF recommendation on aspirin prophylaxis for the prevention of cardiovascular disease. Dr. Budhraja calls attention to subgroups of women in whom the effect of aspirin in preventing cardiovascular disease may differ from that of the general population. In general, the USPSTF is cautious when considering unplanned subgroup analyses of randomized trials, which are the basis for Dr. Budhraja's comment. The authors of the original report from the Women's Health Study (1) mention multiple comparisons as an additional caution in interpreting this subgroup analysis. All subgroup analyses should be considered hypothesis-generating rather than independently persuasive.

The possibility suggested by Dr. Budhraja that the higher risk for peptic ulcer disease in smokers might place them at higher risk for hemorrhage when taking aspirin merits further research.

Dr. Mohan and colleagues raise many valuable points. The inadequate and contradictory information derived from Web-enabled coronary and cardiovascular disease risk calculators has been a matter of great concern for the USPSTF. The calculator referenced in the recommendation was selected primarily because it is easy to use and does not require information about high-density lipoprotein cholesterol concentration. As of this writing, the calculator has been removed from the Medical College of Wisconsin's Web site and reportedly is being revised.

The USPSTF felt that making a recommendation meant to be tailored to estimation of cardiovascular disease risk without making any suggestions to help clinicians use the recommendation would be worse than mentioning an imperfect calculator. Research in this field is sorely needed. The development of a “gold standard” cardiovascular disease risk calculator to aid in predicting contemporary rates of cardiovascular disease in the United States should be a pressing priority for analysis of data derived from large cohort studies done in the past decade, perhaps pooling individual-level data across studies. The Agency for Healthcare Research and Quality has funded a project to evaluate the models currently available for risk calculation for cardiovascular disease. The results will be available soon. The use of risk-prediction models, and the tools based on them, to guide decisions about use of preventive and therapeutic medications, as well as decisions about screening, will become increasingly important in the emerging era of personalized medicine.