F ourteen years ago, Hall and colleagues 1 reported observations that initiated a vigorous debate about a possible interaction between aspirin and angiotensinconverting enzyme (ACE) inhibitors that continues to this day. That debate centers on whether the apparent pharmacological interaction between aspirin and ACE inhibitors might influence clinical outcomes and has been heightened greatly by the provocative suggestion that long-term aspirin use might not be beneficial in patients with chronic coronary heart disease, including those with heart failure. 2 The arguments used in the debate and the studies on which they are based highlight the strengths and weaknesses of many of the main strands of modern cardiovascular therapeutic research and illustrate many useful lessons about the interpretation of these different types of study.
Article p 2572Hall et al 1 carried out what might be called a "mechanistic" study, demonstrating that the short-term hemodynamic effects of an ACE inhibitor in patients with severe heart failure were attenuated by aspirin. Many but not all subsequent studies have supported this finding. [3][4][5][6] Although interesting, the clinical importance of this observation is uncertain. First, whether the effect of aspirin persists in the longer term is unknown. Second, we do not know how an ACE inhibitor exerts its beneficial effect and therefore how important its hemodynamic or other actions are. Third, experience has taught us that however plausible a pharmacological mechanism may seem, it may not lead to the expected effect on clinical outcome. Consequently, studies on the actions of drugs on potentially important biological mechanisms can only be hypothesis generating with respect to clinical outcomes and often have been misleading. 7,8 Despite these limitations, such studies remain the foundation of therapeutic drug development.The report by McAlister and colleagues 9 represents a completely different approach in contemporary cardiovascular research, one that is becoming increasingly common with the availability of relatively inexpensive and powerful statistical software and the growing number of large administrative data sets created as a result of the explosion in information technology in healthcare systems. McAlister et al looked for a potential interaction between ACE inhibitors and aspirin in a large observational study of patients discharged from hospital in Ontario with a primary diagnosis of heart failure, recording death and readmission for heart failure over the subsequent year. 9 The authors did not find evidence of an interaction between the 2 treatments. Can we be reassured by this? The reader should have both specific and general reservations about the study by McAlister and colleagues. Specifically, the proportion of patients with nonischemic heart failure (44%) was unusually high, even assuming that the recorded origin was reliable, and because it is not indicated in nonischemic heart failure, it is unclear why some of these patients were treated with aspirin. Aspirin is ...