Do the results of randomized clinical trials of cardiovascular drugs influence medical practice? The SAVE Investigators

Abstract

Background: Medical practice patterns change in response to a variety of stimuli, one of which may be the publication of the results of randomized clinical trials. We assessed the temporal association between the publication of clinical trials on myocardial infarction and changes in treatment practices for this disorder.

Methods: We analyzed the use of aspirin before and after myocardial infarction and that of calcium antagonists after myocardial infarction in 2231 survivors of myocardial infarction enrolled in the Survival and Ventricular Enlargement (SAVE) study over a three-year period (from January 1987 through January 1990). The proportion of patients using these treatments was analyzed before and after the publication dates of three clinical trials: the Physicians' Health Study, published in January 1988, which supported the use of aspirin to prevent a first myocardial infarction; the Second International Study of Infarct Survival (ISIS-2), published in August 1988, which supported the use of aspirin after myocardial infarction; and the Multicenter Diltiazem Postinfarction Trial, published in August 1988, which reported a deleterious effect of diltiazem in some patients after myocardial infarction.

Results: The use of aspirin before myocardial infarction increased from 16.2 percent to 23.9 percent between January 1987 and January 1990 (P less than 0.001). Enrollment in the study after the publication of the Physicians' Health Study independently predicted aspirin use before myocardial infarction (odds ratio, 1.43; 95 percent confidence interval, 1.11 to 1.85). The use of aspirin after myocardial infarction increased from 38.8 percent to 71.9 percent (P less than 0.001) during the three-year study period. Enrollment in the study after the publication of ISIS-2 independently predicted the use of aspirin after myocardial infarction (odds ratio, 2.28; 95 percent confidence interval, 1.89 to 2.76). The use of calcium antagonists after myocardial infarction decreased from 57.1 percent to 33.1 percent (P less than 0.001) during the study period. Enrollment in the study after the publication of the Multicenter Diltiazem Postinfarction Trial independently predicted the use of calcium antagonists after myocardial infarction (odds ratio, 0.47; 95 percent confidence interval, 0.39 to 0.57).

Conclusions: These observations suggest that randomized clinical trials have a measurable influence on medical practice patterns.