Data monitoring for the aspirin component of the Physicians' Health Study: Issues in early termination for a major secondary endpoint

Abstract

The Physicians' Health Study was a randomized, double-blind, placebo controlled, × 2 factorial primary prevention trial whose primary aims were to test whether aspirin reduces risks of cardiovascular disease (CVD) mortality and beta-carotene decreases the incidence of cancer. The trial was conducted among 22,071 apparently healthy U.S. male physicians aged 40-84 years at entry. After five years of treatment and follow-up, on December 17, 1987, the independent Data and Safety Monitoring Board (DSMB) recommended unanimously the early termination of the aspirin component due principally to the emergence of a statistically extreme (p < 0.00001) 47% reduction in risk of a first myocardial infarction (MI), the major secondary endpoint, in the context of a far lower than anticipated CVD mortality as well as use of aspirin among the vast majority of individuals who experienced a non-fatal event. Several additional factors were involved, including little or no trend in either CVD mortality or stroke, although the numbers of events were too low to distinguish between small benefit, no effect, and small harm. These circumstances suggested clear evidence for aspirin in preventing a first MI, a major outcome of clinical and public health importance in the context of inadequate power to test the primary endpoint of CVD mortality. © 2006 Springer-Verlag New York.