Antiplatelet Trialists' Collaboration Overview

By the early 1990s, dozens of randomized trials had tested aspirin and other antiplatelet agents in patients at risk for vascular events, but most were too small to yield reliable estimates of effect on mortality or stroke. The published literature was a patchwork: different patient populations, different doses, different endpoints, and different durations of follow-up. Clinicians reading individual trials could not extract a coherent message about which patients benefited, by how much, or at what risk. The question called for a different approach.

The Antiplatelet Trialists' Collaboration, coordinated from Oxford, combined 145 randomized trials of antiplatelet therapy for a month or more against control, covering about 70,000 high-risk patients and 30,000 low-risk people from the general population, plus 29 trials comparing regimens. The collaboration included trials of aspirin, dipyridamole, and other agents across patients with prior myocardial infarction, stroke or transient ischemic attack, unstable angina, and peripheral arterial disease. Combining the trials gave the group statistical power no single trial could approach.

In each of four high-risk groups (acute myocardial infarction, past infarction, past stroke or transient ischemic attack, and other vascular conditions), antiplatelet therapy cut vascular events by about a quarter, and among all high-risk patients nonfatal infarction and nonfatal stroke each fell by about a third; overall mortality was significantly reduced. In about 20,000 patients with acute myocardial infarction, vascular events fell from 14 to 10 percent within a month, about 40 avoided for every 1,000 treated. Benefit was consistent across the different high-risk clinical categories; what varied was the absolute gain, which was naturally greatest where baseline risk was highest. Medium-dose aspirin, 75 to 325 mg a day, was the most widely tested regimen, and doses across that range seemed equally effective.

The BMJ published the overview on January 8, 1994. Its conclusions supported long-term aspirin after heart attack, stroke or transient ischemic attack, and benefit was still accumulating between the first and third years of treatment.

The overview showed that moderate, consistent effects could be measured reliably by combining many trials. Reductions were significant in middle and old age, in men and women, and in people with and without hypertension or diabetes. In low-risk people taking aspirin for primary prevention, the overview found a one-third fall in nonfatal heart attack but a nonsignificant rise in stroke, and a much smaller absolute benefit.