Numerous studies, including meta-analyses, have compared the efficacy of oral anticoagulation with warfarin and aspirin for stroke prevention in patients with atrial fibrillation. In this meta-analysis, van Walraven and colleagues reviewed individual patient data, rather than summary data, from six clinical trials. This approach facilitated a more detailed analysis of patient subsets, as well as person-year incidence rates.

The authors included individual patient-level data from the following six trials: Atrial Fibrillation, Aspirin, Anticoagulation Studies 1 and 2; Primary Prevention of Arterial Thromboembolism in Patients with Non-rheumatic Atrial Fibrillation in Primary Care; European Atrial Fibrillation Trial; and Stroke Prevention in Atrial Fibrillation Studies 1, 2, and 3. These trials randomized patients to receive a full dose of an oral anticoagulant (primarily coumarin derivatives such as warfarin or 4-hydroxycoumarin), aspirin, or aspirin with a low dose of oral anticoagulant. Patient characteristics in the trials were somewhat similar, and the mean follow-up period was 1.9 years.

Outcomes considered in the authors' analysis of patient-level data from the six trials included ischemic stroke, hemorrhagic stroke, aggregate cardiovascular events, major bleeding, and all-cause death. The results were expressed in incidence rates (events per 100 person-years) and hazard ratios.

The meta-analysis showed that event rates for all stroke, ischemic stroke, and cardiovascular events decreased significantly with oral anticoagulation compared with aspirin therapy. For example, the event rate for all strokes was 2.4 per 100 person-years in patients receiving an oral anticoagulant as opposed to 4.5 in patients receiving aspirin. The hazard ratio was 0.55 for all stroke, 0.48 for ischemic stroke, and 0.71 for cardiovascular events. The benefit of oral anticoagulation was partially offset by an increase in major bleeds (hazard ratio: 1.71). See accompanying table for a comparison.

The greatest absolute risk reduction in ischemic stroke occurred in patients who had the highest baseline risk for stroke (e.g., hypertension, diabetes, previous cerebral ischemia). For example, the absolute risk reduction was 6.0 percent per year in patients who had a previous history of transient ischemic attack or stroke, compared with 1.2 percent per year in patients who did not have cerebrovascular disease.

The authors conclude that oral anticoagulation is superior to aspirin in preventing all strokes, ischemic strokes, and cardiovascular events in patients with chronic or nonvalvular paroxysmal atrial fibrillation. Using oral anticoagulation rather than aspirin for one year in 1,000 patients with atrial fibrillation would prevent 23 strokes and result in nine additional episodes of major bleeding. With the use of oral anticoagulation, the absolute risk reduction in ischemic stroke is greater in higher-risk patients than in low-risk patients. The authors note that the more significant reduction in stroke risk with oral anticoagulation offsets the risk of major bleeding in patients determined to be at higher risk for stroke. One factor to consider is that patients fear the consequences of stroke more than they do the consequences of a major bleed.