Oral Anticoagulants vs Aspirin in Nonvalvular Atrial Fibrillation: An Individual Patient Meta-analysis | Anticoagulation | JAMA | JAMA Network
[Skip to Navigation]
Access to paid content on this site is currently suspended due to excessive activity being detected from your IP address Please contact the publisher to request reinstatement.
Clinical Cardiology
November 20, 2002

Oral Anticoagulants vs Aspirin in Nonvalvular Atrial Fibrillation: An Individual Patient Meta-analysis

Author Affiliations

Author Affiliations: Clinical Epidemiology Unit, Ottawa Health Research Institute (Dr van Walraven); Institute for Clinical Evaluative Sciences (Drs van Walraven and Laupacis); Canadian Institutes for Health Research (Dr Laupacis), Ottawa, Ontario; Department of Medicine, University of Texas, San Antonio (Dr Hart); Clinical Epidemiology Unit, General Medicine Division, Massachusetts General Hospital, Boston (Drs Singer and Chang); McMaster University, Hamilton, Ontario (Dr Connolly); Hvidovre University Hospital, Copenhagen, Denmark (Dr Petersen); Department of Neurology, Erasmus MC, Rotterdam (Dr Koudstaal) and University of Maastricht, Maastricht (Dr Hellemons), the Netherlands. Dr van Walraven is an Ontario Ministry of Health Career Scientist.


Clinical Cardiology Section Editor: Michael S. Lauer, MD, Contributing Editor.

JAMA. 2002;288(19):2441-2448. doi:10.1001/jama.288.19.2441

Context Patients with nonvalvular atrial fibrillation (AF) have an increased risk of stroke and other vascular events.

Objective To compare the risk of vascular and bleeding events in patients with nonvalvular AF treated with vitamin K –inhibiting oral anticoagulants or acetylsalicylic acid (aspirin).

Design Pooled analysis of patient-level data from 6 published, randomized clinical trials.

Patients A total of 4052 patients with AF randomly assigned to receive therapeutic doses of oral anticoagulant or aspirin with or without low-dose oral anticoagulants.

Main Outcome Measures Ischemic and hemorrhagic stroke, other cardiovascular events, all-cause death, and major bleeding events. Person-year incidence rates were calculated to provide crude comparisons. Relative efficacy was assessed using proportional hazards modeling stratified by study. The variation of the oral anticoagulant's relative effect by pertinent patient factors was explored with interaction terms. All analyses were conducted using the intention-to-treat principle.

Results Patients receiving oral anticoagulant and aspirin were balanced for important prognostic factors. There was no significant heterogeneity between trials in the relative efficacy of oral anticoagulant vs aspirin for any outcome. Patients receiving oral anticoagulant were significantly less likely to experience any stroke (2.4 vs 4.5 events per 100 patient-years; hazard ratio [HR], 0.55; 95% confidence interval [CI], 0.43-0.71), ischemic stroke (HR, 0.48; 95% CI, 0.37-0.63), or cardiovascular events (HR, 0.71; 95% CI, 0.59-0.85) but were more likely to experience major bleeding (2.2 vs 1.3 events per 100 patient-years; HR, 1.71; 95% CI, 1.21-2.41). The reduction in ischemic stroke risk was similar in patients with paroxysmal AF (1.5 vs 4.7 events per 100 patient-years; HR, 0.32; 95% CI, 0.16-0.61; P<.001). Treating 1000 patients with AF for 1 year with oral anticoagulant rather than aspirin would prevent 23 ischemic strokes while causing 9 additional major bleeds. Overall all-cause survival did not differ but appeared to improve for oral anticoagulant patients 3 years after therapy was started.

Conclusions Compared with aspirin, oral anticoagulant significantly decreases the risk of all strokes, ischemic strokes, and cardiovascular events for patients with nonvalvular chronic or paroxysmal AF but modestly increases the absolute risk of major bleeding. The balance of benefits and risks varies by patient subgroup.