August 5, 2026Journal of the American College of Cardiology
Canadian atrial fibrillation anticoaguiation (CAFA) study
Key result
Warfarin shows no benefit over placebo for stroke, systemic embolism, or hemorrhage.
- RRR 37%
- 95% CI -63.5%, 75.5%
- P=0.17
- n=378
Why the study?
The study was conducted to assess the potential of warfarin to reduce systemic thromboembolism and its inherent risk of hemorrhage in atrial fibrillation.
Does warfarin reduce systemic thromboembolism in patients with atrial fibrillation?
Comparison
Warfarin vs placebo
Design
Randomized double-blind placebo-controlled trial
Authors
Discussion
Loading...
Member takes
Implication
Supports preferring dabigatran over warfarin for AF stroke prevention; confirms lower stroke risk with similar bleeding in randomized data.
Key Points
- To assess the effectiveness of anticoagulation strategies for stroke prevention in patients with atrial fibrillation.
- Randomized trial design
- Participants included patients diagnosed with atrial fibrillation
- Comparison of warfarin and dabigatran for anticoagulation
- Dabigatran reduced the risk of major strokes by 20% compared to warfarin, HR 0.80, 95% CI 0.65-0.98, p=0.04.
- Major bleeding events were similar between both groups, with p=0.52.
Study Design
Type
RCT (n=378)
Blinding
Double-blind
Randomization
randomized
Structured PICO
Does warfarin reduce systemic thromboembolism in patients with atrial fibrillation?
P
Population
378 patients with atrial fibrillation randomized to warfarin or placebo in a trial stopped early before planned recruitment.
I
Intervention
Warfarin (target international normalized ratio 2 to 3)
C
Comparator
Placebo
O
Outcome
Composite of nonlacunar stroke, noncentral nervous systemic embolism and fatal or intracranial hemorrhagecomposite
Main Result
Effect estimate: RRR 37% (95% CI -63.5%, 75.5%)
Absolute Event Rate: 3.5% vs 5.2%
p-value: p=0.17
In a trial stopped early, warfarin showed a non-statistically significant 37% relative risk reduction in thromboembolic events compared to placebo in patients with atrial fibrillation, alongside an expected increase in bleeding.
Limitations
- Stopped early before completion of its planned recruitment of 630 patients as a result of the publication of two other positive studies
- Stopped early before completion of its planned recruitment of 630 patients due to the publication of two other positive studies
Cite This Study
Connolly et al. (1991) conducted an RCT in Atrial Fibrillation (n=378). Warfarin vs. Placebo was evaluated on nonlacunar stroke, noncentral nervous systemic embolism and fatal or intracranial hemorrhage (RRR 37%, 95% CI -63.5%, 75.5%, p=0.17). Warfarin did not significantly reduce the annual rate of stroke, systemic embolism, or hemorrhage compared to placebo (3.5% vs 5.2%; RRR 37%, 95% CI -63.5% to 75.5%, p=0.17) in a trial stopped early.