EAFT
(1993)Objective
Assess whether anticoagulation or aspirin reduces recurrent vascular events in patients with non-rheumatic atrial fibrillation and a recent TIA or minor ischaemic stroke (secondary prevention).
Study Summary
• Anticoagulation reduced the risk of stroke alone from 12% to 4% per year (HR 0.34, 95% CI 0.20-0.57) and outperformed aspirin (HR 0.60, 95% CI 0.41-0.87).
• Aspirin 300 mg/day showed a non-significant trend for benefit versus placebo (15% vs 19%/yr; HR 0.83, 95% CI 0.65-1.05).
• Major bleeding was uncommon (anticoagulation 2.8%/yr, aspirin 0.9%/yr) with no intracranial haemorrhages on anticoagulation.
Intervention
Open-label oral anticoagulation (coumarin, target INR 2.5-4.0) or double-blind aspirin 300 mg once daily, versus double-blind placebo.
Inclusion Criteria
Age >25 with non-rheumatic atrial fibrillation and a TIA or minor ischaemic stroke (modified Rankin ≤3) within the preceding 3 months.
Study Design
Arms: Group 1 (anticoagulation-eligible): open-label anticoagulation vs double-blind aspirin 300 mg vs double-blind placebo. Group 2 (contraindication to anticoagulation): double-blind aspirin 300 mg vs placebo.
Patients per Arm: Group 1 (n=669): anticoagulation 225, aspirin 222, placebo 222. Group 2 (n=338): aspirin 171, placebo 167. Total 1007.
Outcome
• Stroke alone in Group 1: 4%/yr anticoagulation vs 12%/yr placebo (HR 0.34, 95% CI 0.20-0.57).
• Anticoagulation vs aspirin for primary composite: HR 0.60 (95% CI 0.41-0.87).
• Aspirin vs placebo (Groups 1+2): 15%/yr vs 19%/yr (HR 0.83, 95% CI 0.65-1.05; NS).
• Major bleeding: 2.8%/yr anticoagulation, 0.9%/yr aspirin, 0.7%/yr placebo; no intracranial bleeds on anticoagulation.
• Absolute benefit: 90 vascular events prevented per 1000 patient-years with anticoagulation; 40/1000 with aspirin.
Clinical Question
In patients with non-rheumatic atrial fibrillation and a recent TIA or minor ischaemic stroke, does oral anticoagulation or aspirin 300 mg daily reduce the risk of recurrent vascular events (vascular death, stroke, MI, or systemic embolism)?
Bottom Line
In non-rheumatic AF patients with a recent TIA or minor stroke, oral anticoagulation (target INR 2.5-4.0) roughly halves the annual risk of recurrent vascular events (HR 0.53) and cuts stroke risk by two thirds (HR 0.34) versus placebo; anticoagulation is significantly more effective than aspirin 300 mg (HR 0.60), while aspirin offers a smaller, non-significant benefit and is a safer alternative when anticoagulation is contraindicated.
Major Points
- First randomised trial to establish oral anticoagulation for SECONDARY prevention in non-rheumatic AF after TIA or minor stroke.
- Anticoagulation prevented ~90 vascular events per 1000 patient-years; aspirin prevented ~40 per 1000 patient-years.
- Anticoagulation vs aspirin head-to-head: HR 0.60 (95% CI 0.41-0.87) for the primary composite, favouring anticoagulation.
- Bleeding risk with target INR 2.5-4.0 was acceptable (2.8%/yr major bleeds) with NO intracranial haemorrhages on anticoagulation.
- Median achieved INR was 2.9 (interquartile range 2.0-3.4); ~half of anticoagulation time was within target range, suggesting benefit is robust despite imperfect INR control.
- Findings shaped guideline recommendations for oral anticoagulation in AF patients with prior cerebral ischaemia and established the framework for later NOAC vs warfarin trials.
Study Design
- Study Type
- Randomised Controlled Trial (multicentre, partly open-label / partly double-blind, placebo-controlled)
- Randomization
- Yes
- Blinding
- Anticoagulation was open-label; aspirin vs placebo was double-blind. Outcome adjudication by a blinded committee.
- Sample Size
- 1007
- Follow-up
- Mean 2.3 years
- Centers
- 108
- Countries
- Austria, Belgium, Denmark, Finland, France, Germany, Greece, Ireland, Italy, Netherlands, Sweden, Switzerland, United Kingdom
Primary Outcome
Definition: Composite of death from vascular disease, non-fatal stroke, non-fatal myocardial infarction, or systemic embolism
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 17% per year (placebo, Group 1) | 8% per year (anticoagulation, Group 1) | 0.53 (0.36-0.79) | 0.001 |
Limitations & Criticisms
- Anticoagulation arm was open-label — introduces potential for ascertainment bias, though endpoint adjudication was blinded.
- Target INR (2.5-4.0) is higher than the modern standard (INR 2.0-3.0); trial-era bleeding rates may overstate current risk with tighter INR control.
- Aspirin dose (300 mg) is higher than current secondary-prevention practice (75-100 mg); comparison may under-represent efficacy of low-dose aspirin.
- Trial predated CT/MRI standardisation and modern imaging-based stroke phenotyping; some 'minor stroke' events would be reclassified today.
- No comparison against newer direct oral anticoagulants (DOACs), which have since been shown to be at least as effective and safer than warfarin in AF.
- Enrolment limited to Europe; generalisability to non-European populations not directly tested.
Citation
Lancet 1993;342(8882):1255-1262. DOI: 10.1016/0140-6736(93)92358-Z. PMID: 7901582.