PDF: AVERROES
(2011)Objective
To determine whether apixaban is superior to aspirin in preventing stroke or systemic embolism in patients with atrial fibrillation who are unsuitable for vitamin K antagonist therapy.
Study Summary
• Apixaban significantly reduced the risk of stroke and systemic embolism compared to aspirin in atrial fibrillation patients not suitable for warfarin, without a significant increase in major bleeding..
Intervention
Apixaban 5 mg BID (2.5 mg BID in select patients) vs. aspirin 81–324 mg daily. Median follow-up: 1.1 years.
Inclusion Criteria
Patients with atrial fibrillation and at least one risk factor for stroke (age ≥75, hypertension, prior stroke/TIA, heart failure, diabetes), and for whom vitamin K antagonists were unsuitable due to bleeding risk, monitoring issues, or patient refusal.
Study Design
Arms: Apixaban vs. Aspirin
Patients per Arm: Apixaban: 2808, Aspirin: 2791
Outcome
• Primary outcome (stroke or systemic embolism): 1.6%/yr (apixaban) vs. 3.7%/yr (aspirin), HR 0.45, 95% CI 0.32–0.62, P<0.001.• Ischemic stroke: 1.1%/yr vs. 3.0%/yr.• Major bleeding: 1.4%/yr (apixaban) vs. 1.2%/yr (aspirin), HR 1.13, 95% CI 0.74–1.75, P=0.57.• Intracranial hemorrhage: 0.4%/yr vs. 0.4%/yr.• All-cause mortality: 3.5%/yr vs. 4.4%/yr, HR 0.79, P=0.07.
Bottom Line
Apixaban significantly reduced stroke/systemic embolism risk compared to aspirin without a significant increase in major bleeding.
Major Points
- AVERROES tested apixaban vs aspirin in AF patients deemed UNSUITABLE for warfarin — filling a critical therapeutic gap for the ~40% of AF patients who don't receive anticoagulation due to perceived warfarin intolerance, high bleeding risk, or patient/physician preference.
- Stopped early by DSMB after 5,599 patients due to overwhelming efficacy: apixaban reduced stroke/SE by 55% vs aspirin (1.6% vs 3.7%/yr; HR 0.45, 95% CI 0.32–0.62, p<0.001). This was the largest relative risk reduction seen in any major AF anticoagulation trial.
- No significant increase in major bleeding: 1.4%/yr apixaban vs 1.2%/yr aspirin (HR 1.13, p=0.57) — remarkable given the 55% stroke reduction. ICH rates were identical (0.4%/yr in both groups).
- All-cause mortality trended lower with apixaban: 3.5% vs 4.4%/yr (HR 0.79, p=0.07) — not significant, but the trial was stopped early before mortality benefit could mature. CV hospitalization was significantly reduced (12.6% vs 15.9%/yr, p<0.001).
- 5,599 patients across 522 centers in 36 countries. Double-blind, double-dummy design (gold standard) — each patient received either real apixaban + sham aspirin, or real aspirin + sham apixaban.
- Reasons for VKA unsuitability: physician judgment (40%), patient refusal (37%), unable to maintain stable INR (22%), prior bleeding on warfarin (13%), poor access to INR monitoring (8%). Many had multiple reasons.
- Mean CHADS2 score was 2.0 — moderate risk population. 14% had prior stroke/TIA (secondary prevention). These patients had higher event rates and derived even greater absolute benefit from apixaban.
- Apixaban dose: 5 mg BID standard, with 2.5 mg BID for patients meeting ≥2 of: age ≥80, weight ≤60 kg, creatinine ≥1.5 mg/dL — the same dose-reduction criteria later used in ARISTOTLE.
- AVERROES complemented ACTIVE A (clopidogrel+aspirin vs aspirin in warfarin-ineligible AF) — together they showed that for warfarin-ineligible patients, apixaban is vastly superior to aspirin, and DAPT adds only modest benefit over aspirin alone.
- Established apixaban as the preferred anticoagulant for VKA-unsuitable AF patients — the 2019 AHA/ACC/HRS AF guidelines gave apixaban a Class I recommendation for warfarin-ineligible patients based primarily on AVERROES.
Study Design
- Study Type
- Randomized, double-blind, double-dummy, placebo-controlled
- Randomization
- Yes
- Blinding
- Double-blind
- Sample Size
- 5599
- Follow-up
- 1.1 years (mean)
- Centers
- 522
- Countries
- 36 countries including North America, Latin America, Europe, Asia, South Africa
Primary Outcome
Definition: Stroke or systemic embolism
| Control | Intervention | HR/OR | P-value |
|---|---|---|---|
| 3.7%/yr | 1.6%/yr | 0.45 (0.32–0.62) | <0.001 |
Limitations & Criticisms
- Trial stopped early (mean 1.1 years) due to DSMB efficacy finding — early stopping systematically overestimates treatment effects (Pocock bias). The true HR may be somewhat less dramatic than 0.45, though still clearly significant.
- VKA 'unsuitability' was loosely defined and physician-assessed — 40% cited 'physician judgment' as the primary reason, making this a heterogeneous population. Some of these patients likely could have tolerated DOACs or even warfarin with better support.
- Wide aspirin dose range (81–324 mg) — though no interaction with dose, this variability means some aspirin-arm patients may have been undertreated (81 mg) while others were on doses associated with more bleeding (324 mg) without additional efficacy in AF.
- No warfarin comparator arm — the trial compared apixaban to aspirin only. AVERROES cannot directly inform the apixaban vs warfarin comparison (that was ARISTOTLE's role). The clinical question is narrow: apixaban vs aspirin in warfarin-unsuitable patients.
- Short follow-up (1.1 years) — long-term safety, bleeding accumulation, and durability of benefit beyond 1 year remain uncertain. Chronic anticoagulation carries cumulative bleeding risk that may differ from short-term data.
- Low rate of dose-reduced apixaban (~6% received 2.5 mg BID) — most patients received full dose, limiting conclusions about the reduced-dose regimen's efficacy in this specific population.
- Industry-sponsored (BMS/Pfizer — manufacturers of Eliquis/apixaban) — with industry involvement in study design, conduct, and analysis.
- No prolonged cardiac monitoring at baseline — some patients classified as 'VKA-unsuitable AF' may actually have had infrequent paroxysmal AF that wouldn't be detected on standard ECG, potentially diluting the true treatment effect.
- Cannot determine if the benefit was from factor Xa inhibition specifically or from any anticoagulant — though RE-LY and ROCKET AF suggest the benefit is a class effect of anticoagulation rather than drug-specific.
Citation
Connolly SJ, et al. N Engl J Med. 2011;364(9):806-817.