Warfarin superior to aspirin for stroke prevention in AF: BAFTA published

Susan Jeffrey

August 10, 2007

Birmingham, UK - Results of the Birmingham Atrial Fibrillation Treatment of the Aged (BAFTA) trial show that among patients over age 75 years with atrial fibrillation (AF), anticoagulation with warfarin was superior to aspirin for primary stroke prevention[1]. The benefit of treatment did not come at the cost of more major hemorrhage, the rates of which were similar between groups.

The results are published in the August 11, 2007 issue of the Lancet. They were presented previously at the 16th European Stroke Conference in Glasgow, Scotland, and reported at that time.

"In summary, these data lend support to the use of anticoagulation for all people aged over 75 years who have atrial fibrillation, unless there are contraindications or the patient decides the size of the benefit is not worth the inconvenience of the treatment," BAFTA investigators, with lead author Dr Jonathan W Mant (University of Birmingham, UK), conclude in their report.

The target international normalized ratio (INR) should be 2 to 3, they add, a range for which there is "clear evidence of benefit and no evidence from this study of harm compared with aspirin. Age itself should not be regarded as a contraindication to anticoagulation therapy."

Elderly primary-care population

Warfarin has been shown to be superior to aspirin for the prevention of stroke in patients with AF, but the large trials that established the efficacy of anticoagulation over aspirin therapy generally enrolled younger patients, in whom the risk for bleeding complications is lower, the authors point out.

The majority of strokes associated with AF occur among those over the age of 75 years. However, Mant said, "there are concerns about the applicability of the evidence both to the elderly, where there is concern that the increased risk of bleeding on warfarin might outweigh the benefits, and also in community-based populations such as primary care, where, of the three trials that have looked at warfarin vs aspirin, two have in fact been negative."

To address these concerns, the BAFTA trial was undertaken, a randomized, controlled trial comparing adjusted-dose warfarin with a target INR of 2.5 (range 2.0-3.0) with aspirin in a dose of 75 mg daily among elderly AF patients taken from primary-care settings.

There were 973 patients with AF older than 75 years enrolled from more than 260 general practices in England and Wales. Patients were followed up at three months after randomization by their general practitioner, then every six months for an average of 2.7 years.

The primary end point was fatal or nonfatal disabling stroke, either ischemic or hemorrhagic, or significant arterial embolism. "We deliberately chose quite a hard outcome measure," Mant said. "Nondisabling stroke was not part of our outcome, but we included hemorrhagic stroke—including intracranial hemorrhage [ICH]—in our primary outcome, so we could make a clear conclusion."

And the conclusion was clear: a significant reduction in the risk for a primary outcome event (p=0.003). There were 24 primary outcome events in the warfarin group, including 21 strokes, two other intracranial hemorrhages, and one systemic embolus, vs 48 such events in the aspirin group, including 44 strokes, one other intracranial hemorrhage, and three systemic emboli.

This resulted in a number needed to treat (NNT) of 50 patients treated for one year to prevent one primary outcome event. The absolute yearly risk reduction was 2% (95% CI 0.7-3.2).

BAFTA: Primary analysis

End point Warfarin Aspirin Hazard ratio (95% CI) NNT
Fatal or nonfatal disabling stroke or significant arterial embolism (%/annum) 1.8 3.8 0.48 (0.28-0.80) 50

There was no difference in major hemorrhage between groups, although the confidence intervals were wide for this end point because there were only 50 events in all, they note. Similarly, there was no difference between other hospital admissions for hemorrhage or the composite of all major hemorrhages, including major ICH.

BAFTA: Bleeding complications with warfarin vs aspirin in AF patients older than 75 years

End point Warfarin Aspirin Hazard ratio (95% CI)
Major extracranial hemorrhage (%/annum) 1.4 1.6 0.87 (0.43-1.73)
All major hemorrhages (%/annum) 1.9 2.2 0.96 (0.53-1.75)

Although the study was not powered to look at subgroups, the authors point out that the risk for hemorrhage rose with age, "but I would point out it rises just as much in the aspirin group, if not more so, than in the warfarin group," Mant noted during his presentation of these data. There was no evidence of any interaction with age in the benefit seen with warfarin or in the harm seen with aspirin, he added, "so it appears the same result is coming through in people aged 85 or older at randomization."

No differences were seen in other secondary outcomes, including all-cause mortality, other vascular mortality, or nonvascular deaths. There were no differences in vascular events between warfarin and aspirin, although one end point, a composite of major vascular events combining stroke, MI, pulmonary embolus, and vascular death, did show a significant reduction with warfarin vs aspirin.

Lack of difference in bleeding risk a surprise

Mant speculated on some of the potential reasons for what he called the "most surprising finding of the BAFTA study," the lack of difference in hemorrhage risk between the groups. Previous studies, however, used higher target INR ranges, up to 4.5, "and it may be that we've overestimated the danger of warfarin because of those studies," he said.

Some 40% of patients in BAFTA had previously been treated with warfarin, which results of the ACTIVE-W trial suggested may be associated with a lower bleeding risk. However, he noted, "we did look at that in a secondary subgroup analysis, and in our trial there was no important difference in the risk for hemorrhage between people who were new to warfarin and those who'd been on warfarin for some time."

At the end of the trial, three-quarters of patients in the aspirin group were still on their assigned therapy, while two-thirds of patients assigned to warfarin were still on the drug to which they had been initially assigned; most of those not on their original treatment assignment crossed over to the other trial medication.

The effect of the treatment crossovers would probably have been to underestimate both the benefits of warfarin in preventing ischemic events and possibly the risk for harm, he noted. "We did do a secondary on-treatment analysis for harm, and that actually made no difference to the results; we still found no difference between warfarin and aspirin when we did an on-treatment analysis looking at hemorrhage as opposed to intention-to-treat."

Important new information for the elderly

In an editorial accompanying the paper[2], Dr David Garcia (University of New Mexico, Albuquerque) and Dr Elaine Hylek (Boston University, MA) discuss whether these results can be extrapolated to the "real-world" population of elderly patients with AF. Because the study was limited to patients for whom clinicians were uncertain as to the best treatment, this probably encouraged recruitment of patients at a lower risk for stroke, they note.

Compared with other study populations, BAFTA participants had a lower prevalence of risk factors for stroke, and just over one-fifth of patients were excluded because warfarin was the only appropriate treatment. The low prevalence of risk factors and the large population of patients already taking a vitamin K antagonist when they entered the study may explain why thrombotic events were lower than anticipated in this study, they write.

"Nevertheless, the fact that BAFTA showed that warfarin is more effective than aspirin, even in this relatively low-risk group of patients, adds to other evidence that, in patients with atrial fibrillation, anticoagulation protects patients against stroke more effectively than antiplatelet therapy."

They agree, they note, with Mant and colleagues, that the lack of difference in major bleeding between the groups is surprising, and the bleeding rate was significantly lower than rates seen in other studies. They speculate on some of the reasons for this, suggesting that more information on the risk factors for hemorrhage in these subjects would be informative.

"Despite these considerations, BAFTA adds important new information for the care of elderly patients with atrial fibrillation," they write. "Mant and colleagues enrolled an unprecedented number of patients in an age group that that has been largely underrepresented in randomized trials. BAFTA firmly establishes the superior efficacy of warfarin as a stroke prevention strategy in elderly patients with atrial fibrillation."

The study was funded by the Medical Research Council. Mant reports no relevant disclosures. Disclosures for other coauthors appear in the paper. Garcia and Hylek have received research funding from Bristol-Myers Squibb and acted as consultants for Bristol-Myers Squibb. Hylek also received research funding from AstraZeneca.


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.