Warfarin vs Aspirin for Atrial Fibrillation in the Elderly

Abstract & Commentary

By Michael H. Crawford, MD, Professor of Medicine, and Chief of Clinical Cardiology, at the University of California, San Francisco. Dr. Crawford is on the speaker's bureau for Pfizer.

Source: Mant J. et al. Warfarin Versus Aspirin for Stroke Prevention in An Elderly Community Population with Atrial Fibrillation (The Birmingham Atrial Fibrillation Treatment of the Aged Study, BAFTA): A Randomized Controlled Trial. Lancet 2007; 370:493-503.

Over half of patients with atrial fibrillation are over age 75 years, yet there is concern about the risk of serious hemorrhage in these patients on warfarin. Thus, the results of the Birmingham Atrial Fibrillation Treatment of the Aged (BAFTA) study are of interest. BAFTA was a prospective randomized open label study with a blinded assessment of end-points. The primary end-point was the frequency of stroke, intracranial hemorrhage or systemic emboli in patients with atrial fibrillation age 75 or greater who were randomized to aspirin 75 mg/day vs warfarin adjusted to an INR of 2-3, target 2.5. Secondary end-points included death, other vascular events and hemorrhage requiring hospitalization. The 973 patients were enrolled in primary care practices over a 3 year period and were followed for an additional 2 years. Average follow up was 2.7 years. Compliance with warfarin was 67% in this intention to treat analysis. Their INR's were therapeutic 67% of the time (19% below, 14% above). There were fewer primary end points in those on warfarin vs those on aspirin (1.8% per year vs 3.8% per year, RR = 0.48, CI 0.28-0.80, number needed to treat 50). No subgroup emerged in which warfarin was not superior to aspirin. The risk of major hemorrhage was small with 50 events (2%) per year and was not different in the 2 treatment groups. Secondary end point analysis showed that all major vascular events combined and strokes alone were less on warfarin. The authors concluded that in those age 75 or older with atrial fibrillation, warfarin is superior to aspirin for preventing embolic events including intracranial hemorrhage. However, about one third of the patients will decide the potential benefits are not worth the inconvenience of therapy with warfarin.


The ACC/AHA/ESC guidelines recommend warfarin for atrial fibrillation patients if they have 2 or more risk factors for stroke (CHADS) where age >75 years is one (A). On the other hand they caution that the risk of major hemorrhage is higher in the elderly, and caution that this risk needs to be considered. This is a mixed message that has resulted in a conservative approach by physicians and patients. Part of the problem is a lack of clinical trial data in patients over 75. Thus, this trial is of considerable interest.

The superiority of adjusted dose warfarin over low dose aspirin in this study is predictable. The surprise was that the risk of major hemorrhage was equal in both groups. This is especially surprising since low dose aspirin was used. The authors discuss several reasons that may help understand these results. Older trials had shown a 2-fold increase in major hemorrhage with warfarin vs aspirin and that major hemorrhage was more frequent as age increased. So this could be an alpha error in this trial. Even if this is the case the frequency of major hemorrhage will not be 2-fold higher. Some older trials used higher INR targets (up to 4.5); this trial used 2-3 (average 2.4). Current warfarin use before study entry was 40%. Therefore, there may have been a selection bias towards patients who could tolerate warfarin. There were significant cross-overs in this intention to treat study design. One third assigned warfarin switched to aspirin and 17% of those assigned to aspirin converted to warfarin. The net effect of these cross-overs would be to decrease the apparent risk of warfarin, but would also decrease the observed benefit of warfarin and the latter was not seen. About 20% of identified patients were excluded because of some contraindication to warfarin, but this does not seem excessive. Finally this was more of a real world study because patients were recruited form primary care practices rather then from hospitalized patients as has been done in other studies. Thus, they may have been a healthier group with fewer co-morbidities. Whatever the reason for the surprising results, this study makes us rethink what is a contraindication to warfarin therapy. Clearly older age alone is not a contraindication.