Warfarin versus aspirin for stroke prevention in the elderly study: Putting fears to rest

The

article by Mant and colleagues

concerning the Birmingham Atrial Fibrillation Treatment of the Aged (BAFTA) study is a welcomed addition to the now vast literature regarding the relative benefits of warfarin anticoagulation in patients with atrial fibrillation.¹ The study goes to the heart (no pun intended) of a paradox: despite most patients with atrial fibrillation being elderly, and many trials having been conducted to assess warfarin versus aspirin for stroke prevention in the setting of atrial fibrillation, the elderly have been considerably underrepresented in these trials. Although at least 29 clinical trials have demonstrated an overwhelming advantage of warfarin over aspirin therapy in preventing atrial fibrillation-related stroke, with Hart and associates reporting an efficacy of 60% compared with 20% for aspirin,² fewer than 50% of elderly patients with atrial fibrillation currently receive warfarin therapy.³

The BAFTA study included 973 patients older than 75 years (mean age, 81 years) with atrial fibrillation. Of these patients, 488 were randomized to warfarin therapy to achieve an international normalized ratio (INR) between 2 and 3, and 485 were randomized to receive aspirin at 75 mg/day. The primary end point was any stroke (ischemic or hemorrhagic) or other embolism, and the patients were followed-up for an average of 2.7 years. The results were conclusive and striking, demonstrating an annual risk of stroke of 1.8% in the warfarin-treated group versus 3.8% in the group treated with aspirin. Contrary to concerns resulting from data reported in previous studies,^4,5 this benefit was not offset by hemorrhagic complications in the warfarin-treated group. Furthermore, even the incidence of extracranial hemorrhage was somewhat lower in the warfarin cohort (1.4% vs 1.7% for aspirin), but this finding was not statistically significant.

A weakness of Mant and colleagues' study is that only 21% of potential candidates identified for the study were randomized, and over 50% of those excluded from the study were kept from participating because their primary care physician favored a particular treatment, most commonly warfarin. It would have been interesting had these patients been followed-up along with the randomized cohort and included in a supplemental analysis of benefit versus risk. However, when looking at this shortfall from another perspective, it may have been considered a strength because the study showed that even in patients with caregivers who had misgivings regarding warfarin or aspirin as the best therapy, warfarin use demonstrated a clear relative benefit.

Another weakness of the study is that the degree of brain imaging used to ascertain the primary outcome is not specified. The differential diagnosis of hemorrhagic or ischemic stroke seems to be left only to the discretion of the attending neurologist. Given the clinical similarity of the 2 conditions, I cannot conclude from this study what the true rate of intracranial hemorrhage (both hemorrhagic stroke and subdural hematoma) would be in this elderly population; however even this may be viewed as a strength, because if one focuses on the end point of disabling stroke, it is clear that warfarin's relative benefit is not offset by hemorrhage, either extracranial or intracranial.

Finally, it is worthwhile noting that warfarin treatment was about 50% effective for those intended to treat: only 67% of those patients continued warfarin therapy through their observation period, and those treated with warfarin yielded a therapeutic INR only 67% of the time. Whether or not this experience accurately reflects that of elderly populations treated in a primary care setting, it does suggest that the benefits of warfarin may be greater with closer monitoring.

My experience with atrial fibrillation and stroke is likely quite different from that of primary care physicians. The majority of patients who present to my institution with stroke secondary to atrial fibrillation are older than 75 years, and most of these patients have not received therapeutic anticoagulation previously. This finding was observed in a stroke study that several colleagues and I conducted, but our primary finding was that strokes produced by atrial fibrillation are generally much more severe than those attributed to other causes.⁶ Atrial fibrillation often produces large emboli that lodge in the stem of the middle cerebral artery, causing a highly lethal and dehumanizing stroke. Because strokes resulting from atrial fibrillation are generally so severe, prevention is imperative.

In a study by Hylek and associates, the most common reason given by caregivers for not using warfarin in elderly patients included falling risk and a history of other hemorrhages.⁷ Many other reasons can be added to this list, including compliance concerns in elderly shut-ins or in those with cognitive decline. While such concerns are clearly valid, they must always be weighed against the risk of severe ischemic stroke secondary to atrial fibrillation. Mant and colleagues' study assuages a major concern by demonstrating that warfarin can be relatively safe even in the very elderly.