What’s New in the Patient Safety World


October 2009

Warfarin for Atrial Fibrillation – Too Much? Or Just the Wrong Patients?


For years we have heard over and over that anticoagulation with warfarin is underutilized in patients with atrial fibrillation. We’ve heard that most patients with atrial fibrillation other than those with uncomplicated atrial fibrillation should be fully anticoagulated unless contraindications were present. We’ve even had complex calculations tell us the number needed to treat and the financial implications of strokes that could have been avoided by use of warfarin.


Now a new study (Singer 2009) has challenged conventional wisdom by looking at a new risk stratification of patients with atrial fibrillation and showing the benefit of warfarin is not uniform in patients with atrial fibrillation. The study used data from a large healthplan database (the ATRIA study cohort) to look at outcomes in patients with atrial fibrillation who were stratified by the CHAD2 score. The CHAD2 score assigns one point for CHF, diabetes, hypertension, age 75 or older and two points for a history of prior stroke or TIA. It also utilized an outcome formula that included not only the outcomes of ischemic stroke and systemic emboli but also the risk of intracranial hemorrhages attributable to anticoagulation (thus the net benefit).


As expected there was a net clinical benefit for the whole group on warfarin. But the subgroup analysis revealed some surprising findings. The net benefit was greatest for those patients having a history of prior stroke and those age 85 and older. The latter group, of course, is one in which the argument of increased risk for intracranial hemorrhage has often used to justify not anticoagulating.


There was a clearcut benefit for warfarin in patients with the highest CHAD2 score (4-6) but there was no net benefit in those with a CHAD2 score of 0-1. Clinical benefit began for those with a CHAD2 score of 2. Overall, almost half of the patients anticoagulated with warfarin had no clinical benefit. Though the study was retrospective and used administrative data to determine outcomes, it confirmed the value of using CHAD2 scores to stratify patients and confirmed the benefit of anticoagulation in the very elderly that had also been seen in the BAFTA trial.


Lessons from this study?

1)      Don’t ignore the very elderly with atrial fibrillation. They have the most to gain from anticoagulation.

2)      Patients with prior stroke or TIA also have the most to gain from anticoagulation.

3)      Patients with few clinical risk factors (CHAD2 score of 0-1) are not likely to have a net benefit from anticoagulation.


So not only have we probably been anticoagulating many patients in whom there is no net benefit, we have probably also been ignoring the group in which the most net benefit is seen.





Singer DE, Chang Y, Fang MC, et al. The Net Clinical Benefit of Warfarin Anticoagulation in Atrial Fibrillation. Ann Intern Med 2009; 151(5): 297-305













Patient Safety Tip of the Week Archive


What’s New in the Patient Safety World Archive