Comparing the Imperfect With the Imperfect
The Imprecise Science of Assessing the Risk and Benefits of Anticoagulation in Atrial Fibrillation
Atrial fibrillation (AF) is the most common arrhythmia in the United States, with ≈7 million Americans estimated to have AF by 2020.1,2 A major cause of morbidity and mortality in AF is stroke. Pharmacological therapy for the prevention of stroke has undergone a renaissance with the advent of newer oral anticoagulants that are safe and effective alternatives to warfarin. However, the decision to initiate anticoagulation remains a subjective assessment of risks versus benefits. Although guided by well-validated risk scores for stroke and bleeding,3–5 real-world decisions on anticoagulation continue to differ significantly from guideline recommendations, with many patients at high risk not receiving anticoagulation because of a perceived high risk of bleeding and many low-risk patients being anticoagulated as a result of a perceived low risk of bleeding, the so-called risk-treatment paradox.6 This phenomenon, thought to account for the continued underuse of oral anticoagulation therapy,7 has yet to be validated in a large outpatient-based practice. Therefore, the present study by Steinberg et al8 in this issue of Circulation is a timely effort to better understand and address some of the reasons underlying this risk mismatch in thromboembolic assessment and anticoagulation therapy in a community outpatient-based group of patients with stable AF.
Article see p 2005
The authors examined 10 094 AF patients enrolled in the Outcomes Registry for Better Informed Treatment of Atrial Fibrillation (ORBIT-AF) between June 2010 and August 2011.9 This US prospective registry of incident and prevalent AF is formed by a multispecialty collaboration of healthcare providers, including primary care physicians, cardiologists, and electrophysiologists. It is the largest clinical registry of its kind in the United States, enrolling ≈10 000 patients from 200 US outpatient practices. The patients are followed up for at least 2 years to characterize real-world …