In their retrospective follow-up study of all patients attending a hospital anticoagulant clinic for atrial fibrillation, Copland and colleagues1 found that the elderly population has both good anticoagulant control and a low incidence of hemorrhage while receiving warfarin therapy. Although interesting, from our point of view, these data cannot support the authors' general conclusion in favor of anticoagulant use. In fact, in the "Patients, Materials, and Methods" section, the authors note that prosthetic heart valve replacement was the only exclusion criterion, implicitly admitting the possibility of "a selection bias at the level of referral to the clinic."