Despite a vast body of research on antithrombotic therapy for patients with cardiac disease, there are few clinical settings where robust evidence of their benefit exists. Patients with heart failure often have vascular disease and atrial fibrillation contributing to their poor prognosis. For patients with heart failure and atrial fibrillation, anticoagulants are appropriate. For patients with heart failure in sinus rhythm, the weight of evidence suggests that doctors should generally avoid using any antithrombotic agent even if the patient has coronary artery disease. If there is a compulsion to treat, then there is less evidence of harm with clopidogrel or warfarin than with aspirin, although most receive aspirin. More research is required for this "evidence-light" problem. For those with the opportunity, engaging with a randomized trial is clinically and scientifically appropriate. The dilemma for such studies is the comparator. Should it be against or in addition to "standard of care" or both?