Abstract

Atrial fibrillation is said to predispose to embolic stroke. Patients with rheumatic heart disease in atrial fibrillation (AF) are usually treated with warfarin, even though no trial has ever shown that this reduces the risk. Nowadays AF usually has a non-rheumatic cause and is said to carry an increased risk of stroke, but anticoagulation is much less widely practised. Perhaps any association is not causal but simply reflects the coincidence of coronary artery and cerebrovascular disease. Anticoagulation would be worthwhile if it were clear it reduced the risk of embolic stroke and that the benefits exceeded the risks. How clear is the evidence?