Article a b s t r a c t C h r o n i c atrial fibrillation (AF) as a precursor of stroke was assessed over 24 years offollow-up of the general population sample a t Framingham, Massachusetts. Persons with chronic established AF, with or without rheumatic heart disease (RHD), are at greatly increased risk of stroke, and the stroke is probably due to embolism. Chronic AF in the absence of RHD is associated with more than a fivefold increase in stroke incidence, while AF with RHD has a 17-fold increase. Stroke occurrence increased as duration of AF increased, with no evidence of a particularly vulnerable period. Chronic idiopathic AF is an important precursor of cerebral embolism. Atrial fibrillation (AF) in rheumatic heart disease (RHD), particularly with mitral stenosis, is accepted as a factor that predisposes to systemic embolism.' Embolism from the left atrium occurs frequently and cerebral infarction is a common cause of death among RHD patients with m i t r a l stenosis.' There is no such agreement about risk of cerebral embolism in persons with chronic AF without rheumatic valvular disease. Beer and Ghitman3 found only one stroke among 50 patients with AF due to ischemic heart disease, a rate of cerebral embolism not appreciably different than the 2 percent occurring in ischemic heart disease patients without this arrhythmia. However, others have found systemic embolism to be as common in chronic AF with coronary and hypertensive heart disease as in RHD."."Since i t is likely that progress will come from prevention rather than from improved medical management of completed embolic strokes, it is important to determine if chronic AF predisposes to stroke. The least distorted view of the relationship of AF to stroke can be obtained through prospective epidemiologic study of a general population which is free of the biases of selection that exist in clinical and autopsy populations. We have studied the development of stroke in a population followed prospectively since 1950, and we have related stroke incidence to antecedent cardiac rhythm and disease. Methods.We evaluated the development of stroke in 5184 men and women, aged 30 to 62, and free of stroke at entry, followed for 24 years. Sampling procedure, criteria, and methods of examination have been described e l~e w h e r e .~,~ Subjects were examined every 2 years. Follow-up was good, with 81 percent taking all possible examinations and less than 5 percent of the original cohort lost to mortality follow-up.On each of the 13 biennial examinations, the subject was routinely questioned by a physician concerning habits, medications, and illnesses during the preceding 2 years. Physical examination and laboratory studies were made, and details surrounding all interim illnesses were sought. For stroke, including transient ischemic attacks (TIAs), surveillance was maintained by daily monitoring of all admissions to the only general hospital in town. If a stroke was suspected, the patient was seen in the hospital by the study neurologist. Neurologic symptoms or signs...

No abstract