Atrial fibrillation and stroke: new ideas, persisting dilemmas.

The risk of systemic embolism associated with atrial fibrillation (AF) in patients with rheumatic mitral valve disease has long been appreciated. In only the past decade has it become clear that nonrheumatic (nonvalvulopathic) AF is a marker of increased risk for ischemic stroke.Several recent clinical and epidemiologic surveys have confirmed this association (Table 1) and suggest a fivefold increased risk of stroke. At least 15% of all ischemic strokes and more than a third of ischemic strokes in the elderly are associated with AF."About one in three people with AF will experience a stroke during their lifetime.'An estimated 75,000 strokes occur each year among the 1-1.5 million North Americans with AF. In addition to these clinical strokes, AF has been associated with an undue risk of subclinical, "silent" strokes.Silent infarcts detected by computed tomography (CT) have been found in 35-37% of nonrheumatic AF patients with no history of stroke.In a preliminary report including patients with rheumatic and nonrheumatic AF but without previous stroke, CT evidence of previous stroke was 2.9-4.5 times as frequent as in non-AF patients. Kempster and colleagues found CT evidence of remote infarct in 13% of patients with AF-related stroke compared with 4% of non-AF controls. While these CT-defined infarcts are labeled asymptomatic or "silent," it is likely that they take a subtle, but cumulative, toll on cognition in elderly people. Considering the combined risk of clinical and subclinical stroke, AF becomes a substantial threat to the brain. While the importance of AF-associated stroke is not in doubt, preventive strategies and management have remained empiric and controversial. Despite vigorous ongoing work by several groups of investigators, major unanswered questions remain in three related areas: the mechanism(s) of AF-