Background: While atrial fibrillation (AF) increases the risk of cardioembolic stroke, some ischemic strokes in AF patients are noncardioembolic. Objectives: To assess ischemic stroke mechanisms in AF and to compare their responses to antithrombotic therapies. Methods: On-therapy analyses of ischemic strokes occurring in 3,950 participants in the Stroke Prevention in Atrial Fibrillation I–III clinical trials. Strokes were classified by presumed mechanism according to specified neurologic features by neurologists unaware of antithrombotic therapy. Results: Of 217 ischemic strokes, 52% were classified as probably cardioembolic, 24% as noncardioembolic, and 24% as of uncertain cause (i.e., 68% of classifiable infarcts were deemed cardioembolic). Compared to those receiving placebo or no antithrombotic therapy, the proportion of cardioembolic stroke was lower in patients taking adjusted-dose warfarin (p = 0.02), while the proportion of noncardioembolic stroke was lower in those taking aspirin (p = 0.06). Most (56%) ischemic strokes occurring in AF patients taking adjusted-dose warfarin were noncardioembolic vs. 16% of strokes in those taking aspirin. Adjusted-dose warfarin reduced cardioembolic strokes by 83% (p < 0.001) relative to aspirin. Cardioembolic strokes were particularly disabling (p = 0.05). Conclusions: Most ischemic strokes in AF patients are probably cardioembolic, and these are sharply reduced by adjusted-dose warfarin. Aspirin in AF patients appears to primarily reduce noncardioembolic strokes. AF patients at highest risk for stroke have the highest rates of cardioembolic stroke and have the greatest reduction in stroke by warfarin.

No abstract

Ischemic strokes occurring in patients with nonrheumatic atrial fibrillation are due to a variety of mechanisms, not exclusively to cardiogenic embolism. Without knowledge of antithrombotic therapy assignment, we categorized strokes in the Stroke Prevention in Atrial Fibrillation Study as presumed cardioembolic or noncardioembolic. We then compared patient clinical and echocardiographic variables, as well as the efficacy of aspirin prophylaxis, for each stroke type. Of 71 ischemic strokes, we categorized 46 (65%) as cardioembolic, 13 (18%) as noncardioembolic, and 12 (17%) as of uncertain cause. Patients developing noncardioembolic strokes, relative to cardioembolic strokes, were more commonly men (p = 0.005) and were more likely to have left ventricular wall motion abnormalities by two-dimensional echocardiography (p = 0.002). Aspirin reduced the occurrence of strokes categorized as noncardioembolic significantly more than it did those categorized as cardioembolic (p = 0.01). These results emphasize the value of considering stroke mechanisms in therapeutic trials of antithrombotic agents and suggest a differential effect of aspirin according to mechanism.

No abstract

No abstract