Objective: The incidence of stroke associated with atrial fibrillation, even in high-risk patients, can be reduced significantly by adequate anticoagulation. However, anticoagulation does not abolish the stroke rate, and unfortunately only 40% of patients with atrial fibrillation actually receive anticoagulant therapy, even in areas where adequate health care is available. Methods: During the past 11.5 years, we have performed the maze procedure for the treatment of medically refractory atrial fibrillation in 306 patients, 58 of whom presented with a history of having had a stroke (n = 40) or transient ischemic attack (n = 18) before surgery. All patients with atrial fibrillation are at an increaSed risk for these complications, but they are especially prevalent in those patients with previous thromboembolic events and those with other recognized risk factors. Results: Among the 306 patients who had surgery, there were only 2 perioperative strokes (0.7%), and in the 265 patients followed for up to 11.5 years after the maze procedure, there has been only 1 late minor stroke that has now completely resolved. Conclusions: The ability of the maze procedure to decrease the risk of stroke associated with atrial fibrillation so dramatically is likely due to the restoring of sinus rhythm and atrial transport function in combination with surgical removal or obliteration of the left atrial appendage, where most thrombi associated with atrial fibrillation develop.