Two hundred ninety-three carotid endarterectomies were performed with electroencephalogram (EEG) monitoring and without the use of a shunt. Two hundred sixteen patients had contralateral carotid stenosis of less than 70%; 45 had contralateral stenosis of 70 to 99%; and 32 had contralateral occlusion. There were six perioperative strokes (2.0%) and two deaths (0.7%). Major EEG changes were seen in 11 of the 77 patients (14.3%) with significant contralateral stenosis or occlusion versus 11 of the 216 patients (5.1%) in those without (P less than 0.025). The risk of immediate postoperative deficit was significantly higher in the subgroup with major EEG changes (4 of 22, 18.2%) than in those without such changes (5 of 271, 1.8%) (P less than 0.005). The risk in patients with less than 70% contralateral stenosis (7 of 216, 3.2%) was not significantly different from those with greater contralateral stenosis or occlusion (2 of 77, 2.6%). Carotid endarterectomy can be safely performed without a temporary shunt. Contralateral stenosis or occlusion alone does not confer increased risk. Major EEG changes are infrequent, but they identify a subgroup with significantly higher risk of intraoperative stroke.