Carotid endarterectomy for nonhemispheric symptoms: predictors of success.

Over a 4-year period 335 patients underwent 402 carotid endarterectomies: 227 (56%) for carotid territory symptoms, 107 (27%) for nonhemispheric symptoms, and 68 (17%) for asymptomatic lesions. In 317 four-vessel arteriograms, proximal subclavian and/or vertebral lesions were found more frequently in the nonhemispheric group (80 of 97, or 82%) than in the carotid territory group (110 of 220, or 50%; p less than 0.05). EEG changes after carotid clamping occurred more often in the nonhemispheric group (15 of 65, or 23%) than in the carotid territory group (16 of 140, or 11%; p less than 0.001). The perioperative stroke rate was independent of whether nonhemispheric or carotid territory symptoms were present preoperatively (2.8% vs. 3.5%). Follow-up ranged from 1 month to 4 years. Carotid endarterectomy was successful in ameliorating symptoms in patients meeting the criteria for "classic" vertebrobasilar insufficiency more often than in patients not meeting these criteria (73% vs 43% asymptomatic at 24 months). Carotid endarterectomy was successful in patients with carotid stenoses of greater than 60% diameter reduction more often than in patients with smaller stenoses (77% vs. 36% asymptomatic at 24 months). Carotid endarterectomy appears justified in patients with nonhemispheric symptoms when classic vertebrobasilar insufficiency and/or hemodynamically significant carotid stenoses are present.