In a collaborative prospective study from two institutions, we reviewed the clinical course of 969 consecutive patients who had 1200 carotid endarterectomies (CEs) for the treatment of occlusive arterial disease during the period 1977 to 1987. The indications for CE comprised transient ischemic attacks (TIAs) in 581 cases (48.4%), cerebral infarction (CI) in 170 (14.2%), monocular blindness in 166 (13.8%), and asymptomatic stenosis in 283 (23.6%). Neurologic monitoring of the awake patient provided more reliable indication of the need for brain protection during operative arterial clamping than did electroencephalography or carotid stump pressure measurement. Of the 1200 cases, 113 (9%) required a shunt as determined by this method. Patients with contralateral carotid occlusion or severe stenosis required shunting six times more frequently than those with a unilateral lesion. Among all procedures, there were nine cases of transient neurologic deficit (0.9%), 11 cases of permanent neurologic deficit (0.9%), and eight deaths (0.67%). Among 283 CEs performed to treat asymptomatic lesions, no strokes and only one death (0.4%) occurred. One hundred sixty-six cases with amaurosis fugax were operated on without stroke or death. In the age group of 70 to 90 years, 508 procedures were carried out with four deaths (0.8%) and three strokes (0.6%). We conclude that CE performed with the patient under local anesthesia is safe and effective and permits satisfactory management of old and high-risk patients.