The results following carotid endarterectomy were prospectively evaluated in 134 patients (145 sides) by repeat ultrasonic duplex scanning and clinical evaluation extending for a period of 4 years. There were 107 men and 27 women in the study group. The perioperative stroke rate was 1.3% and the mortality rate, 0.7%. There were 9 late deaths, of which two were stroke related (1.4%). Focal symptoms occurred in 12 patients on the ipsilateral side, six of which were strokes (one lacunar). The remaining symptoms developed in the presence of moderate degrees of carotid stenosis (less than 50%). There were seven patients who had transient ischemic attacks (TIAs) referable to the operated side, but only two of these were associated with a recurrent high-grade stenosis. During follow-up 32 (22%) patients had recurrent high-grade stenosis. Restenosis regressed in seven, giving a persistent rate of 17.1%. The incidence of restenosis was significantly higher in women (p less than 0.01). By life-table analysis, restenosis occurred early, the majority within 24 months. There was no consistent association between the development of symptoms and the occurrence of restenosis. Therefore, it is concluded that there is no justification for reoperation based on the degree of narrowing observed to prevent subsequent TIAs and strokes.