Saphenous vein patch angioplasty is reported to yield superior results for carotid endarterectomy. In order to evaluate an alternative technique, which leaves the saphenous vein intact for other possible graft purposes, 200 carotid endarterectomies were included in a prospective randomized comparative study. Patients were randomized to two statistically equivalent groups: one group underwent classical carotid endarterectomy through a longitudinal incision with saphenous vein patch angioplasty; the other had endarterectomy through an oblique division of the internal carotid followed by in situ anastomosis. Cross-clamping time was approximately 5min shorter with the division-endarterectomy-anastomosis technique. The overall perioperative (< 30 days) mortality rate was 2.5% and cumulative mortality-morbidity rate 8% in the patch group compared with 4% in the other (P > 0.05). There were significantly more cranial nerve injuries in the patch group, most of which were transient (P < 0.01). The mean follow-up was 365 days. The late mortality rate was 5.5%. There were no late permanent or fatal strokes, but 3% of patients sustained mild transient neurological events. Only three significant (> 60%) stenoses developed during follow-up, all within 9 months. Dilatation and disturbed flow were more pronounced in the patch group (P < 0.05). There were no statistically significant differences between both techniques on mortality, disabling neurological morbidity and recurrent stenosis. In conclusion, the results of the division-endarterectomy-anastomosis technique are equivalent to those with patch angioplasty, leaving the patient's venous capital intact.