The clinical course of 45 patients treated by carotid endarterectomy over the past 5 years is described with emphasis on the following three points: 1) Diagnostic methods, namely digital subtraction angiography (DSA) and B-mode Doppler imaging technique; 2) surgical procedure using an improved shunt tube and surgical instruments; and 3) monitoring before and during surgery. All operations were conducted using a shunt. Morbidity and mortality rates were both 0%. Postoperative transient hemiparesis lasting for 6 hours was recognized in only four cases. The total percentage of correct diagnoses using intravenous DSA compared with conventional angiography was approximately 80%. The accuracy of the non-invasive B-mode Doppler technique in measuring the degree of constriction compared with conventional angiography was 84%. The shunt was made of silicone tubing and was based on a tube 30 cm in length and 3.5 mm in diameter which was a T-shaped loop. Different sized bulbs were fixed to each end of the tube to prevent extravascular deviation. Modified bulldog clamps and Sugita clips were used for fixation in the vessel. Regional cerebral blood flow (rCBF) measurement and electroencephalography (EEG) under contralateral Matas procedure were conducted before surgery, and cross circulation during short-term occlusion of the common carotid artery was evaluated. The emergence or increase of delta waves in EEG during occlusion was observed in six cases. The rCBF of the affected middle cerebral artery territory in these patients was lower than that in patients with no increase of delta waves. Furthermore, the mean stump pressure during surgery in cases with preoperative EEG changes was 40 mmHg and that in cases without changes was 63 mmHg; these values were significantly different.(ABSTRACT TRUNCATED AT 250 WORDS)