To test the hypothesis that the transcranial Doppler is a useful technique for intraoperative neuromonitoring, we prospectively used it to study 83 patients undergoing carotid end-arterectomy. A 2 MHz pulsed-wave, range-gated transcranial Doppler was positioned at the transtemporal window after induction of general anesthesia. Mean middle cerebral artery velocity, intraarterial blood pressure, end-tidal PCO2, heart rate, and a spectral array of electroencephalographic activity were recorded continuously throughout the operation. Internal carotid artery back pressure was measured routinely. On completion of the endarterectomy, duplex ultrasound examinations and arteriograms were uniformly obtained to assess technical adequacy. Forty-nine of the 83 patients (60%) had complete preoperative and intraoperative transcranial Doppler examinations. Eleven (13%) had incomplete assessments because of small or absent transtemporal windows. Twenty-three (27%) had unsuccessful monitoring because of technical difficulties, primarily because of inability to maintain probe position--with loss of mean middle cerebral artery velocity recording. In the patients with complete studies, transcranial Doppler failed to provide information that altered surgical therapy. All monitoring modalities were normal in the one patient (1.2%) who sustained an operative stroke. We conclude that at this time, transcranial Doppler has not been useful to routinely monitor the intraoperative events during carotid endarterectomy.