OBJECTIVE Studies comparing transcranial Doppler ultrasonography (TCD) with other intraoperative monitoring techniques for detecting clamping ischaemia during carotid endarterectomy under general anaesthesia suggest that a reduction of > two-thirds in the mean middle cerebral artery velocity (mMCAv) or a reduction of > 0.4 in the preclamping mMCAv: clamping mMCAv ratio warrants cerebral protection. Our aim was to study the relationship between mMCAvs and clamping ischaemia during carotid endarterectomy in awake patients. METHODS In a consecutive series of 57 patients undergoing carotid endarterectomy under locoregional anaesthesia 51 were monitored by intraoperative TCD, continuous EEG, and neurologic awake testing. RESULTS Five of the 51 (9.8%) patients had transient clamping ischaemia, which carotid shunting reversed. TCD showed that these five patients had significant lower mean mMCAvs than the other 46 patients, who had no deficits (1.8 +/- 1.1 cm/s vs. 26.2 +/- 8.5, p = 0.0003). Current TCD criteria indicated that four other patients (7.8%) should have been shunted. All four had significantly higher clamping mMCAvs than the five shunted patients (11.5 +/- 1.9 vs. 1.8 +/- 1.1, p = 0.0012). CONCLUSIONS Intraoperative TCD detected cerebral ischaemia and yielded no false-negative. An mMCAv of 10 cm/s or less may indicate the risk of clamping ischaemia better than the higher threshold currently proposed. This would avoid unnecessary shunting due to false-positives.