Abstract Study Objective To evaluate the relationship between continuous noninvasive monitoring of cerebral saturation (regional cerebral oxygen saturation [rS o 2]) and occurrence of clinical and electroencephalographic (EEG) signs of cerebral ischemia during carotid cross-clamping. Design Prospective clinical study. Setting University hospital. Patients Fifty ASA physical status II and III inpatients undergoing elective carotid endarterectomy with a cervical plexus block. Interventions rS o 2 was continuously monitored throughout surgery, while an independent neurologist evaluated the occurrence of both clinical and EEG signs of cerebral ischemia induced during carotid cross-clamping. Measurements and Main Results rS o 2 was recorded 1 and 3 minutes after clamping the carotid artery during a 3-minute clamping test. In 5 patients (10%), the carotid clamping test was associated with the occurrence of clinical and EEG signs of cerebral ischemia. All these patients were treated with the placement of a Javid shunt, which completely resolved the symptoms. In no patient was permanent neurological injury reported at hospital discharge. In 4 of these patients, EEG signs of cerebral ischemia were present at both observation times, and in one of them, the duration of cerebral ischemia was less than 2 minutes. The percentage rS o 2 reduction from baseline during the carotid clamping test was 17% ± 4% in patients requiring shunt placement and only 8% ± 6% in those who did not require it ( P = .01). A decrease in rS o 2 15% or greater during the carotid clamping test was associated with a 20-fold increase in the odd for developing severe cerebral ischemia (odds ratio, 20; 95% confidence interval, 6.7-59.2) ( P = .001); however, this threshold had a 44% sensitivity and 82% specificity, with only 94% negative predictive value. Conclusions Continuous rS o 2 monitoring is a simple and noninvasive method that correlates with the development of clinical and EEG signs of cerebral ischemia during carotid cross-clamping; however, we could not identify an rS o 2 threshold that can be used alone to predict the need for shunt placement because of the low sensitivity and specificity.