Cerebral monitoring during carotid endarterectomy--a comparison between electroencephalography, transcranial cerebral oximetry and carotid stump pressure.

OBJECTIVE Various modalities are used for cerebral monitoring during carotid endarterectomy (CEA). The aim of this study was to evaluate whether transcranial cerebral oximetry (TCO) and carotid stump pressure (SP) are as accurate as electroencephalography (EEG) for monitoring cerebral ischaemia during carotid cross-clamping. METHODS One hundred consecutive patients who underwent CEA were studied with continuous and simultaneous EEG and TCO. SP was measured for each patient. The percentage decrease of oxygenation on TCO was calculated during cross-clamping and surgery. EEG findings were used as the benchmark to detect cerebral ischaemia and were the indication for insertion of a temporary shunt. The relationship with TCO was observed in terms of percentage decrease in oxygenation. RESULTS A total of 6 patients were shunted on the basis of their EEG changes. TCO changed more than 20% in these 6 patients, but an additional 12 patients had TCO changes with a normal EEG. This correlated with a decrease in blood pressure (BP) and was corrected by increasing the BP. The positive predictive values (PPVs) and negative predictive values (NPVs) for shunting based on TCO (as compared with EEG) were 33% and 100% respectively. Thirty-four patients had SP <50 mmHg, of whom 4 were shunted based on EEG changes. Two of 66 patients with SP >50 mmHg were shunted based on EEG changes. If a shunting policy had been based on a SP of 50 mmHg, 30 patients would have been shunted unnecessarily (PPV 12%), whereas the non-requirement for a shunt was predicted correctly in 64 of 66 patients (NPV 97%). There were 2 major strokes: 1 contralateral on day 3 in a patient with bilateral severe stenoses, and 1 ipsilateral in a nonshunted patient with normal EEG, TCO and SP >50 mmHg. CONCLUSION Compared with EEG, TCO is a practical and non-invasive monitoring system with a high sensitivity (100%) but a low specificity. TCO is more sensitive to a drop in BP and responds earlier to these changes than EEG. SP should not be used as the sole predictor for shunting during CEA.