Near-infrared spectroscopy to indicate selective shunt use during carotid endarterectomy.

OBJECTIVES This study assessed the value of cerebral near-infrared spectroscopy (NIRS) and transcranial Doppler (TCD) in relation to electroencephalography (EEG) changes for the detection of cerebral hypoperfusion necessitating shunt placement during carotid endarterectomy (CEA). METHODS This was a prospective cohort study. Patients with a sufficient TCD window undergoing CEA from February 2009 to June 2011 were included. All patients were continuously monitored with NIRS and EEG. An intraluminal shunt was placed, selectively determined by predefined EEG changes in alpha, beta, theta, or delta activity. Relative changes in regional cerebral oxygen saturation (rSO2) in the frontal lobe and mean blood flow velocity (Vmean) 30 seconds before carotid cross-clamping versus 2 minutes after carotid cross-clamping were related to shunt placement. Receiver operating characteristic curve analysis was performed to determine the optimal thresholds. Diagnostic values were reported as positive and negative predictive value (PPV and NPV). RESULTS Of a cohort of 151 patients, 17(11%) showed EEG changes requiring shunt placement. The rSO2 and Vmean decreased more in the shunt group than in the non-shunt group (mean ± standard error of the mean) 21 ± 4% versus 7 ± 5% and 76 ± 6% versus 12 ± 3%, respectively (p < .005), Receiver operating characteristic curve analysis revealed a threshold of 16% decrease in rSO2 (PPV 76% and NPV 99%) and 48% decrease in Vmean (PPV 53% and NPV 99%) as the optimal cut-off value to detect cerebral ischemia during CEA under general anesthesia. CONCLUSIONS Compared with EEG, we found moderate PPV but high NPV for NIRS and TCD to detect cerebral ischemia during CEA under general anesthesia, meaning that both techniques independently may be suitable to exclude patients for unnecessary shunt use and to direct the use of selective shunting. However, the optimal thresholds for NIRS remain to be determined.