Peripheral Vascular Surgery Society

#8 DETECTION OF INTRAOPERATIVE CEREBRAL ISCHEMIA DURING CAROTID ENDARTERECTOMY (CEA).

Roy I. Davidovitch, BA, Jean-Paul Spire, MD, Lewis B. Schwartz, MD, Hisham S. Bassiouny, MD, Diane Suarez, REEGT, Sol Aronson, MD, Bruce L. Gewertz, MD, and James F. McKinsey, MD

The University of Chicago, Chicago, IL

Processed electroencephalography (EEG) is widely used to detect cerebral hypoperfusion (CH) during CEA; however, EEG requires trained technical support and may be influenced by anesthetic agents and patient temperature. To evaluate more economical and less subjective detectors of CH during CEA, we compared EEG to transcutaneous venous cerebral oximetry (TVCO) using a near-infrared spectroscopy sensor placed on the ipsilateral scalp and jugular venous oximetry (JVO2) via a continuous oximetric catheter at the jugular bulb.

We prospectively monitored 54 patients undergoing 57 CEA's with EEG, JVO2 and TVCO. EEG changes during carotid occlusion were stratified into three levels of severity (Grade 0,1,2) representing no interval change, mild change (sustained EEG change of <50% power reduction) or severe change (>50% power reduction), respectively. A TVCO reduction of >25% from baseline was considered indicative of CH. Intraoperative interventions including shunting and blood pressure manipulations were based solely on EEG changes.

In 57 CEA's, EEG detected 30 CH events (Grades 1 and 2). A TVCO change >25% accompanied 27 of these 30 events (Sensitivity 90%, Negative Predictive Value (NPV) 96%). Of 17 episodes of Grade 2 EEG events, a TVCO change of >25% was noted in 16 (Sensitivity 94%, NPV 98%). Additionally, maximum TVCO change from baseline was significantly greater during EEG Grades 1 and 2 events than during intervals with no EEG change, with respective mean differences equaling 11.26 (se=1.93, p=0.0001) and 6.92 (se=2.05, p=0.0014), adjusted for repeated measures on patients. TVCO change falsely reported 14 CH events (Specificity 82%). In contrast, JVO2 decrease from baseline was highly variable and did not differentiate between those patients experiencing any EEG changes (mean ± SD; 12.2 ± 11) and those without EEG changes (11.8 ± 14).

EEGD vs.	TVCOD	EEG changes
		Grade 1	Grade 2
TVCOD	>25%	11	16
	>25%	2	1

Our data shows that TVCO monitoring can reliably detect CH and can serve as an alternative to EEG or a monitor of shunt patency when EEG is not available. The ease of use and low cost of the TVCO offsets the potential for a slightly higher use of intraoperative shunts.

Return to Program

Last Updated 5/28/99