Neurologic Derangement and Regional Cerebral Oxygen Desaturation Associated with Patency of the Circle of Willis during Carotid Endarterectomy

Objectives To explore the relationship between the maximal fractional decrease of regional cerebral oxygen saturation (rSO₂) in neurologic derangement and the patency of the circle of Willis and contralateral carotid artery stenosis. Design A prospective observational study. Settings A tertiary-care university hospital Participants This study enrolled 307 patients undergoing carotid endarterectomy under regional anesthesia. Interventions No interventions. Measurements and Main Results Magnetic resonance angiography and carotid color-duplex ultrasound were performed, and the rSO₂ was recorded. The relationship between the maximal fractional decrease of rSO₂ from preclamp baseline against shunt insertion and patency of the circle of Willis was analyzed by a 2-way analysis of variance. Receiver operating characteristic analysis of the maximal fractional decrease of rSO₂ also was performed to calculate the cut-off value for detecting neurologic derangement. In addition, probability of shunt insertion was estimated by logistic regression. Patency of the circle of Willis did not influence the maximal fractional decrease of rSO₂. When both anterior and posterior circulations were nonpatent, the degree of contralateral carotid artery stenosis (Contra) was 54.7%±29.0% versus 40.7%±26.0% in patients with versus without shunting, respectively (p<0.05). The cut-off value of rSO₂ for predicting shunt insertion was 25.8%, regardless of the patency of the circle of Willis. Probability of shunt insertion for nonpatent anterior circulation = exp(-2.02+0.02×Contra)/[1+exp(-2.02+0.02×Contra)]. Conclusions The rSO₂ can be used to predict shunt insertion, regardless of the patency of the circle of Willis. The probability of shunt insertion increased with increasing degree of contralateral carotid artery stenosis in the absence of anterior circulation in the circle of Willis.