Neurologic Monitoring and Cerebral Perfusion

Intraoperative monitoring for cerebral ischemia, hypoperfusion, and cerebral emboli during carotid endarterectomy is controversial (see Chapter 38 ). Monitoring techniques include internal carotid artery stump pressure determinations, rCBF measurements, EEG and SSEP monitoring, and transcranial Doppler (TCD) scanning. Newer modalities, such as cerebral oximetry, are being evaluated. ^[637] The rationale for the use of such monitoring is based on the need to prevent intraoperative strokes. The primary clinical utility of cerebral monitoring is to identify patients in need of carotid artery shunting; secondarily, cerebral monitoring is used to identify patients who may benefit from blood pressure augmentation or a change in surgical technique.

The internal carotid artery stump pressure represents the back-pressure resulting from collateral flow through the circle of Willis through the contralateral carotid artery and the vertebrobasilar system. The advantages of carotid stump pressure monitoring are that it is inexpensive, relatively easy to obtain, and continuously available during carotid clamping (i.e., dynamic stump pressure). Despite these advantages, few centers employ stump pressure monitoring because its accuracy in determining adequacy of collateral flow and the need for selective shunting have been questioned.^{[629] [638] [639] [640] [641]} Although the critical stump pressure is unknown, pressures below 50 mm Hg are thought to be associated with hypoperfusion.^[642]

The rCBF measurements during carotid endarterectomy are obtained by intravenous or ipsilateral carotid artery injection of radioactive xenon and analysis of decay curves obtained from detectors placed over the area of the ipsilateral cortex supplied by the middle cerebral artery. Measurements are typically obtained before, during, and immediately after carotid clamping. This technology, combined with EEG monitoring, has provided important insight into the relationship between rCBF and EEG evidence of cerebral ischemia and the critical rCBF associated with various anesthetics. ^{[593] [594]} The critical rCBF varies depending on the volatile anesthetic used. In patients receiving nitrous oxide plus a volatile anesthetic, the rCBF is approximately 20, 15, 10, and 10 mL/100 g of brain tissue per minute for halothane, enflurane, isoflurane, and sevoflurane, respectively.^{[593] [594] [595] [643]} The expense and the expertise required to make and interpret these blood flow measurements have limited the use of this technology to only a few centers.

Many centers advocate the intraoperative use of EEG monitoring for the detection of cerebral ischemia and subsequent selective shunting.^{[644] [645] [646]} The full 16-channel strip-chart EEG and the processed (compressed spectral

2104

When EEG is used for cerebral ischemia monitoring during carotid endarterectomy, a stable physiologic and anesthetic milieu is mandatory. The preferred anesthetic technique (discussed earlier) uses sufentanil, 50% nitrous oxide in oxygen, and less than 0.5 MAC volatile agent with a short-acting neuromuscular blocking agent. This combination has minor effects on the EEG and requires no significant intraoperative adjustment. Isoflurane, desflurane, and sevoflurane produce similar EEG changes at equipotent levels,^{[595] [653] [654]} and, when used at approximately 0.5 MAC, allow for reliable EEG cerebral ischemia monitoring.

The clinical usefulness of intraoperative EEG as an ischemia monitor during carotid endarterectomy is limited by several factors.^[648] First, EEG monitoring may not detect subcortical or small cortical infarcts. Second, falsenegative results (i.e., neurologic deficit with no ischemic EEG changes intraoperatively) are not uncommon. Patients with preexisting stroke or reversible neurologic deficits may have a particularly high incidence of such results. Third, the EEG is not ischemia specific and may be affected by changes in temperature, blood pressure, and anesthetic depth. Fourth, false-positive results (i.e., no perioperative neurologic deficit with significant ischemic EEG changes intraoperatively) occur because not all cerebral ischemia uniformly proceeds to infarction. Fifth, intraoperative EEG monitoring is inherently limited because most intraoperative strokes are felt to be thromboembolic and most perioperative strokes occur postoperatively. No consistent data demonstrate that EEG monitoring is clearly superior to other methods of intraoperative cerebral monitoring or that the use of EEG monitoring improves outcome.

Somatosensory evoked potential monitoring is based on the response of the sensory cortex to electrical impulses from peripheral sensory nerve stimulation. The sensory cortex, being primarily supplied by the middle cerebral artery, is at risk during carotid artery clamping. SSEP monitoring, unlike EEG, is able to detect subcortical sensory pathway ischemia. Characteristic SSEP tracings (i.e., decrease in amplitude or increase in latency, or both) occur with decreased regional cerebral blood flow and are abolished in primates when flow decreases to less than 12 mL/100 g of brain tissue per minute.^[655] No specific amplitude reduction or increase in latency has been established as a physiologic marker of impaired rCBF under operative conditions in humans. Anesthetics, hypothermia, and blood pressure may affect SSEP significantly, and false-negative results have been reported.^[656] The validity of SSEP as an intraoperative monitor of cerebral ischemia during carotid endarterectomy has not been definitively established.

TCD ultrasonography allows continuous measurement of mean blood flow velocity and detection of microembolic events in the middle cerebral artery. These parameters have important clinical implications because most perioperative neurologic deficits are felt to be thromboembolic in origin.^[657] With TCD, intraoperative embolization has been detected in more than 90% of patients undergoing carotid endarterectomy.^{[658] [659]} Most intraoperative emboli are characteristic of air and are not associated with adverse neurologic outcome. TCD may provide useful information regarding shunt function, malfunction, and incidence of emboli during shunt insertion. Embolization during carotid artery dissection may indicate plaque instability and the need for early carotid artery clamping.^[659] Embolization during dissection and wound closure are associated with operative stroke.^[660] One center reported that combined TCD monitoring and completion angiography resulted in a reduction in intraoperative stroke rate from 4% to 0%.^[661] Early postoperative embolization has been detected in more than 70% of patients after carotid endarterectomy^[659] and is exclusively particulate in nature.^[662] Most TCD-detected emboli occur in the first 2 to 3 hours after surgery.^[663] Persistent particulate embolization in the early postoperative period has been shown to predict thrombosis and development of major neurologic deficit.^[658] Frequent early postoperative TCD embolic signals have been shown to be highly predictive of early postoperative ipsilateral focal cerebral ischemia.^[659] Intervention with dextran has been shown to reduce and ultimately stop sustained embolization after carotid endarterectomy.^[663] Perioperative microembolization is more common in women and patients with symptomatic carotid disease.^{[664] [665]} TCD monitoring has been reported to detect early asymptomatic carotid artery occlusion and hyperperfusion syndrome after carotid endarterectomy.^{[666] [667]} Although TCD monitoring holds much promise, conclusive evidence demonstrating improved outcome has not been reported.