Electroencephalographic monitoring for ischemia during carotid endarterectomy: visual versus computer analysis

QEEG changes during carotid clamping in carotid endarterectomy: spectral edge frequency parameters and relative band power parameters

Intraoperative monitoring is needed to identify accurately those patients in need of a shunt during carotid endarterectomy. EEG can be used for this purpose, but there is no consensus on the variables to use. Using a database consisting of 149 EEGs recorded from patients during carotid endarterectomy under isoflurane (n=61) or propofol (n=88) anesthesia and who did or did not receive a shunt, the authors investigated which of 16 derivations (common reference, Cz) and 12 parameters (relative and absolute powers and spectral edge frequencies [SEFs]) singly or in combination could best distinguish between the shunt and the nonshunt groups for the two anesthesia regimens. Receiver operating characteristic curves were used to select derivation/parameter combinations for three types of trend computation: (1) values of relative powers and SEFs during clamping (C) only, (2) clamp minus preclamp (baseline) differences (C-B), and (3) C-B differences in absolute logarithmic power (DeltalogP). For both anesthesia regimens, C-B computation distinguished best between the shunt and nonshunt groups. For isoflurane anesthesia, SEF parameters were the best, and for propofol anesthesia the relative power parameters. Discriminant analysis, in which additional derivation/parameter combinations were added, increased the discriminative power of the DeltalogP computation but not of the C or C-B computations. For isoflurane anesthesia, SEF 90% was the best single parameter for distinguishing between patients who did and did not need a shunt and the four best derivations were F3-Cz, P4-Cz, C4-Cz, and F7-Cz. For the propofol anesthesia, the relative power (C or C-B computations) of the delta band was the best and the four best derivations were F8-Cz, T4-Cz, C4-Cz, and F4-Cz.

Electroencephalography (EEG) and somatosensory evoked potentials (SEP) to prevent cerebral ischaemia in the operating room

We review the principal aspects of EEG and SEP to detect and prevent cerebral ischaemia in the operating room during interventions at risk. EEG and SEP are variables that indirectly reflect cerebral blood flow (CBF) provided that anaesthetic regimen, body temperature, and arterial blood pressure of the patient are stable. When CBF decreases and reaches the functional threshold, slowing and/or attenuation of EEG occurs while the amplitude and the latency of cortical SEP are, respectively decreased and lengthened. Based on these changes, numerous criteria corresponding to critical thresholds have been defined. A decrease in EEG amplitude greater than 30% or EEG changes lasting more than 30 s have been considered as significant by clinicians. The main criteria resulting from computerized EEG analysis were a reduction in total power and/or in spectral edge frequency. Regarding SEP, a more than 50% decrease in N20 amplitude and/or a more than 1 ms increase in central conduction time were the most frequently used criteria. According to the bulk of literature, it may be concluded that processed EEG analysis is more sensitive than visual EEG analysis to detect cerebral ischaemia, and that SEP are not less sensitive than conventional EEG. Moreover, literature shows that SEP are as specific as computerized EEG analysis to disclose ischaemia during carotid endarterectomy.

The development of spectral EEG changes during short periods of circulatory arrest

The EEG was monitored in 56 patients during implantation of an internal cardioverter defibrillator. The purpose of this study was to determine the main EEG frequency ranges that represent ischemic changes during short periods of circulatory arrest. The EEG was recorded with a 16-channel common reference montage (Cz). After onset of circulatory arrest, the log spectral changes of three-epoch moving averages were calculated relative to the baseline spectrum. For factor analysis, 17 EEG periods were selected that showed changes progressing to an isoelectrical period. Topographic differences and the time course of quantitative EEG (qEEG) changes were studied in all 56 patients. For each patient the EEG period with the longest duration of circulatory arrest was chosen. Factor analysis revealed four factors that represented the spectral EEG changes occurring during circulatory arrest and recovery. The frequency intervals of these factors were 0 to 0.5 Hz, 1.5 to 3 Hz, 7.5 to 9.5 Hz, and 15 to 20 Hz for all channels. Only minor topographic differences were found in the power of the spectral changes; the sequence of events was similar for all electrode positions. The first EEG change after circulatory arrest was an initial increase in alpha power and a decrease in beta power. On average, after approximately 15 seconds alpha power started to decrease, beta power decreased further, delta-1 power started to increase, and delta-2 power started to decrease. After approximately 25 seconds, the delta-1 power increase appeared to plateau or to decrease. A circulatory arrest longer than approximately 30 seconds resulted in an isoelectrical EEG. After restoration of the circulation, there was a fast transient increase in delta-1 and delta-2 power, followed by a decrease to baseline. alpha and beta power showed a more gradual increase in power toward baseline and were the last to restore after 60 to 90 seconds. In general, the spectral changes in the alpha and beta frequency ranges were most pronounced and consistent. In conclusion, to detect intraoperative cerebral ischemia, monitoring of changes in the four frequency ranges found is preferable to monitoring changes in the classically defined frequency bands. Furthermore, these results stress the importance of the alpha and beta ranges in detecting cerebral ischemia.

Quantitative EEG during early recovery from hypoxic-ischemic injury in immature piglets: burst occurrence and duration

This study examined the course of EEG recovery in an animal model of hypoxic-ischemic injury. The model used periods of hypoxia, room air and asphyxia to induce cardiac arrest. One-week-old piglets (n = 16) were exposed to a period of hypoxia, room air and complete asphyxia for 7 minutes. After cardiac arrest and resuscitation, two EEG features were evaluated as prognostic indicators of behavioral outcome as assessed by a neuroscore at 24 hours after insult. A prominent EEG feature was the number and duration of bursts evident during recovery. Episodes of bursting were detected through the thresholds on sustained periods of elevated power. After the animal was resuscitated, the EEG was monitored continuously for 4 hours. To assess outcome in the recovering animal, a behavioral testing scale was used to test the animal's neurological capabilities. Trends of EEG burst counts were measured through thresholds on sustained power changes. Bursts are energy transients in the EEG record. High degrees of bursting were characteristic of animals having good neurological condition whereas piglets having low burst counts had poor 24 hr neuroscores. At 100 min the average burst rate of the good neuroscore outcome group was more than 8 per min and was significantly different from the poor outcome group's level of 2.7 (p < or = 0.05). When these counts were weighted by their total duration, differences between groups increased (p < or = 0.02). This study showed that the QEEG measure of burst counts and duration together provided a strong prognostic indication of the 24 hour outcome after asphyxic injury in a neonatal animal model. The critical determinant of the bursting character was the time when bursting occurred. Bursting occurring early in recovery was a good gauge of outcome. We conclude that quantitative EEG analysis and interpretation can be an important tool for the outcome determination during recovery from cerebral injury states.

Carotid endarterectomy monitoring: patterns of spectral EEG changes due to carotid artery clamping

OBJECTIVE

The purpose of the study was to make an objective and quantitative analysis of the EEG changes caused by carotid artery clamping during carotid endarterectomy (CEA) monitoring.

METHODS

Factor analysis was used to study the intraoperative spectral EEG changes in 94 patients during clamping of the carotid artery. In addition, the relation between the extracted factors and the changes in blood pressure and blood flow velocity in the middle cerebral artery during clamping was studied.

RESULTS

Two factors were extracted with factor analysis. The first factor represented a change in power in the alpha and beta frequency ranges in combination with a less pronounced opposite change in power in the delta frequency range. The second factor represented a change in power restricted to the delta and theta frequencies. With the first factor, two types of spectral EEG changes could be distinguished: changes indicative of cerebral ischemia (decrease in fast activity and increase in slow activity) and the opposite changes suggesting cerebral activation (arousal). With the two factors combined, the changes indicative of minor ischemia (decrease in fast activity only) could also be distinguished.

CONCLUSION

Further study is required to test whether patients showing the EEG changes indicative of activation or minor ischemia actually require shunting.

Continuous electroencephalographic monitoring and selective shunting reduces neurologic morbidity rates in carotid endarterectomy

PURPOSE

The role of continuous electroencephalographic (EEG) monitoring during carotid endarterectomy was evaluated in this retrospective review.

METHODS

We analyzed data from 902 consecutive carotid endarterectomy procedures performed with vein patch angioplasty. In 591 operations from 1980 to 1988 we did not use intraoperative EEG monitoring or shunting (non-EEG group). Continuous intraoperative EEG monitoring and selective shunting were used in 311 procedures from 1988 to 1994 (EEG group). The patients' mean age was higher in the EEG group (68.8 years; range, 41 to 87 years) than in the non-EEG group (66.2 years; range, 34 to 90 years; p < 0.001). There was also a significantly higher incidence of hypertension (56.2% vs 41.9%) and redo operations (5.4% vs 2.54%) in the EEG group than in the non-EEG group (p < 0.05). The operative technique was identical in both groups. We defined a significant EEG change as a greater than 50% reduction of the amplitude of the faster frequencies, a persistent increase of delta activity, or both.

RESULTS

In the EEG group, acute EEG changes occurred in 40 patients (12.8%); 31 (77.5%) unilateral and ipsilateral to the operated carotid artery, and nine (22.5%) bilateral. In five patients (12.5%) the changes correlated with an intraoperative episode of hypotension, and after normal blood pressure was restored the EEG returned to normal. In 35 procedures (87.5%) a carotid shunt was inserted. In 33 of those patients the EEG returned to baseline, in one patient there was a significant improvement, and in one patient the EEG changes persisted. Postoperative hospital strokes occurred in one patient (0.32%) in the EEG group and in 13 patients (2.19%) in the non-EEG group (p < 0.05). All strokes (n = 14) were ipsilateral to the operated carotid artery. Of the 13 strokes in the non-EEG group nine were major and four were minor. The one stroke in the EEG group was embolic in origin and occurred before carotid cross-clamping; it was associated with profound EEG changes that did not reverse after placement of a shunt. In the total group (n = 902), intraoperative EEG monitoring was inversely associated with postoperative stroke (p < 0.05).

CONCLUSION

The overall neurologic morbidity rate was significantly lower in the EEG group than in the non-EEG group, therapy demonstrating the value of intraoperative EEG monitoring in carotid endarterectomy.

The reliability of quantitative electroencephalography as an indicator of cerebral ischemia

The electroencephalogram (EEG) has been used to detect episodes of cerebral ischemia during various surgical procedures. Recently, computerized systems for recording and interpreting the quantitative EEG (QEEG) have been used by anesthesiologists because of their ease of application, clarity of display, and reported ability to identify ischemic EEG changes. However, the extent to which automated techniques of QEEG interpretation reliably differentiate cerebral ischemia from the confounding effects of anesthetics and other sources of "artifact" is not completely established. In this study, EEGs were recorded before and after defibrillator testing in patients undergoing implantable cardioverter defibrillator (ICD) placement and during analogous time periods in control patients undergoing abdominal surgery. EEGs were subjected to standard visual inspection by an experienced electroencephalographer and QEEG analysis with a commercially available system was used for automated EEG interpretation in order to evaluate the reliability of this quantitative technique. The CIMON technique identified episodes which met previously defined criteria for QEEG cerebral dysfunction and ischemic pattern in both groups, despite the presumed absence of cerebral ischemia in the control patients. Since there was no evidence of cerebral ischemia in the raw EEGs of either the ICD patients or the controls, these QEEG changes were not confirmed by conventional techniques of EEG interpretation. Our results suggest that caution is warranted when using automated systems for intraoperative interpretation of EEG.

Implantable cardioverter-defibrillator. Evaluation of clinical neurologic outcome and electroencephalographic changes during implantation

During placement of implantable cardioverter-defibrillators, ventricular arrhythmias are induced to test the function of the devices. Although cerebral hypoperfusion and ischemic electroencephalographic changes occur in patients while implantable cardioverter-defibrillators are being tested, no investigation has assessed neurologic outcome in these patients. Nine patients having either implantation or change of an implantable cardioverter-defibrillator underwent neurologic examination and neuropsychometric tests before and after the operation. After induction of general anesthesia and insertion of implantable cardioverter-defibrillator leads (when needed), ventricular fibrillation, ventricular flutter, or ventricular tachycardia, was induced by means of programmed electrical stimulation. Implantable cardioverter-defibrillator testing continued until satisfactory lead placement was confirmed. The intraoperative electroencephalographic recording was analyzed for evidence of ischemic change. In all, an electroencephalogram was recorded during 50 periods of circulatory arrest. Mean duration of the arrest periods was 13.6 seconds. By means of conventional visual inspection of the raw electroencephalogram, high-amplitude rhythmic delta or theta, voltage attenuation, or loss of fast frequency activity was observed in 30 of the arrests. By means of an automated technique of electroencephalographic interpretation based on power spectral analysis, electroencephalographic changes were correctly identified in 26 of the arrests. The incidence of these electroencephalographic changes was dependent on the arrest duration. The mean interval from arrest onset to electroencephalographic change was 7.5 seconds (standard deviation +/- 1.8 seconds). In patients with electroencephalographic changes during multiple arrests, no downward trend in this interval was detected in later arrests and no evidence of persistent ischemic change was observed in electroencephalograms recorded after the conclusion of implantable cardioverter-defibrillator testing. Postoperative neurologic and neuropsychometric testing was completed in eight patients, none of whom exhibited a new neurologic deficit, exacerbation of a preexisting neurologic condition, or significant deterioration in neuropsychometric performance. We conclude that the brief arrest of cerebral circulation induced during insertion of an implantable cardioverter-defibrillator is not associated with permanent neurologic injury.

Computer-derived density spectral array in detection of mild analog electroencephalographic ischemic pattern changes during carotid endarterectomy

The purpose of this prospective study was twofold: 1) to determine the sensitivity and specificity of computer-derived density spectral array in detecting analog electroencephalographic (EEG) ischemic pattern changes during carotid artery cross-clamping in patients undergoing carotid endarterectomy; and 2) to assess the ability of density spectral array to identify such changes in comparison with the degree and type of change seen in the analog EEG ischemic pattern. Sixteen channels of anteroposterior bipolar and two to four channels of referential electroencephalography with four channels of density spectral array were monitored simultaneously during carotid endarterectomy in 103 patients under general anesthesia. Two "observers" interpreted the density spectral array and the analog electroencephalograms, one during and immediately after the operations and the other 6 months after completion of all surgery. Analyses were conducted to establish both the number of patients with analog EEG ischemic changes and the number of ischemia events during carotid artery cross-clamping. Observer A indicated that the density spectral array identified analog EEG ischemic changes in 21 of 29 patients, for a sensitivity of 72% (specificity 99%), whereas Observer B's results showed that the density spectral array identified analog EEG ischemic changes in 16 of 27 patients, for a sensitivity of 59% (specificity 96%). Density spectral array detection of analog EEG ischemic changes based on severity classifications were 61% and 18% in the mild group, 70% and 71% in the moderate group, and 95% in the severe group, indicating a relationship between density spectral array sensitivity and severity of analog EEG ischemic change, with p = 0.02 and p = 0.004 for the two observers. The kappa statistics for observer reproducibility were highly significant, with k = 0.95 for analog EEG ischemic changes and 0.85 for density spectral array changes. It is concluded that density spectral array does not reliably detect mild analog EEG pattern changes of cerebral ischemia and is not a reliable substitute for 16-channel analog EEG monitoring of cerebral ischemia during carotid endarterectomy.