Cerebral Oximetry in Patients Undergoing Carotid Endarterectomy Under Regional Anesthesia
Background and Purpose Near-infrared spectroscopy is a technique that can potentially monitor changes in cerebral oxygenation. There are at present limited clinical data regarding the value of this technology in relating neurological outcome to cerebrovascular hemoglobin oxygen saturation (Sco2). This investigation reports changes in Sco2 due to carotid cross-clamping during carotid endarterectomy in awake patients.
Methods Sco2 was monitored in 38 adult patients undergoing 41 carotid endarterectomies under regional anesthesia. Ipsilateral and contralateral hemispheres were monitored simultaneously during 36 operations, with ipsilateral monitoring alone in the remaining 5 operations.
Results No significant difference was detected between ipsilateral and contralateral Sco2 during preclamp or postclamp periods. Carotid cross-clamping caused a statistically significant (P<.01) decrease in the ipsilateral Sco2, which decreased from 71.8±6.91% to 65.8±8.2%, while the contralateral Sco2 remained stable at 70.5±7.5% and 70.3±7.9%. The change in ipsilateral Sco2 ranged from +2.6% to −28.6% of the preclamp value. The difference between ipsilateral and contralateral Sco2 during cross-clamping was statistically significant (P<.001). The duration of cross-clamping was 39±11 minutes (range, 18 to 89 minutes). The decrease in ipsilateral Sco2 was highly variable from patient to patient and did not correlate with the duration of cross-clamping.
Conclusions These results suggest that carotid artery occlusion causes a statistically significant but variable decrease in Sco2 in the majority of patients. Data in this investigation provide a range of Sco2 values that was not associated with a clinically detectable neurological dysfunction.
Two recent studies have documented that carotid endarterectomy (CEA) is beneficial for both symptomatic and asymptomatic patients with high-grade carotid artery stenoses.1 2 The impact of these studies has been a resurgence in the number of these operations.3 4 Unfortunately, an inherent risk of CEA is that surgery itself may result in stroke. Commonly accepted causes of stroke in the perioperative period include compromised cerebral perfusion or embolization during operation and thrombosis or occlusion of the carotid artery in the immediate postoperative period. Several monitoring modalities have been used to assess the adequacy of cerebral perfusion during CEA, including electroencephalography, somatosensory-evoked potentials, jugular venous oxyhemoglobin saturation measurements, stump pressure monitoring, and sonographic measurements of middle cerebral artery flow velocity. These monitoring techniques have one or more inherent problems: they may be invasive or cumbersome, the data may be difficult to interpret, and most important, they may be unreliable because of the high number of false-negative and false-positive results.5 6 It has therefore been suggested that neurological evaluation of the awake patient during carotid cross-clamping should be the gold standard for evaluation of all other monitors of adequacy of cerebral perfusion.7
Near-infrared spectroscopy, first described by Jobsis,8 is a technique that can potentially monitor changes in cerebral oxygenation and tissue oxygen utilization. Limited clinical experience with this monitoring technique exists, and the critical level of cerebrovascular hemoglobin oxygen saturation (Sco2), below which neuronal damage takes place, has not been established. However, Sco2 has been shown to have an excellent correlation with jugular venous oxyhemoglobin saturation and middle cerebral artery blood flow velocity in patients during CEA under general anesthesia.9 It has been suggested that this technology may be clinically useful during open heart surgery, neurosurgical procedures, and management of patients with head injury.9 10 11 Changes in Sco2 during cardiac surgery with hypothermic cardiac arrest have been shown to correlate with postoperative neurological outcome in a small group of pediatric patients.12
The present clinical investigation was designed to study the relationship of Sco2 with clinically detectable neurological dysfunction or deficits during carotid cross-clamping in awake patients. The overall objective was to determine whether Sco2 changes could identify patients who might benefit from a shunt placement during CEA under general anesthesia.
Subjects and Methods
Thirty-eight adults (26 men, 12 women) between the ages of 36 and 90 years were studied during 41 carotid endarterectomies. All patients gave informed consent to participate in the study in accordance with the guidelines approved by the Institutional Review Board of the University of Michigan Hospital. Carotid angiograms were available in 12 patients. Carotid duplex sonograms were available in all 38 patients. All patients had high-grade (>80%) internal carotid stenoses. Three patients had bilateral CEA at intervals of 1, 6, and 8 months. All patients were scheduled to undergo CEA under regional anesthesia. An ipsilateral cervical plexus block was done with 0.375% bupivacaine, and a radial artery was cannulated for continuous monitoring of blood pressure. Electrocardiogram (lead V5), peripheral hemoglobin saturation (model N100, pulse oximeter, Nellcor Inc), and Sco2 were continuously monitored during surgery with the use of a cerebral oximeter (model INVOS-3100, Somanetics Inc). The cerebral oximeter relies on principles of near-infrared spectroscopy similar to those used for peripheral pulse oximetry, with one important difference. Pulse oximeters measure a pulse-gated change in optical density and hence monitor arterial hemoglobin saturation. Cerebral oximeters measure total optical density and hence monitor hemoglobin oxygen saturation in the total tissue bed including capillaries, arterioles, and venules.
The cerebral oximeter sensors used in this investigation consisted of a near-infrared light transmitter and two detectors (placed at a distance of 30 and 40 mm from the transmitter) housed within an adhesive strip. Near-infrared light at 730 and 810 nm wavelengths (selected for maximum tissue penetration) was scattered by the tissues in two parabolic curves. The detector placed 30 mm from the transmitter receives light scattered predominantly from the scalp and skull, while that at 40 mm receives light scattered from the scalp, skull, and a large section of the brain tissue. This separation distance of two detectors is estimated to allow intracerebral penetration greater than 15 mm. The computer in the oximeter subtracts the returned signal from the superficial structures from that of the deeper tissues, thereby emphasizing the oxygen saturation of blood and brain tissue. Since 75% of the blood volume in tissue beds is in the venous circulation, Sco2 approximates tissue venous blood oxygen saturation. Oximeter sensors were placed on both sides of the forehead such that the light transmitters were at least 3 cm away from the midline. Sensors were covered with an adhesive cover to shield them from ambient light.
Two cerebral oximeters were used in this study to monitor simultaneous Sco2 from both ipsilateral and contralateral hemispheres during 36 CEA operations. In five operations ipsilateral Sco2 alone was monitored because of the unavailability of a second cerebral oximeter. The oximeter sensors were applied to the forehead, one on either side of the midline, and monitored Sco2 in the region of the frontal lobes. In 27 patients ipsilateral Sco2 monitoring was continued in the recovery room, including bilateral Sco2 monitoring in all but 4 of these patients.
The numerical Sco2 readings were recorded at 1-minute intervals and stored on floppy disks for later analysis. The entire duration of the CEA procedure was divided into (1) preclamp, (2) cross-clamp, and (3) postclamp periods, based on the time of carotid cross-clamping. Duration of carotid cross-clamping (clamp time) in each patient was noted. Mean values of Sco2 for preclamp, cross-clamp, postclamp, and recovery room periods for each patient were calculated. The relation between the duration of carotid cross-clamping and change in ipsilateral Sco2 was examined by normalizing the Sco2 data for each patient such that mean Sco2 during the preclamp period was assigned a value of 100% and mean values for later periods were calculated as a percentage of the preclamp value. Normalizing the data in this manner allowed us to study the change in Sco2 (rather than an absolute reading) and make a comparison across the patients (with varying values of Sco2) possible. The patients were followed until the time of discharge from the hospital (usually 48 hours). The development of any new neurological deficits was recorded.
Numerical data for mean ipsilateral and contralateral Sco2 during preclamp, cross-clamp, and postclamp time intervals for 36 CEAs with bilateral monitoring were subjected to a repeated-measures ANOVA. A value of P<.05 was considered significant. The Newman-Keuls test was used for post hoc testing for pairwise comparisons to study the interaction between the two hemispheres and the three time intervals. Data for duration of cross-clamping and percent change in ipsilateral Sco2 were subjected to a correlation analysis. A value of P<.05 was considered significant. A commercially available statistical package, Statistics for Macintosh, version 4.1 (Statsoft, Inc), was used for performing the statistical analyses.
CEA was successfully completed with regional anesthesia in all patients. None of the patients developed hypotension during surgery. Mean arterial pressure fluctuated less than 10% during surgery and often increased during the period of cross-clamping compared with the preclamp period. A shunt was inserted in two cases to restore antegrade cerebral blood flow. One patient had a recent history of stroke, which prompted the surgeon to use the shunt as a “prophylactic” measure. A second patient became somewhat obtunded 3 minutes after cross-clamping. She was a 76-year-old woman in whom altered neurological status was not absolutely convincing for the insertion of shunt because she did not develop a motor deficit. Unfortunately, this patient had only ipsilateral monitoring, and therefore a comparison with contralateral Sco2 could not be made. Her ipsilateral Sco2 graph (Fig 4⇓) did not appear different from that of another elderly female patient (Fig 3⇓) who remained alert during carotid occlusion. Although no motor deficit developed, the surgeon chose to insert a shunt. One patient developed confusion with difficulty in contralateral grip and a mild motor drift 16 hours after CEA and required reexploration under general anesthesia. All other patients left the hospital without a clinically detectable neurological deficit.
Changes (mean±SD) in ipsilateral and contralateral Sco2 in 41 CEAs are shown in Fig 1⇓. The effect of carotid cross-clamping on Sco2 was statistically significant (P<.01). The difference between ipsilateral and contralateral Sco2 was significant only during the cross-clamp interval (P<.001), while differences during the preclamp (P=.46) and postclamp (P=.5) periods were not significant. In 23 cases of bilateral monitoring in the recovery room, the difference between ipsilateral and contralateral Sco2 was not significant (P=.51). There was remarkable variability in the ipsilateral Sco2 change with carotid occlusion from one patient to another, as is evident in recordings from a patient with minimal changes (Fig 2⇓) and one with maximal changes (Fig 3⇓). The Sco2 in the patient in whom a shunt was used (Fig 4⇓) was higher than the Sco2 in the patient with maximal Sco2 change.
The mean duration of carotid cross-clamping was 39±11 minutes, with a range of 18 to 89 minutes. A scatterplot of duration of carotid cross-clamping and ipsilateral Sco2 calculated as a percentage of preclamp value (Fig 5⇓) did not reveal any consistent relationship (P=.49). The ipsilateral Sco2 decreased after cross-clamping in 37 of 41 CEA procedures. This decrease varied from −2.5% to −28.6% of the preclamp Sco2. In three cases Sco2 readings increased 1%, and in one case it increased 2.6%.
The purpose of this investigation was to study the effect of carotid cross-clamping on Sco2. A secondary aim was to determine whether the changes in ipsilateral Sco2 were associated with a change in the neurological status of the patient. The overall objective of evaluating this technology was to determine whether the Sco2 change could identify patients undergoing CEA under general anesthesia who might benefit from insertion of a shunt during carotid cross-clamping. This investigation showed a significant change in ipsilateral Sco2 during cross-clamping compared with preclamp and postclamp values as well as compared with contralateral Sco2, suggesting that this monitoring tracks carotid occlusion in the majority of patients (37 of 41).
An additional finding of our study is the lack of correlation between duration of carotid occlusion and decrease in Sco2. We observed an abrupt decrease in Sco2, reaching a new plateau within 1 to 2 minutes (Figs 2⇑ and 3⇑). This phenomenon is in direct contrast to Sco2 change during deep hypothermic cardiac arrest,12 13 when there is a steady continued decrease after circulatory arrest until the circulation is reestablished. This difference most probably is due to the presence of collateral circulation during carotid occlusion, which quickly restores oxygen supply to the brain. During circulatory arrest there is complete cessation of oxygen supply while oxygen consumption continues, resulting in a slow decrease in Sco2 over time.
Williams and coworkers,9 using an INVOS-3100 oximeter with an oxygen sensor, similar to the one we used, have shown that Sco2 changes correlate with jugular bulb oxyhemoglobin saturation and middle cerebral artery flow velocity measured by transcranial Doppler ultrasonography. They suggested that since cerebral oximetry is a noninvasive technique that is capable of continuously monitoring Sco2, it has an important clinical application in monitoring patients undergoing CEA under general anesthesia. Similar suggestions based on anecdotal experience with a few cases have been made by others.14 15 16 Our data suggest that it may be premature to recommend routine use of Sco2 as a monitor of adequacy of cerebral perfusion during CEA under general anesthesia. We observed Sco2 changes after carotid cross-clamping that were highly variable from patient to patient and unrelated to neurological dysfunction, such that a critical value of Sco2 change, which might justify placement of a shunt, could not be defined. A large amount of data on awake patients who develop neurological symptoms would have to be collected to determine the clinical value of Sco2 monitoring. Our findings are similar to those of Mascia et al,17 who studied Sco2 changes in eight patients during CEA under regional anesthesia. They monitored only the ipsilateral hemisphere, and therefore comparative changes in the contralateral hemisphere could not be evaluated. They too observed a mean decrease of 8.3% in ipsilateral Sco2, and none of the patients in their study required insertion of a shunt. Ausman and coworkers13 suggested an absolute Sco2 value of 35% as the critical value based on their experience of seven adult patients undergoing intracranial aneurysm clipping with deep hypothermia and circulatory arrest. Five of their patients in whom Sco2 remained above 35% had good neurological outcome, and two patients who had Sco2 below 35% showed postmortem evidence of cerebral hypoxia. In a recent study involving pediatric patients undergoing cardiac surgery with deep hypothermic circulatory arrest, Kurth and coworkers12 observed a decrease of 62±5% in Sco2 compared with prearrest values without a postoperative neurological deficit in the majority of their patients. They did not report actual values of Sco2 in their study, but assuming 70% as a normal value for Sco2, a 62% reduction translates to Sco2 readings below 35%. It should be emphasized that these studies involved patients during deep hypothermia; as such, their findings may not be applicable to anesthetized, normothermic patients undergoing CEA.
An observation that deserves comment is that during hypoventilation, such as after sedation, there is a bilateral decrease in Sco2 (Fig 3⇑) accompanied by simultaneous decrease in peripheral oxyhemoglobin saturation. However, during carotid cross-clamping only the ipsilateral Sco2 decreased. This finding emphasizes the importance of bilateral monitoring of Sco2 during CEA. This information may also be helpful in the immediate postoperative period in distinguishing sensorium changes due to other causes (sedation, systemic hypoxemia, hypotension, vasogenic cerebral edema caused by postoperative hypertension), which may be accompanied by either bilateral change or no change in Sco2, from sensorium changes due to carotid thrombosis or embolism, which should show only an ipsilateral Sco2 decrease.
Data from the current investigation suggest that the change in Sco2 after carotid cross-clamping is highly variable and that normothermic, awake patients can tolerate a wide range of decreased Sco2 for a period of 30 to 60 minutes without clinically detectable neurological dysfunction. There are two possible explanations for this phenomenon: (1) the existence of an adequate collateral circulation and (2) inaccuracy of Sco2 readings as a result of contamination by extracranial blood flow. Considerable variability of intracranial blood flow exists in the presence of carotid artery stenosis, particularly in patients with unilateral disease.18 Carotid clamping in a patient with adequate blood flow from the contralateral hemisphere will not have a significant hemodynamic effect on the ipsilateral hemisphere, reflected by no change or only negligible change in the Sco2. In patients with minimal blood flow from the contralateral circulation, changes in the Sco2 will be greater. Unfortunately, the state of collateral circulation is difficult to evaluate preoperatively. Therefore, Sco2 monitoring could become very useful if further data to establish a critical value of change in Sco2, associated with neurological symptoms, can be collected.
Another limitation of currently available Sco2 monitoring technology is that the oxygen sensor can be applied to only the hair-free area of the scalp. We chose to apply the sensors to the forehead, thus monitoring oxygen saturation in the frontal lobes. Focal cerebral ischemia in other parts of the brain may develop without a decrease in Sco2 registered by the sensors placed on the forehead. This issue may have to be addressed before the use of Sco2 monitoring during CEA under general anesthesia is advocated.
A second factor of importance in patient-to-patient variability is that of contamination by extracranial blood flow. Germon and coworkers19 raised this issue in a recent investigation using the INVOS-3100 monitor. They evaluated the contribution of the extracranial circulation by producing scalp ischemia with application of a tourniquet around the forehead and by frontalis muscle exercise by rapid and repetitive wrinkling of the forehead. They reported a decrease in Sco2 from 72% to 59% with scalp ischemia and 73% to 68% with frontalis muscle exercise and concluded that while this technology was capable of detecting cerebral hypoxia in volunteers breathing hypoxic gas mixtures, it was also significantly affected by extracranial blood flow. Certain problems may be inherent to the model used by Germon and coworkers. Venous congestion caused by scalp tourniquet inflation may increase the distance between oxygen sensor and brain surface, thereby increasing the chances of measuring oxygen saturation in extracranial tissues. Frontalis muscle exercise results in constant motion of the sensor, which causes both movement artifact and “light piping,” allowing the light to travel on the surface of the skin rather than penetrating to deeper tissues. Preliminary clinical work to establish the validity of Sco2 monitoring by INVOS-3100 included the studies of Sco2 in response to hypocapnia.16 Sco2 decreased in response to hypocapnia, and it was concluded that since hypocapnia is known to decrease cerebral blood flow and not extracranial blood flow, a decrease in Sco2 indicates that the INVOS-3100 is monitoring intracranial rather than extracranial saturation. Germon and coworkers used the oxygen sensors in which the detectors were 10 and 27 mm away from the light transmitter. This separation distance allows intracerebral penetration of 5 to 15 mm. Williams et al,9 using sensors with different detector separation distances, studied the contribution of extracranial blood flow by selective injection of indocyanine green into external and internal carotid arteries. They found that elimination of extracranial blood flow was ensured when detector separation distance was increased to 30 and 40 mm because such separation allows a deeper penetration of brain tissue. The issue of extracranial contamination may best be resolved by selective internal carotid artery and external carotid artery cross-clamping and monitoring bilateral Sco2, but to the best of our knowledge such a study has not yet been done.
Near-infrared spectroscopy to monitor Sco2 is a promising technology. However, more data need to be collected to determine its value in clinical practice. This investigation demonstrated that a significant but variable decrease in ipsilateral Sco2 in the absence of neurological dysfunction occurs in response to carotid cross-clamping. A critical value of Sco2 or Sco2 change that will necessitate the use of a shunt was not identified. Change in Sco2 is abrupt and is not related to the duration of cross-clamping. Bilateral monitoring of Sco2 can help to distinguish between carotid occlusion and cerebral hypoxemia due to systemic causes. Monitoring of Sco2 in the immediate postoperative period in patients who exhibit Sco2 decrease in response to carotid cross-clamping during surgery may assist in the early resolution of carotid occlusion.
Currently a decision to reexplore the patient after CEA is based on changes in the patient’s neurological status, often involving the sensorium, and valuable time is sometimes lost in reevaluating the carotid circulation with either angiography or sonography. Data from our study have shown that an appreciable decrease in Sco2 usually occurred in response to carotid occlusion without neurological changes. A postoperative decrease in Sco2 after CEA might demonstrate carotid occlusion before the sensorium is altered, thereby reducing the time needed to make a decision to reexplore the vessels. This may provide a valuable use of Sco2 measurements in patients undergoing CEA, particularly because it is a noninvasive monitor without complications. At present this may be an important, albeit limited, value of this monitoring technique in clinical practice.
- Received August 31, 1995.
- Revision received October 5, 1995.
- Accepted October 5, 1995.
- Copyright © 1996 by American Heart Association
National Institute of Neurological Disorders and Stroke, National Institutes of Health, Department of Health and Human Services. Clinical advisory: carotid endarterectomy for patients with asymptomatic internal carotid artery stenosis. Stroke. 1994;25:2523-2524.
Dyken ML. Controversies in stroke: past and present: the Willis lecture. Stroke. 1993;24:1251-1258.
Horsch S, De Vleeschauwer P, Ktenidis K. Intraoperative assessment of cerebral ischemia during carotid surgery. J Cardiovasc Surg.. 1990;31:599-602.
Evans WE, Hays JP, Waltke EA, Vermilion BD. Optimal cerebral monitoring during carotid endarterectomy: neurologic response under local anesthesia. J Cardiovasc Surg. 1985;2:775-777.
Jobsis FF. Noninvasive, infrared monitoring of cerebral and myocardial oxygen sufficiency and circulatory parameters. Science. 1977;198:1264-1267.
Dujovny M, Slavin KV, Cui W, Lewis G, Ausman JI. Somanetics INVOS 3100 cerebral oximeter: instrumentation, technique, and technology. Neurosurgery. 1994;34:935-936.
Williams IM, McCollum C. Cerebral oximetry in carotid endarterectomy and acute stroke. In: Greenhalgh RM, Hollier LH, eds. Surgery for Stroke. London, England: WB Saunders Co; 1993:129-138.
Hart WH, Janssen I, Furst H. Effect of carotid endarterectomy on patterns of cerebrovascular reactivity in patients with unilateral carotid artery stenosis. Stroke. 1994;25:1952-1957.
Germon TJ, Kane NM, Manara AR, Nelson RJ. Near-infrared spectroscopy in adults: effects of extracranial ischaemia and intracranial hypoxia on estimation of cerebral oxygenation. Br J Anaesth.. 1994;73:503-506.