Hemodynamic ischemic stroke during carotid endarterectomy: An appraisal of risk and cerebral protection J. E. Frawley, FRACS, R. G. Hicks , MSc, M. Beaudoin , F A N Z C A , and IL Woodey , F A N Z C A , Sydney, Australia
Purpose: The purpose of this study was to validate the commonly accepted indicators of risk of ischemic stroke that indicate the necessity for cerebral protection during carotid endarterectomy (CEA), and to examine the efficacy of high-dose thiopentone sodium (thiopental) as a cerebral protection method in patients who are at high risk of intraop- erative ischemic stroke. Method: In a prospective study of 37 CEAs performed for symptomatic stenosis >70%, functional and clinical indicators of risk of ischemic stroke during carotid cross-clamping were identified. Functional indicators of risk were the development of ischemic electro- encephalogram (EEG) changes and stump pressure <25 mm Hg. Clinical indicators of risk were previous ischemic hemispheric stroke and severe bilateral disease. These indica- tors were correlated in all patients, some of whom had two or three coexisting indicators of risk. The EEG and stump pressure were monitored continuously during carotid occlusion in all operations. Carotid occlusion times were recorded. Intraluminal shunting was eliminated in favor of high-dose thiopental cerebral protection in all patients. Neurologic outcome was deemed to measure the efficacy of thiopental protection in patients who are identified to be at risk and, hence, in need of cerebral protection. The validity of the indicators used to identify risk of ischemic stroke during CEA was assessed. Results: The absolute stroke risk was found to be 29.7% for the whole group (37 patients) and 57.9% in 19 patients who had commonly accepted indications for protective shunt- ing. The correlation ofischemic EEG changes with stump pressure <25 mm Hg was only 27.3%, whereas the expected correlation based on well-documented reports in the literature was 100%. The lack of correlation may have been related to the prevention of ischemic EEG changes by thiopental. There were no neurologic deficits in the series. Conclusions: The absence of neurologic deficit in the study indicated that thiopental protection was effective in preventing ischemic stroke in high-risk patients and safely replaced intralmninal shunting. (J Vase Surg 1997;25:611-9.)
It is generally accepted that carotid artery occlu- sion during carotid endarterectomy (CEA) may re- sult in focal or regional hypoperfusion that in some patients may precipitate an ischemic stroke. Since the inception of CEA 40 years ago, two major principles have been applied to provide protection against isch- emic cerebral damage as a result o f hypoperfusion, namely, prevention ofhypoperfusion by intraluminal shunting and protection of cerebral cells against the
From the Departments of Vascular and Transplantation Surgery, Neurophysiology, and Anaesthesia and Intensive Care, The Prince Henry and Prince of Wales Hospitals, University of New South Wales.
Reprint requests: John E. Frawley, FRACS, 23 Riddles Ln., Pymble, N.S.W. 2073, Australia.
effects of ischemia. Intraluminal shunting has been the most enduring and widely practiced and has been further refined by techniques designed to identify those patients who are at risk of ischemic stroke during carotid occlusion and to shunt selectively only in those patients. In our practice, shunting has been abandoned in favor o f routine pharmacologic cere- bral protect ion using high-dose th iopentone so- dium (thiopental) . l"2 We recently repor ted our ex- perience o f 259 CEAs for symptomat ic stenosis >70% associated with contralateral critical stenosis (>70%) or contralateral occlusion; a major stroke rate o f 0.4% was the produc t o f one postoperat ive cerebral hemorrhage. 2 In that report , patients who had a defined specific risk o f ischemic stroke were not identified, and a l though there were no intra- operative ischemic strokes in the series, we were unable to state unequivocally that thiopental cere-
611
JOURNAL OF VASCULAR SURGERY 612 Frawley et al. April 1997
bral protection safely replaced intraluminal shunt- ing in patients who were at risk of intraoperative ischemic stroke. To clarify the validity of thio- pental cerebral protection, a prospective study was undertaken between November 1994 and Sep- tember 1995 in 36 consecutive symptomatic pa- tients who underwent CEA as a result of stenosis >70% in whom the risk of ischemic stroke during carotid occlusion was identified. The study was based on the premise that the identification of ischemic stroke risk during carotid occlusion de- mands some form of cerebral protection against the effects of hypoperfusion ischemia to prevent ischemic stroke.
PATIENTS AND METHODS Definitions
The term "ischemic stroke" used in the text im- plies hemodynamic ischemic stroke that is associated with carotid cross-clamping and excludes ischemic stroke that results from intraoperative cerebral em- bolus or operative site occlusion.
For purposes of clarity, stump pressure up to and including 25 mm Hg is referred to as "stump pres- sure <25 mm Hg."
Patients Thirty-six symptomatic patients with docu-
mented histories of hemispheric transient ischemic attacks underwent 37 CEAs. Two patients with cre- scendo transient ischemic attacks arrived in the oper- ating room for surgery with neurologic deficit, one with dysphasia and the other with hemiparesis. One patient underwent bilateral CEA with a period of 8 weeks between operations. There were 22 men and 14 women in the study, with ages ranging from 51 to 83 years (mean, 68.4 years).
Identification of ischemic stroke risk
Patients were identified as being at risk of isch- emic stroke according to both the functional and clinical criteria widely reported in the literature. Functional indicators of risk were the development of asymmetrical EEG changes appearing with carotid cross-clamping, and mean stump pressure <25 mm Hg during the carotid occlusion period. Clinical in- dicators of risk were a history of previous completed ischemic hemispheric stroke, and the presence of severe bilateral carotid disease, defined as stenosis >70% (the operative lesion) associated with con- tralateral stenosis >70%, or contralateral occlusion. In addition, carotid occlusion exceeding 25 minutes was deemed to constitute increased risk. "Potential
stroke risk" was defined by the identification of both clinical and functional indications for intraluminal shunting, either alone or in combination. "Absolute stroke risk" was defined by the identification ofisch- emic electroencephalographic changes, stump pres- sure <25 mm Hg persisting during carotid occlu- sion, or both. Patients who were identified to be at risk of cross-clamping ischemic stroke, and who thus required cerebral protection, were protected during surgery using high-dose thiopental rather than in- traluminal shunting.
Anesthesia and cerebral protection All operations were performed with the patient
under general anesthesia using thiopental induction and an oxygen-nitrous oxide mixture with isoflurane for maintenance. High-dose thiopental was used for cerebral protection in all patients. Thiopental was titrated by hand injection in doses sufficient to pro- duce burst-suppression activity on continuous elec- troencephalogram (EEG) monitoring, with mini- mum interburst intervals of 60 seconds. The total dose of thiopental required to produce this burst- suppression pattern and maintain it during the ca- rotid occlusion period by incremental bolus doses varied from 1350 mg to 4000 mg (average, 1800 mg). Ischemic EEG changes that appeared with carotid occlusion were managed by increasing the interburst interval to as long as 5 minutes by ad- ministration of further bolus doses of thiopental. No further thiopental was administered after com- pletion of the endarterectomy and clamp release. During the period of thiopentaI administration, nitrous oxide and isoflurane were discontinued, and anesthesia relied on oxygen and thiopental. Intraluminal shunts were not used in any of the operations.
Specific intraoperative monitoring
In all cases the raw EEG was monitored continu- ously, using subcutaneous platinum electrodes in a 10-20 montage, on a Medelec/Teca DG32 electro- encephalograph (Medelec Ltd., Surrey, U.K.). EEG abnormalities, or changes appearing after carotid cross-clamping, were interpreted during monitoring in the operating room by a specialist neurophysiolo- gist (R. G. H.). Unilateral change in burst activity or diminution in burst amplitude greater than 30% were interpreted as ischemic in origin. Systemic arterial pressure was monitored continuously via a radial artery catheter. Internal carotid artery back pressure was monitored continuously during carotid occlu- sion (dynamic stump pressure).
J O U R N A L O F V A S C U L A R S U R G E R Y
V o l u m e 2 5 , N u m b e r 4 Frawley et al. 6 1 3
i ! + l ~ + ; . ~ / i i : . i : i i 5 ~ ~ : : : i!P,P,P,i :li ~ ~ i + + + +"+ " t=-SE++ : : : ; + . :~ ++
!?? :? : ? ? : : : : :
=~+. +:+, ..... : + . . . . i ~- : iJ -+_-- : i_=.-.-
W.=+'.'i ~++~++ ::++i ;+ ::+i++~
Fig. 1. Dynamic stump pressure. Simultaneous continuous recording of mean systemic arterial pressure (upper recording) an d mean stump pressure (lower recording). Stump pressure mirrors arterial systemic pressure precisely and varies only when systemic pressure varies. With adminis- tration of thiopental bolus doses (arrows) both pressures fall slightly (5 to 10 mm Hg) and recover rapidly (<1 minute), sometimes requiring 50 to 100 Izg bolus doses of metaraminol administered via a computerized syringe infusion pump (58% of cases). Scale: 1 vertical division = 5 mm Hg; 1 horizontal division = 24 seconds. MAP, Mean systemic arterial pressure; MSP, mean stump pressure; B-V, back-venting of internal carotid artery,; C-R, clamp release.
Dynamic s t u m p pressure
Stump pressure was measured using a 4F wedge pressure catheter (Arrow International Inc., Read- ing, Pa.) connected via a silicon elastic line to a pressure transducer. The system was fluid-filled, with exclusion o f air bubbles, and connected to a twin- channel recorder. The transducer was calibrated for zero at the level o f the internal carotid artery. After control o f the carotid arteries, the wedge catheter was passed via the carotid arteriotomy into the inter- nal carotid artery for a distance of 2 cm, and the balloon was inflated via a tap to provide occlusion. Balloon inflation was maintained by closing the tap. A continuous recording of mean stump pressure was obtained in parallel with mean systemic arterial pres- sure recorded on the second channel (Fig. 1). Spot readings were taken at 2-minute intervals and were averaged to provide a figure for mean stump pressure across the carotid occlusion period. Mean systemic arterial pressure was maintained at the pre-thiopental level during carotid occlusion using 50 to 100 vtg doses o f metaraminol if necessary (58% of cases).
Pos topera t ive care
All patients were transferred to the intensive care unit immediately after surgery and were released to the open ward on the morning after surgery (mean,
22 hours). Recovery from thiopental anesthesia to the equivalent o f a Glasgow Coma Scale assessment of 15 occurred in 30 minutes to 3 hours (mean, 1.9 hours). Endotracheal intubation varied from 1 to 10 hours (mean, 2.9 hours). Hospital discharge was on the third or fourth day, dictated largely by the timing of formal Depar tment ward rounds. The details o f postoperative care and monitoring have been report- ed. 1,2
Assessment
The validity of thiopental cerebral protection was assessed by the occurrence of transient or permanent neurologic deficit. The study also allowed an assess- ment of the validity of the indicators used to identify the risk ofischemic stroke.
R E S U L T S Funct iona l indicators o f ischemic s t roke risk
In the whole group (37 patients) there were seven EEG asymmetries and seven stump pressures <25 m m H g identified in 11 patients. EEG asymme- try and stump pressure <25 m m H g coexisted in three patients, and four EEG asymmetries and four stump pressures <25 m m H g occurred alone. O f 23 patients who had no clinical indicators of risk, four EEG asymmctries and three stump pressures <25
JOURNAL OF VASCULAR SURGERY 614 Frawley el: al. April 1997
Table I. Potential ischemic stroke risk--37 patients
Stump pressures o f 26 to 50 m m H g were not considered indicators o f risk. Stump pressure not recorded in the table were all > 5 0 m m Hg. Average carotid occlusion time was 41 minutes (range, 26 to 57 minutes). Potential stroke risk, 51% (19 o f 37).
Table II . Stroke risk associated with disease
Disease Patients Risk indicators Absolute risk
Bilateral 12 5 41.7% Unilateral 25 6 24%
mm Hg were identified in five patients. EEG asym- metry corresponded with stump pressure <25 mm Hg in two (Table I).
Clinical indicators o f ischemic stroke risk
Clinical indicators of risk were identified in 14 patients. Ten of these had severe bilateral disease, two had both bilateral disease and previous ischemic stroke, and two had previous ischemic stroke. Seven functional indicators of risk, three EEG asymmetries and four stump pressures <25 mm Hg, coexisted with clinical indicators in six patients. The concur- rence of functional with clinical indicators of risk was 43% (6 of i4; Table I).
Comparison o f bilateral and unilateral disease
Functional indicators of risk were identified in five of 12 patients with severe bilateral disease (41.7%) a n d in six of 25 patients with unilateral disease (24%; Table II).
Ischemic stroke risk
A potential ischemic stroke risk, as defined fbr this study, indicated the necessity for cerebral protec- tion against ischcmia in 19 of the 37 patients (51%). In these 19 patients carotid occlusion time averaged 41 minutes (range, 26 to 57 minutes), and 11 of the i9 had multiple (two or three) indicators of risk (Table I). The absolute ischemic stroke risk, as de- fined for this study, in the 19 patients who had commonly accepted indications for protective shunt- ing was 57.9% (11 of 19). In these 11 patients ca- rotid occlusion time averaged 37 minutes (range, 26 to 57 minutes). For the whole group the absolute ischemic stroke risk was 29.7% (11 of 37; Table I).
The potential risk was almost certainly higher be- cause of the long carotid occlusion times and the possibility of the failure to detect some ischemic EEG changes prevented by thiopental. I f no effective cere- bral protection was used, an intraoperative ischemic stroke rate of at least 29.7% could reasonably have been expected in the whole group, with stroke pre- dominating in those who had accepted indications for protective shunting, in whom a stroke rate of 57,9% could have been anticipated. All patients in the study were symptomatic with >70% stenosis, and it is now well recognized that symptomatic patients are at greater risk of stroke than are asymptomatic patients. 3,4 Although this study has not specifically assessed other risk factors such as systemic hyperten- sion, cardiac disease, and cerebrovascular disease, such factors did in fact coexist in a number of the patients. Clearly, the patients in this study represent an extremely high-risk group for perioperative stroke.
Neurologic outcomes
There were no neurologic deficits, transient or permanent, in the series. Both patients who were admitted for operation with neurologic deficit made full recoveries within 24 hours after surgery.
D I S C U S S I O N
Hemodynamic ischemic stroke results from a critical reduction of cerebral blood flow. Much of the practice of CEA revolves around the prevention of ischemia associated with hypopeffusion by maintain- ing blood flow with an intraluminal shunt. Tech- niques that are designed to detect reduced cerebral blood flow during surgery, and thus indicate a poten- tial risk of ischemic stroke, have allowed selective shunting in those patients identified to be at risk. The development during carotid occlusion for CEA of ischemic EEG changes, neurologic symptoms or signs in the awake patient, and stump pressures <25 mm Hg, have been investigated and correlated with
JOURNAL OF VASCULAR SURGERY Volume 25, Number 4 Frawleyetal. 615
cerebral blood flow and are the commonly used indi- cators of risk and thus for intraluminal shunting. In some centers, selective shunting is based on the clin- ical criteria of severe bilateral disease and previous completed hemispheric stroke without supporting evidence of decreased cerebral blood flow. The alter- native to shunting is to provide protection against the effects of reduced blood flow at the neuronal cellular level or to enhance blood flow in an ischemic area by pharmacologic or physiologic means.
Risk o f ischemic stroke
The risk of ischemic stroke during CEA has not bccn clearly defined, although ischemia during ca- rotid occlusion has bccn widely investigated. Studies that correlate ischemic changes detected by monitor- ing (ischemic EEG changes, neurologic symptoms or signs in the awake patient, and low stump pressures) with cerebral blood flow indicate a potential risk for ischemic stroke during carotid occlusion across the spectrum ofcxtracranial vascular disease of as high as 48% when patients with contralateral occlusion are singled o u t . 5,6 This does not mean that ischemic stroke will inevitably occur unless a shunt or some other tbrm of cerebral protection is used, because such ischemia may be detectable by monitoring within a range of cerebral blood flow reduction from 24 ml /100 g m / m i n down to 10 ml /100 gin/rain and not cause permanent ischemic ncuropathologic damage. 7,8 The risk of stroke, theretbrc, does not necessarily equate with the incidence of ischemic changes detected by monitoring. However, Kass et al. 9 have demonstrated that the development of neu- ronal cellular changes associated with cell death (so- dium and calcium influx and potassium etBux) is related to the duration of ischemia. Prolonged cross- clamping, as in the patients in this study, may also, therefore, be an important determinant of risk and, within the range of blood flow reduction producing ischemic EEG changes (24 to 10mls /100 gm/min) , the risk of permanent ncurologic damage is likely to increase with prolonged clamp times. Ferguson, m equating risk of ischemic stroke with intraoperative stroke, has established a risk of 1.4% in nonshuntcd and otherwise unprotected patients, and identified patients with severe EEG changes and stump pres- sures <25 mm Hg as being at high risk o fhcmody- namic ischemic stroke. Risk determined across the full spectrum of patients subjected to CEA does not equate with risk in an individual patient. In some patients there may be no risk ofischemic stroke, and, in others, such as Ferguson's l° group with severe EEG changes and stump pressure <25 mm Hg, the
risk may approximate 100%. It is clear that there is little or no validity in equating the risk of ischemic stroke with the intraoperative stroke rates in groups of patients who arc submitted to operation with a wide diversity in symptoms and severiw of disease. The risk o f hemodvnamic ischemic stroke during CEA exists only when carotid cross-clamping pro- duccs a critical reduction of cerebral blood flow, can be determined only tbr each individual patient, and can only be defined by and equated with a demon- strated reduction of cerebral blood flow. Functional indicators of reduced cerebral blood flow are, there- tbre, the only valid indicators of risk of ischcmic stroke.
St roke risk associated wi th cerebral p ro tec t ion
The prevention of ischemic stroke is essential for successful CEA and provides the rationale for intralu- minal shunting or some other form of intraoperative cerebral protection. It is axiomatic that any cerebral protection method must prevent ischemic stroke and carry no inherent risk of stroke. Analysis of routinely shunted patients provides no direct indication o f the stroke-prevention capability of shunting because pa- tients who are at definite risk of ischemic stroke during carotid cross-clamping are not identified. However, excellent results have been reported in routinely shunted patients, with stroke rates between 1.1% and 1.4%, ~ 13 well within the acceptable prac- tice guidelines recently advised by Moore et al. 3 Such results from centers o f recognized excellence with experience of large numbers of operations are well below the stroke rates achieved in the majority of centers that perform CEA ~4a5 and provide indirect evidence that shunting must prevent some ischemic strokes. Riles c ta l . 16 have demonstrated improving stroke rates with increasing experience of shunting, which suggests that technical familiarity with shunt- ing can result in reduction of stroke rate. This may indicate that shunting has an inherent stroke risk, albeit small, related to limited technical experience.
Selective shunting is specifically based on the identification of patients who are believed to be at risk and provides a model tbr the study o f the efficacy of intraluminal shunting in the prevention of isch- cmic stroke. If shunting removes the risk of operative ischcmic stroke and has no inherent risk, then stroke rates and pathologic stroke profiles in any one cohort of selectively shunted patients should be similar in both shunted and nonshunted patients. Stroke rates in the nonshunted patients in the selectively shunted series for same surgeons and same institutions re- corded in Table II117-22 fall well within the accepted
JOURNAL OF VASCULAR SURGERY 6 1 6 Frawley et al. April 1997
Table I lL Stroke rates in selectively shunted series
Factorial in crease
Patients No with Author (n) Ch,erall shunt Shunt shunt
Davies et al. 17 389 ~ 2.5% 1.1% 6.5% 5.9 Archie is 665 0.8% 0.2% 5.2% 26 .00 F r i edman et al. 19 145 6.3% 3.7% 14% 3.78 Mackey et al. z° 598 2.8% 2.3% 5% 2 .17 Perler et al. 21 205 3.9% 2.1% 5.4% 2 .57 Graham et al. 22 212 3.3% 2.0% 6.3% 3.15
practice guidelines 3 and indicate that the surgeons involved were technically competent and capable of above-average results as recently defined by Rothwell et al. 4 (stroke rates of 2.51% for asymptomatic dis- ease and 4.28% for symptomatic disease). Across these series the shunted patients, compared with the nonshunted, sustained stroke rates increased by mul- tiples of 5.90, 26, 3.78, 2.17, 2.57, and 3.15, respec- tively, despite a presumably constant level of surgical experience and technical ability. In a multicenter transcranial Doppler study reported by Halsey 23 that assessed stroke associated with shunting in patients without evidence of cross-clamping ischemia, shunted patients sustained a 6.3 multiple increase in stroke rate over nonshunted patients (4.4% vs 0.7%). The argument that a higher stroke rate is to be expected in selectively shunted patients because they are identified to be at risk ofischemic stroke is invalid and unsustainable if shunting removes the threat of ischemia. It is difficult to attribute the higher stroke rate with shunting in the series in Table III to any- thing other than the shunt, either because of techni- cal failure of shunt flow, damage to the vessels, em- bolization through the shunt, or, most likely, imposition of the shunt in the operative field pre- venting adequate endarterectomy. It would seem that shunting does carry an inherent stroke risk de- spite good surgical technique and therefore is not an ideal protection method. Elimination of shunting, with its inherent stroke risk, could be expected to reduce perioperativc stroke rates provided ischemic damage was prevented by some other effective pro- tection method. However, any reduction in stroke rate by replacement of shunting with some other effective protection method is likely to be marginal in centers of technical expertise and wide experience of shunting but may possibly achieve some significance in centers of limited experience or in centers where shunting is used infrequently on a selective basis.
In this study, intraluminal shunting was replaced
by routine high-dose thiopental cerebral protection, and ischcmic EEG changes and low stump pressures were detected after cerebral protection had been implemented. In our experience of 1135 operations using this technique (1060 CEAs and 75 complex arch and great vessel reconstructions), in which the causes of neurologic deficit have been documented prospectively by neurologic examination, operative technique review, computed tomographic cerebral scanning, carotid duplex scanning, and autopsy, there have been no neurologic deficits, transient or permanent, that could be attributed to the use of high-dose thiopental. There are no reports in the experimental or clinical literature that suggest that thiopental carries an inherent risk of stroke. Pharma- cologic thiopcntal cerebral protection, unlike shunt- ing, therefore represents an ideal protection method provided it effectively protects against, or minimizes, ischemic cerebral neuronal damage. Both the exper- imental 14-27 and clinical 1,2,2s,29 literature indicate that thiopental does provide such protection.
Functional indicators o f ischemic stroke risk
Ferguson, 1° Moore and Hall, 3° and Hobson et al.31 have all demonstrated that stump pressure <25 mm Hg is an accurate indicator of ischemic stroke risk because such pressures occur inside the critical range of cerebral blood flow reduction (18 to 10 mls/100 gm/min). Importantly, they also demon- strated a 100% correlation between ischemic EEG changes and stump pressures <25 mm Hg. This combination, ischemic EEG change with stump pressure <25 mm Hg, would seem to represent the most accurate indicator of risk ofischemic stroke. In the current study there were seven patients who had stump pressure <25 mm Hg. However, the correla- tion between ischemic EEG changes and stump pres- sure <25 mm Hg was only 43% (3 of 7). The lower correlation between ischemic EEG changes and stump pressure <25 mm Hg may well be related to the prevention of ischemia, and hence of ischemic EEG changes, by thiopental at the neuronal level without any corresponding effect on stump pressure. The study also indicates that when stump pressure is <25 mm Hg, indicating critical reduction of cerebral blood flow, ischcmic EEG changes may persist de- spite the effects of thiopental. Because there were no ncurologic deficits, thiopental would appear to pro- vide protection against ischemic stroke in the face of critical blood flow reduction associated with persis- tent ischemic EEG changes and stump pressure <25 mm Hg.
The development of ischemic EEG changes dur-
JOURNAL OF VASCULAR SURGERY Volume 25, Number 4 Frawley et al. 617
ing carotid cross-clamping has been shown to have a high correlation (above 90%) with decreased cerebral blood flow. s,3°,32 Such changes can occur with stump pressures outside the range indicating critical reduc- tion of flow (i.e. above 25 mm Hg). McKay et al., 32 in an excellent study correlating ischemic EEG changes and low stump pressures with cerebral blood flow, found a 46% correlation between ischemic EEG changes and stump pressures <50 mm Hg. Some of these changes were seen outside the range of critical reduction of flow indicated by stump pressure <25 mm Hg. In the current study, two patients had ischemic EEG changes with stump pressures be- tween 26 mm H g and 50 mm Hg (43 mm Hg and 28 mm Hg) and two with stump pressure >50 mm Hg. In this situation it may be that the changes are related to focal or regional ischemia associated with intracerebral arterial disease, rather than global re- duction of cerebral blood flow associated with cross- clamping and reflected in stump pressure <25 mm Hg. It would seem safest to consider such EEG changes as indicating risk of ischemic stroke. The protective effect of thiopental may also have pre- vented the appearance of ischemic EEG changes in some patients with stump pressures >25 mm Hg.
It would seem that if stump pressure is used in isolation to detect risk, stump pressure <25 mm H g will fail to identify some patients at risk who have pressures >25 mm Hg, and stump pressure <50 mm Hg will overidentify some patients not at risk who have pressures of 26 to 50 mm Hg and will fail to identify some patients at risk with pressures > 50 mm Hg. Reliance on stump pressure measurement may thus result in higher stroke rates because of failure to identify all patients at risk and unnecessary shunting in others. Furthermore, the technique commonly used to measure stump pressure relies on clamping of the common and external carotid arteries with a patent internal carotid artery and may carry an inher- ent risk of embolization. The technique for stump pressure measurement described here eliminates this risk. Stump pressure measurement, as an indicator of stroke risk and for protective shunting, is clearly less than ideal and is invalid as a universal indicator of risk.
Although we have not been able to assess the awake patient, the occurrence of clinical symptoms or signs in the awake patient has been shown to almost invariably correlate with ischemic EEG changes 33 and therefore represents a valid indicator of the presence ofischemia during carotid occlusion. Its value as an indicator o f risk o f ischemic stroke and of the need for protective shunting is thus
similar to that o f the development o f ischemic EEG changes.
Clinical indicators o f ischemic s t roke risk
Clinical indicators of the risk of ischemic stroke bear no direct relationship to critical reduction of cerebral blood flow and cannot be universally re- garded as valid indicators of risk. Obviously, some patients will have critical reduction of flow but can only be identified by the development of functional changes that indicate reduction of flow with cross- clamping. In this study, six of the 14 patients with clinical indicators of risk developed functional indica- tors of risk: ischemic EEG changes alone (2), stump pressure <25 mm Hg alone (3), and both ischemic EEG changes and stump pressure <25 mm Hg (1). The correlation was 43% (6 of 14). Again, some ischemic EEG changes may have been prevented by thiopental. Nevertheless, it is clear that clinical indi- cations tbr protection are by no means accurate and are likely to result in unnecessary shunting with its potential for increasing perioperative stroke rates.
Risk associated wi th severity o f disease
This study has identified a potential risk of isch- emic stroke of 41.7% in patients who have severe bilateral disease compared with 24% in patients who have predominantly unilateral disease (Table II). This finding supports the commonly held perception that patients who have severe bilateral disease are at greater risk of ischemic stroke, and it is common practice to offer such patients cerebral protection. We have reported a 0.4% major stroke rate in 259 pa- tients with severe bilateral disease using thiopental protection, 2 and we have also indicated that when such protection is used during surgeu, patients who have severe bilateral disease are not at increased risk of stroke compared with patients who have unilateral disease? Other investigators, using intraluminal shunting for protection, have also reported that pa- tients who have severe bilateral disease are not at increased risk during protected surgery. 2°'21'34 It would seem, therefore, that shunting and thiopental protection both offer effective protection against ischemic stroke.
C O N C L U S I O N
The literature generally concedes that there is a risk of ischemic stroke during carotid occlusion for endarterectomy. Patients with previous stroke associ- ated with ipsilateral internal carotid occlusions, se- vere bilateral disease, ischemic EEG changes with clamping, and low stump pressures have been selec-
JOURNAL OF VASCULAR SURGERY 618 Frawley et aL April 1997
tively identified as being at risk and as requiring cerebral protection against ischemia. The 19 patients reported here, with commonly accepted indications for protective shunting, all fall into one of these risk categories with multiple (2 or 3) risk categories in more than half of them. An absolute risk of ischemic stroke was identified in 11, for an expected stroke rate of 57.9% (11 of 19) if no cerebral protective measures were used, particularly with the average carotid occlusion time of 41 minutes. There were no neurologic deficits. If the described risk factors and the indications for the need for cerebral protection are indeed valid, then the absence of perioperative stroke indicates that thiopental protection has pro- vided effective and safe protection against ischemic stroke. If, however, thiopental has not provided ef- fective cerebral protection, then ischemic EEG changes and stump pressure <25 mm Hg that persist during carotid occlusion, as indicators o f ischemic stroke risk and the need for cerebral protection, can be dismissed as invalid and protection itself dismissed as unnecessary. Such a conclusion cannot be sus- tained under the weight of evidence to the contrary, widely documented in the literature.
This study, therefore, allows the conclusion that thiopental cerebral protection prevents intraopera- tive ischemic stroke, safely replaces intraluminal shunting as a means of ischemic stroke prevention during CEA, and has no inherent stroke risk. There- fore, thiopental protection may be preferable to shunting with its inherent risk, and consequently, to techniques such as awake patient monitoring under local cervical block anesthesia and stump pressure measurement, which are designed to indicate the necessity for shunting.
REFERENCES
1. Frawley JE, Hicks RG, Horton DA, Gray LJ, Niesche JW, Matheson JM. Thiopental sodium cerebral protection during carotid endarterectomy: perioperafive disease and death. J Vasc Surg 1994;19:732-8.
2. Frawley JE, Hicks RG, Gray LJ, Niesche ]W. Carotid endar- terectomy without a shunt for symptomatic lesions associated with contralateral severe stenosis or occlusion. J Vase Surg 1996;23:421-7.
3. Moore WS, Barnett HJM, Beebe HG, Bernstein EF, Brener BJ, Brott T, et al. Guidelines for carotid endarterectomy: a multidisciplinary consensus statement from the Ad Hoc Com- mittee, American Heart Association. Stroke 1995;26:188- 201.
4. Rothwell PM, Slattery J, Warlow CP. A systematic compari- son of the risks of stroke and death due to carotid endarterec- tomy for symptomatic and asymptomatic stenosis. Stroke 1996;27:266-9.
5. Cherry KJ, Roland CF, Hallett JW, Gloviczki P, Bower TC, Toomey, BJ, et al. Stump pressure, the contralateral carotid
artery, and electroencephalographic changes. Am J Surg 1991;162:185-9.
6. Phillips MR, Johnson WC, Scott RM, Vollman RW, Levine H, Nabseth DC. Carotid endarterectomy in the presence of contralateral carotid occlusion. Arch Surg 1979;114:1232-9.
7. Brewster DC, O'Hara PJ, Darling RC, Hallett JW. Relation- ship of intraoperative EEG monitoring and stump pressure measurements during carotid endarterectomy. Circulation 1980;62:14-6.
9. Kass IS, Abramowicz AE, Cottrell JE, Chambers G. The barbiturate thiopental reduces ATP levels during anoxia but improves electrophysiological recovery and ionic homeostasis in the rat hippocampal slice. Neuroscience 1992;49:537-43.
10. Ferguson GG. Carotid endarterectomy: to shunt or not to shunt? Arch Neurol 1986;43:615-7.
11. Edwards WH, Edwards WH Jr, Jenkins JM, Mulherin JL Jr. Analysis of a decade of carotid reconstructive operations. J Cardiovasc Surg (Torino) 1989;30:424-9.
12. Javid H, Julian OC, Dye WS, Hunter JA, Najafi H, Goldin MD, et al. Seventeen-year experience with routine shunting in carotid artery surgery. World J Surg 1979;3:167-77.
13. Thompson JE, Talkington CM. Carotid endarterectomy. Adv Surg 1993;26:99-131.
15. Brott TG, Labretra RJ, Kempczinski RF. Changing patterns in the practice of carotid endarterectomy in a large metropol- itan area. JAMA 1986;255:2609-12.
16. Riles TS, Imparato AM, Jacobowitz GR, Lamparello PJ, Gi- angola G, Adelman MA, Landis R. The cause ofperioperative stroke after carotid endarterectomy. J Vasc Surg 1994;19: 206-16.
17. Davies MJ, Mooney PH, Scott "DA, Silbert BS, Cook RJ. Neurologic changes during carotid endarterectomy under cervical block predict a high risk of postoperative stroke. Anaesthesiology 1993;78:829-33.
18. Archie JP Jr. Technique and clinical results of carotid stump back-pressure to determine selective shunting during carotid endarterectomy. J Vasc Surg 1991;13:319-27.
19. Friedman SG, Riles TS, Lamparello PJ, Imparato AM, Sakwa MP. Surgical therapy for the patient with internal carotid artery occlusion and contralateral stenosis. J Vasc Surg 1987; 5:856-61.
20. Mackey WC, O'Donnell TF Jr, Callow AD. Carotid endarter- ectomy contralateral to an occluded carotid artery: periopera- tive risk and late results. J Vase Surg 1990;11:778-85.
21. Perler BA, Burdick JF, Williams GM. Does contralateral inter- nal carotid artery occlusion increase the risk of carotid endar- terectomy? J Vase Surg 1992;16:347-53.
22. Graham AR, Lord RSA. Validity of internal carotid back pressure (CBP) measurement in unilateral carotid stenosis. Abstracts, Royal Australasian College of Surgeons, General Scientific Meeting, Sydney 1985;168.
23. Halsey JH Jr. Risks and benefits of shunting in carotid endar- terectomy. Stroke 1992;23:1583-7.
24. Michenfelder JD, Milde JA, Sundt TM. Cerebral protection by barbiturate anaesthesia: use after middle cerebral artery occlusion in Java monkeys. Arch Neurol 1976;33:345-50.
25. Boarini DJ, Kassel NF, Coester HC. Comparison of sodium
JOURNAL OF VASCULAR SURGERY Volume 25, Number 4 Frawley et al. 619
thiopental and methohexital for high-dose barbiturate anaes- thesia. J Neurosurg 1984;60:602-8.
26. Guo J, White JA, Batjer HH. The protective effects of thio- pental on brain stem ischaemia. Neurosurgery 1995;37:490-4.
27. Drummond JC, Cole DJ, Patel PM, Reynolds LW. Focal cerebral ischaemia during anaesthesia with etomidate, isoflu- rane, or thiopental: a comparison of the extent of cerebral injury. Neurosurgery 1995;37:742-9.
28. Gross CE, Adams HP Jr, Sokoll MD, Yamada T. Use of anticoagulants, electroencephalographic monitoring, and barbiturate cerebral protection in carotid endarterectomy. Neurosurgery 1981;9:1-5.
29. Spetzler RF, Martin N, Hadley MN, Thompson RA, Wilkin- son E, Raudzens PA. Microsurgical endarterectomy under barbiturate protection: a prospective stud),. J Neurosurg 1986;65:63-73.
30. Moore WS, Hall AD. Carotid artery back pressure: a test of cerebral tolerance to temporary carotid occlusion. Arch Surg 1969;99:702-10.
31. Hobson RW, Rich NM, Wright CB, Fedde CW. Operative assessment of carotid endarterectomy: internal carotid arterial back pressure, carotid arterial blood flow, and carotid arte- riography. Am Surg 1975;41:603-10.
32. McKay RD, Sundt TM, Michenfelder JD, Gronert GA, Mes- sick JM, Sharbrough FW, et al. Internal carotid artery stump pressure and cerebral blood flow during carotid endarterec- tomy: modification by halothane, enflurane, and immovar. Anaesthesiology 1976;45:390-9.
33. Harris EJ, Brown WH, Parr), RN, Anderson WW, Stone DW. Continuous encephalographic monitoring during carotid ar- te n, endarterectomy. SurgeD' 1967;62:441-6.
34. Moore DJ, Modi JR, Finch WT, Sumner DS. Influence of the contralateral carotid artery on neurologic complications tbl- lowing carotid endarterectomy. J Vase Surg 1984;1:409-14.
Submitted May 31, 1996; accepted Sep. 13, 1996.
O N THE MOVE? d us your n e w address at least six w e e k s ahead
Don't miss a single issue of the journal! To ensure prompt service when you change your address, please photocopy and complete the form below.
Please send your change of address notification at least six weeks before your move to ensure continued service. We regret we cannot guarantee replacement of issues missed due to late notification.
JOURNAL T I T L E : Fill in the title of the journal here.
OLD ADDRESS: Affix the address label from a recent issue of the journal here.
NEW ADDRESS: Clearly print your new address here.
Name
Address
City/State/ZIP
C O P Y AND MAIL THIS FORM T O : J o u r n a l S u b s c r i p t i o n S e r v i c e s M o s b y - Y e a r Book , Inc. 11830 W e s t l i n e I n d u s t r i a l Dr. St. Lou i s , M O 63146-3318
OR FAX T O : 3 1 4 4 3 2 - 1 1 5 8
[~v~ Mosby
OR PHONE: 1-800453-4351 O u t s i d e t h e U.S., ca l l 314-453-4351
![[Accessory] [9361] Rock of Bral](https://static.documents.pub/doc/80x56/577cd5761a28ab9e789ad89a/accessory-9361-rock-of-bral.jpg)