Intracerebral hemorrhage after carotid endarterectomy: Incidence, contribution to neurologic morbidity, and predictive factors

Transcription

1 Intracerebral hemorrhage after carotid endarterectomy: Incidence, contribution to neurologic morbidity, and predictive factors Kenneth Ouriel, MD, Cynthia K. ShorteU, MD, Karl A. Illig, MD, Roy IC Greenberg, MD, and Richard M. Green, MD, Rochester, NY Purpose: With a diminishing rate of cardiac and neurologic events after carotid endarterectomy, intracerebral hemorrhage is gaining increasing importance as a cause of perioperative morbidity and mortality. To date, information has been largely anecdotal, and there has been no comparison with a control group of patients. Methods: The records of all patients experiencing symptomatic intracerebral hemorrhage after carotid endarterectomy were reviewed and compared with data from 50 randomly selected patients who did not experience intracranial bleeding. Univariate analyses were performed, using the Fisher exact test for dichotomous data and the Student t test for continuous data. Results: During a 6-year period, symptomatic intracranial hemorrhage developed in 11 (0.75%) of 1471 patients undergoing carotid endarterectomy, accounting for 35% of the 31 total perioperative neurologic events. Hemorrhage occurred a median of 3 days postoperatively (range, 0 to 18 days). Signs and symptoms included hypertension in all 11 patients, headache in 7 conscious patients (64%), and bradycardia in 6 patients (55%). Massive hemorrhage with herniation and death occurred in 4 patients (36%). Moderate hemorrhage developed in 5 patients (45%); 3 of these patients had partial recovery, and 2 had complete recovery. Petechial hemorrhage occurred in the remaining 2 patients (18%), 1 with partial and 1 with complete recovery. In comparison with the control group, there were no differences in respect to sex, indication for operation, smoking or diabetic history, and antiplatelet therapy or perioperative heparin management. Patients with intracranial hemorrhage were, however, younger, more frequently hypertensive, had a higher degree of ipsilateral and contralateral carotid stenosis, and had a higher rate of contralateral carotid occlusion. Conclusion: Intracranial hemorrhage occurs with notable frequency after carotid endarterectomy and accounts for a significant proportion of neurologic morbidity and mortality. Younger patients, hypertensive patients, and patients with severe cerebrovascular occlusive disease appear to be at greatest risk for the complication. (J Vasc Surg 1999;29:82-9.) Carotid endarterectomy is one of the most commonly performed peripheral vascular procedures, with a well-demonstrated benefit in symptomatic and asymptomatic patients with high-grade bifurcation stenoses.1, 2 Although ischemic cerebral infarction is the most widely publicized perioperative complication, intracranial hemorrhage is a compli- From the Department of Surgery, the University of Rochester. Presented at the Fifty-second Annual Meeting of The Society for Vascular Surgery, San Diego, Calif, June 9-10, Reprint requests: Kenneth Ouriel, MD, Box 652, University of Rochester, 601 Elmwood Ave, Rochester, NY Copyright 1999 by The Society for Vascular Surgery and International Society for Cardiovascular Surgery, North American Chapter /99/$ /6/ cation that occurs in association with the repair of critically stenotic carotid lesions, previous stroke, hypertension, and anticoagulation. As refinements in preoperative cardiac screening, operative technique, and perioperative patient treatment effect a reduction in ischemic stroke and myocardial complications, 3 intracranial hemorrhage will attain increasing importance as a perioperative complication. The first step in decreasing the risk of the complication is to identify those factors associated with its occurrence. Parameters associated with intracranial hemorrhage after carotid endarterectomy were studied, by means of comparing a group of patients who had intracerebral bleeding with a control group of patients who did not experience this complication.

2 Volume 29, Number 1 Ouriel et al 83 Fig 1. Magnetic resonance image of a "petechial" hemorrhage, with patchy bleeding following the pattern of the gyri. Table I. Proposed classification of postcarotid endarterectomy intracranial hemorrhage Size of Number of hemorrhagic hemorrhagic Mass Class areas areas effect Herniation Petechial Small Many None None Moderate Moderate Usually 1 None to None mild Massive Large Usually 1 Significant Impending or ~ank METHODS Patients undergoing carotid endarterectomy between January 1, 1992, and November 1, 1997, in whom intracerebral hemorrhage associated with neurologic deterioration developed within 30 days of operation were identified. During this period, 1471 total carotid endarterectomy procedures were performed in 1208 patients; 263 of the procedures (17.9%) were bilateral. A control group of 50 patients who did not experience intracranial bleeding was randomly selected from registry databases using a computer-generated random number sequence, so that the representation by year in the control group matched the ratio of procedures performed per year in the overall group. The medical records were reviewed, and data on baseline variables and operarive details were tabulated. The head computed tomographic or magnetic resonance imaging scans of the patients with intra- Fig 2. An example of a "moderate" intracranial bleed, with a localized collection of blood in the absence of significant mass effect. cerebral hemorrhage were reviewed, and the events were placed into 3 proposed, novel classifications: massive, moderate, and petechial (Table I). Massive hemorrhage was defined when the amount of bleeding produced cerebral herniation or severe mass effect with impending herniation (Fig 1). Petechial hemorrhage was defined by the presence of numerous, nonconfluent areas of intracerebral blood with a pattern following the gyri, without evidence of mass effect (Fig 2). Moderate hemorrhages comprised those bleeds that fell between the criteria of petechial and massive hemorrhage; patients with moderate events had areas of confluent hemorrhage with or without mass effect, but without cerebral herniation (Fig 3). Bleeding into a previously documented area of cerebral infarction was classified as a postoperative hemorrhagic event, but only if the absence of preoperative bleeding could be documented. Patients with extension of intracranial hemorrhage that was present preoperatively were not included in the series. The hemorrhagic events were classified into 3 novel groups defined by time of onset. Early events comprised those bleeds that occurred within 6 hours of operation. Intermediate events occurred more than 6 hours after endarterectomy, but before the

3 84 Ouriel et al January 1999 Table II. Baseline characteristics of patients undergoing carotid endarterectomy (unless otherwise stated, the denominator is the number of patients) Randomly selected patients Patients with Parameter All patients without hemorrhage without hemorrhage hemorrhage Procedures Patients Age (years + SEM) _+ 2.3 Male : Female 57.4% : 42.6% 52.0% : 48.0% 64.0% : 36.0% Smoking history 691 (57.2%) 30 (60.0%) 8 (72.7%) Diabetes 160 (13.2%) 11 (22.0%) 2 (18.2%) Hypertension 678 (56.1%) 32 (64.0%) 9 (81.8%) Coronary artery disease 543 (45.0%) 27 (54.0%) 5 (45.5%) Chronic pulmonary disease 266 (22.0%) 5 (10.0%) 0 (0.0%) Hyperlipidemia 200 (16.6%) 8 (16.0%) 2 (18.2%) Indication* Asymptomatic 497 (34.0%) 11 (22.0%) 4 (36.4%) Transient event 549 (37.6%) 23 (46.0%) 3 (27.2%) Completed event 414 (28.4%) 16 (32.0%) 4 (36.4%) *Denominator is number of procedures SEM, Standard error of the mean. seventh postoperative day. Late events occurred on or after the seventh postoperative day. Continuous data was compared with the Student t test. The Fisher exact test was used to analyze categorical data. Significance was assumed when the 2- tailed P value was less than.05. RESULTS Baseline characteristics of the patient groups were comparable (Table II). Intracranial hemorrhage with neurologic deficit occurred in 11 of the 1471 patients (0.75%) who underwent carotid endarterectomy procedures in the 6-year period. All events were ipsilateral to the operated carotid artery. In the same period, 20 bland postoperative infarctions occurred; thus, hemorrhagic events accounted for 35% of the 31 total perioperative (less than 30 day) fixed neurologic events. Two asymptomatic intracranial hemorrhages were identified on computed tomographic scans performed for other reasons; both were petechial in nature. Bleeding occurred a median of 3 days after surgery (range, 0 to 18 days). The bleeding events were early in 4 of the cases; 2 of these patients awoke with dense neurologic deficits culminating in death. Hemorrhage was detected at an intermediate time point in 3 cases. The remaining 4 patients were discharged from the hospital and were without new neurologic deficits at 1 week, but had signs and symptoms of intracranial bleeding subsequently. Massive hemorrhage occurred in 4 patients, each of whom died within 72 hours of the event. Moderate hemorrhage developed in 5 patients, 3 of whom had partial recoveries and 2 of whom had complete neurologic recoveries. Petechial hemorrhage occurred in 2 patients, 1 with partial recovery and I with complete recovery. No patient with moderate or petechial hemorrhage died of the event. Early hemorrhagic events were often massive, whereas late events were usually less severe (Fig 4). Thus, there appeared to be an association between the time of onset of the bleed and its magnitude; earlier bleeds tended to be more severe, whereas late bleeds were more often moderate or petechial in character. Intracranial hemorrhage was associated with headache in 7 of 7 conscious patients, severe hypertension (systolic blood pressure greater than 180 mm Hg, diastolic blood pressure greater than 100 mm Hg) in all 11 patients, and bradycardia (heart rate less than 60) in 6 of the patients. No patient experienced seizures before hemorrhage, but 3 of the patients had seizures after the event. In comparison with the control group of 50 patients, there were no differences in sex distribution, indication for operation (asymptomatic, transient ischemia, or fixed neurologic event), or smoking or diabetic history. Recent stroke was not associated with an increased risk of intracranial hemorrhage, but the number of patients was too small to conclude this with certainty. There were no differences in the use of preoperative antiplatelet therapy or perioperative heparin management in patients with or without intracranial hemorrhage, but the statistical power of this observation is low, given the ubiquitous use of perioperative aspirin, intraoperative heparin, and protamine reversal. Patients with hemorrhage were, however, younger and more often hypertensive

4 Volume 29, Number 1 Ouriel et al 85 Table III. Parameters associated with an increased risk of intracranial bleeding after carotid endarterectomy Hemorrhage No hemorrhage Parameter (n = 11) (n = 50) P value Age (years) 66 _ _ History of hypertension 82% 64%.05 Ipsilateral stenosis 92% _+ 1.0% 77% %.006 Contralateral stenosis 78% % 51% ± 3.7%.003 Contralateral occlusion 64% 12%.05 Stenosis, percent diameter reduction as measured by duplex ultrasound examination. Fig 3. A massive intracranial bleed that occurred immediately after carotid endarterectomy. The patient awoke from anesthesia, but became comatose 1 hour after surgery and died on the first postoperative day. (Table III). A greater degree of preoperative carotid stenosis, both ipsilateral and contralateral, was also noted in patients with hemorrhage. The frequency of contralateral carotid occlusion was greater in the patients who experienced hemorrhagic events than in those who did not. The treatment of intracranial hemorrhage was based on the magnitude of the bleed. Neurosurgical intervention was considered in patients with massive bleeds, but open evacuation of the hematoma was declined in each case on the basis ofrapidiy progressive, irreversible neurologic deficits. An intracranial pressure monitor was used in two of these patients, but herniation progressed despite modifications in fluid status and ventilation. DISCUSSION The first reports of carotid revascularization appeared in the early 1950s, with Eastcott, Picketing, and Rob's 4 landmark communication describing the successfial resection and reanastomosis of the carotid bifurcation in a patient with transient ischemic attacks. A decade later, Breutman and associates 5 reported 6 cases of cerebral hemorrhage developing after carotid endartercctomy, and the next year, Wylie 6 reported an association between carotid repair for acute stroke and subsequent intracranial hemorrhage, describing four ipsilateral and one contralateral intracranial bleeding events. To date, there have been 57 cases of postendarterectomy intracranial hemorrhage reported in the English language literature; the present series brings this number to 68. However, the number of cases in the literature dramatically underestimates the frequency of this complication. Assuming a risk of just under 1% 749 and an annual frequency of carotid endarterectomy of more than 100,000 procedures annually in Medicare bencficiaries, l intracranial hemorrhage likely occurs in almost 1000 patients undergoing carotid endarterectomy in the US Medicare population alone. A variety of risk factors have been identified for postendarterectomy intracranial bleeding. The performance of carotid revascularization in the setting of an acute cerebral infarction was perhaps the first risk factor defined. 6 Hypertension also was implicated by Caplan et al, 11 when these investigators reported fatal hemorrhagic strokes occurring shortly after endarterectomy. Severe cerebrovascular occlusive disease has been suggested as another factor predisposing patients to hemorrhage. Hafner and colleagues 12 described a significant association between the repair of critical carotid stenoses and postoperative hemorrhage. The present series of 11 patients with intracranial hemorrhage after carotid endarterectomy confirms the association between postoperative bleeding, hypertension, and severe extracranial cerebrovascular occlusion. Hemorrhage occurred in patients with those symptoms generally associated with increased intracranial pressure, including headache, severe hypertension, and bradycardia. In contrast with previous series, a younger age was found to bc an additional risk factor in the present study. It is possible that the design of previous studies did not allow identification of age as a factor, because most of

5 86 Ouriel t al January 1999 Massive I i Early Intermediate Late Fig 4. Timing of the hemorrhagic events. Immediate cvents were manifest at the conclusion of the surgical procedure. Early events occurred after patients awoke normally from anesthesia, but betbre the seventh postoperative day. Late events occurred more than 1 week after the procedure. these studies were anecdotal in nature and did not include a control group for comparison. The explanation for this finding remains obscure, and the relatively small number of patients with intracranial hemorrhage precluded the performance of a meaningful multifactor analysis, an analysis that would have elucidated whether age alone Was associated with bleeding or whether it was merely a confounding variable. Our findings and the findings of others implicate postoperative "hypcrperfusion" as the common etiologic mechanism underlying intracranial hemorrhage after carotid repair. Sundt 13 documented an augmentation in cerebral blood flow after carotid endarterectomy and suggested that the "hyperperfusion syndrome" was the cause of such symptoms as headache and seizures, symptoms that may be harbingers of impending cerebral hemorrhage. Piepgras, 14 from the same institution, documented ipsilateral cerebral blood flows more than twice baseline in 14 patients with postoperative intracerebral hemorrhage. Jansen and associates 7 documented increased transcranial Doppler velocity measurements and related these to postendarterectomy intracranial bleeding. Two separate groups documented intracranial arterial histologic changes resembling those observed in malignant hypertension in patients who died of intracranial hemorrhage after carotid endartercctomy.9,15 Hemorrhage occurs within the healthy brain tissue, and not the frankly infarcted tissue, adding further credence to the hyperperfusion hypothesis, with vascular disruption at the site of a diseased, hyperperfuscd vessel. These observations arc consistent with the following etiologic hypothesis for postoperative intracerebral hemorrhage. First, long-standing cxtracranial carotid occlusive disease and hypoperfusion cause compensatory intracranial vascular dilatation and loss of autorcgulation. Carotid endartcrectomy relieves the stcnosis, but hypcrpcrfusion results from increased pressurc into a distal bed fixed in maximal dilatation, not unlike the hypcremic foot after lower extremity revascularization. In its mild-to-moderate form, the hyperperfusion syndrome produces cerebral edema, headache, and occasionally seizures. In its scverc form, rupture of an abnormal, hyperperfused intracranial vessel may occur with resultant intracranial hemorrhage. Identification of patients prone to intracranial hemorrhage will allow one to take precautions that may reduce the risk of the complication. Maintenance of blood pressure at slightly lower than normal levels, avoidance of postoperative anticoagulation and possibly antiplatelet agents, and relative fluid restriction appear to be rational in patients with several risk factors. Transcranial Doppler velocity measurements,

6 Volume 29, Number 1 Ouriel et al 87 although possibly helpful in predicting impending hemorrhage, are not within the repertoire of most vascular laboratories. Patients should be informed of the possibility of this potential complication and should be instructed to notify the vascular team immediately if a severe, throbbing headache develops. Unfortunately, tittle can be done to salvage the patient once a massive bleed has occurred. Evacuation of the hematoma is unnecessary when the bleeding is mild to moderate, as recovery without intervention is likely. Further, evacuation will not be fruitful when the bleeding is massive, because these patients manifest irreversible cerebral injury at the outset. It is clear that the complication of intracranial hemorrhage after carotid endarterectomy cannot be eliminated entirely, especially in high-risk patients in whom perioperative antithrombotic therapy and avoidance ofhypotension are beneficial. Nevertheless, aggressive prophylactic measures and appropriate patient education will probably effect a noticeable decrease in its incidence. REFERENCES 1. North American Symptomatic Endarterectomy Trial Collaborators. Beneficial effect of carotid endarterectomy in symptomatic patients with high grade stenosis. N Engl J Med 1991;325: Executive Committee for the Asymptomatic Carotid Atherosclerosis Study. Endarterectomy for asymptomatic carotid artery stenosis. JAMA 1995;273: Till JS, Toote JF, Howard VJ, Ford CS, Williams D. Declining morbidity and mortality of carotid endarterectomy. The Wake Forest University Medical Center experience. Stroke 1987;18: Eastcott HHG, Picketing GW, Rob CG. Reconstruction of internal carotid artery in a patient with intermittent attacks of hemiplegia. Lancet 1954;2: Breutman ME, Fields WS, Crawford ES, DeBakey ME. Cerebral hemorrhage in carotid artery surgery. Arch Neurol 1963;9: Wylie EJ, Hein EF, Adams JE. Intracranial hemorrhage following surgical revascularization for treatment of acute strokes. J Neurosurg 1964;21: Jansen C, Sprengers AM, Moll FL, Vermeulen FE, Hamerlijnck RP, van GI, et al. Prediction of intracerebral hemorrhage after carotid endarterectomy by clinical criteria and intraoperative transcranial Doppler monitoring. Eur J Vasc Surg 1994;8: Reith HB, Edelmann M, Reith C. Spontaneous intracerebral hemorrhage following carotid endarterectomy. Experience of 328 operations from Int Surg 1992;77: Mansoor GA, White WB, Grunnet M, Ruby ST. Intracerebral hemorrhage after carotid endarterectomy associated with ipsilateral fibrinoid necrosis: a consequence of the hyperperfusion syndrome? J Vasc Surg 1996;23: Hsia DC, Moscoe LM, Krushat WM. Epidemiology of carotid endarterectomy among Medicare beneficiaries: update. Stroke 1998;29: Caplan LR, Skillman J, Ojemann R, Fields WS. Intracerebral hemorrhage following carotid endarterectomy: a hypertensive complication? Stroke 1978;9: Hafner DH, Smith RB, King OW, Perdue GD, Stewart MT, Rosenthal D, et at. Massive intracerebral hemorrhage following carotid endarterectomy. Arch Surg 1987;122: Sundt TM j'r, Sharbrough FW, Piepgras DG, Kearns TP, Messick JM Jr, O'Malley MK. Correlation of cerebral blood flow and electroencephalographic changes during carotid endarterectomy, with results of surgery and hemodynamics of cerebral ischemia. Mayo Clin Proc 1981;56: Piepgras DG, Morgan MK, Sundt TM Jr, Yanagihara T, Neschis DG. Intracerebral hemorrhage after carotid endarterectomy. J Neurosurg 1988;68: Bernstein M, Fleming JFR, Deck JH. Cerebral hyperperfusion after carotid endarterectomy: a cause of cerebral hemorrhage. Neurosurgery 1984;15:50-6. Submitted Jun 11, 1998; accepted Aug 31, DISCUSSION Dr William R. Flinn (Baltimore, Md). Intracerebral hemorrhage after carotid endarterectomy almost always has a very morbid or fatal outcome. A 36% mortality rate was observed in the present series, and there was a 57% rate of permanent neurologic deficit among the survivors. It is most frustrating for the surgeon that these events almost always occur in the setting of an otherwise uneventful operation. In the authors' experience, these events were gratifyingly rare, less than 1% of all cases, but they represented more than a third of all the neurologic deficits after carotid endarterectomy during their study period. Based on their analysis, the authors concluded that postoperative intracerebral hemorrhage was more likely in younger patients, patients with severe hypertension, and those with the most severe ipsilateral and/or contralateral carotid disease. Although their control group may have arguably been statistically unreliable, these findings are almost identical to those of previous reports from 1, 2, or even 3 decades ago, so they're probably reasonably reliable retrospective observations. Unfortunately, however, such retrospective observations have done little in the past to expand our understanding of these events, much less prevent them. It's hardly likely at the present time that we would consider withholding operations from patients because they were younger, hypertensive, or had more severe carotid disease. How then can we make sense of all this? One must assume that none of these patients had intracranial aneurysms as a source of hemorrhage, although the authors did not mention whether all 11 patients with

7 88 Ouriel et al January 1999 bleeds underwent complete preoperative arteriography. The authors, like most of us, presume that these events arise because of a reperfusion injury. They state that patients at risk have cerebral hypoperfusion and incur hyperperfusion after carotid endarterectomy. One would assume from this line of reasoning that these patients with pre-existent cerebral hypoperfusion would be at higher risk for the development of cerebral ischemia during cross-clamping and, thus, would be more likely to require shunting. What techniques did the authors use in this series to detect intraoperative cerebral ischemia during carotid endarterectomy? Was cerebral ischemia universal in the patients who had subsequent bleeds, or was it significantly more common than in their overall patient population? The authors also note that the most morbid intracerebral bleeds occurred very early. In this series, 36%, or 4 cases, 2 of which were fatal, occurred in the first 24 hours, and 2 occurred intraoperatively, which in our experience has been exquisitely rare. Can we learn anything more from the analysis of these so-called worst cases? Were these patients who had previous strokes? Were they unstable neurologically? Did they have electroencephalographic (EEG) changes and the shunt couldn't be placed? Was there anything that could be identified in this worst-case scenario that would help us understand this? Or, sadly, were these just cases that were normal, straightforward procedures until they became catastrophic? Finally, how do we make these things go away? The authors seem to imply that we know the pathophysiology. Can't we simply do a test to identify the patients at risk? Should we do transcranial Doppler examinations on all the patients with severe carotid stenosis? Should we give them acetezolamide or have them breathe into a paper bag to see if they autoregulate? Based on this review, what significant changes have the authors made in their diagnostic or treatment algorithms in the treatment of their patients, their young, hypertensive patients with the most severe carotid disease? I truly enjoyed reviewing this manuscript and, of course, was honored to be invited by the Society to discuss it. I hope the authors have an answer to these questions, because I don't, but I truly doubt they do. Thank you. Dr Kenneth Ouriel. The goal of our review was to identify those factors associated with intracranial bleeding. Unfortunately, we did not come up with an answer for how to prevent these problems. We believe that the most susceptible patients are those with preoperative cerebral hypoperfusion. We do not, however, have a good test for cerebral hypoperfusion. Such tests as xenon computed tomography scans have not been completely reliable in these patients, because the hypoperfusion syndrome may be so borderline that it is only unmasked through provocative testing. What these patients do manifest are cerebral vessels that are fixed in dilatation, vessels that do not respond normally with constriction once normal blood flow is re-established. Your first question related to the techniques that we use intraoperatively and preoperatively to assess whether we have hyp0perfusion. We do use EEGs in virtually every patient; unfortunately, it's been completely insensitive in our hands for predicting hypoperfusion. We do shunt on the basis of EEG changes; unfortunately, with the small number of patients in this study, we were unable to show any difference between the requirement for a shunt and intracerebral bleeding. In the same context, you asked about patients who had a previous stroke and unstable patients. The numbers are small, but we were unable to come up with any correlation between preoperative symptoms and intracranial hemorrhage. How do we decrease the frequency of this complication? We are much more aggressive now when a patient has a perioperative headache. This will not diminish the frequency of the very early recovery room bleeds, but close hemodynamic control and the discontinuation of anticoagulants and antiplatelet agents may help when patients complain of throbbing headaches. Dr Keith D. Calligaro (Philadelphia, Pa). Ken, I enjoyed your talk. However, from a practical standpoint, do you treat the patients of whom you are very suspicious of being more prone to intracerebral bleeds differently postoperatively? Specifically, what I'm referring to is that in many centers patients are transferred to the floor the same day and discharged home the next day. We tend to be overly cautious with patients in whom there is a concern about intracerebral bleeds. And when we're very worried, we'll keep these patients in the intensive care unit to specifically keep an eye on their blood pressures and start antihypertensive medications early. The problem with that approach is that a significant percent of patients undergoing this operation are going to have bilateral severe stenosis and have preoperative hypertension, and I'm just wondering how you've changed your postoperative management. Dr Ouriel. We, too, have been very aggressive in getting patients to the floor without having them spend an inordinate amount of time in the postanesthesia care unit. They do not go to the intensive care unit unless they have a severe postoperative headache or labile blood pressure. In these patients, computed tomography is obtained, and the patients are transferred to the intensive care unit. We haven't been proactive in this regard. When preoperative risk factors have been identified, direct admission to the intensive care unit may be advisable. Dr Ramon Berguer (Detroit, Mich). Dr. Ouriel, if the pathophysiology of this problem is hypoperfusion, then in those patients who have a contralateral carotid occlusion the bleed should be equally frequent in both hemispheres, or perhaps even a little higher on the contralateral hemisphere. Can you tell us if this is so in your series? Dr Ouriel. There are at least 2 cases in the literature that describe contralateral bleeds after carotid endarterectomy. I'm not sure that I agree with you about where the bleed is going to occur. Reperfusion of hypoperfused hemispheres should result in a greater increase in blood flow on the ipsilateral side, because of the requirement for flow to traverse the small collaterals within the circle of Willis to get to the

8 Volume 29, Number 1 Ouriel et al 89 contralateral hemisphere. This is, however, pure speculation to explain an aneurysmal empiric finding. Dr Stephen C. Nicholls (Seattle, Wash). You've alluded to the pathophysiology of this problem and the possible use of preoperative single photon emission computed tomography (SPECT) scans to identify patients at risk, although, as you stated, they are qualitative rather than quantitative tests. Another option is to perform CO 2 reactivity with TCD testing preoperatively in those patients you think are at risk, particularly the group with contralaterai occlusion, which, as you've shown, has a higher incidence of events. You can also do on-line TCD middle cerebral artery monitoring during the surgery, when you would expect to get reactive hyperemia after unclamping the carotid. If this is sustained and does not subside within a minute or 2, you can then treat the hyperdynamic circulation in that ipsilateral middle cerebral artery pharmacologically. This monitoring can eas- fly be continued in the recovery room. I would suggest that the option of CO 2 testing of vasomotor reactivity with TCD is one option for quantitative assessment of this problem preoperatively, and that 3-methylcholanthrene can be used to monitor the hyperemic response intraoperatively and postoperatively, and intervention made when indicated. Another option for a quantitative test preoperatively is xenon, but that is less available and more expensive. Would you comment on these options? Dr Ouriel. We have not used transcranial Doppler intraoperatively or immediately postoperatively to address this problem. As I'm sure you're alluding to, there are reports in the literature, specifically reports from Europe, in which TCD was done on virtually every patient undergoing carotid endarterectomy. There is a well-documented relationship between sustained increased TCD velocities and the development of the hypoperfusion syndrome.