Intracranial hemorrhage after carotid endarterectomy

Intracranial hemorrhage after carotid endarterectomy Frank B. Pomposelli, M.D., Patrick J. Lamparello, M.D., Thomas S. Riles, M.D., Claude C. Craighead, M.D., Gary Giangola, M.D., and Anthony M. Imparato, M.D.,

New York, N.Y. Among 1500 carotid endarterectomies performed between 1975 and 1984, 11 ipsilateral intracranial hemorrhages (IH) occurred between the first and tenth postoperative days for an incidence of 0.7%. The mortality rate among these patients was 36%. The only recognizable predisposing factor was relief of high-grade carotid stenosis (greater than 90%) whereas other factors such as age (58 to 81 years), preoperative hypertension (systolic blood pressure 120 to 160 nun Hg), preoperative head CT scans showing recent infarction (only one in five positive), and preoperative cerebral infarction (only 1 of 11 patients) did not play a role. All patients had normal coagulation studies. No patient required a shunt because all tolerated cross-damping of the carotid artery. Postoperative systolic blood pressures were 200 to 240 mm Hg in 6 of 11 patients. The time of occurrence of IH extended from the immediate postoperative period to the tenth postoperative day (mean interval 3.3 days). Treatment consisted ofcraniotomy in five patients; four survived and one recovered completely. Of the six patients treated nonoperatively, three survived and two completely recovered. IH shares equal incidence with recurrent thrombosis, cross-clamping ischemia, and embolization as a cause ofperioperative stroke. Although all except IH can be prevented by current practice, the means of preventing IH are not apparent; however, careful monitoring of blood pressure to prevent uncontrolled hypertension deserves consideration. (J VASe SuItG 1988;7:248-55.)

Reports from centers with large experiences in carotid surgery have shown that carotid endarterectomy can be performed with an overall stroke rate of 2% to 3%. At New York University, the postoperative stroke incidence was 2.5%. 1 The routine use of local anesthesia with careful observation of the awake patient together with an aggressive diagnostic approach including angiography, CT scanning of the brain, and ultimately exploration of the operative sites has allowed us to identify the cause of the strokes in 80% of patients. Embolization, carotid thrombosis, hypotension, intracranial hemorrhage, and uncertain causes each accounted for approximately 0.5% of perioperative stroke? The early experience of Wylie et al.s and others ~-6 revealed the high likelihood of catastrophic intracerebral hemorrhage in patients who had recently suffered an acute stroke and underwent subsequent carotid endarterectomy to relieve total occlusion or

From the Division of Vascular Surgery, New York University Medical Center. Presented at the Forty-first Annual Meeting of the Society for Vascular Surgery,Toronto, Ontario, Canada, June 9d0, 1987. Reprint requests: Patrick ]. Lamparello, M.D., New York UniversityMedical Center, Division of Vascular Surgery, 550 First Avenue, New York, NY 10016.

248

high-grade stenotic lesions of the ipsilateral internal carotid artery. Animal studies of Meyer 7 and Harvey and Rassmussen8 supported the contention by Wylie et al. that endarterectomy in those circumstances converted an ischemic infarct into a hemorrhagic one, principally because of the damage to arterioles and capillaries by major cerebral arterial occlusions. On the basis of these observations, the attempts to open totally occluded internal carotid arteries xi, the presence of recently completed stroke were largely abandoned. Empirically, carotid surgery was delayed 4 to 6 weeks after completed stroke in patients with stenotic lesions. This proved to be successful in decreasing the incidence of cerebral hemorrhage after endarterectomy in these patients; however, occasional cases were still seen. Patients with very stenotic carotid lesions, anticoagulation with heparin 9 or warfarin, 1° or uncontrolled hypertension 11 had risk factors previously discussed in the medical literature. Recently there have been reports of intracerebral hemorrhage in patients without apparent previous cerebral infarctions, who were not given anticoagulants and who were normotensive after relief of high-grade stenosis of the ipsilateral internal carotid artery.~2,1sThis prompted review of our experience with intracranial hemorrhage after carotid endarterectomy in a series of 1500 procedures per,,"

Volume 7 Number 2 February 1988

Intracranial hemorrhage after carotid endarterectomy 2 4 9

Table I. Preoperative clinical data summary and procedure Preoperative Patient Presenting angiography (age~sex) symptoms (% Stenosis) 59/F

(L) TIA

99 (L)

Preoperative CT scan

Operation

Normal

(L) CE

Normal

(L) CE

(R) Cortex infarct (recent) (L) Cortex infarct (old) ND

(R) CE (L) CE

Normal

(R) CE

ND

(R) CE

ND

(R) GE

ND

(R) CE

ND

(L) CE

Normal

(L) CE

50 (R) 62/F 58/M

Deteriorating mental status (R) Infarct

63/M

(L) TIA

81/M

(R) TIA

62/M

(R) TIA

70/F

(R) TIA

68/F

(R) & (L) TIA

61/M

(R) TIA

71/F

Vertigo

67/F

(L) AF

99 99 99 50 99 50 95 60 99 80 99 40 90 70 90 60 100

(L) (R) (R) (L) (L) (R) (L) (R) (R) (L) (R) (L) (R) (L) (R) (L) (R)

(L) CE

90 (L) 95 (L) 60 (R)

TIA = transient ischemic attack; AF = amaurosis fugax; CE = carotid endarterectomy; ND = no data.

formed over a recent 10-year period (1975 to 1984) at New York University Medical Center. The purpose of this article is to identify the factors that place patients most at risk for intracranial hemorrhage after carotid endarterectomy, discuss possible pathophysiologic mechanisms, and describe our principles for diagnosis and management of this difficult problem. MATERIAL A N D M E T H O D S

Among 1500 carotid endarterectomies consecutively performed between 1975 and 1984 on the vascular surgical service at New York University Medical Center, 11 patients were identified who sustained a perioperative stroke because of an ipsilateral intracranial hemorrhage. The hospital charts of these patients were reviewed. The age, sex, indication for surgery, symptom-free interval before endarterectomy, and preoperative CT scan findings were recorded. All patients underwent preoperative biplane arterial angiography. The degree of stenosis in the carotid arteries was determined by measurement of the transverse luminal diameter of the involved vessels as visualized on the arteriogram. Ten operations were done with cervical block anesthesia and one with general endotracheal anesthesia. No shunts were used. All patients remained neurologically intact during carotid cross-clamping. The one patient operated

on while under general anesthesia was initially given regional cervical block anesthetic. This patient was neurologically intact with carotid cross-clamping and was placed under general anesthesia halfway through the carotid endarterectomy because of discomfort. All patients received 3000 units of intravenous heparin before carotid cross-clamping; this was not reversed. All patients receive aspirin 325 mg/day and dipyridamole 75 mg/day preoperatively. Five patients received aspirin 325 mg/day and dipyridamole 75 rag/day postoperatively. One patient was given heparin before surgery, but this was discontinued in the postoperative period. The blood pressure at admission, maximal intraoperative blood pressure, and maximal blood pressure before occurrence of cerebral hemorrhage were noted. All patients had radial intra-arterial pressure monitoring performed in the operating room and recovery room. The time period between surgery and hemorrhage influenced subsequent management. Patients who had stroke in the recovery room were immediately operated on again. If the carotid arteries appeared pulsatile, intraoperative carotid angiograms were obtained to rule out irregularities at operative sites that could have been the source of emboli. Patients whose deficits occurred later underwent noninvasive evaluation with ocular photoplethysmography and/or duplex scan, angiography, CT scan, or

250

Journal of VASCULAR SURGERY

Pomposelli et al.

240 l

240

220 "

220

--

2O0

200"

IIiit 1t I

180-

160 ~80

160"

mm Hg

mm Hg

HO"

1-

~40 120

120 "

I00 IO0 "

80 80-

60 60-

40 40 i 2

I

, 3

i 4

, 5

, 6

, 7

I

, 8

9

I0

2 0 0

-1

l

180160-

mmHg

140,

120 " I00 806040, I

, 2

, 3

i 4

, 5

, 6

, 7

,

,

i

,

,

,

,

J

4

5

6

7

8

9

I0

II

PATIENT

Fig. 1. Preoperative blood pressure at admission.

220"

,

3

II

PATIENT

240-

,

2

, 8

, 9

, I0

, II

PATIENT

Fig. 2. Maximally reached intraoperative blood pressure. some combination o f these. Some patients underwent craniotomy. Eventual outcomes in all cases were determined. RESULTS The patients ranged in age from 58 to 81 years. There were five men and six women. The presenting symptoms were transient ischemic attack in seven patients, previous cerebral infarct in one, amaurosis fugax in another, and nonfocal symptoms in two. Prothrombin time, partial thromboplastin time, and platelet count were normal in each patient. All had 90% or greater stenosis of the appropriate internal carotid artery and six had 99% stenosis. Several had significant lesions in the contralateral internal carotid artery. Six patients had preoperative CT scans, four of which were normal. One scan was consistent with

Fig. 3. Maximal blood pressure in postoperative period before intracerebral hemorrhage.

recent infarction in the ipsilateral hemisphere and carotid endarterectomy was delayed 30 days. One CT scan revealed an old ipsilateral infarct. Five patients operated on early in the series had no CT scan preoperatively, but all were neurologically intact at the time of operation. The symptom-free interval before surgery ranged from 1 to 30 days but was at least 1 week in eight of nine patients with focal cerebral ischemia (Table I). Three patients were taking antihypertensive medications preoperatively, but none had a preoperative blood pressure greater than 160/90 m m Hg (Fig. 1). Intraoperatively 9 of 11 patients became moderately to severely hypertensive (systolic 200 m m Hg or greater) with carotid cross-clamping (Fig. 2). Blood pressure was not artificially elevated with drugs. No patient had a neurologic deficit intraoperatively. During the postoperative period 5 o f !1 patients remained moderately or severely hypertensive before diagnosis of cerebral hemorrhage (Fig. 3). Hemorrhage occurred between 1 and 10 days postoperatively for a mean of 3.45 days. One patient showed symptoms in the recovery r o o m and was immediately returned to the operating room. An intraoperative arteriogram was normal in this patient. One other patient had repeat surgery after symptoms developed o n the third postoperative day and also had a normal intraoperative arteriogram. Five other patients also underwent cerebral angiography and all had patent carotid arteries (Table II). The presenting symptoms of cerebral hemorrhage are shown in Table II. Hemiparesis was the most

Volume 7 Number 2 February 1988

fntracranial hemorrhage after carotid endarterectomy 251

Table II. Summary of course of patients after cerebral hemorrhage Patient (age/sex)

Time to hemorrhage

59/F

4 days

62/F

Symptoms

Postoperative angiography

Postoperative CT scanning

Normal

(L) Parietal hem

1 day

HA (R) Hemi (R) Henri

Normal

(L) Temporal hem

58/M

RR

(L) Hemi

Normal

(R) Temporal hem

63/M

4 days

Normal

(L) Hem

81/M

10 days

(R) Hemi Aphasia Sz (R) Hemi

Normal

(L) Subdural hem

62/M

2 days

HA (L) Hemi

Normal

(R) Parietal hem

70/F

5 days

ND

(R) Temporal hem

68/F

3 days

ND

(R) Occipital hem

61/F

1 day

HA Sz (L) Hemi (L) Visual field Defick (L) Hemi

ND

ND

71/F

5 days

ND

ND

67/F

1 day

HA (R) Hemi Sz Coma HA (L) Hemi Coma

Normal

(L) Temporal hem

Management/ outcome Crani Surv def Crani Surv def Reexploration crani Surv def Crani Died Med Rx Surv Comp rec Med Rx Surv Comp rec Crani Comp rec Med Rx Surv def Med Rx Died Med Rx

Reexploration Med Rx Died

HA = headache; Hemi = hemiparesis; Sz = seizure; def = deficit; Comp rec = complete recovery; Crani = craniotomy; Hem = hematoma; Surv = survived.

common. Only five complained of headache. Three patients suffered seizures. In two patients early in this series the diagnosis was made by postmortem examination in one patient and by a combination of findings of hypertension, coma, dilated pupils, flaccid exremities after a seizure, and bloody cerebral spinal fluid on a lumbar puncture in the other. Nine of the 11 patients had CT scans postoperatively. Nine CT scans demonstrated intracranial hematoma ipsilateral t0 the side of surgery; eight showed intracerebral hematoma and one scan demonstrated a subdural hematoma (Table II). Four patients died. Five patients underwent craniotomy and four survived, but only one completely recovered. The indications for craniotomy were based on neurosurgical criteria. These included a progressing neurologic deficit despite medical therapy with mass effect on the CT scan from a surgically accessible hematoma. Only three of the seven survivors had minimal or no deficits (Table III). Overall the survival rate of postoperative cerebral hemorrhage was 64% (seven patients).

Table III. Eventual outcome of patients on the basis of type of therapy No. ofpatients Recovery Deficit Death Craniotomy Medical therapy

5 6

1 2

3 1

1 •3

DISCUSSION Reports from centers with large experiences in carotid surgery have consistently demonstrated that carotid endarterectomy can be performed with a perioperative stroke rate of 2% to 3%, but rarely is the mechanism of perioperative stroke described. We have found that performing carotid endarterectomy in the awake patient with continuous assessment of neurologic status and selective use of shunting on the basis of the patient's ability to tolerate carotid crossclamping together with aggressive evaluation of postoperative deficits has identified embolization, thrombosis, hypotension, cerebral hemorrhage, and uncertain origin as equally occurring causes of peri-

252 Pomposelli et al.

operative strokes. 2 Although presently available techniques can prevent strokes caused by most of these mechanisms, the means of prevention of cerebral hemorrhage afterendarterectomy remains unknown. Normal brain possesses the capability of resisting changes in blood flow caused by variations in blood pressure by the reflex contraction ofintracerebral vessels as pressure increases and their relaxation as it decreases? 4 In ischemic brain this autoregulatory mechanism is lost and vessels tend to remain maximally dilated. The use of positron emission tomography in focal cerebral infarction has shown a zone ofischemic, but viable brain surrounding the infarct, which has been labeled the "ischemic penumbra. "is Vasodilation in this area leads to enhancement on contrast CT scanning. Lassen 16 dubbed this "luxury perfusion" and suggested that it might be duc to local metabolic alterations, that is, acidosis from anaerobic metabolism. Animal studies performed by Waltz ~7 suggest that in these areas of cerebral vasodilation blood flow becomes passively dependent on blood pressure. Harvey and Rassmusscns showed that greatly fluctuating blood pressure in ischemic areas of brain could lead to hemorrhage, suggesting that the combination of autoregulatory paralysis and hypertension can greatly increase perfusion and precipitate vascular disruption in these areas. Several authors have suggested that a similar mechanism is responsible for paralysis of autoregulatory mechanisms in cerebral hemispheres of patients with near-occlusive carotid stenosis. Cerebral perfusion studies performed by Sundt et al.18 in patients with high-grade carotid stenosis revealed abnormally low blood flow to the ipsilateral cerebral hemisphere. Removal of the near-occlusive lesions by endarterectomy resulted in increases in blood flow of three to four times normal flow in some cases. This same category of patients occasionally manifests postoperative neurologic dysfunction characterized by severe temporal headache, nausea, vomiting, and seizures. Reigel et al?9 recently reported 10 patients with this symptom complex after relief of high-grade carotid stenosis. These patients uniformly demonstratcd periodic lateral izing epileptiform discharges on the ipsilateral side of the brain on electroencephalograms obtained in the postoperative period and suggested that these were a marker for its occurrence. Bernstein et al.~2 reported a case of a man who died of a massive cerebral hemorrhage 6 days after correction of an asymptomatic high-grade ipsilateral carotid stenosis. Preoperative neurologic examination and CT scan were normal. In the postoperative period he was normotensive and not taking heparin

Journal of VASCULAR SURGERY

or warfarin, but complained of severe headaches. Postmortem examination of the brain revealed histopathologic changes consistent with malignant hypertension in the hemisphere distal to the carotid endarterectomy whereas in the contralateral hemisphere pathologic findings were normal. They suggested that cerebral hyperperfusion led to these histologic findings, although others 2° have disputed this. It has been our policy to delay carotid surgery for 1 month in patients with acute frank infarction and carotid stenosis. Previously, patients with infrequent transient ischemic attacks, amaurosis fugax, or who were asymptomatic did not necessarily have preoperative CT scans if found to be clinically intact with a neurologic examination. In recent years most of our patients have had preoperative CT scans, although the discovery of an occult cerebral infarct ha~ not necessarily delayed surgery. In five patients operated on during the earlier part of this study, no CT scan was obtained preoperatively, but all patients were neurologically intact with stable symptoms at the time of operation. All of our patients except those with specific contraindications, receive 325 mg of aspirin and 75 nag of dipyridamole daily postoperatively. Heparin and warfarin are not used postoperatively except where specific strong indications exist (i.e., patients with prosthetic heart valves). No patient in this group was taking heparin or warfarin after operation. Hypertensive patients with blood pressure consistently greater than 160/90 m m Hg on multiple determinations who require carotid endarterectomy have surgery delayed until blood pressure is controlled. Diuretics, beta blockers, calcium channel blockers, and vasodilators are used to stabilize these~ patients preoperatively. Despite this, we frequently see intraoperative hypertension with carotid crossdamping in the conscious patient (75% of patients). Postoperatively, hypertension has been uncommon (less than 5%) in our patients, although when it does occur, we attempt to keep the systolic blood pressure below 180 m m Hg. From 1975'to 1983, intravenous hydralazine in doses of 5 mg was used for blood pressure control after carotid endarterectomy. From 1984 until the present, intravenous metoprolol 10 to 15 nag every 12 hours is our drug regimen. Five of 11 patients (Fig. 3) had postoperative blood pressures greater than 180 m m Hg despite use of this drug regimen. We note that these patients were the exception to our usual experience. Four of the patients in this study were both neurologically intact and had normal preoperative CT

Volume 7 Number 2 February 1988

scans. All tolerated surgery and initially did well. Two were never hypertensive postoperatively. Two complained of headache before showing signs of intracranial bleeding. It becomes apparent from these observations that normotensive, neurologically intact patients without demonstrable cerebral infarction or hypertension can also have cerebral hemorrhage after endarterectomy. The single factor that appears to link all 11 patients was the relief of high-grade (90%) or critical (99%) carotid stenosis. From these observations it would appear that the blood brain barrier is overcome from rapid increases in blood flow that occur after carotid endarterectomy done for high-grade stenosis. Moreover, severe hypertension (Fig. 3) may serve to increase the likelihood of disruption of the blood-brain barrier with resultant intracranial hemorrhage. It remains unknown why only a small fraction of patients with high-grade carotid stenosis have cerebral hemorrhage after endarterectomy, although the apparent frequency of contrallateral high-grade lesions or total occlusions that we and others ls'19 have observed suggests that the status of the contralateral carotid artery may bc a necessary factor. It is important to recognize that cerebral hemorrhage after endarterectomy, although rare, is an important cause of postoperative neurologic deficit that may occur in the absence of headache, nausea, seizures, and other signs associated with intracranial bleeding. Since effective treatment is available, and patients can recover without permanent disability, it is important that all new postoperative neurologic deficits, no matter how trivial, be expeditiously investigated. Our present rationale for evaluation of perioperative strokes has been described previously.2 Briefly, all major deficits occurring in the recovery room mandate immediate reexploration. A clinical assessment is then made of arterial patency. If the artery appears patent, intraoperative angiography is performed to look for correctable technical errors or nonoccluding thrombus. Major deficits occurring later require emergency angiography or reexploration and intraoperative angiography if a significant time delay (greater than 6 hours) is anticipated before a routine angiogram can be performed. Minor transient symptoms can be initially evaluated by noninvasive studies if available, but angiography should still be performed if they are not, and obviously in all patients in whom they show abnormalities. CT scans are also obtained in these patients. Patients with surgically correctable technical defects or thrombosis of their arteries are usually operated on unless CT scan reveals a new infarct or hemorrhage. Patients

Intracranial hemorrhage after carotid endarterectomy 253

with normal angiograms who have symptoms of stroke and hypertension should be suspected of having hemorrhage even in the absence of headache. When an intracranial hemorrhage is discovered, neurosurgical consultation should be obtained. Initially, patients are treated medically with blood pressure control. Patients with a progressing deficit, with mass effect demonstrated on the CT scan, from a hematoma in a surgically accessible location, are candidates for craniotomy and evacuation of the hematoma. In our series this strategy resulted in a survival rate of 66% and complete recovery in 27% of patients. In summary, we have observed postoperative intracranial hemorrhage in 0.5% to 0.7% of patients undergoing carotid endarterectomy. It accounts for approximately 20% of perioperative strokes. The most prominent risk factor other than recent cerebral infarction is surgical relief of high-grade carotid stenosis. Patients who are hypertensive, both intraoperatively and postoperatively, or where there is difficulty controlling postoperative hypertension are at added risk. Normotensive, neurologically intact patients without demonstrable cerebral infarction are also at risk with relief of high-grade carotid stenosis. Postoperative intracranial hemorrhage mandates the same expeditious and aggressive evaluation as other causes of perioperative stroke. Although morbidity and mortality rates remain high, careful medical management or neurosurgical intervention can save selected patients even in the face of severe neurologic deficits. REFERENCES

1. Imparato AM, Ramirez A, Riles TS, Mintger R. Cerebral protection in carotid surgery. Arch Surg 1982; 117:1073-8. 2. Imparato AM, Riles TS, Lamparello PJ, Ramirez A. The management of TIAs and acute strokes after carotid endarterectomy. In: Bernhard VM, Towne JF, eds. Complications in vascular surgery. New York: Grune & Stratton, Inc, 1985:725-38. 3. Wylie EJ, Hein MF, Adams JE. Intracranial hemorrhage following surgical revascularizatioan for treatment of acute strokes. )" Neurosurg 1954;21:212-5. 4. Breutman ME, Fields WS, Crawford ES, DeBakey ME. Cerebral hemorrhage in carotid artery surgery. Arch Neurol 1963;9:458-67. 5. Gonzalez L, Carson ML. Cerebral hemorrhage following successful endarterectomy of the internal carotid artery. Surg Gynecol Obstet 1966;122:773-7. 6. Hass WK, Claus RH, Goldberg AF, Johnson AL, Imparato AM, Ransohoff J. Special problems associated with surgical and thrombolytic treatment of strokes. Arch Surg i966; 92:27-31. 7. Meyer JS. Importance ofischemic damage to small vessels in experimental cerebral infarction. Neuropath Exp Neurol 1958;17:571-84.

254

Pomposelli et al.

Journal of VASCULAR SURGERY

8. Harvey J, Rassmussen T. Occlusion of the middle cerebral artery. Arch Neurol Psychiatr 1951;66:20-9. 9. Sundt TM. The ischemic tolerance of neural tissue and the need for monitoring and selective shunting during carotid endarterectomy. Stroke 1983;14:93-7. 10. Takolander RJ, Bergqvist D. Intracerebral hemorrhage after internal carotid endarterectomy. Acta Chir Scand 1983; 149:215-20. 11. Caplan LR, Skillman I, Ojemann R, Fields WS. Intracerebral hemorrhage following carotid endarterectomy. A hypertensive complication? Stroke 1978;9:457-60. 12. Bernstein M, Fleming JFR, Beck JHN. Cerebral hyperperfusion after carotid endarterectomy. A cause of cerebral hemorrhage. Neurosurgery 1984;15:50-6. 13. Solomon RA, Loftus CM, Quest DO, Lorrel JW. Incidence and etiology of intracerebral hemorrhage following carotid endarterectomy. J Neurosurg 1986;64:29-34. 14. Hachinski V, Norris JW. The vascular infrastructure. In: Hachinski V, Norris JW. The acute stroke. Philadelphia: FA Davis Co, 1985:27-40.

15. Hachinsld V, Norris JW. The reversibility of cerebral ischemia. In: Hachinski V, Norris JW. The acute stroke. Philadelphia: FA Davis Co, 1985:41-63. 16. Lassen WA. The luxury-perfusion syndrome and its possible relation to acute metabolic acidosis localizedwithin the brain. Lancet 1966;2:1113-5. 17. Waltz AG. Effect of blood pressure on blood flow in ischemic and non-ischemic cerebral cortex. Neurology 1968;18: 613-21. 18. Sundt TF, Sharbrough FW, Piepgras DG, Kearns TP, Messick IM, O'Fallon WM. Correlation of cerebral blood flow and electroencephalographic changes during carotid endarterectomy. Mayo Clin Proc 1981;56:533-43. 19. Reigel MM, Hollier LH, Sundt TF, Piepgras DG, Sharbrough FW, Cherry KJ. Cerebral hyperperfusion syndrome: a cause of neurologic dysfimction after carotid endarterectomy. J VASe SURG 1987;5:628-34. 20. Schroeder T, Sillesen H, Boesen J, Lanrsen H, Sorenson PS. Intracerebral hemorrhage after carotid endarterectomy. Eur J Vasc Surg 1987;1:51-60.

DISCUSSION

The most constant factor in these cases was relief of high-grade stenosis. 'Sundt at the Mayo Clinic has demonstrated hyperperfusion with cerebral blood flows increasing three or four times baseline levels after removal of such stenoses. We have documented such hyperperfusion with arteriography in a patient with TIAs and a tight stenosis who had severe headaches and seizures after operation. His postoperative arteriogram showed marked increased vascularity and hyperperfusion. Another most important factor associated with cerebral hemorrhage is postoperative hypertension. Thus in cases with possible ischemic vascular damage, loss of autoregulation, reactive hyperemia, and hyperperfusion, the superimposition of severe hypertension may well result ir~ vascular rupture and cerebral hemorrhage. At present, with careful monitoring and vigorous therapy, we can control postoperative hypertension. H o w to protect or enhance the integrity of the blood-brain barrier is the problem. Dr. Pomposelli, do you think the use of a shunt would have any value? Would use of steroids during crossdamping add any protection? Should we avoid postoperative aspirin therapy? Finally, is routine preoperative CT scanning in every case worthwhile or justifiable? Dr. Pomposelli. I thank Dr. Thompson for his remarks. This study was undertaken because of our impression that most causes of perioperative neurologic deficit after carotid endarterectomy are avoidable by presently available techniques. However, cerebral hemorrhage, which represents 20% of perioperative neurologic deficits in our patients, is not preventable at the present time. Indeed, despite a substantial amount of experimental animal work and clinical observation by Dr. Sundt and others,

Dr. Jesse E. Thompson (Dallas, Tex.). Cerebral hemorrhage now is a rare complication of carotid endarterectomy. Dr. Imparato's group has documented only 11 cases in 10 years, an average of one per year. However, when it occurs, it is a devastating event. About 25 years ago we all learned that patients with acute profound strokes with totally occluded or tightly stenotic carotid arteries operated on early suffered a high rate of hemorrhagic infarction with a prohibitive mortality rate. These patients are no longer operated on and are not the patients we are discussing here. Interestingly, of the 11 patients, Dr. Pomposelli has analyzed only one who had a previous remote infarct, whereas all the others had TIAs or nonlocalizing symptoms. All had intraoperative hypertension, whereas 6 of the 11 patients had postoperative hypertension. I analyzed a recent 10-year experience of our group of 918 endarterectomies and could identify with certainty only four cases of cerebral hemorrhage, two in patients with remote strokes and two with TIAs. All had tight stenoses, three had severe hypertension, and all four died. Our operative routine differs from the authors' in that we use general anesthesia, employ a shunt routinely, and inject dexamethasone (Decadron) intravenously immediately before carotid cross-clamping. In none of the 11 cases reported by Dr. Pomposelli was a shunt used. Although no demonstrable neurologic deficit was noted during carotid cross-clamping with the patients awake, I wonder whether some ischemia may have occurred. Prolonged ischemia followed by reperfusion results in an increase in oxygen-dependent enzyme metabolites, which may be detrimental to the blood-brain barrier.

Volume 7 Number 2 February 1988

cerebral hemorrhage has an undetermined mechanism in patients who have not had previous -strokes. Dr. Thompson, we do not believe that the routine use of shunts will prevent this complication. Experiences reported from other institutions that have used both selective and routine shunting have shown results similar to ours. In addition, we have found the development of cerebral ischemia with carotid cross-damping is not a subtle finding; patients usually manifest severe symptoms,, either losing consciousness or having a profound neurologic deficit within seconds of being clamped. Most of our patients who have had surgery without a shunt and do not show signs of cerebral ischemia have ultimately done well even with significant stenosis. We have, not routinely given our patients steroids. I am not aware of any data to suggest that it will prevent cerebral hemorrhage postoperatively. We routinely give our patients aspirin postoperatively. 'Iowever, half of the patients in this series were not given aspirin or dipyridamole postoperatively principally because the hemorrhage occurred before the therapy would have

Intracranial hemorrhage after carotid endarterectomy 255

routinely been started. In other studies that I am aware of, cerebral hemorrhage has occurred in patients who were and were not taking antiplatelet agents. It is unclear whether or not aspirin usage should be continued routinely in the postoperative period, although we have continued to do so to avoid the possible deposition ofplatclet thrombus on the endarterectomized area of the carotid artery. We have not routinely obtained preoperative CT scans on our patients because we believe that the neurologic status of the patient and the clinical presentation are more important factors in determining risk and ultimate outcome. In addition, we have not found that the discovery of an occult cerebral infarct in an otherwise asymptomatic or stable patient has altered our treatment plan. I am aware of studies that have suggested that patients with asymptomatic stenosis or transient cerebral ischemia who have occult infarction shown on the preoperative CT scan are • at an increased risk of stroke. However, I do not know of any studies that suggested there is an increased risk of cerebral hemorrhage in these circumstances.